systematic review and meta synthesis of qualitative studies

• Open access
• Published: 15 December 2015

Qualitative and mixed methods in systematic reviews

• David Gough 1  

Systematic Reviews volume  4 , Article number:  181 ( 2015 ) Cite this article

22k Accesses

46 Citations

23 Altmetric

Metrics details

Expanding the range of methods of systematic review

The logic of systematic reviews is very simple. We use transparent rigorous approaches to undertake primary research, and so we should do the same in bringing together studies to describe what has been studied (a research map) or to integrate the findings of the different studies to answer a research question (a research synthesis). We should not really need to use the term ‘systematic’ as it should be assumed that researchers are using and reporting systematic methods in all of their research, whether primary or secondary. Despite the universality of this logic, systematic reviews (maps and syntheses) are much better known in health research and for answering questions of the effectiveness of interventions (what works). Systematic reviews addressing other sorts of questions have been around for many years, as in, for example, meta ethnography [ 1 ] and other forms of conceptual synthesis [ 2 ], but only recently has there been a major increase in the use of systematic review approaches to answer other sorts of research questions.

There are probably several reasons for this broadening of approach. One may be that the increased awareness of systematic reviews has made people consider the possibilities for all areas of research. A second related factor may be that more training and funding resources have become available and increased the capacity to undertake such varied review work.

A third reason could be that some of the initial anxieties about systematic reviews have subsided. Initially, there were concerns that their use was being promoted by a new managerialism where reviews, particularly effectiveness reviews, were being used to promote particular ideological and theoretical assumptions and to indirectly control research agendas. However, others like me believe that explicit methods should be used to enable transparency of perspectives driving research and to open up access to and participation in research agendas and priority setting [ 3 ] as illustrated, for example, by the James Lind Alliance (see http://www.jla.nihr.ac.uk/ ).

A fourth possible reason for the development of new approaches is that effectiveness reviews have themselves broadened. Some ‘what works’ reviews can be open to criticism for only testing a ‘black box’ hypothesis of what works with little theorizing or any logic model about why any such hypothesis should be true and the mechanisms involved in such processes. There is now more concern to develop theory and to test how variables combine and interact. In primary research, qualitative strategies are advised prior to undertaking experimental trials [ 4 , 5 ] and similar approaches are being advocated to address complexity in reviews [ 6 ], in order to ask questions and use methods that address theories and processes that enable an understanding of both impact and context.

This Special Issue of Systematic Reviews Journal is providing a focus for these new methods of review whether these use qualitative review methods on their own or mixed together with more quantitative approaches. We are linking together with the sister journal Trials for this Special Issue as there is a similar interest in what qualitative approaches can and should contribute to primary research using experimentally controlled trials (see Trials Special Issue editorial by Claire Snowdon).

Dimensions of difference in reviews

Developing the range of methods to address different questions for review creates a challenge in describing and understanding such methods. There are many names and brands for the new methods which may or may not withstand the changes of historical time, but another way to comprehend the changes and new developments is to consider the dimensions on which the approaches to review differ [ 7 , 8 ].

One important distinction is the research question being asked and the associated paradigm underlying the method used to address this question. Research assumes a particular theoretical position and then gathers data within this conceptual lens. In some cases, this is a very specific hypothesis that is then tested empirically, and sometimes, the research is more exploratory and iterative with concepts being emergent and constructed during the research process. This distinction is often labelled as quantitative or positivist versus qualitative or constructionist. However, this can be confusing as much research taking a ‘quantitative’ perspective does not have the necessary numeric data to analyse. Even if it does have such data, this might be explored for emergent properties. Similarly, research taking a ‘qualitative’ perspective may include implicit quantitative themes in terms of the extent of different qualitative findings reported by a study.

Sandelowski and colleagues’ solution is to consider the analytic activity and whether this aggregates (adds up) or configures (arranges) the data [ 9 ]. In a randomized controlled trial and an effectiveness review of such studies, the main analysis is the aggregation of data using a priori non-emergent strategies with little iteration. However, there may also be post hoc analysis that is more exploratory in arranging (configuring) data to identify patterns as in, for example, meta regression or qualitative comparative analysis aiming to identify the active ingredients of effective interventions [ 10 ]. Similarly, qualitative primary research or reviews of such research are predominantly exploring emergent patterns and developing concepts iteratively, yet there may be some aggregation of data to make statements of generalizations of extent.

Even where the analysis is predominantly configuration, there can be a wide variation in the dimensions of difference of iteration of theories and concepts. In thematic synthesis [ 11 ], there may be few presumptions about the concepts that will be configured. In meta ethnography which can be richer in theory, there may be theoretical assumptions underlying the review question framing the analysis. In framework synthesis, there is an explicit conceptual framework that is iteratively developed and changed through the review process [ 12 , 13 ].

In addition to the variation in question, degree of configuration, complexity of theory, and iteration are many other dimensions of difference between reviews. Some of these differences follow on from the research questions being asked and the research paradigm being used such as in the approach to searching (exhaustive or based on exploration or saturation) and the appraisal of the quality and relevance of included studies (based more on risk of bias or more on meaning). Others include the extent that reviews have a broad question, depth of analysis, and the extent of resultant ‘work done’ in terms of progressing a field of inquiry [ 7 , 8 ].

Mixed methods reviews

As one reason for the growth in qualitative synthesis is what they can add to quantitative reviews, it is not surprising that there is also growing interest in mixed methods reviews. This reflects similar developments in primary research in mixing methods to examine the relationship between theory and empirical data which is of course the cornerstone of much research. But, both primary and secondary mixed methods research also face similar challenges in examining complex questions at different levels of analysis and of combining research findings investigated in different ways and may be based on very different epistemological assumptions [ 14 , 15 ].

Some mixed methods approaches are convergent in that they integrate different data and methods of analysis together at the same time [ 16 , 17 ]. Convergent systematic reviews could be described as having broad inclusion criteria (or two or more different sets of criteria) for methods of primary studies and have special methods for the synthesis of the resultant variation in data. Other reviews (and also primary mixed methods studies) are sequences of sub-reviews in that one sub-study using one research paradigm is followed by another sub-study with a different research paradigm. In other words, a qualitative synthesis might be used to explore the findings of a prior quantitative synthesis or vice versa [ 16 , 17 ].

An example of a predominantly aggregative sub-review followed by a configuring sub-review is the EPPI-Centre’s mixed methods review of barriers to healthy eating [ 18 ]. A sub-review on the effectiveness of public health interventions showed a modest effect size. A configuring review of studies of children and young people’s understanding and views about eating provided evidence that the public health interventions did not take good account of such user views research, and that the interventions most closely aligned to the user views were the most effective. The already mentioned qualitative comparative analysis to identify the active ingredients within interventions leading to impact could also be considered a qualitative configuring investigation of an existing quantitative aggregative review [ 10 ].

An example of a predominantly configurative review followed by an aggregative review is realist synthesis. Realist reviews examine the evidence in support of mid-range theories [ 19 ] with a first stage of a configuring review of what is proposed by the theory or proposal (what would need to be in place and what casual pathways would have to be effective for the outcomes proposed by the theory to be supported?) and a second stage searching for empirical evidence to test for those necessary conditions and effectiveness of the pathways. The empirical testing does not however use a standard ‘what works’ a priori methods approach but rather a more iterative seeking out of evidence that confirms or undermines the theory being evaluated [ 20 ].

Although sequential mixed methods approaches are considered to be sub-parts of one larger study, they could be separate studies as part of a long-term strategic approach to studying an issue. We tend to see both primary studies and reviews as one-off events, yet reviews are a way of examining what we know and what more we want to know as a strategic approach to studying an issue over time. If we are in favour of mixing paradigms of research to enable multiple levels and perspectives and mixing of theory development and empirical evaluation, then we are really seeking mixed methods research strategies rather than simply mixed methods studies and reviews.

Noblit G. Hare RD: meta-ethnography: synthesizing qualitative studies. Newbury Park NY: Sage Publications; 1988.

Google Scholar  

Barnett-Page E, Thomas J. Methods for the synthesis of qualitative research: a critical review. BMC Med Res Methodol. 2009;9:59.

Article   PubMed   PubMed Central   Google Scholar  

Gough D, Elbourne D. Systematic research synthesis to inform policy, practice and democratic debate. Soc Pol Soc. 2002;2002:1.

Moore GF, Audrey S, Barker M, Bond L, Bonell C, Hardeman W, et al. Process evaluation of complex interventions: Medical Research Council guidance 2015. BMJ. 2015;350:h1258

Candy B, Jone L, King M, Oliver S. Using qualitative evidence to help understand complex palliative care interventions: a novel evidence synthesis approach. BMJ Support Palliat Care. 2014;4:Supp A41–A42.

Article   Google Scholar  

Noyes J, Gough D, Lewin S, Mayhew A, Michie S, Pantoja T, et al. A research and development agenda for systematic reviews that ask complex questions about complex interventions. J Clin Epidemiol. 2013;66:11.

Gough D, Oliver S, Thomas J. Introduction to systematic reviews. London: Sage; 2012.

Gough D, Thomas J, Oliver S. Clarifying differences between review designs and methods. Syst Rev. 2012;1:28.

Sandelowski M, Voils CJ, Leeman J, Crandlee JL. Mapping the mixed methods-mixed research synthesis terrain. J Mix Methods Res. 2012;6:4.

Thomas J, O’Mara-Eves A, Brunton G. Using qualitative comparative analysis (QCA) in systematic reviews of complex interventions: a worked example. Syst Rev. 2014;3:67.

Thomas J, Harden A. Methods for the thematic synthesis of qualitative research in systematic reviews. BMC Med Res Methodol. 2008;8:45.

Oliver S, Rees R, Clarke-Jones L, Milne R, Oakley AR, Gabbay J, et al. A multidimensional conceptual framework for analysing public involvement in health services research. Health Exp. 2008;11:72–84.

Booth A, Carroll C. How to build up the actionable knowledge base: the role of ‘best fit’ framework synthesis for studies of improvement in healthcare. BMJ Qual Saf. 2015. 2014-003642.

Brannen J. Mixed methods research: a discussion paper. NCRM Methods Review Papers, 2006. NCRM/005.

Creswell J. Mapping the developing landscape of mixed methods research. In: Teddlie C, Tashakkori A, editors. SAGE handbook of mixed methods in social & behavioral research. New York: Sage; 2011.

Morse JM. Principles of mixed method and multi-method research design. In: Teddlie C, Tashakkori A, editors. Handbook of mixed methods in social and behavioural research. London: Sage; 2003.

Pluye P, Hong QN. Combining the power of stories and the power of numbers: mixed methods research and mixed studies reviews. Annu Rev Public Health. 2014;35:29–45.

Harden A, Thomas J. Mixed methods and systematic reviews: examples and emerging issues. In: Tashakkori A, Teddlie C, editors. Handbook of mixed methods in the social and behavioral sciences. 2nd ed. London: Sage; 2010. p. 749–74.

Chapter   Google Scholar  

Pawson R. Evidenced-based policy: a realist perspective. London: Sage; 2006.

Book   Google Scholar  

Gough D. Meta-narrative and realist reviews: guidance, rules, publication standards and quality appraisal. BMC Med. 2013;11:22.

Download references

Author information

Authors and affiliations.

EPPI-Centre, Social Science Research Unit, University College London, London, WC1H 0NR, UK

David Gough

You can also search for this author in PubMed   Google Scholar

Corresponding author

Correspondence to David Gough .

Additional information

Competing interests.

The author is a writer and researcher in this area. The author declares that he has no other competing interests.

Rights and permissions

Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License ( http://creativecommons.org/licenses/by/4.0/ ), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver ( http://creativecommons.org/publicdomain/zero/1.0/ ) applies to the data made available in this article, unless otherwise stated.

Reprints and permissions

About this article

Cite this article.

Gough, D. Qualitative and mixed methods in systematic reviews. Syst Rev 4 , 181 (2015). https://doi.org/10.1186/s13643-015-0151-y

Download citation

Received : 13 October 2015

Accepted : 29 October 2015

Published : 15 December 2015

DOI : https://doi.org/10.1186/s13643-015-0151-y

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

Systematic Reviews

ISSN: 2046-4053

• Submission enquiries: Access here and click Contact Us
• General enquiries: [email protected]

Systematic Reviews and Meta-Analyses: Synthesis & Discussion

• Get Started
• Exploratory Search
• Where to Search
• How to Search
• Grey Literature
• What about errata and retractions?
• Eligibility Screening
• Critical Appraisal
• Data Extraction
• Synthesis & Discussion
• Assess Certainty
• Share & Archive

One of the final steps in a systematic review is the synthesis of evidence  and writing the  discussion.

Your team began working toward this stage in the protocol  when you clearly identified the comparisons of interest. The work you've done in  data extraction  and critical appraisal  phases will feed directly into the synthesis.

Qualitative Synthesis

Qualitative synthesis in systematic reviews and/or meta-analyses.

Selecting the best approach for synthesis will depend on your scope , included material, field of research, etc. Therefore, it is important to follow methodological guidance that best matches your scope and field (e.g., a heath-focused  review guided by the Cochrane Handbook ). It can also be helpful to check out the synthesis and discussion of systematic reviews published by journals to which you plan to submit your review. 

In almost all cases, a qualitative synthesis of some kind will be part of your systematic review. A quantitative synthesis (e.g., meta-analysis ) should only be pursued as appropriate.

Meta-synthesis  and  Qualitative Evidence Synthesis are term sometimes used to describe a systematic review with only a qualitative synthesis . 

Guidance for Qualitative Synthesis

In some methodological guidance , this stage may effectively be described as a separate methodology altogether.

For example, the Cochrane Handbook,  Part 2: Core Methods covers synthesis through the lens of conducting a meta-analysis and/or quantitative synthesis. In Part 3: Specific perspectives in reviews,  Cochrane goes into more detail about qualitative evidence synthesis in Chapter 21 : Qualitative Evidence. Similarly, the JBI Manual for Evidence Synthesis contains a stand-alone chapter, Chapter 2: Systematic Reviews of Qualitative Evidence

Considerations and Decisions

• How you will group data for your synthesis   and how grouping decisions are made , whether you're pursuing just a qualitative synthesis or both a qualitative synthesis and a meta-analysis, is an important consideration prior to starting the synthesis.
• Assess heterogeneity  between studies, even if you don't plan to pursue a meta-analysis. Consider variability in participants studied, the definitions/measurements/frequency/etc. of interventions, or exposures, or outcomes, etc. This is part of the process to determine which studies are reasonable to synthesize.
• Selection of a formal qualitative synthesis approach ( optional )

Qualitative Data and Analysis Tools

Check out this Library Guide for more information about tools for qualitative data anlaysis at Virginia Tech.

Qualitative Synthesis Approaches

This is not a comprehensive list of approaches.  However, it can be a jumping off point for your team as you plan. The selection of approaches listed here is partially informed by Barnett-Page & Thomas (2009)

Note: Many of these approaches are also  stand-alone qualitative research methods. 

Content Analysis

"In the case of qualitative systematic reviews, raw data consist of qualitative research findings (i.e. text) that have been systematically extracted from existing research reports...The manner in which these findings are coded is largely guided by the research topic and questions and the data that are available for analysis." ( Finfgeld-Connett, 2014 )

• Identification of data segments
• Memoing & diagramming

Resources for Content Analysis

• Finfgeld-Connett D. Use of content analysis to conduct knowledge-building and theory-generating qualitative systematic reviews .  Qualitative Research . 2014;14(3):341-352. doi:10.1177/1468794113481790
• Elo, S., & Kyngäs, H. (2008). The qualitative content analysis process . Journal of Advanced Nursing, 62(1), 107–115. https://doi.org/10.1111/j.1365-2648.2007.04569.x 
• Mayring, P. (2015). Qualitative content analysis: Theoretical background and procedures . In A. Bikner-Ahsbahs, C. Knipping, & N. Presmeg (Eds.), Approaches to Qualitative Research in Mathematics Education: Examples of Methodology and Methods (pp. 365–380). Springer Netherlands. https://doi.org/10.1007/978-94-017-9181-6_13

Thematic Synthesis 

"Developed out of a need to conduct reviews that addressed questions relating to intervention need, appropriateness, acceptability, [and effectiveness] without compromising on key principles developed in systematic reviews"( Barnett-Paige & Thomas 2009 )

According to Thomas & Harden (2008) :

• Code text (line-by-line) 
• Develop descriptive themes
• Generate analytic themes

Resources for Thematic Synthesis 

Thomas J, Harden A. Methods for the thematic synthesis of qualitative research in systematic reviews . BMC Med Res Methodol. 2008 Jul 10;8:45. doi: 10.1186/1471-2288-8-45. PMID: 18616818; PMCID: PMC2478656.

Framework Synthesis

The "rationale [behind framework synthesis] is that qualitative research produces large amounts of textual data in the form of transcripts, observational fieldnotes etc. The sheer wealth of information poses a challenge for rigorous analysis. Framework synthesis offers a highly structured approach to organising and analysing data (e.g. indexing using numerical codes, rearranging data into charts etc)." ( Barnett-Page & Thomas, 2009 )

According to Brunton & James (2020) :

• Familiarization (with existing literature)
• Framework selection 
• Indexing & charting
• Mapping & interpretation

Resources for Framework Synthesis 

• Brunton, G., Oliver, S., & Thomas, J. (2020). Innovations in framework synthesis as a systematic review method . Research Synthesis Methods, 11(3), 316–330. https://doi.org/10.1002/jrsm.1399
• Dixon-Woods, M. Using framework-based synthesis for conducting reviews of qualitative studies .   BMC Med   9,  39 (2011). https://doi.org/10.1186/1741-7015-9-39

Grounded Theory

Grounded theory is defined as "a specific methodology developed by Glaser and Strauss (1967) for the purpose of building theory from data . In this book the term grounded theory is used in a more generic sense to denote theoretical constructs derived from qualitative analysis of data." ( Strauss & Corbin, 2008 )

According to Barnett-Paige & Thomas, 2009 , "key methods and assumptions...include":

• " simultaneous phases of data collection and analysis;
• inductive approach to analysis, allowing the theory to emerge from the theory ;
• the use of constant comparison method ;
• the use of theoretical sampling to reach theoretical saturation; and the generation of new theory"

Resources for Grounded Theory

• Glaser, B. G., & Strauss, A. L. (1967). The Discovery of Grounded Theory: Strategies for Qualitative Research . Aldine.
• Corbin, J., & Strauss, A. (2008).  Basics of qualitative research (3rd ed.): Techniques and procedures for developing grounded theory . SAGE Publications, Inc. https://dx. doi. org/10.4135/9781452230153

Meta-Ethnography 

This is proposed as an alternative to "Meta-Analysis" (Nolbit & Hare, 1998;  Barnett-Paige & Thomas 2009 ) and "should be interpretive rather than aggregative . We make the case that is should take the form of reciprocal translations of studies into one another" (Nolbit & Hare, 1998)

• Reciprocal translational analysis (RTA) - translate concepts; evolve overarching concepts
• Refutational synthesis - explore and explain contradictions between studies 
• Lines-of-argument (LOA) synthesis - building up a picture of a whole from the parts (the individual studies) 

Reporting Guideline

Improving reporting of meta-ethnography: The eMERGe reporting guidance (documents the development of eMERGe)

Resources for Meta-Ethnography

• Sattar, R., Lawton, R., Panagioti, M.  et al.   Meta-ethnography in healthcare research: a guide to using a meta-ethnographic approach for literature synthesis .  BMC Health Serv Res   21,  50 (2021). https://doi-org.ezproxy.lib.vt.edu/10.1186/s12913-020-06049-w
• France, E.F., Wells, M., Lang, H.  et al.   Why, when and how to update a meta-ethnography qualitative synthesis .  Syst Rev   5,  44 (2016). https://doi-org.ezproxy.lib.vt.edu/10.1186/s13643-016-0218-4
• Noblit, G. W., & Hare, R. D. (1988).  Meta-ethnography . SAGE Publications, Inc. https://dx. doi. org/10.4135/9781412985000
• Barnett-Page, E., Thomas, J. Methods for the synthesis of qualitative research: a critical review .  BMC Med Res Methodoly   9,  59 (2009). https://doi-org.ezproxy.lib.vt.edu/10.1186/1471-2288-9-59 
• Flemming, K., & Noyes, J. (2021). Qualitative Evidence Synthesis: Where Are We at?   International Journal of Qualitative Methods .  https://doi.org/10.1177/1609406921993276

Meta-Analysis

• Presenting Results
• Alternative Quantitative Synthesis

Meta-analysis

“The statistical analysis of a large collection of analysis results from individual studies for the purpose of integrating the findings .” ( Glass, 1976 )

“A statistical analysis which combines the results of several independent studies considered by the analyst to be ‘combinable’. ” (Huque, 1988)

“Meta-analysis is the statistical combination of results from two or more separate studies .” (Cochrane Handbook for Systematic Reviews of Interventions version 6.3, Chapter 10 )

The Cochrane Handbook ( Chapter 10.1 ) states:

"Do not start here!" ...results of meta-analyses can be very misleading if suitable attention has not been given to formulating the review question; specifying eligibility criteria; identifying and selecting studies; collecting appropriate data; considering risk of bias; planning intervention comparisons; and deciding what data would be meaningful to analyse. 

Choosing to pursue a Meta-Analysis

Reasons to pursue a meta-analysis.

Meta-analyses are a desirable end-goal as a this kind of synthesis can:

• Increase statistical power / improve precision
• Result in a summary estimate of the direction and size of the effect or association 
• Determine consistent across studies and explore why studies found different results
• Address questions that can’t be addressed by the individual studies (related to factors that differ across studies)
• Potentially resolve uncertainties if disagreement in literature and identify areas where evidence is insufficient

Reasons  not  to pursue a Meta-Analysis

Despite the appeal of the meta-analytic approach, it is vital that studies in the meta-analysis measure the same thing in the same way - that the studies themselves are reasonable to combine statistically .

According to Cochrane Chapter 12.1 , "Legitimate reasons [for not conducting a meta-analysis] include limited evidence ; incompletely reported outcome/effect estimates, or different effect measures used across studies; and bias in the evidence."  Table 12.1.a describes scenarios that may preclude meta-analyses, with possible solutions

Likewise, a synthesis is only as good as the studies included . In other words, a meta-analysis cannot improve poor quality studies.

This is not a comprehensive list - as with any analysis, you'll need to select specific approaches based on the kind of data you have.

• How you will  group data  for your synthesis   and  how grouping decisions are made , is an important consideration prior to starting the synthesis.
• Effect size measures  must be comparable across included studies and/or computable given the information available in the primary studies. For example, in a review of weight loss studies, you may convert all effects to pounds of lost weight. 
• Fixed-Effects: "assumes (1) all studies are measuring the same common (true) effect size (why we call it fixed), [and] (2) the observed results would be identical expect for random (sampling error)" ( Borenstein, 2009 )
• Random-Effects:  "assumes (1) there are multiple population effects that the studies are estimating - different effect sizes underlying different studies, [and] (2) variability between effect sizes is due to sampling error + variability in population of effects" ( Borenstein, 2009 )
• There are some additional analyses you'll need to run to determine  heterogeneity  (how different studies are from each other). A  sensitivity analysis  or meta-regression  is used to evaluate the effects of including or excluding certain groups of studies in your analysis, for example studies rated as low quality or high-risk of bias during the critical appraisal. You can also consider  publication bias  in your sample using a funnel plot (although there are valid critiques of the reliability of this practice).
• Glass, Gene V. “ Primary, Secondary, and Meta-Analysis of Research .”   Educational Researcher , vol. 5, no. 10, 1976, pp. 3–8, https://doi.org/10.2307/1174772.
• Borenstein, M., Hedges, L. V., Higgins, J. P. T., & Rothstein, H. R. (2009). Introduction to Meta-Analysis . John Wiley & Sons, Ltd. https://doi.org/10.1002/9780470743386
• Pigott, T. D., & Polanin, J. R. (2020). Methodological Guidance Paper: High-Quality Meta-Analysis in a Systematic Review .  Review of Educational Research ,  90 (1), 24–46.  https://doi.org/10.3102/0034654319877153

Tools for Meta-Analyses

Several tools exist for running your own meta-analyses. If you need further support, check out the help tab in this box.

Graphical User Interface (no programming required) 

• RevMan | Developed by the Cochrane Collaboration; good for beginners
• PyMeta | Built from PythonMeta package for command line interface in python  
• Comprehensive Meta-Analysis  |  fee-based
• MedCalc   |  fee-based

Command Line Interface (programming required)

• Metafor | R package; introduction from creator, Wolfgang Viechtbauer
• xmeta | R package; toolbox for multivariate meta-analyses
• PythonMeta | Python package; graphical interface available as PyMeta
• Polanin, J. R., Hennessy, E. A., & Tanner-Smith, E. E. (2017). A Review of Meta-Analysis Packages in R . Journal of Educational and Behavioral Statistics , 42 (2), 206–242. https://doi.org/10.3102/1076998616674315
• Video for using R for Meta package

Present Meta-Analysis Results

A meta-analysis is most commonly presented as a Forest Plot.

Forest Plot

If you are new to the concept of forest plots, check out Dr. Terry Shaneyfelt from UAB School of Medicine How to interpret a forest plot .

Alternative Quantitative Synthesis Methods

According to Cochrane Chapter 9.5 , "There are circumstances under which a meta-analysis is not possible, however, and other statistical synthesis methods might be considered, so as to make best use of the available data."

Table 9.5.a from the Cochrane Handbook , represented below, outlines some alternative synthesis method (and one  summary  method in the first row).

While the Evidence Synthesis Services (ESS) team at the University Libraries is available to support the other stages of a systematic review and/or meta-analysis,

we recommend reaching out to the Statistical Applications and Innovations Group (SAIG) for support in the statistical synthesis / meta-analysis. 

Methodological Guidance

• Health Sciences
• Animal, Food Sciences
• Social Sciences
• Environmental Sciences

Cochrane Handbook  - Part 1: About Cochrane Reviews

Chapter III : Reporting the Review  (specifically part  III.III );  Note: if you are not conducting a Cochrane Review, use this resource as a guidepost

Cochrane Handbook  -  Part 2: Core Methods

Chapter 9 : Summarizing study characteristics and preparing for synthesis

• 9.2 A general framework for synthesis 
• 9.3 Preliminary steps of a synthesis 
• 9.4 Checking data before synthesis
• 9.5  Types of synthesis

Chapter 10 : Analyzing data and undertaking meta-analyses  

• 10.1 Do not start here!
• 10.2 Introduction to meta-analysis
• 10.3 A generic inverse-variance approach to meta-analysis 
• 10.4 Meta-analysis of dichotomous outcomes
• 10.5 Meta-analysis of continuous outcomes 
• 10.6 Combining dichotomous and continuous outcomes
• 10.7 Meta-analysis of ordinal outcomes and measurement scales
• 10.8 Meta-analysis of counts and rates
• 10.9 Meta-analysis of time-to-event outcomes
• 10.10 Heterogeneity
• 10.11 Investing heterogeneity
• 10.12 Missing data
• 10.13 Bayesian approaches to meta-analysis 
• 10.14 Sensitivity analyses  
• 10.S1 Supplementary material: Statistical algorithms in Review Manager 5.1

Chapter 12 : Synthesizing and presenting findings using other methods

• 12.1 Why a meta-analysis of effect estimates may not be possible 
• 12.2 Statistical synthesis when meta-analysis of effect estimates is not possible
• 12.3 Visual display and presentation of the data

Chapter 13 : Assessing risk of bias due to missing results in a synthesis

• 13.2 Minimizing risk of bias due to missing results
• 13.3 A framework for assessing risk of bias due to missing results in a synthesis

Chapter 15: Interpreting results and drawing conclusions 

• 15.2 Issues of indirectness and applicability 
• 15.3 Interpreting results of statistical analyses 
• 15.4 Interpreting results from dichotomous outcomes
• 15.5 Interpreting results from continuous outcomes (including standardized mean differences)
• 15.6 Drawing conclusions 

Cochrane Handbook  - Part 3: Specific Perspectives in Reviews

Chapter 21 : Qualitative Evidence 

• 21.2 Designs for synthesizing and integrating qualitative evidence with intervention reviews
• 21.3 Defining qualitative evidence and studies
• 21.4 Planning qualitative evidence synthesis linked to an intervention review
• 21.5 Question development 
• 21.13 Methods for integrating the qualitative evidence synthesis with an intervention review

SYREAF Tutorials

Step 5 . data synthesis.

Conducting systematic reviews of intervention questions III: Synthesizing data from intervention studies using meta-analysis.  O’Connor AM, Sargeant JM, Wang C. Zoonoses Public Health. 2014 Jun;61 Suppl 1:52-63. doi: 10.1111/zph.12123. PMID: 24905996

Meta-analyses  including data from observational studies.  O’Connor AM, Sargeant JM. Prev Vet Med. 2014 Feb 15;113(3):313-22. doi: 10.1016/j.prevetmed.2013.10.017. Epub 2013 Oct 31. PMID: 24268538

Step 6. Presenting the results &  Step 7. Reaching a conclusion

Conducting systematic reviews of intervention questions II: Relevance screening, data extraction, assessing risk of bias, presenting the results and interpreting the findings.  Sargeant JM, O’Connor AM. Zoonoses Public Health. 2014 Jun;61 Suppl 1:39-51. doi: 10.1111/zph.12124. PMID: 24905995

Campbell -  MECCIR

C59. Addressing risk of bias / study quality in the synthesis ( review / final manuscript )

C60 . Incorporating assessments of risk of bias ( review / final manuscript )

C61. Combining different scales  ( review / final manuscript )

C62. Ensuring meta-analyses are meaningful  ( review / final manuscript )

C63. Assessing statistical heterogeneity  ( protocol &   review / final manuscript )

C64. Addressing missing outcome data  ( review / final manuscript )

C65. Addressing skewed data  ( review / final manuscript )

C66. Addressing studies with more than two groups  ( protocol &   review / final manuscript )

C67. Comparing subgroups  ( protocol &   review / final manuscript )

C68. Interpreting subgroup analyses  ( protocol &   review / final manuscript )

C69. Considering statistical heterogeneity when interpreting the results ( review / final manuscript )

C70. Addressing non-standard designs  ( protocol &   review / final manuscript )

C71. Conducting sensitivity analysis  ( protocol &   review / final manuscript )

C72. Interpreting results  ( review / final manuscript )

C73. Investigating reporting biases  ( review / final manuscript )

C77. Formulating implications for practice  ( review / final manuscript )

C78. Avoiding recommendations  ( review / final manuscript )

C79. Formulating implications for research  ( review / final manuscript )

CEE  -  Guidelines and Standards for Evidence synthesis in Environmental Management

Section 9. data synthesis.

CEE Standards for conduct and reporting

9.1 Systematic Reviews

9.1.1 Narrative Synthesis

9.1.2 Quantitative Data Synthesis 

9.1.3 Qualitative Data Synthesis

Section 10. Interpreting findings and reporting conduct

10.1 The interpretation of evidence syntheses

10.2 Reporting conduct of evidence synthesis

10.3 Reporting findings of evidence syntheses

Reporting in Protocol and Final Manuscript

• Final Manuscript

In the Protocol |  PRISMA-P

Data synthesis   (item 15), qualitative synthesis only.

If quantitative synthesis is not appropriate, describe the type of summary planned   (Item 15d)

all of the above plus:

Describe criteria under which study data will be quantitatively synthesised   (Item 15a) ...quantitative synthesis, describe planned summary measures , methods of handling data and methods of combining data from studies , including any planned exploration of consistency (such as I2 , Kendall’s τ)  (Item 15b) ...describe any proposed additional analyses (such as sensitivity or subgroup analyses, meta-regression)  (Item 15c)

In the Final Manuscript |  PRISMA

Synthesis methods (item 13; report in  methods ), essential items.

• Describe the processes used to decide which studies were eligible for each synthesis .  (Item 13a)
• Report any methods required to prepare the data collected from studies for presentation or synthesis, such as handling of missing summary statistics or data conversions  (Item 13b)
• Report chosen tabular structure(s) used to display results of individual studies and syntheses, along with details of the data presented  (Item 13c)
• Report chosen graphical methods used to visually display results of individual studies and syntheses  (Item 13c)
• If it was not possible to conduct a meta-analysis, describe and justify the synthesis methods ...or summary approach used   (Item 13d)
• If a planned synthesis was not considered possible or appropriate, report this and the reason for that decision   (Item 13d)

Additional Items

• If studies are ordered or grouped within tables or graphs based on study characteristics (such as by size of the study effect, year of publication), consider reporting the basis for the chosen ordering/grouping   (Item 13c)
• If non-standard graphs were used, consider reporting the rationale for selecting the chosen graph   (Item 13c)

Meta-Analysis (or other quantitative methods used)

• ...reference the software, packages, and version numbers used to implement synthesis methods (such as metan in Stata metafor (version 2.1-0) in R118)  (Item 13d)
• the meta-analysis model (fixed-effect, fixed-effects, or random-effects) and provide rationale for the selected model.
• the method used (such as Mantel-Haenszel, inverse-variance).
• any methods used to identify or quantify statistical heterogeneity (such as visual inspection of results, a formal statistical test for heterogeneity, heterogeneity variance (τ2), inconsistency (such as I2), and prediction intervals) 
• the between-study (heterogeneity) variance estimator used (such as DerSimonian and Laird, restricted maximum likelihood (REML)).
• the method used to calculate the confidence interval for the summary effect (such as Wald-type confidence interval, Hartung-Knapp-Sidik-Jonkman) 
• If a Bayesian approach to meta-analysis was used, describe the prior distributions about quantities of interest (such as intervention effect being analysed, amount of heterogeneity in results across studies)  (Item 13d)
• If multiple effect estimates from a study were included in a meta-analysis...describe the method(s) used to model or account for the statistical dependency. .. (Item 13d)
• If methods were used to explore possible causes of statistical heterogeneity , specify the method used (such as subgroup analysis, meta-regression)  (Item 13e)
• which factors were explored, levels of those factors, and which direction of effect modification was expected and why (where possible)  (Item 13e)
• whether analyses were conducted using study-level variables (where each study is included in one subgroup only), within-study contrasts (where data on subsets of participants within a study are available, allowing the study to be included in more than one subgroup), or some combination of the above ( Item 13e)
• how subgroup effects were compared (such as statistical test for interaction for subgroup analyses)  (Item 13e)
• If other methods were used to explore heterogeneity because data were not amenable to meta-analysis of effect estimates, describe the methods used (such as structuring tables to examine variation in results across studies based on subpopulation, key intervention components, or contextual factors) along with the factors and levels  (Item 13e)
• If any analyses used to explore heterogeneity were not pre-specified, identify them as such  (Item 13e)
• If sensitivity analyses were performed, provide d etails of each analysis (such as removal of studies at high risk of bias, use of an alternative meta-analysis model)  (Item 13f)
• If any sensitivity analyses were not pre-specified , identify them as such  (Item 13f)

If a random-effects meta-analysis model was used, consider specifying other details about the methods used, such as the method for calculating confidence limits for the heterogeneity variance  (Item 13d)

Reporting Bias Assessment (Item 14; report in methods )

• Specify the methods ... used to assess the risk of bias due to missing results in a synthesis (arising from reporting biases).
• If risk of bias due to missing results was assessed using an existing tool, specify the methodological components/domains/items of the tool, and the process used to reach a judgment of overall risk of bias .
• If any adaptations to an existing tool to assess risk of bias due to missing results were made (such as omitting or modifying items), specify the adaptations.
• If a new tool to assess risk of bias due to missing results was developed for use in the review, describe the content of the tool and make it publicly accessible.
• Report how many reviewers assessed risk of bias due to missing results in a synthesis, whether multiple reviewers worked independently, and any processes used to resolve disagreements between assessors.
• Report any processes used to obtain or confirm relevant information from study investigators.
• If an automation tool was used to assess risk of bias due to missing results, report how the tool was used , how the tool was trained , and details on the tool’s performance and internal validation

Results of Synthesis (Item 20; report in results )

• Provide a brief summary of the characteristics and risk of bias among studies contributing to each synthesis (meta-analysis or other). The summary should focus only on study characteristics that help in interpreting the results (especially those that suggest the evidence addresses only a restricted part of the review question, or indirectly addresses the question). If the same set of studies contribute to more than one synthesis, or if the same risk of bias issues are relevant across studies for different syntheses, such a summary need be provided once only  (Item 20a)
• Indicate which studies were included in each synthesis (such as by listing each study in a forest plot or table or citing studies in the text)  (Item 20a)
• Report results of all statistical syntheses described in the protocol and all syntheses conducted that were not pre-specified  (Item 20b)

Meta-Analysis (or other quantitative methods used)

• the summary estimate and its precision (such as standard error or 95% confidence/credible interval).
• measures of statistical heterogeneity (such as τ2, I2, prediction interval).
• If other statistical synthesis methods were used (such as summarising effect estimates, combining P values), report the synthesised result and a measure of precision (or equivalent information, for example, the number of studies and total sample size)  (Item 20b)
• If the statistical synthesis method does not yield an estimate of effect (such as when P values are combined), report the relevant statistics (such as P value from the statistical test), along with an interpretation of the result that is consistent with the question addressed by the synthesis method (for example, “There was strong evidence of benefit of the intervention in at least one study (P < 0.001, 10 studies)” when P values have been combined)  (Item 20b)
• If comparing groups , describe the direction of effect (such as fewer events in the intervention group, or higher pain in the comparator group)  (Item 20b)
• If synthesising mean differences , specify for each synthesis, where applicable, the unit of measurement (such as kilograms or pounds for weight), the upper and lower limits of the measurement scale (for example, anchors range from 0 to 10), direction of benefit (for example, higher scores denote higher severity of pain), and the minimally important difference , if known. If synthesising standardised mean differences and the effect estimate is being re-expressed to a particular instrument, details of the instrument, as per the mean difference, should be reported  (Item 20b)
• present results regardless of the statistical significance, magnitude, or direction of effect modification  (Item 20c)
• identify the studies contributing to each subgroup   (Item 20c)
• report results with due consideration to the observational nature of the analysis and risk of confounding due to other factors  (Item 20c)
• If subgroup analysis was conducted, report for each analysis the exact P value for a test for interaction as well as, within each subgroup, the summary estimates , their precision (such as standard error or 95% confidence/credible interval) and measures of heterogeneity . Results from subgroup analyses might usefully be presented graphically  (Item 20c)
• If meta-regression was conducted, report for each analysis the exact P value for the regression coefficient and its precision  (Item 20c)
• If informal methods (that is, those that do not involve a formal statistical test) were used to investigate heterogeneity —which may arise particularly when the data are not amenable to meta-analysis— describe the results observed . For example, present a table that groups study results by dose or overall risk of bias and comment on any patterns observed  (Item 20c)
• report the results for each sensitivity analysis  (Item 20d)
• comment on how robust the main analysis was given the results of all corresponding sensitivity analyses   (Item 20d)
• If subgroup analysis was conducted, consider presenting the estimate for the difference between subgroups and its precision  (Item 20c)
• If meta-regression was conducted, consider presenting a meta-regression scatterplot with the study effect estimates plotted against the potential effect modifier   (Item 20c)
• the summary effect estimate , a measure of precision (and potentially other relevant statistics, for example, I2 statistic) and contributing studies for the original meta-analysis;
• the same information for the  sensitivity analysis ; and
• details of the original and sensitivity analysis assumptions   (Item 20d)
• presenting results of sensitivity analyses visually using forest plots   (Item 20d)

Reporting Biases (Item 21; report in results )

• Present assessments of risk of bias due to missing results (arising from reporting biases) for each synthesis assessed.
• If a tool was used to assess risk of bias due to missing results in a synthesis, present responses to questions in the tool, judgments about risk of bias, and any i nformation used to support such judgments to help readers understand why particular judgments were made.
• If a funnel plot was generated to evaluate small-study effects (one cause of which is reporting biases), present the plot and specify the effect estimate and measure of precision used in the plot (presented typically on the horizontal axis and vertical axis respectively). If a contour-enhanced funnel plot was generated, specify the “milestones” of statistical significance that the plotted contour lines represent (P=0.01, 0.05, 0.1, etc).
• If a test for funnel plot asymmetry was used, report the exact P value observed for the test and potentially other relevant statistics, such as the standardised normal deviate, from which the P value is derived.
• If any sensitivity analyses seeking to explore the potential impact of missing results on the synthesis were conducted, present results of each analysis (see item #20d), compare them with results of the primary analysis, and report results with due consideration of the limitations of the statistical method.
• If studies were assessed for selective non-reporting of results by comparing outcomes and analyses pre-specified in study registers, protocols, and statistical analysis plans with results that were available in study reports, consider presenting a matrix (with rows as studies and columns as syntheses) to present the availability of study results.
• If an assessment of selective non-reporting of results reveals that some studies are missing from the synthesis, consider displaying the studies with missing results underneath a forest plot or including a table with the available study results (for example, see forest plot in Page et al)

Discussion (Item 23)

• Provide a  general interpretation of the results  in the context of other evidence  (Item 23a)
• Discuss any  limitations of the evidence  included in the review  (Item 23b)
• Discuss any  limitations of the review processes  used and comment on the  potential impact  of each limitation  (Item 23c)
• Discuss  implications of the results  for practice and policy  (Item 23d)
• Make  explicit recommendations  for  future research  (Item 23d)
• << Previous: Data Extraction
• Next: Assess Certainty >>
• Last Updated: May 15, 2024 8:44 AM
• URL: https://guides.lib.vt.edu/SRMA

Advertisement

Meta-Analysis and Meta-Synthesis Methodologies: Rigorously Piecing Together Research

• Original Paper
• Published: 18 June 2018
• Volume 62 , pages 525–534, ( 2018 )

Cite this article

• Heather Leary   ORCID: orcid.org/0000-0002-2487-578X 1 &
• Andrew Walker 2  

5504 Accesses

33 Citations

7 Altmetric

Explore all metrics

For a variety of reasons, education research can be difficult to summarize. Varying contexts, designs, levels of quality, measurement challenges, definition of underlying constructs, and treatments as well as the complexity of research subjects themselves can result in variability. Education research is voluminous and draws on multiple methods including quantitative, as well as, qualitative approaches to answer key research questions. With increased numbers of empirical research in Instructional Design and Technology (IDT), using various synthesis methods can provide a means to more deeply understand trends and patterns in research findings across multiple studies. The purpose of this article is to illustrate structured review or meta-synthesis procedures for qualitative research, as well as, novel meta-analysis procedures for the kinds of multiple treatment designs common to IDT settings. Sample analyses are used to discuss key methodological ideas as a way to introduce researchers to these techniques.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price includes VAT (Russian Federation)

Instant access to the full article PDF.

Rent this article via DeepDyve

Institutional subscriptions

Similar content being viewed by others

Trends and Issues in Qualitative Research Methods

Educational design research: grappling with methodological fit

Methodological Approaches to Literature Review

Attride-Stirling, J. (2001). Thematic networks: An analytic tool for qualitative research. Qualitative Research, 1 (3), 385–405.

Article   Google Scholar  

Barnett-Page, E., & Thomas, J. (2009). Methods for the synthesis of qualitative research: a critical review. ESRC National Centre for Research Methods, Social Science Research Unit, Institute of Education, University of London, London, 01/09.

Barton, E. E., Pustejovsky, J. E., Maggin, D. M., & Reichow, B. (2017). Technology-aided instruction and intervention for students with ASD: A meta-analysis using novel methods of estimating effect sizes for single-case research. Remedial and Special Education, 38 (6), 371–386.

Beal, C. R., Arroyo, I., Cohen, P. R., & Woolf, B. P. (2010). Evaluation of AnimalWatch: An intelligent tutoring system for arithmetic and fractions. Journal of Interactive Online Learning, 9 (1), 64–77.

Google Scholar  

Belland, B., Walker, A., Kim, N., & Lefler, M. (2017a). Synthesizing results from empirical research on computer-based scaffolding in STEM education: A meta-analysis. Review of Educational Research, 82 (2), 309–344.

Belland, B., Walker, A., & Kim, N. (2017b). A bayesian network meta-analysis to synthesize the influence of contexts of scaffolding use on cognitive outcomes in STEM education. Review of Educational Research, 87 (6), 1042–1081.

Bhatnagar, N., Lakshmi, P. V. M., & Jeyashree, K. (2014). Multiple treatment and indirect treatment comparisons: An overview of network meta-analysis. Perspectives in Clinical Research, 5 (4), 154–158. https://doi.org/10.4103/2229-3485.140550 .

Boote, D. N., & Beile, P. (2005). Scholars before researchers: On the centrality of the dissertation literature review in research preparation. Educational Researcher, 34 (6), 3–15. https://doi.org/10.3102/0013189X034006003 .

Borenstein, M., Hedges, L. V., Higgins, J. P. T., & Rothstein, H. R. (2009). Introduction to meta-analysis . Chichester: Wiley.

Book   Google Scholar  

Bulu, S., & Pedersen, S. (2010). Scaffolding middle school students’ content knowledge and illstructured problem solving in a problem-based hypermedia learning environment. Educational Technology Research & Development, 58 (5), 507–529.

Cooper, H. (2009). Research synthesis and meta-analysis: A step-by-step approach (applied social research methods) . Thousand Oaks: Sage Publications.

Cooper, H. (2016). Research synthesis and meta-analysis: A step-by-step approach (5th ed.). Thousand Oaks: Sage publications.

Cooper, H., & Hedges, L. V. (1994). The handbook of research synthesis . New York: Russell Sage Foundation.

Finfgeld, D. L. (2003). Metasynthesis: The state of the art - so far. Qualitative Health Research, 13 (7), 893–904.

Finfgeld-Connett, D. (2010). Generalizability and transferability of meta-synthesis research findings. Journal of Advanced Nursing, 66 (2), 246–254.

Finfgeld-Connett, D. (2014). Metasynthesis findings: Potential versus reality. Qualitative Health Research, 24 , 1581–1591. https://doi.org/10.1177/1049732314548878 .

Finfgeld-Connett, D. (2016). The future of theory-generating meta-synthesis research. Qualitative Health Research, 26 (3), 291–293.

Finlayson, K., & Dixon, A. (2008). Qualitative meta-synthesis: A guide for the novice. Nurse Researcher, 15 (2), 59–71.

Glass, G. V. (1976). Primary, secondary, and meta-analysis of research. Educational Researcher, 5 (10), 3–8.

Glass, G. V. (2000). Meta-analysis at 25 . http://www.gvglass.info/papers/meta25.html

Gurevitch, J., Koricheva, J., Nakagawa, S., & Stewart, G. (2018). Meta-analysis and the science of research synthesis. Nature, 555 , 175–182.

Harden, A. (2010). Mixed-methods systematic reviews: Integrating quantitative and qualitative findings. Focus Technical Brief, 25 , 1–8.

Heller, R. (1990). The role of hypermedia in education: A look at the research issues. Journal of Research on Computing in Education, 22 , 431–441.

Higgins, S. (2016). Meta-synthesis and comparative meta-analysis of education research findings: Some risks and benefits. Review of Education, 4 (1), 31–53.

Higgins, J. P., & Green, S. (2008). Cochrane handbook for systematic reviews of interventions . Chichester: Wiley. https://doi.org/10.1002/9780470712184 .

Higgins, J. P. T., Altman, D. G., Gøtzsche, P. C., Jüni, P., Moher, D., Oxman, A. D., … Sterne, J. (2011). The Cochrane Collaboration’s tool for assessing risk of bias in randomised trials. BMJ, 343 . Retrieved from http://www.bmj.com/content/343/bmj.d5928

Karich, A. C., Burns, M. K., & Maki, K. E. (2014). Updated meta-analysis of learner control within educational technology. Review of Educational Research, 84 (3), 392–410.

Lachal, J., Revah-Levy, A., Orri, M., & Moro, M. R. (2017). Metasynthesis: An original method to synthesize qualitative literature. Frontiers in Psychiatry, 8 , 1–9.

Lam, S. K. H., & Owen, A. (2007). Combined resynchronisation and implantable defibrillator therapy in left ventricular dysfunction: Bayesian network meta-analysis of randomised controlled trials. BMJ: British Medical Journal, 335 (7626), 925–928.

Leary, H., Shelton, B. E., & Walker, A. (2010). Rich visual media meta-analyses for learning: An approach at meta-synthesis. Paper presented at the American Educational Research Association annual meeting, Denver, CO.

Lin, H. (2015). A meta-synthesis of empirical research on the effectiveness of computer-mediated communication (CMC) in SLA. Language Learning & technology, 19 (2), 85–117 Retrieved from http://llt.msu.edu/issues/june2015/lin.pdf .

Lipsey, M., & Wilson, D. (2001). Practical meta-analysis (Applied social research methods) . Thousand Oaks: Sage Publications.

Liu, Y., Cornish, A., & Clegg, J. (2007). ICT and special educational needs: Using meta-synthesis for bridging the multifaceted divide. In Y. Shi, G. D. van Albada, J. Dongarra, & P. M. A. Sloot (Eds.), Computational science–ICCS 2007. ICCS 2007. Lecture notes in computer science (Vol. 4490). Berlin, Heidelberg: Springer.

Ludvigsen, M. S., Hall, E. O. C., Meyer, G., Fegran, L., Aagaard, H., & Uhrenfeldt, L. (2016). Using Sandelowski and Barroso’s meta-synthesis method in advancing qualitative evidence. Qualitative Health Research, 26 (3), 320–329.

Lumley, T. (2002). Network meta-analysis for indirect treatment comparisons. Statistics in Medicine, 21 (16), 2313–2324. https://doi.org/10.1002/sim.1201 .

Lunn, D., Spiegelhalter, D., Thomas, A., & Best, N. (2009). The BUGS project: Evolution, critique and future directions. Statistics in Medicine, 28 (25), 3049–3067.

McKenney, S., & Reeves, T. C. (2012). Conducting educational design research . London: Routledge.

Mumtaz, S. (2000). Factors affecting teachers’ use of information and communications technology: A review of the literature. Journal of Information Technology for Teacher Education, 9 (3), 319–342.

Noblit, G. W., & Hare, R. D. (1988). Meta-ethnography: Synthesizing qualitative studies . London: Sage.

Paterson, B. L., Thorne, S. E., Canam, C., & Jillings, C. (2001). Meta-study of qualitative health research: A practical guide to meta-analysis and meta-synthesis . London: Sage.

Puhan, M. A., Schünemann, H. J., Murad, M. H., Li, T., Brignardello-Petersen, R., Singh, J. A., … Guyatt, G. H. (2014). A GRADE Working Group approach for rating the quality of treatment effect estimates from network meta-analysis. BMJ, 349, g5630.

Reeves, T., & Oh, E. (2017). The goals and methods of educational technology research over a quarter century (1989–2014). Educational Technology Research & Development, 65 (2), 325–339.

Salanti, G., Giovane, C. D., Chaimani, A., Caldwell, D. M., & Higgins, J. P. T. (2014). Evaluating the quality of evidence from a network meta-analysis. PLOS ONE, 9 (7), e99682.

Sandelowski, M., & Barroso, J. (2003). Toward a metasynthesis of qualitative findings on motherhood in HIV-positive women. Research in Nursing and Health, 26 (2), 153–170.

Sandelowski, M., Docherty, S., & Emden, C. (1997). Quality metasynthesis: Issues and techniques. Research in Nursing and Health, 20 , 365–371.

Taquero, J. M. (2011). A meta-ethnographic synthesis of support services in distance learning programs. Journal of Information Technology Education: Innovations in Practice, 10 , 157–179.

Thorne, S., Jensen, L., Kearney, M. H., Noblit, G., & Sandelowski, M. (2004). Qualitative metasynthesis: Reflection on methodological orientation and ideological agenda. Qualitative Health Research, 14 (10), 1342–1365.

Tondeur, J., van Braak, J., Sang, G., Voogt, J., Fisser, P., & Ottenbreit-Lewftwich, A. (2012). Preparing pre-service teachers to integrate technology in education: A synthesis of qualitative evidence. Computers & Education, 59 (1), 134–144.

Torraco, R. J. (2016). Writing integrative literature reviews: Using the past and present to explore the future. Human Resource Development Review, 15 (4), 404–428.

Walker, A., Belland, B., Kim, N., & Lefler, M. (2016). Searching for structure: A bayesian network meta-analysis of computer-based scaffolding research in STEM education. Paper presented at the American Educational Research Association, San Antonio, TX.

Walsh, D., & Downe, S. (2005). Meta-synthesis method for qualitative research: A literature review. Methodological Issues in Nursing Education, 50 (2), 204–211.

Young, J. (2017). Technology-enhanced mathematics instruction: A second-order meta-analysis of 30 years of research. Educational Research Review, 22 , 19–33.

Download references

Author information

Authors and affiliations.

Brigham Young University, 150-G MCKB, Provo, UT, 84602, USA

Heather Leary

Utah State University, 2830 Old Main Hill, Logan, UT, 84322, USA

Andrew Walker

You can also search for this author in PubMed   Google Scholar

Corresponding author

Correspondence to Heather Leary .

Ethics declarations

Conflict of interest.

The authors declare that they have no conflict of interest.

Human Studies

This article does not contain any studies with human participants performed by any of the authors.

Rights and permissions

Reprints and permissions

About this article

Leary, H., Walker, A. Meta-Analysis and Meta-Synthesis Methodologies: Rigorously Piecing Together Research. TechTrends 62 , 525–534 (2018). https://doi.org/10.1007/s11528-018-0312-7

Download citation

Published : 18 June 2018

Issue Date : September 2018

DOI : https://doi.org/10.1007/s11528-018-0312-7

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Meta-analysis
• Meta-synthesis
• Synthesis methods
• Education research
• Find a journal
• Publish with us
• Track your research

• Meta-Analysis/Meta-Synthesis

Meta Analysis

Meta-analysis   is a set of statistical techniques for synthesizing data across studies. It is a statistical method for combining the findings from quantitative studies. It evaluates, synthesizes, and summarizes results. It may be conducted independently or as a specialized subset of a systematic review.  A systematic review attempts to collate empirical evidence that fits predefined eligibility criteria to answer a specific research question. Meta-analysis is a quantitative, formal, epidemiological study design used to systematically assess the results of previous research to derive conclusions about that body of research (Haidrich, 2010). Rigorously conducted meta-analyses are useful tools in evidence-based medicine . Outcomes from a meta-analysis may include a more precise estimate of the effect of a treatment or risk factor for disease or other outcomes. Not all systematic reviews include meta-analysis , but all meta-analyses are found in systematic reviews (Haidrich, 2010).

A Meta analysis is appropriate when a group of studies report quantitative results rather than qualitative findings or theory, if they examine the same or similar constructs or relationships, if they are derived from similar research designs and report the simple relationships between two variables rather than relationships that have been adjusted for the effect of additional variables (siddaway, et al., 2019).

Haidich A. B. (2010). Meta-analysis in medical research.  Hippokratia ,  14 (Suppl 1), 29–37.

Siddaway, A. P., Wood, A. M., & Hedges, L. V. (2019). How to do a systematic review: A best practice guide for conducting and reporting narrative reviews, meta-analyses, and meta-syntheses.  Annual Review of Psychology, 70 , 747–770.

Meta Synthesis

A meta synthesis is the systematic review and integration of findings from qualitative studies (Lachal et al., 2017). Reviews of qualitative information can be conducted and reported using the same replicable, rigorous, and transparent methodology and presentation. A meta-synthesis can be used when a review aims to integrate qualitative research.  A meta-synthesis attempts to synthesize qualitative studies on a topic to identify key themes, concepts, or theories that provide novel or more powerful explanations for the phenomenon under review (Siddaway et al., 2019).

Lachal, J., Revah-Levy, A., Orri, M., & Moro, M. R. (2017). Metasynthesis: An original method to synthesize qualitative literature in psychiatry.  Frontiers in Psychiatry, 8 , 269 . 

Siddaway, A. P., Wood, A. M., & Hedges, L. V. (2019). How to do a systematic review: A best practice guide for conducting and reporting narrative reviews, meta-analyses, and meta-syntheses.  Annual Review of Psychology, 70 , 747–770 .

• << Previous: Rapid Review
• Next: Selecting a Review Type >>
• Types of Questions
• Key Features and Limitations
• Is a Systematic Review Right for Your Research?
• Integrative Review
• Scoping Review
• Rapid Review
• Selecting a Review Type
• Reducing Bias
• Guidelines for Student Researchers
• Training Resources
• Register Your Protocol
• Handbooks & Manuals
• Reporting Guidelines
• PRESS 2015 Guidelines
• Search Strategies
• Selected Databases
• Grey Literature
• Handsearching
• Citation Searching
• Study Types & Terminology
• Quantitative vs. Qualitative Research
• Critical Appraisal of Studies
• Broad Functionality Programs & Tools
• Search Strategy Tools
• Deduplication Tools
• CItation Screening
• Critical Appraisal Tools
• Quality Assessment/Risk of Bias Tools
• Data Collection/Extraction
• Meta Analysis Tools
• Books on Systematic Reviews
• Finding Systematic Review Articles in the Databases
• Systematic Review Journals
• More Resources
• Evidence-Based Practice Research in Nursing
• Citation Management Programs
• Last Updated: May 3, 2024 5:52 PM
• URL: https://libguides.adelphi.edu/Systematic_Reviews

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here .

Loading metrics

Open Access

Peer-reviewed

Research Article

Care and support when a baby is stillborn: A systematic review and an interpretive meta-synthesis of qualitative studies in high-income countries

Contributed equally to this work with: Margareta Persson, Carina Berterö

Roles Conceptualization, Data curation, Formal analysis, Methodology, Validation, Writing – original draft

Affiliation Department of Nursing, Umeå University, Umeå, Sweden

Roles Conceptualization, Methodology, Writing – review & editing

¶ ‡ IH, MH, MKF, NP and RAS also contributed equally to this work.

Affiliation Department of Women’s and Children’s Health, Uppsala University, Uppsala, Sweden

Roles Conceptualization, Methodology, Project administration, Resources, Writing – review & editing

Current address: HTA Region Stockholm, Centre for Health Economics, Informatics and Health Services Research (CHIS), Stockholm Health Care Services, Sweden

Affiliations Swedish Agency for Health Technology Assessment and Assessment of Social Services (SBU), Stockholm, Sweden, Department of Learning, Informatics, Management and Ethics, Karolinska Institute, Stockholm, Sweden

Roles Methodology, Resources, Writing – review & editing

Affiliation Swedish Agency for Health Technology Assessment and Assessment of Social Services (SBU), Stockholm, Sweden

Affiliations Department of Women’s and Children’s Health, Uppsala University, Uppsala, Sweden, Centre for Crisis Psychology, University of Bergen, Bergen, Norway

* E-mail: [email protected]

Affiliation Division of Nursing Sciences and Reproductive Health, Department of Health, Medicine and Caring Sciences, Linköping University, Linköping, Sweden

• Margareta Persson, 
• Ingegerd Hildingsson, 
• Monica Hultcrantz, 
• Maja Kärrman Fredriksson, 
• Nathalie Peira, 
• Rebecca A. Silverstein, 
• Josefin Sveen, 
• Carina Berterö

• Published: August 15, 2023
• https://doi.org/10.1371/journal.pone.0289617
• Peer Review
• Reader Comments

Introduction

Approximately 2 million babies are stillborn annually worldwide, most in low- and middle-income countries. Present review studies of the parental and healthcare providers’ experiences of stillbirth often include a variety of settings, which may skew the findings as the available resources can vary considerably. In high-income countries, the prevalence of stillbirth is low, and support programs are often initiated immediately when a baby with no signs of life is detected. There is limited knowledge about what matters to parents, siblings, and healthcare providers when a baby is stillborn in high-income countries.

This systematic review and interpretive meta-synthesis aim to identify important aspects of care and support for parents, siblings, and healthcare professionals in high-income countries from the diagnosis of stillbirth throughout the birth and postpartum period.

A systematic review and qualitative meta-synthesis were conducted to gain a deeper and broader understanding of the available knowledge about treatment and support when stillbirth occurred. Relevant papers were identified by systematically searching international electronic databases and citation tracking. The quality of the included studies was assessed, and the data was interpreted and synthesised using Gadamer’s hermeneutics. The review protocol, including qualitative and quantitative study approaches, was registered on PROSPERO (CRD42022306655).

Sixteen studies were identified and included in the qualitative meta-synthesis. Experiences of care and support were interpreted and identified as four fusions. First, Personification is of central importance and stresses the need to acknowledge the baby as a unique person. The parents became parents even though their baby was born dead: The staff should also be recognised as the individuals they are with their personal histories. Second, the personification is reinforced by a respectful attitude where the parents are confirmed in their grief; the baby is treated the same way a live baby would be. Healthcare professionals need enough time to process their experiences before caring for other families giving birth. Third, Existential issues about life and death become intensely tangible for everyone involved, and they often feel lonely and vulnerable. Healthcare professionals also reflect on the thin line between life and death and often question their performance, especially when lacking collegial and organisational support. Finally, the fusion Stigmatisation focused on how parents, siblings, and healthcare professionals experienced stigma expressed as a sense of loneliness, vulnerability, and being deviant and marginalised when a baby died before or during birth. GRADE CERQual ratings for the four fusions ranged from moderate to high confidence.

Conclusions

The profound experiences synthesised in the fusions of this meta-synthesis showed the complex impacts the birth of a baby with no signs of life had on everyone involved. These fusions can be addressed and supported by applying person-centred care to all individuals involved. Hence, grief may be facilitated for parents and siblings, and healthcare professionals may be provided with good conditions in their professional practice. Furthermore, continuing education and support to healthcare professionals may facilitate them to provide compassionate care and support to affected parents and siblings. The fusions should also be considered when implementing national recommendations, guidelines, and clinical practice.

Citation: Persson M, Hildingsson I, Hultcrantz M, Kärrman Fredriksson M, Peira N, Silverstein RA, et al. (2023) Care and support when a baby is stillborn: A systematic review and an interpretive meta-synthesis of qualitative studies in high-income countries. PLoS ONE 18(8): e0289617. https://doi.org/10.1371/journal.pone.0289617

Editor: Pracheth Raghuveer, Kasturba Medical College Mangalore, Manipal Academy of Higher Education, INDIA

Received: February 8, 2023; Accepted: July 22, 2023; Published: August 15, 2023

Copyright: © 2023 Persson et al. This is an open access article distributed under the terms of the Creative Commons Attribution License , which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Data Availability: All relevant data are within the paper and its Supporting information files.

Funding: The author(s) received no specific funding for this work.

Competing interests: The authors have declared that no competing interests exist.

From a global perspective, about one stillbirth occurs every 16 th second, resulting in nearly 2 million annually [ 1 ]. Most (84%) stillbirths occur in low- and middle-income countries. The World Health Organization (WHO) further states that the experience of having a stillborn baby is often overlooked in national and international documents and agendas despite the psychological and economic consequences for the affected women and their families [ 2 , 3 ].

The WHO defines stillbirth as “the death of a baby after 22 gestational weeks, and death occurred before or during birth”, but the WHO also recommends using the definition of stillbirth after 28 weeks of pregnancy to facilitate international comparisons. However, many high-income countries use a lower cut-off, where a baby born without signs of life after 20–22 gestational weeks is defined as stillbirth [ 3 ]. In 2021, Hug and co-authors [ 4 ] show that the global reduction of stillbirths at least 28 weeks gestational age is still slow compared to the worldwide improvement of the survival of children younger than five years, especially in Sub-Saharan countries. For example, high-income countries show a stillbirth rate of 3.0 stillbirths per 1000 total births, compared with 7.0 in upper-middle-income countries, 17.1 in lower-middle-income countries, and 22.7 in low-income countries in 2019 [ 4 ].

Several international research findings have stressed the need to reduce the stigma of stillbirth. A scoping review including 23 studies covering countries of low, middle, and high incomes show that the shared experiences of grieving parents are that their identities are confronted, and feelings of shame, guilt, and blame are common [ 5 ]. Furthermore, the healthcare provided for the best outcomes for bereaved parents is a limited field within the research field of pregnancy and childbirth. In their review, Ellis and co-authors [ 6 ] include 52 studies of various study designs. They show that parents’ experiences are often mirrored by the staff’s experiences when attending to the woman and her partner during a stillbirth. What the healthcare workers do affects the parents’ memories, and concurrently, the staff indicates several barriers to providing care as wanted. Both staff and affected parents highlight measures such as better training and continuity of care. At the same time, parents also wanted supportive measures and care pathways to improve the care for parents experiencing a stillbirth [ 6 ]. Similar findings are highlighted from studies from low and middle-income countries, where basic actions such as educational interventions to the public may reduce the societal stigma of stillbirth. Additionally, respectful, supportive healthcare providers and properly investigating the causes of stillbirth may improve the experiences of affected women and their families [ 7 ].

In summary, much of today’s evidence of care for bereaved parents includes a variety of countries and contexts where the stillbirth prevalence may vary considerably. There is limited information about what matters to parents, siblings, and healthcare providers when a baby is born without signs of life in a high-income context with a low prevalence of stillbirths internationally. High-income countries may have the financial resources in healthcare services to provide quality care and support programs after stillbirth, influencing the experiences of parents, siblings, and healthcare professionals. This systematic review and interpretive meta-synthesis aim to identify important aspects of care and support from the time a stillbirth is diagnosed through the birth and into the postpartum period from the perspectives of parents, siblings, and healthcare professionals in high-income countries.

This article is part of a larger health technologies assessment covering both qualitative and quantitative aspects of care and support after stillbirth conducted by the Swedish Agency for Health Technology Assessment and Assessment of Social Services (SBU). This extensive assessment was registered in 2022 in the PROSPERO database under registration number CRD42022306655 [ 8 ]. PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-analyses) guidance was followed throughout this review [ 9 ]. Please see S1 Checklist for the PRISMA checklist.

This paper is a systematic review and meta-synthesis covering a common term for an interpretive integration of qualitative findings. This integration method offers more than the sum of the individual data sets because it provides an innovative interpretation of the findings [ 10 – 12 ]. These interpretations are conclusions from examining all the studies in a sample as a collective group and presenting overarching interpretations not found in single studies [ 10 ]. Therefore, approximately 10 to 20 studies represent an ideal meta-synthesis [ 13 ]. This meta-synthesis followed the three processes: meta-data analysis, meta-method, and meta-theory, outlined by Paterson et al. [ 12 ]. During meta-data analysis, the researchers critically compare the descriptions of a phenomenon to reveal similarities and discrepancies among studies. In the meta-method analysis, the researchers determine the accuracy and soundness of the research methods used in each of the studies reviewed. Finally, during the meta-theory analysis, the researchers scrutinise the underlying theoretical perspectives of each study to understand how the theories directed or influenced the findings to ensure that the findings were interpreted appropriately [ 12 ].

Selection criteria

Before deciding on the criteria for inclusion, we sought input from representatives for non-profit organisations that support and advocate for families who have experienced stillbirth, as well as clinicians actively working or conducting research in the field, to ensure that our study question would be as relevant to them as possible. The following criteria were agreed on: qualitative and mixed-method studies addressing parents’, siblings or healthcare professionals’ experience of supportive care and reception from the detection and diagnosis of the stillbirth to six months postpartum were included. To be included, studies should have been conducted in high-income contexts/countries with a stillbirth rate below 5 per 1000 live births, see S1 File . Another criterion was a defined gestational age of 22 to 41 weeks at stillbirth, reflecting international definitions. Studies were excluded on the following conditions: i) addressing miscarriage, foetal anomaly, and neonatal death alone; ii) other languages than English, Danish, Norwegian, or Swedish; and iii) review articles, opinion pieces, or books.

Search strategy

An information specialist designed the search strategy, in collaboration with the review team, based on the criteria for inclusion. After that, another information specialist at SBU reviewed the search strategy using the PRESS Checklist. A comprehensive database search was performed in December 2021 and was updated in August 2022. We searched CINAHL (EBSCO), Cochrane Library (Wiley), EMBASE (Embase.com), Medline (Ovid), and PsycINFO (EBSCO). Records retrieved from multiple databases were checked for duplicates in EndNote [ 14 ]. Hand-searching related reviews and the reference lists of the included studies supplemented the database searches. In addition, a citation search was performed via Scopus (Elsevier) to identify literature that cited any of the included studies. The entire search strategy is presented as ( S2 File ).

Study selection

In pairs of two, three researchers (MP, RS, and CB) independently screened titles and abstracts using Rayyan [ 15 ], an online platform for systematic review collaboration. Two researchers (MP and CB) also independently screened the full texts according to the inclusion and exclusion criteria. Discussions within the group of authors resolved disagreements. Many of the identified articles included a mix of experiences of miscarriage, stillbirth, and early neonatal death. Still, these studies were included if the findings specific to stillbirths could be identified (e.g., citations or parts of the findings explicitly addressed stillbirth experience) and if at least 50% of participants (i.e., the majority of study participants) had experienced a stillbirth.

One of the excluded studies was a patient-led co-designed focus group study focusing on identifying improvement opportunities in healthcare. This paper was retained and later used to validate our results. Fig 1 presents the PRISMA flow diagram of the procedures [ 13 ].

• PPT PowerPoint slide
• PNG larger image
• TIFF original image

https://doi.org/10.1371/journal.pone.0289617.g001

Critical appraisal of full-text studies

After relevance screening, the methodological quality of the studies was appraised using the Primary Research Appraisal Tool-Qualitative (PRAT -Q) (12). The studies were systematically and independently evaluated on several aspects, such as study sample, research designs and methodology, data collection rigour and analysis, and theoretical frameworks by MP and CB. The appraisal of the studies followed a reading guide, which resulted in an individual assessment of each possible study considered for inclusion (yes/no/uncertain). All uncertainties or disagreements in the critical appraisals were discussed to reach a consensus. Only studies with no or few flaws unlikely to influence the study’s trustworthiness were included. Furthermore, as all eligible studies were appraised similarly, a comparative assessment across studies was possible. This procedure allowed the development of a cross-study display, summarising key features of the studies. A final sample of 16 articles for the analysis was agreed upon for the meta-synthesis.

Reflexive note

The researcher’s active role in qualitative research is well-known since the researcher is the tool for collecting, analysing, and managing their own experiences, knowledge, and assumptions. Therefore, it is essential to present the research team. The authors differed in disciplines and career stages and were all women. Before initiating the analysis, MP and CB discussed their pre-understanding of having researched and cared for parents of a stillborn child and their non-professional experiences of the studied phenomenon. This was done to make the researchers’ potential values and perspectives visible. By applying a hermeneutical analysis approach, own experiences and understandings are an advantage or a prerequisite for the analysis [ 20 ].

Data extraction and synthesis

Data were independently extracted from the included studies in duplicate by two authors (MP and CB) using a standardised data extraction form developed by CB. The data extraction form was piloted before use and used in other meta-thesis studies [ 16 – 19 ], and no changes were made to the form for this data extraction. Text related to the aim of this study was extracted, and the authors repeatedly discussed data extraction to ensure consensus and rigour.

The extracted data varied as most papers presented raw data as thematic surveys and/or direct quotations from participant interviews. We used the technique of hermeneutic appraisal to extract statements from the studies’ findings to evaluate the text’s horizon [ 20 ]. A horizon is a previous knowledge/prejudice found in the text or the person. We then interpreted these statements within the context of a guiding question: Is this about care and support when babies are born dead ? We could identify possible fusions by working inductively, using our horizons to interpret the text, and then merging the horizons of the text and the author. The authors reviewed and discussed the fusions and agreed that a deepened and broadened understanding was obtained. A fusion is when both horizons from the text and the interpreter are expanding and giving a broader and deeper understanding than just answering a question.

Confidence in synthesis findings

Once the fusions were agreed on, the level of confidence in each fusion was assessed using the Grading of Recommendations Assessment, Development, and Evaluation–Confidence in Evidence from Reviews of Qualitative Research (GRADE-CERQual) [ 21 ]. GRADE-CERQual comprises four domains that assess uncertainties in the data and include: methodological limitations, relevance, adequacy of data, and coherence. This assessment aims to evaluate and describe through a transparent procedure how much confidence there is in the findings. Therefore, an overall confidence rating of “high,” “moderate,” “low,” or “very low” was assigned to each fusion, considering each of the four GRADE CERQual components.

All fusions except Stigmatisation were judged to have a high confidence level ( Table 1 ). This is because they were based on rich material where the context was relevant, the data was rich, had high coherence, and only had minor methodological limitations. On the other hand, stigmatisation was considered moderate confidence as the data was not clearly described as stigmatisation in a few articles. Still, several aspects constituting the concept of stigmatisation were mentioned in these data, supporting the fusion to be labelled “stigmatisation”.

https://doi.org/10.1371/journal.pone.0289617.t001

Included studies

In sum, 22 020 records were identified based on the search strategy, of which 7408 were semi-automatically removed using EndNote. With the addition of 598 abstracts retrieved from citation and hand searches, 15 200 unique abstracts were screened for relevance. Three hundred ninety-two articles were retrieved and assessed for eligibility in full text. Forty-four scientific articles were considered relevant to the study question. However, 28 of those were excluded due to methodological limitations. Please see S3 File for a list of excluded articles. Sixteen articles published between 2004 and 2020 were included in the data analysis [references 22–37].

Characteristics of included studies

Table 2 shows the primary key features of included articles. Of the 16 included articles, eleven dealt with the research question from a parental perspective [ 23 – 26 , 29 – 30 , 33 – 37 ], one study had a sibling perspective [ 22 ], and four studies covered the perspectives of healthcare professionals [ 27 – 30 , 31 , 32 ]. The studies included 380 parents, preferably mothers [ 23 – 26 , 29 , 30 , 33 – 37 ], thirteen siblings aged 13 to 17 years [ 22 ], and 51 healthcare workers of various professions [ 27 – 32 ], which added to a total of 421 participants. The studies used varying qualitative analysis methods such as qualitative content analysis [ 22 , 23 , 26 , 29 , 30 , 35 , 37 ], phenomenology [ 24 , 31 – 34 , 37 ], Grounded Theory [ 25 ], and qualitative analysis according to Parse [ 28 ] and thematic analysis [ 27 ]. The studies were conducted in Ireland [ 31 , 32 ], Canada [ 26 , 28 ], Norway [ 37 ], Spain [ 24 , 29 , 30 ], the United Kingdom [ 25 , 34 ], and Sweden [ 22 , 28 , 30 , 35 , 36 ]; all developed countries with a prevalence of < 5 stillbirths in 1000 live births. The study participants shared their experiences deriving from maternal health care, antenatal clinics, and other care or support bodies. In addition, the siblings’ experiences included receiving information about their stillborn sibling and its subsequent consequences.

https://doi.org/10.1371/journal.pone.0289617.t002

Meta-synthesis

Our qualitative meta-synthesis resulted in four fusions: Personification , Respectful attitude , Existential issues , and Stigmatisation . The fusions are interdependent and interacting, thus bridging over to other fusions, therefore presented in the following order, reflecting a more profound and broadened understanding of what matters to parents, siblings, and healthcare professionals when a baby is born without signs of life. Please, see Fig 2 for an illustrative overview of the findings. In addition, the following presentation includes quotes from the studies that contributed to the creation of each fusion.

https://doi.org/10.1371/journal.pone.0289617.g002

Personification

The fusion “Personification” implied that the baby was considered unique, even in cases where the medically correct term would be a deceased foetus. The parents had experienced pregnancy and waited for their baby, who was now born without signs of life. Parents needed to meet their baby just as if the baby was alive to acknowledge and honour the baby. By personification, the parents could also care for their baby as parents do and have time to get to know their baby as a unique human being. Siblings and other family members could also see and connect to the baby if the parents wished. The stillborn baby was considered a unique individual and part of a family, thus treated as a family member. This personification also meant that the healthcare professionals treated the baby like any other baby, alive or dead, by, for example, addressing the baby by name, caring for it, or referring to it as a person.

"They spoke of him as they might speak about any of the babies . They put a diaper on him (Loss in 2008) . " [ 23 ].

Creating memories of the baby had the potential to strengthen personification. Memories of this unique individual could be created by, for example, encouraging parents to take photos of the baby on its own and with the parents and make footprints or handprints. Memories were also strengthened when parents could care for their babies for as long as they wished. Healthcare professionals could initiate and support the parents’ memory creation and enable and strengthen the parental relationship with the baby. By participating in this personification, siblings could understand that a little brother or sister was born without signs of life. Furthermore, various rituals and ceremonies could also honour the baby’s existence, empowering parents, siblings, relatives, and healthcare professionals. Thus, such creations of memories of the baby might facilitate the process of grief as the memories confirm that it is a unique person, a baby, who is being mourned.

"You’ve got to cram a lifetime of memories into a few hours . " (Interview #16) [ 25 ]. "I kept five photos that I took and with the prints of her little feet … I look at the photos , I put them away , and I am left in peace . At first , I could not bear seeing , but now it is a memory that helps me in this great pain … " (Participant 2) [ 24 ].

Personification also included that parents were confirmed and identified as parents. Their baby was born, and the baby was dead, but they are still parents and had a personified loss and grief associated with their dead baby. Their parenthood to the dead baby was lifelong but invisible—deficiencies in recognising the unique individual and the parenthood complicated parents’ and siblings’ handling of the loss.

Personification also included the health care professionals’ perspective. They were expected to be competent and supportive professionals providing care and support to parents in this challenging situation. At the same time, they were people with feelings, their own families, and life situations, which meant the case could become personal and affect their lives. Lack of understanding or resources in the organisation meant that healthcare professionals sometimes did not get the support they needed. This could lead to repressed emotions and distancing themselves from the care instead of having the courage and energy to stay present with the family.

"It puts stress on you . You don’t really feel like working afterwards , that’s for sure because it does take a lot out of you . " [ 28 ]. "At first , what I was doing was hiding . Fleeing from the anxiety , and fear and ignorance . Then , I would leave to avoid everything I could , and when it was my turn , I had no choice but to stay with the woman . I had no skills or weapons to overcome my own anxiety . I gave the woman a technically correct treatment : put in her IV , give her oxytocin , with her husband or her companion , and then , leave . Because I was unable to look the woman in the face , incapable , incapable" . (I8 — Midwife) [ 27 ].

Respectful attitude

The fusion ‘Respectful attitude showed how crucial the caregivers’ respectful attitude and care of the parents and their dead baby were for the personification and the parents’ understanding and management of their loss. But the synthesis also indicated how deficiencies in treatment and lack of information and guidance made the parents’ strained situation even harder.

Respectful attitude and care of parents and the dead baby meant that the parents were treated as parents, informed, guided, and supported as in all births, regardless of whether the baby is alive or dead. A respectful attitude also included that the baby was treated and handled with the same respect and care as a live baby by the healthcare professionals. A respectful and supportive approach also covered that the woman should have the opportunity to talk about her childbirth experience after the birth and feel pride and strength that her body was able to give birth to a baby and that she was able to handle the situation. The caregivers’ support, adapted information, and respectful guidance through the birth process and in the first meeting with the baby were essential to the parents. The parents were in unimaginable chaos after being informed that the baby was dead; hence decisions were difficult or impossible. They depended on being supported and guided by the healthcare professionals’ empathic treatment and care. This respectful attitude, treatment and care also covered that the caregivers’ compassion, respect, and empathy facilitated the meeting with the dead baby and that the parents had to say goodbye to their baby.

“The caregivers were very competent , skilled , warm , and accommodating as they instructed us and , at the same time , kind but firm . They guided us in making the right decisions . The brief moment with the baby was very poignant . I experienced my attachment to the baby as being strengthened in that moment . It is also the only time I wept over the baby’s death (father) . ” [ 36 ].

Being cared for by the same staff throughout the process meant security for the parents; they were cared for by people who knew their situation and with whom they could create a relationship.

"All parents spoke of the relationships they had with the staff who cared for them during their pregnancy and following the birth of their baby . " I could feel the kindness off her [consultant] . I knew she really cared . " (2013P2) [ 33 ].

The challenge for healthcare professionals was to provide equal care and a respectful attitude to all expectant parents, regardless of whether the baby was alive, following standard routines and treating parents and babies with warmth and respect. Thus, healthcare professionals must also be treated with respect by the organisation and colleagues when caring for parents and their stillborn babies. In addition, healthcare professionals needed time to process their experiences before caring for other birthing women and families. This respectful support was required from colleagues and the organisation.

"Because it’s so emotionally draining , sometimes you need time afterward just to regroup or talk about what you’re going through . " [ 28 ].

The disrespectful and callous treatment and care of parents or handling of the stillborn baby made the parents’ meeting with the baby more difficult, gave the parents less opportunity to be and act as parents, and complicated the process of saying goodbye to the baby and mourning it.

"The delivery was just awful from beginning to end . They almost treated me like ’The Woman With The Dead Baby’[mother’s emphasis] . There was no sympathy . When I asked to see a doctor , this particular doctor came in and said , ’we’re very busy . ’ And his exact words , I’ll never forget them ’Well , with all due respect , your baby’s dead already . ’ Which was just the most awful thing you could say . " (Interview #9) [ 25 ].

Existential issues

This fusion was characterised by a storm of emotions in which parents are thrown between hope and despair and between life and death. When the parents were informed that the baby was dead, a chaotic flood of feelings emerged—shock reactions, feelings of being thrown into an abyss, and panic wells up. At the same time, feelings of miscalculation and distrust of the information emerged and hopes for new assessments with information that the baby was alive would be presented. However, the information about the baby’s death was characterised by uncertainty; the healthcare staff whispered with each other, and the parents did not seem to receive explicit information, or they were not receptive to the provided information due to the chaotic situation. Everything was chaos in the parents’ world; what was to be a joyful event, a new life to be born, was now exchanged for death. Parents became mentally blocked and could not comprehend information or understand what was said. Life and justice were questioned. Why did an innocent baby die in the womb? Why was this happening to us? The questions were many but with no or few answers. This uncertainty caused frustration and anger versus healthcare professionals as the parents believed it was better to get relevant but scant information than no information as they needed something to hold on to. The information that the baby had died left the parents with enormous sadness. Concurrently, anger and resentment directed at life, the meaning of life, and the people around were experienced. The offence could be amplified when the nursing staff started to inform about funerals and rituals before the baby was born.

"He [husband] said ’there’s no heartbeat’ and I said , ’we’ll wait; we’ll see’ . I still continued in labour as if my baby could still be alive . I said , ’I’m not going to accept there’s no heartbeat until I see my baby’ . And then the baby came and he wasn’t alive so I had no words , just no words . " [ 33 ]. "At the time I just wanted to die , I can’t put it any other way , I just wanted to get away from the whole world , I couldn’t believe all this was happening . I was in shock , I got hysterical and kept saying I had to have a cesarean and get rid of him . I didn’t want to … but they said I had to give birth the usual way . And because of that—I didn’t understand what they meant by that—I thought they were being mean to me and wanted to hurt me . " [ 37 ].

The synthesis showed that questions of life and death became apparent to all persons in contact with the situation. First, the long-awaited baby, the new life, and the future were shattered, and then, the parents had to encounter other parents with their newborns and be noticeably reminded that their arms were empty. Even the siblings of the stillborn baby experienced chaos and uncertainty, and the thoughts of life and death became palpable. Furthermore, the synthesis showed that the siblings’ existential thoughts about the meaning of life and what mattered in life and death raised questions about their future. At the same time, the siblings might miss talking about their existential concerns as their parents were grieving.

"As I held Ruby , I promised her that dying wouldn’t be in vain . That I would do everything I could to try and stop somebody else going through the same sort of pain . " (Interview #15) , [ 25 ]. "In the scope of things , death is a part of life . And it is a part of a natural occurrence in life . It just happens; it is out of our control . I think it makes you more aware of life and death and the meaning of it . And to value life more than you would otherwise" (girl aged 15) ,[ 22 ].

As birth was regarded as a predominantly joyful event—not grief and death, it contributed to a duality among the nurses and midwives, which made caring for a stillbirth difficult. Balancing being a professional healthcare worker and an empathic fellow human being could be straining. Caring for parents with a stillborn baby was regarded as a great responsibility, but it could also be a great honour to share a unique birth and have time afterwards with the family. The caregivers could express gratitude for contributing and assisting in making the birth bearable by providing empathetic care and support. Among healthcare professionals, existential issues arose, and a great awareness of the fragile dividing line between life and death was expressed. This awareness of the delicate line between life and death also aroused gratitude and joy for their family and life. However, attending to a stillbirth could be so painful and emotionally intense that healthcare professionals avoid being involved and engaged. The emotions could overwhelm them, and the whole situation lacked meaning, which affected their work and private lives and could result in questioning one’s professional ability and place in life.

"I find it difficult to separate the job at work … and coming home and just switching off . " [ 32 ].

There were also concerns and uncertainty among healthcare professionals about what awaits parents afterwards; how should parents deal with the grief they experienced and what would wait for them after returning home from the maternity unit? Would there be any follow-up or support for the grieving parents? The nursing staff expressed that they thought about how the grieving parents could prepare to meet other pregnant women or parents with infants daily without being strongly reminded of their loss.

Stigmatisation

This fusion presents how parents, siblings, and caregivers experienced stigma expressed as a sense of loneliness, vulnerability, and being deviant and marginalised when the baby dies before birth.

The parents felt alienated and alone, which started when they were informed about the death of their baby. This alienation continued throughout the birth process and after the baby was born. They have become parents, but their parenthood was invisible to others and society. They left the hospital with empty arms. Encountering parents with live infants reminded them of their loss and their deviation from the societal norm and other parents.

Siblings experienced a similar stigma, being alone in their grief, feeling odd and sometimes misunderstood among their peers. Additionally, they could feel lonely being part of the family as the parents were more absent due to their grief.

“I cannot explain , but really , I thought it was terribly difficult to leave . And then we went home . And it was unbearable to leave the delivery room with only our bags and nothing else . ” (Mother to a baby born in week 35) ,[ 29 ].

Caregivers might also experience loneliness and vulnerability when caring for parents with a stillborn baby. When collegial or organisational support was lacking, the strain was immense on each healthcare provider.

"There was no recognition that it might be difficult … there was no training … there was no debriefing … and I think that’s bad . You did it yourself … nobody cared if you got so psychiatrically disturbed you threw yourself off the roof the following week . " [ 32 ].

Furthermore, feelings of guilt and shame appeared for the parents. These feelings originated from thinking that something they did might have contributed to the death or that they missed warning signs. It also happened that women might feel shame as their bodies failed to give birth to a live baby. Also, siblings could feel guilt when they had mixed feelings about having a new sibling, but also feelings of guilt when they wanted to live "a normal" life despite themselves and the family grieving. Healthcare professionals could feel guilty when there were suspicions that something significant in care or treatment had been missed or mistreated. These feelings were often related to thoughts about responsibility and what responsibility came with the profession.

Furthermore, society’s incomprehension of parental loss further contributed to the stigmatisation of parents. Being met with belittling of one’s grief from relatives, friends, healthcare professionals, or others in the society or having to hear that "it was just a foetus" or "you can have new babies" was traumatising and marginalised parental loss and grief. The extent and meaning of grief and loss were preferably shared with other parents with similar experiences.

"Society does not understand it , it measures the pain by the size of the coffin , and you have to listen to people say things like : "It’s okay , you’ll have more children" … it was my daughter , and she’s not going to come back , and I’ve lost her … " (Participant 2) .[ 24 ].

Validation of the findings using a patient-led co-designed focus group study

To triangulate and validate our findings, we used the results of a patient-led co-designed focus group study, where eleven participants had sessions talking about their experiences of care when their babies were born dead. Later, they reflected on the findings that would form the basis for designing practical recommendations that facilitate grief processing. After these sessions, the participants listed 15 recommendations to facilitate the parents’ grieving process when affected by stillbirth [ 38 ]. These recommendations mirrored the experiences presented in our meta-synthesis to a large extent. Moreover, since the Canadian study [ 38 ] was conducted with another study focus and other study participants in a high-income context, their findings and recommendations can be suitable to compare with the results of meta-synthesis (triangulation).

Gillis and co-authors [ 38 ] summarise the parents’ experiences with the words "Stillbirth, still life,"; emphasising the co-occurring parental experiences of life and death when babies are born dead. Thus, parents need the opportunity and support to handle their experiences and grief. Our fusion "Personification" corresponded with the study findings of the parents’ stories and is reflected in the recommendations provided. The study stresses the necessity that the baby is honoured as an individual and recognised as a baby. Parents also need to be supported to create memories of its existence in the brief time they share with their dead baby. By honouring the baby’s existence, parents also receive a confirmation of their parenting, which coincides with the fusion of “Personification” in our meta-synthesis. Furthermore, the study [ 38 ] describes how parents are left in silence and need the empathetic treatment of healthcare professionals both in acute and long-term situations. The parents are completely unprepared for the baby’s death and require time, support, and empathic guidance through all the decisions that must be taken. Suppose the health care professionals can respectfully guide and show possibilities and secure that parents are given time with their baby; in such case, the parents may not regret taking advantage of the opportunities presented to care for their baby [ 38 ]. The study also shows that the parents need to feel like a parent in connection with their baby’s birth (i.e., to receive the same treatment as if the baby were alive). Simultaneously as the parents want to be treated like other expectant parents at birth, it also emphasises the need for competent and respectful nursing staff and not having to meet other parents whose babies are alive. The need for support goes beyond the acute stage, and both parents’ well-being is essential to consider. Attitude is crucial for parents and their handling of grief, and they clearly remember what is said and done from the time the baby’s death is confirmed and on throughout the entire period of care and follow-up. These experiences and recommendations provided by Gillis et al. [ 38 ] fit well with the fusion of "Respectful attitude" in our findings.

Life and death exist collaterally, hence raise existential issues can also be seen in the parents’ narratives at Gillis et al. [ 38 ]. Parents are forced to deal with the death of their waited baby, but at the same time, the death can also raise hope about the future, but the parents have been fundamentally changed by their experience. In Gillis’ study, the parents express the situation as their existence being "mid-between" during a period, which corresponds to the synthesis’ interpretation of falling into chaos and needing a respectful attitude and the opportunity to express existential issues.

Our fusion "Stigmatisation" resembles the experiences described by Gillis et al. [ 38 ], who show that the parents talk about how they internalise guilt and shame. Their loss is made invisible by family and friends who cannot face the grief the parents show. They also think that they are treated differently than before. Even healthcare professionals and society contribute to reducing and making the parent’s loss invisible because stillbirth has no place in society. A stillborn baby becomes the "secret" of parents, which they do not share with others [ 38 ].

Based on the meta-synthesis of 16 studies [ref 22–37], four fusions related to care and attitude were identified and interpreted. The four fusions identified spanned the countries, the people involved, and the period in the included papers. These vital aspects of care and attitude were characterised by the personification of the baby, its parents and siblings, and the involved healthcare professionals. A respectful attitude was essential to enable personification. Furthermore, the birth of a baby without signs of life raised several existential questions. It also implied stigmatisation and loneliness as the stillborn babies were perceived as a silenced and concealed event in society.

Other studies have stressed what mothers and fathers of stillborn babies find essential when interacting with healthcare professionals, as shown in a meta-synthesis of parents’ experiences of perinatal loss [ 39 ]. Like our findings, the study indicates that perinatal loss is transformative, whereas the parents experience multiple losses and complex emotions. The parents experience evolving relationships with their baby, which they consider a person with an identity. Healthcare professionals are an integral part of the parents’ pregnancy experience and perinatal loss experience, which shows healthcare professionals’ unique position to influence parents’ overall experiences. The parents feel that their loss makes them traverse the social sphere in isolation and that people around ignore their pregnancy and loss [ 39 ]. Further, acknowledgement of the parenthood and parental grief, an understanding of the stillbirth’s trauma, and support are emphasised even for a more extended period. Finally, the findings show how perinatal loss sets perinatal grief apart from other forms of suffering, such as grieving the loss of a parent or spouse [ 39 ], i.e., findings that align with several aspects of our fusions.

A previous review by Burden et al. [ 40 ] included 144 studies with a global approach. They show that despite country and parental gender, stillbirth experience may devastate physical and psychological health, affecting social costs and relationships. Furthermore, it influences the subsequently born children. But this review also shows that the traumatic experience may have positive effects in the form of personal growth through the development of resilience, new skills, and capacities, which resemble the existential issues described in our meta-synthesis.

Very few studies have applied the sibling’s perspectives; the present studies often cover parents’ need for support to siblings. Avelin et al. [ 41 ] show that it is straining for parents to try to balance their grief and simultaneously function as parents and manage everyday life. Parents try and want to include siblings in the memory creating of the dead sibling, but they need guidance and support from healthcare professionals. In a recent review [ 42 ], 25 studies were included in an integrative review of bereavement in siblings, with most focusing on loss due to the medical illness of the sibling. They conclude that the loss of a sibling is a significant childhood event with an increased risk of negative impact on physical and psychological health. In the Lancet Series of Stillbirth, Goldenberg et al. [ 43 ] mentioned bereavement support as a key action for families and communities. Still, more than a decade later, guidance and support to siblings and parents may be inadequate. Furthermore, bereavement support for children and teenagers must be age-appropriate [ 44 ].

Ellis et al. [ 6 ] show in their review that parents’ and healthcare professionals’ experiences in stillbirth often mirror each other. For example, parents describe how healthcare professionals hide behind tasks and routines, while staff indicates that distancing themselves from parents is their coping strategy to cope with the situation. Our findings also revealed how healthcare professionals were emotionally affected by caring for parents having a stillborn baby. Perinatal loss has a psychological impact on healthcare providers, as shown in a systematic review [ 45 ], including 20 studies focusing on the psychological impact of perinatal loss on healthcare providers. The findings show that the providers are forced to handle a variety of feelings that may, in the long term, lead to, for example, acute stress and depression. A Swedish survey [ 46 ] shows that several obstetricians and midwives report partial symptoms of post-traumatic stress syndrome after experiencing traumatic birth events, such as stillbirth, perinatal loss, etc. Staff who exhibit these symptoms often leave obstetric care and switch to work in primary care more often than their colleagues without these symptoms. Therefore, the authors conclude that targeted professional training is necessary to support healthcare professionals in managing their strains and stress. Providing empathic and quality care to parents having a stillborn child requires training and continued education. The lack of training for healthcare professionals and obstetricians has been highlighted in previous studies [ 47 , 48 ]. A recently published study [ 49 ] shows that applying drama techniques in obstetrician workshop training significantly improves confidence and communication skills in breaking bad news and caring for families experiencing stillbirth. Furthermore, the workshop training also enhances the ability to recognise own emotional reactions and support colleagues. Hence, there are reasons to emphasise that training and continued education can be essential clinical interventions to aid healthcare workers in providing quality care to parents giving birth to a stillborn baby.

Our fusion, Personification was the crucial element to perceived care and support, declares that person-centred care could be a strategy to improve care and support at stillbirths further as it focuses on the aspects of care, support, and treatment that matter most to the patient/parents, their family, and carers. The patients, in our case, parents experiencing the unexpected loss of their baby, are not just a set of diagnoses or symptoms; instead, they are humans with emotions and social and practical needs. Person-centred care is vital for patients, but the benefits of the person-centred practice can only be fulfilled if it includes the personhood and well-being of healthcare professionals [ 50 , 51 ].

Strengths and limitations

A strength of the study is the systematic and transparent approach to the procedure and analysis, which strengthens credibility. The eligible studies were systematically and independently appraised using a Swedish PRAT-Q based on the international PRAT [ 12 ], entailing the design, selection, data collection, analysis, ethics, and theoretical frame of reference before inclusion. This appraisal also implied that only studies of no or few methodological flaws were included in the analysis. Any differences in the reviewers’ assessments were discussed until consensus was established, further strengthening the result’s relevance and reliability. Through interpretation, we performed a meta-synthesis to lift research results/findings to the next level and not only repeat what we already knew with some certainty [ 7 ]. The meta-synthesis used in this study was grounded in hermeneutic theories [ 12 , 20 ]. In this way, we clarified the experience of the parents, siblings, and healthcare professionals when there is a stillbirth. To triangulate and validate our fusions, we used the results of a patient-led co-designed focus group study [ 38 ]. This triangulation and validation showed excellent agreement between the study findings and those in the meta-synthesis.

Some limitations in this study need addressing. Fewer studies examining the experiences of care and support from the perspectives of siblings and caregivers are available, hence in the minority of the included studies. Four of the 16 included studies deal with healthcare professionals’ experiences. Only one study covers the siblings’ experiences. Therefore, caregivers’ and siblings’ experiences may not be fully explored. Furthermore, children and young people may have different experiences and needs than their parents. Still, the siblings’ experiences contributed to three of four fusions as the article contained rich information and descriptions of their experiences. Furthermore, few studies deal with the experience of health professionals in assisting parents when children are born dead. Also, in this case, the included studies contributed rich and informative data to the synthesis. Hence, the articles of siblings and health professionals significantly contributed to each fusion despite being outnumbered by parental experiences. However, the rich data from the included articles imply that the experiences of parents, siblings and caregivers may be transferable to other high-income settings. This is also supported by the triangulation and validation performed.

Another noted limitation is that the included articles have not indicated the extent to which minority groups have been represented. For example, it appears that foreign-born are included in some of the studies, but without stating whether they spoke the native language of the country or if the interview was conducted in another language or through an interpreter. Speaking the language of the country was not mentioned as an inclusion criterion for studies, which may seem relevant if interviews or surveys are conducted. When minority groups are excluded from participating in studies due to language difficulties, these groups’ critical perspectives and experiences are lost. This means that research may risk only highlighting the experiences of the majority population or among immigrant women who have been in the country long enough to master the country’s language. Thus, excluding minority groups may limit the transferability of our findings to the majority populations in high-income countries.

Implications for research and practice and directions for future research

The meta-synthesis showed that it is essential that personification works well. In the case of the stillborn baby, the baby is treated with respect as a unique individual born into a family. The siblings should be involved in care to create a relationship with the child based on their stage of development. According to the personification perspective, parents are confirmed and respected in their parenting, and other relatives and friends are welcomed to the hospital to see the baby and family. Early involvement of others may be beneficial for continued support for the family.

The meta-synthesis findings also show that healthcare professionals’ need for support from colleagues and organisations must be respected. For example, they are given good conditions to focus on caring for the parents having a stillborn baby and not having to care for other patients simultaneously. Continuity with the same staff during the care period must be sought, both in the short and long term. The organisation needs to ensure that there is time and resources for reflecting conversations for the healthcare workers when they have been involved in caring for these families, in processing their feelings and getting support from the organisation in their work.

Regular training of health professionals is also vital to make them feel safe and confident in assisting the needs of parents in connection with birth. Furthermore, individual care plans can be drawn up in consultation with the parents so that the support can be designed according to the needs of the particular family regardless of, e.g., ethnic origin or belief, i.e., applying person-centred care of the family.

The existential questions of life and death arising in parents and healthcare professionals can also be essential to address and dare to respond to, as they affect the treatment of the family, parents and siblings who have lost a baby/siblings. Greater openness in society could lead to parents receiving improved societal care and support and diminish the stigma of having a stillborn baby.

Future qualitative research should endeavour to care for and support parents from different socio-economic and demographic groups to expand the current homogeneous understanding of the importance of care and support in stillbirth. These studies would include men, same-sex partners and participants from minority groups. There is also a lack of studies on care and support for siblings to a stillborn baby, as well as reflections on care and support for healthcare professionals in these situations.

With increased qualitative knowledge, the creation of relevant and adapted support measures based on theory and previous research aimed at all involved and touched when a baby is born dead. These support measures should then be systematically evaluated.

The profound experiences synthesised in the fusions of this meta-synthesis showed the complex impacts the birth of a baby with no signs of life had on everyone involved. These four fusions, Personification , Respectful attitude , Existential issues , and Stigmatisation can be addressed and supported by applying person-centred care to all individuals involved. Hence, grief may be facilitated for parents and siblings, and healthcare professionals may be provided with good conditions in their professional practice. Furthermore, continuing education and support to healthcare professionals may facilitate them to provide compassionate and person-centred care and support to affected parents and siblings. The fusions should also be considered when implementing national recommendations, guidelines, and clinical practice.

Supporting information

S1 checklist. prisma checklist..

https://doi.org/10.1371/journal.pone.0289617.s001

S1 File. Developed countries.

https://doi.org/10.1371/journal.pone.0289617.s002

S2 File. Search strategies for qualitative articles.

https://doi.org/10.1371/journal.pone.0289617.s003

S3 File. Excluded studies.

https://doi.org/10.1371/journal.pone.0289617.s004

Acknowledgments

The authors express many thanks to Anna Attergren Granath at SBU for administrative support during the study.

• View Article
• PubMed/NCBI
• Google Scholar
• 2. World Health Organization (WHO). Stillbirth accessed December 2022, https://www.who.int/health-topics/stillbirth#tab=tab_1
• 3. Report of the UN Inter-agency Group for Child Mortality Estimation. A Neglected Tragedy The global burden of stillbirths. A neglected tragedy: The global burden of stillbirths 2020 | UNICEF 2020
• 8. Hultcrantz M. Care for parents and siblings of stillborn children. National Institute for Health and Care Research (NIHR); 2022. PROSPERO https://www.crd.york.ac.uk/prospero/display_record.php?ID=CRD42022306655
• 12. Paterson B, Thorne S, Canam C, Jillings C. Meta-Study of Qualitative Health Research. Thousand Oaks, California; 2001. https://methods.sagepub.com/book/meta-study-of-qualitative-healthresearch
• 20. Gadamer HG. Truth and Method. Shed &Wards: London, UK; 1975/2001.
• 51. McCormack B., & McCance T. Person-centred practice in nursing and health care : Theory and practice . John Wiley & Sons, Incorporated. 2017. http://ebookcentral.proquest.com/lib/uow/detail.action?docID=4605105

• Open access
• Published: 14 May 2024

Socio-cultural beliefs and perceptions influencing diagnosis and treatment of breast cancer among women in Ghana: a systematic review

• Agani Afaya 1 , 2 ,
• Emmanuel Anongeba Anaba 3 ,
• Victoria Bam 4 ,
• Richard Adongo Afaya 5 ,
• Ahmed-Rufai Yahaya 6 ,
• Abdul-Aziz Seidu 7 &
• Bright Opoku Ahinkorah 8  

BMC Women's Health volume  24 , Article number:  288 ( 2024 ) Cite this article

161 Accesses

Metrics details

Breast cancer is currently the most commonly diagnosed cancer in Ghana and the leading cause of cancer mortality among women. Few published empirical evidence exist on cultural beliefs and perceptions about breast cancer diagnosis and treatment in Ghana. This systematic review sought to map evidence on the socio-cultural beliefs and perceptions influencing the diagnosis and treatment of breast cancer among Ghanaian women.

This review was conducted following the methodological guideline of Joanna Briggs Institute and reported in accordance with the Preferred Reporting Items for Systematic reviews and Meta-Analyses. The literature search was conducted in PubMed, CINAHL via EBSCO host , PsycINFO, Web of Science, and Embase. Studies that were conducted on cultural, religious, and spiritual beliefs were included. The included studies were screened by title, abstract, and full text by three reviewers. Data were charted and results were presented in a narrative synthesis form.

After the title, abstract, and full-text screening, 15 studies were included. Three categories were identified after the synthesis of the charted data. The categories included: cultural, religious and spiritual beliefs and misconceptions about breast cancer. The cultural beliefs included ancestral punishment and curses from the gods for wrongdoing leading to breast cancer. Spiritual beliefs about breast cancer were attributed to spiritual or supernatural forces. People had the religious belief that breast cancer is a test from God and they resorted to prayers for healing. Some women perceived that breast cancer is caused by spider bites, heredity, extreme stress, trauma, infections, diet, or lifestyle.

This study adduces evidence of the socio-cultural beliefs that impact on the diagnosis and treatment of breast cancer among women in Ghana. Taking into consideration the diverse cultural and traditional beliefs about breast cancer diagnosis and treatment, there is a compelling need to intensify nationwide public education on breast cancer to clarify the myths and misconceptions about the disease. We recommend the need to incorporate socio-cultural factors influencing breast cancer diagnosis and treatment into breast cancer awareness programs, education, and interventions in Ghana.

Peer Review reports

Introduction

Breast cancer is a global public health concern due to its increasing incidence coupled with the high mortality rate among women in low- and high-income countries [ 1 ]. In 2020, it was estimated that 2.3 million breast cancer cases were newly diagnosed with approximately 685,000 deaths globally [ 1 ]. In Ghana, breast cancer is the most commonly diagnosed cancer and the leading cause of cancer mortality among women [ 2 ]. In 2020, breast cancer accounted for approximately 31.8% of all cancer cases in Ghana [ 3 ].

Evidence shows that cultural factors such as conceptualizations of health, illness, beliefs, and values influence breast cancer screening among women in certain populations [ 4 , 5 , 6 ]. Breast cancer screening is reported to be relatively low among women living in Ghana. A nationwide study revealed that only 4.5% of Ghanaian women aged 50 years and older had undergone mammography screening [ 7 ]. The low levels of breast cancer screening lead to undetected breast cancer symptoms, contributing to the late-stage diagnosis of breast cancer and subsequent poorer outcomes and mortality [ 8 ]. There have been low levels of awareness and knowledge about breast cancer among women in Ghana [ 9 ]. Also, there is a lack of understanding of the perceptions and beliefs toward breast cancer diagnosis and treatment in Ghana.

Culture is considered a multidimensional set of shared beliefs and socially transmitted ideologies about the world, which are passed on from generation to generation [ 10 , 11 ]. Cultural beliefs within certain communities across the globe are considered a determinant of health risk perceptions and behaviors in promoting or seeking health care in diverse populations [ 12 ]. In traditional Ghanaian communities, good health is recognized as a suitable relationship between the living and the dead and being in harmony with the individuals’ environment. Thus, disease is conceptualized as a malfunctioning of the body system which is probably due to a lack of harmony with supernatural/ancestral forces [ 13 ]. This belief influences how diseases are treated and the steps taken to manage the disease and ultimately how the disease is experienced [ 13 , 14 ]. Cultural beliefs connected to breast cancer are among the key determinants in women’s decision-making regarding breast cancer screening practices in traditional societies [ 14 , 15 ]. In most Ghanaian communities, breast cancer is believed to be associated with supernatural powers, hence, women seek alternative treatments (healing/prayer camps) first and only report to health facilities in advanced stages of breast cancer [ 16 ].

It is therefore important to consider how socio-cultural factors impact breast cancer diagnosis and treatment because these factors influence cancer care in resource-limited settings. To the best of our knowledge, no review has been conducted in Ghana specifically to address the cultural, religious, and spiritual beliefs influencing timely diagnosis and treatment of breast cancer among women. To fill this gap, this systematic review sought to map evidence on the cultural beliefs and perceptions that influence the timely diagnosis and treatment of breast cancer among women.

This systematic review was conducted following the updated methodological guideline of Joanna Briggs Institute (JBI) [ 17 , 18 ] and reported in accordance with the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) statement. The updated JBI methodological guidance regarding conducting a mixed methods systematic review recommends that reviewers use a convergent approach to synthesize and integrate both qualitative and quantitative studies [ 18 ]. Therefore, using a mixed methods systematic review involving both quantitative and qualitative studies was deemed the most appropriate study design because this is the first evidence synthesis on the cultural, religious, and spiritual beliefs that influence breast cancer diagnosis and treatment in Ghana.

Inclusion and exclusion criteria

Studies conducted among women and explored the cultural beliefs and perceptions about breast cancer were included.

Studies that were only limited to Ghanaian communities were included.

Empirical studies published in peer-review journals.

Observational studies, using qualitative and/or quantitative methods were also included.

The exclusion criteria involved review studies, conference papers, editorials and abstracts.

Studies published before 2012 were also excluded.

Search strategy

This review adopted the triple-step search strategy proposed by the JBI for all types of reviews [ 19 ]. The first step involved an initial limited search in PubMed for already existing published research articles on sociocultural beliefs and perceptions about breast cancer in Ghana. The initial limited search ensured the identification of relevant keywords used in developing the preliminary search terms. Step two involved a formal search after finalizing and combining the following keywords (‘breast cancer’, ‘cultural beliefs’, ‘religious beliefs’, ‘traditional beliefs’, ‘perception’, and ‘Ghana’) using Boolean operators. A comprehensive search was conducted in PubMed, CINAHL via EBSCO host , PsycINFO, Web of Science, and Embase from 2012–2022. The final step involved manual tracing of the reference list of studies for additional studies. This was done up to the point of saturation where no new information emanated from the subsequent manual search of articles.

Study selection

Following the searches, the identified records were exported into EndNote 2020 reference manager for duplicate removal. After the duplicate removal, the reviewers ensured consistency in screening through the following process: (1) joint screening by two reviewers was conducted until they felt confident to start independent screening, (2) independent blinded screening of titles/abstracts followed by a meeting and discussion of discrepancies and (3) repetition of step 2 until an acceptable agreement was met. Following the screening of the titles/abstracts, full-text review was conducted following a two-step process. The first step involved two reviewers who screened all the articles identified after the title/abstract screening. Thereafter, two independent reviewers assessed the full-text articles for inclusion or exclusion. In the course of the full-text screening, any disagreements that emerged were discussed for consensus. Throughout the screening of the abstracts, full-texts, and data extraction, the reviewers regularly met to discuss and solve emerging issues.

Data extraction

A data extraction form was developed in line with the aim of this review. Two authors independently extracted the relevant information from the included articles. The following information was extracted from the articles: first author’s name, year of publication, study location, study type, aim, study population, and key findings. Disagreements during the data extraction process were resolved by a discussion and where a resolution was not reachable, the last author resolved it through further adjudication. Study selection and data extraction were conducted manually.

Data analysis

A convergent integrated approach [ 20 ] was employed to transform the data into narrative form because the extracted information was from quantitative and qualitative studies. The analysis followed JBI recommendation where we qualitized quantitative data for data transformation because this is less prone to error when codified than when qualitative data is given numerical values. Qualitizing entails taking data from quantitative studies, translating or converting it into textual descriptions so that it can be integrated with qualitative data, and providing a narrative interpretation of the quantitative results [ 18 ]. Following the convergent synthesis of the transformed data, the reviewers undertook repeated, detailed examination of the assembled data to identify categories on the basis of similarity in meaning [ 18 ]. Out of these, three categories were derived from the analysis.

Assessment of methodological quality

Using the Mixed Methods Appraisal Tool (MMAT)  version 2018, two researchers (AA and RAA) evaluated each included study’s quality separately [ 21 ]. After discussing disagreements between the two reviewers (AA and RAA), BOA helped to forge a consensus. Methodological quality standards for evaluating research using mixed methodologies, quantitative, and qualitative approaches are included in the MMAT. The MMAT assesses the suitability of the research objective, study design, technique, participant recruitment, data collection, data analysis, results presentation, author comments, and conclusions. Hong et al. [ 21 ] discourages the overall quality scoring of the included studies, therefore, the methodological quality of the studies was evaluated using the recommended guidelines.

Flow Chart of evidence selection

Literature search

Our search yielded a total of 176 records from the electronic databases. After duplicates were automatically removed through the EndNote ( n  = 76), 100 records were reviewed independently by two authors based on the title and abstract. Records that did not meet the inclusion ( n  = 75) were removed after holding discussions to identify discrepancies in the review process. Thereafter, full texts of the remaining 25 articles were assessed for eligibility. Hand-search of the included study references yielded no results. In total, we included 15 studies [ 22 , 23 , 24 , 25 , 26 , 27 , 28 , 29 , 30 , 31 , 32 , 33 , 34 , 35 , 36 ]. The article selection process is shown in the PRISMA flow diagram (Fig.  1 ).

Characteristics of the included studies and quality

The majority of the studies [ 22 , 23 , 24 , 26 , 27 , 28 , 29 , 30 , 31 , 32 , 33 , 34 , 35 ] were conducted in the southern part of Ghana where there are better health infrastructures compared to the northern part of Ghana. Eight of the included studies were qualitative while the rest employed quantitative study designs. The summary of the characteristics of the 15 studies is shown in Table  1 . The appraisal of the included studies was assessed using the MMAT. All the studies were included, and none were excluded due to poor methodological quality. All 15 studies met the screening criteria and provided clear research questions. The studies included clearly stated and described research design, and target population, and used appropriate measurements.

Cultural beliefs

Breast cancer is believed by some sections of Ghanaians to be a curse or a punishment from the lesser gods for sins committed by the individual [ 22 ]. Some women believed that an extra-marital immoral lifestyle provokes God’s retribution for breast cancer development [ 29 ]. Some people believed that it is an ancestral punishment for the woman’s refusal to give birth in order to continue the ancestral lineage [ 23 ] and because of this, they are given spiritual babies to suckle the breast which then causes cancer [ 23 ]. It is also believed some women have been pronounced cursed due to some wrongdoings [ 25 ]. Due to the cultural belief, some women prayed to their ancestors so that traditional medicine will heal them of the breast cancer [ 26 ].

“…when it started, my uncles came to my aid, they took me to the village to see a “Tim Lana” (referring to a traditional healer). He was very good. He told me everything about my problem. So, there was no need for visiting the hospital…” [ 36 ].

Spiritual and religious beliefs

Some studies in Greater Accra, Tamale, and Kumasi indicated that breast cancer was a spiritual attack from humans or family members that sought to kill them while some believe it emanated from evil forces [ 29 , 31 , 36 ]. Participants in some studies indicated that breast cancer is attributed to some spiritual or supernatural forces [ 32 , 33 , 36 ] and can only be cured through spiritual means [ 33 ]. Due to the spiritual beliefs, some women went to traditional healers for treatment [ 26 , 36 ]. A study in the northern part of Ghana revealed that women who suffer from breast cancer are witches and have used their breasts for ritual purposes [ 25 ] while in the southern part of Ghana some participants believed that breast cancer is caused by witches [ 22 ]. For example, a narration from a participant stated:

“I believe my condition is spiritual and I realized it is coming from my mother’s side” [ 31 ].

“The problem is that my disease is a spiritual attack, so it has to be treated spiritually; the hospital drugs cannot get this out of me…” [ 36 ].

Some studies in the southern and northern part of Ghana stated that participants had a religious belief that the disease was a test from God and resulted in prayers for healing [ 31 , 36 ] and also believed that God had the supernatural powers to miraculously melt the breast lump [ 29 , 32 ] and completely cure them [ 32 ]. Some women also believed that it was their fate to get breast cancer [ 36 ]. Due to these religious beliefs some women had to resort to prayer camps for healing which leads to delay in diagnosis and treatment of breast cancer [ 26 ].

Misconceptions about breast cancer

Some women perceived that breast cancer is caused by spider bites [ 24 ], heredity, extreme stress [ 22 , 32 ], trauma, infections [ 22 ], diet, or lifestyle [ 22 , 35 ]. Some perceived risk factors of breast cancer as stated by some women included non-breastfeeding women, obesity, or overweight [ 25 , 30 , 33 ], and contraceptive use [ 30 ]. Some women had the perception that male health practitioners would not be allowed to examine or see their breasts while some preferred male doctors to examine their breasts [ 27 ]. A study in Accra conducted among female nonmedical students revealed that suckling the breast by a male caused breast cancer [ 28 ]. It is also perceived that putting money in the brassieres could be a possible cause of breast cancer among females [ 23 , 35 ]. A study by Iddrisu et al. [ 31 ] and Agbokey [ 23 ] revealed that breast cancer is a disgraceful disease, dangerous, and a fast killer. Some people also believed that breast cancer can be cured [ 27 , 32 ] by herbal treatment or medicine [ 25 ] while some believed that it is not curable [ 27 ]. Some people also believed that breast cancer was contagious and transmissible and avoided sharing equipment with breast cancer survivors [ 31 ]. A breast cancer survivor narrated:

“…my mum believes the disease can be transmitted so she does not allow me to eat with my son. I have separate bowls, spoons, and cups from that of the family…” [ 31 ].

This study reviews the existing literature on socio-cultural beliefs influencing the timely diagnosis and treatment of breast cancer among women, and this revealed diverse cultural, spiritual, and religious beliefs across the regions of Ghana. The current findings emphasize critical issues that lead to misguidance and share ignorance about breast cancer and its treatment among a section of Ghanaian communities which is rooted in their personal beliefs. Cultural beliefs are key in the decision-making process for the treatment of ailments depending on their knowledge level about the condition. This could probably lead to making the right decision or the wrong treatment decision. The diverse cultural, spiritual, and religious beliefs about breast cancer could affect the health seeking behavior of women diagnosed with breast cancer within the Ghanaian communities.

Consistent with a systematic review findings [ 13 ] it is believed that breast cancer emanates as a result of supernatural forces, curses, and punishment from lesser gods/ancestors for wrongdoings. Though not all Africans hold this traditional belief in ancestral spirits, some believe that health and illness are in the hands of a higher power such as God or Allah [ 13 ]. Hence, in most African communities it is common practice to seek traditional medicine for the treatment of diseases which is in line with their beliefs [ 37 ]. Due to the cultural/traditional belief systems and practices, most women report to health facilities with advanced stages of breast cancer which adversely impacts the breast cancer diagnosis and treatment [ 36 ]. Most women resort to traditional or spiritual healing because this method of treatment combines body, soul, and spirit. In some African settings, traditional healers are trusted to treat diseases including cancer because women believe they look for both scientific and metaphysical causes of the disease. It is possible that breast cancer patients who combine both traditional and modern methods of treatment may experience treatment interference. This dual approach can impact treatment effectiveness and lead to adverse effects or complications. The provision of culturally sensitive care by recognizing unique cultural, religious, and social beliefs and practices is of paramount importance for early detection and treatment of breast cancer among women [ 38 , 39 , 40 ]. Globally, women’s cultural beliefs and perceptions towards breast cancer should be examined to optimize timely breast cancer diagnosis and treatment.

Religious fanaticism coupled with lack of knowledge about the disease condition could impede the utilization of medical treatment, especially when religious beliefs impact negatively on people’s health-seeking behaviors [ 36 ]. A study in Nigeria revealed that religious beliefs about breast cancer were observed to be a barrier to breast cancer screening among women [ 41 ]. This review found that some women in the southern part of Ghana believed that breast cancer was a test from God and resorted to prayers because they believed that God had supernatural powers to heal them from the disease. Though religious beliefs are considered to be a source of spiritual strength and help people to cope with the disease, the religious misconceptions, and mistaken beliefs are thought to contribute to delayed heath-seeking attitudes and lack of breast cancer screening among women [ 42 ]. In the current review, it was reported that some women stayed in prayer camps for almost one year seeking healing and later reported to health facilities with advanced breast cancer which has dire consequences on the survival rate of women. Efforts to sensitize women and religious leaders about the early presentation of breast disease to health facilities for diagnosis and treatment would be key to reduce the number of breast cancer cases detained in religious camps. It is also imperative for religious bodies to discuss health related issues including breast cancer to create much awareness about the condition.

This review identified varied perceptions of breast cancer where breast cancer has been attributed to spider bites and putting money in the brassieres among others. Some believed that breast cancer was a contagious and transmissible disease. These findings show poor knowledge level among women concerning breast cancer. Even though in this review most women had heard or were aware of breast cancer, the varied perceptions about breast cancer suggests low knowledge level of breast cancer. The low knowledge level of breast cancer among women have been associated with late presentation of breast cancer to health facilities [ 40 ]. Women presenting to health facilities with advanced stage breast cancer have been associated with low survival rate in the African region as compared to high income countries [ 43 ]. A study conducted in Ghana revealed that the breast cancer survival rate among women was below 50% which was probably due to late presentation and lack of breast cancer screening [ 44 ]. We recommend intensification of public health education campaigns on breast cancer in order to improve women’s knowledge of the disease which will subsequently enhance early presentation, diagnosis, and treatment.

Implication for policy and practice

Metaphors such as spider bites, supernatural forces, witchcraft, and many other beliefs are associated with breast cancer in Ghana which impact the understanding of the disease and whether or not to seek medical treatment. Therefore, culturally sensitive intervention programs targeted at improving breast cancer awareness among women, religious and traditional leaders are imperative. These intervention programs could entail community engagement, workshops, or educational materials tailored to address specific cultural beliefs and misconceptions.

Taking into consideration the diverse cultural beliefs about breast cancer, there is a compelling need for nationwide public education on breast cancer to clarify the myths and misconceptions about the disease. The education program should be culturally tailored to address the myths and misconceptions. It is important that considerations are given to these issues, not only focusing on how these issues affect women’s lives post-treatment but also on how these issues can be resolved to improve diagnosis and treatment of the disease. We recommend that socio-cultural factors influencing breast cancer diagnosis and treatment should be incorporated into breast cancer awareness programs, education, and intervention programs in Ghana. We believe these would help inform women and encourage them to report to health facilities early with breast cancer symptoms to initiate timely diagnosis and treatment to improve the outcomes of the disease in Ghana.

Further research is required to explore appropriate and effective multidimensional culturally sensitive intervention research that integrates cultural beliefs and breast cancer treatment especially, in different Ghanaian communities.

Strengths and limitations of the study

This study has several strengths, one major strength is the extensive and comprehensive search in various electronic databases following the methodological guideline of JBI and reported in accordance with the PRISMA guidelines. Also, the inclusion of both qualitative and quantitative studies, allowed for a more comprehensive understanding of the socio-cultural beliefs influencing breast cancer diagnosis and treatment in Ghana.

The review considered only published studies and possibly may have overlooked unpublished or gray literature that could contribute to a more comprehensive understanding of the subject matter. Most of the studies were concentrated in the southern part of Ghana and therefore the results might not represent all the regions in Ghana.

This study adduces evidence on the socio-cultural beliefs that impact diagnosis and treatment of breast cancer among women in Ghana. As policy makers, clinicians and other stakeholders strive to improve breast cancer diagnosis and treatment, there is a need to address the socio-cultural beliefs to improve breast cancer outcomes in Ghana and potentially reduce breast cancer-related mortality.

Data availability

The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Sung H, Ferlay J, Siegel RL, Laversanne M, Soerjomataram I, Jemal A, Bray F. Global Cancer statistics 2020: GLOBOCAN estimates of incidence and Mortality Worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 2021;71(3):209–49.

Article   PubMed   Google Scholar  

Afaya A, Seidu AA, Sang S, Yakong VN, Afaya RA, Shin J, Ahinkorah BO. Mapping evidence on knowledge of breast cancer screening and its uptake among women in Ghana: a scoping review. BMC Health Serv Res. 2022;22(1):526.

Article   PubMed   PubMed Central   Google Scholar  

International Agency for Research on Cancer. Global Cancer Observatory. 2021: https://gco.iarc.fr/ .

Donnelly TT, Al Khater A-H, Al-Bader SB, Al Kuwari MG, Al-Meer N, Malik M, Singh R, Chaudhry S, Fung T. Beliefs and attitudes about breast cancer and screening practices among arab women living in Qatar: a cross-sectional study. BMC Womens Health. 2013;13(1):1–16.

Article   Google Scholar  

Pasick RJ, Burke NJ. A critical review of theory in breast cancer screening promotion across cultures. Annu Rev Public Health. 2008;29:351–68.

Azaiza F, Cohen M. Between traditional and modern perceptions of breast and cervical cancer screenings: a qualitative study of arab women in Israel. Psychooncology. 2008;17(1):34–41.

Agyemang AF, Tei-Muno AN, Dzomeku VM, Nakua EK, Duodu PA, Duah HO, Bentil AB, Agbadi P. The prevalence and predictive factors of breast cancer screening among older Ghanaian women. Heliyon. 2020;6(4):e03838.

Rivera-Franco MM, Leon-Rodriguez E. Delays in breast Cancer detection and treatment in developing countries. Breast Cancer (Auckl). 2018;12:1178223417752677.

PubMed   Google Scholar  

Boamah Mensah AB, Mensah KB, Aborigo RA, Bangalee V, Oosthuizen F, Kugbey N, Clegg-Lamptey JN, Virnig B, Kulasingam S, Ncama BP. Breast cancer screening pathways in Ghana: applying an exploratory single case study methodology with cross-case analysis. Heliyon. 2022;8(11):e11413.

d’Andrade RG, Strauss C. Human motives and cultural models. Volume 1. Cambridge University Press; 1992.

Koltko-Rivera ME. The psychology of worldviews. Rev Gen Psychol. 2004;8(1):3–58.

Iwelunmor J, Newsome V, Airhihenbuwa CO. Framing the impact of culture on health: a systematic review of the PEN-3 cultural model and its application in public health research and interventions. Ethn Health. 2014;19(1):20–46.

White P. The concept of diseases and health care in African traditional religion in Ghana. HTS: Theological Stud. 2015;71(3):1–7.

Google Scholar  

Tetteh DA, Faulkner SL. Sociocultural factors and breast cancer in sub-saharan Africa: implications for diagnosis and management. Womens Health (Lond). 2016;12(1):147–56.

Article   CAS   PubMed   Google Scholar  

Kim JG, Hong HC, Lee H, Ferrans CE, Kim EM. Cultural beliefs about breast cancer in Vietnamese women. BMC Womens Health. 2019;19(1):74.

Clegg-Lamptey JN, Dakubo JC, Attobra YN. Psychosocial aspects of breast cancer treatment in Accra, Ghana. East Afr Med J. 2009;86(7):348–53.

CAS   PubMed   Google Scholar  

Peters MDJ, Marnie C, Tricco AC, Pollock D, Munn Z, Alexander L, McInerney P, Godfrey CM, Khalil H. Updated methodological guidance for the conduct of scoping reviews. JBI Evid Synth. 2020;18(10):2119–26.

Stern C, Lizarondo L, Carrier J, Godfrey C, Rieger K, Salmond S, Apóstolo J, Kirkpatrick P, Loveday H. Methodological guidance for the conduct of mixed methods systematic reviews. JBI Evid Synthesis. 2020;18(10):2108–18.

Peters MD, Godfrey CM, McInerney P, Soares CB, Khalil H, Parker D. The Joanna Briggs Institute reviewers’ manual 2015: methodology for JBI scoping reviews. 2015.

Hong QN, Pluye P, Bujold M, Wassef M. Convergent and sequential synthesis designs: implications for conducting and reporting systematic reviews of qualitative and quantitative evidence. Syst Rev. 2017;6(1):61.

Hong QN, Pluye P, Fàbregues S, Bartlett G, Boardman F, Cargo M, Dagenais P, Gagnon M-P, Griffiths F, Nicolau B. Mixed methods appraisal tool (MMAT), version 2018. Registration Copyr 2018, 1148552(10).

Addae-Korankye A, Abada A, Imoro F. Assessment of breast cancer awareness in Tamale Metropolis in Ghana. World Sci Res. 2014;3(1):1–5.

Agbokey F. Health seeking behaviours for breast cancer among breast cancer patients at the Komfo Anokye Teaching Hospital. Kumasi, Ghana: University of Ghana; 2014.

Agbokey F, Kudzawu E, Dzodzomenyo M, Ae-Ngibise KA, Owusu-Agyei S, Asante KP. Knowledge and Health Seeking Behaviour of Breast Cancer Patients in Ghana. Int J Breast Cancer 2019, 2019:5239840.

Asobayire A, Barley R. Women’s cultural perceptions and attitudes towards breast cancer: Northern Ghana. Health Promot Int. 2015;30(3):647–57.

Asoogo C, Duma SE. Factors contributing to late breast cancer presentation for health care amongst women in Kumasi, Ghana. Curationis 2015, 38(1).

Azumah FD, Onzaberigu NJ. Community knowledge, perception and attitude toward breast cancer in Sekyere East District Ghana. Int J Innov Educ Res. 2017;5(6):221–35.

Boafo IM, Tetteh PM. Self-efficacy and perceived barriers as determinants of breast self-examination among female nonmedical students of the University of Ghana. Int Q Community Health Educ. 2020;40(4):289–97.

Bonsu AB, Ncama BP. Recognizing and appraising symptoms of breast cancer as a reason for delayed presentation in Ghanaian women: a qualitative study. PLoS ONE. 2019;14(1):e0208773.

Article   CAS   PubMed   PubMed Central   Google Scholar  

Dadzi R, Adam A. Assessment of knowledge and practice of breast self-examination among reproductive age women in Akatsi South district of Volta region of Ghana. PLoS ONE. 2019;14(12):e0226925.

Iddrisu M, Aziato L, Ohene LA. Socioeconomic impact of breast cancer on young women in Ghana: a qualitative study. Nurs Open. 2021;8(1):29–38.

Kugbey N, Oppong Asante K, Meyer-Weitz A. Illness perception and coping among women living with breast cancer in Ghana: an exploratory qualitative study. BMJ Open. 2020;10(7):e033019.

Opoku SY, Benwell M, Yarney J. Knowledge, attitudes, beliefs, behaviour and breast cancer screening practices in Ghana, West Africa. Pan Afr Med J. 2012;11:28.

PubMed   PubMed Central   Google Scholar  

Osei E, Osei Afriyie S, Oppong S, Ampofo E, Amu H. Perceived breast Cancer risk among female undergraduate students in Ghana: a cross-sectional study. J Oncol. 2021;2021:8811353.

Osei-Afriyie S, Addae AK, Oppong S, Amu H, Ampofo E, Osei E. Breast cancer awareness, risk factors and screening practices among future health professionals in Ghana: a cross-sectional study. PLoS ONE. 2021;16(6):e0253373.

Salisu WJ, Mirlashari J, Seylani K, Varaei S, Thorne S. Fatalism, distrust, and breast cancer treatment refusal in Ghana. Can Oncol Nurs J. 2022;32(2):198–205.

Omonzejele PF. African concepts of health, disease, and treatment: an ethical inquiry. Explore (NY). 2008;4(2):120–6.

Wright SV. An investigation into the causes of absconding among black African breast cancer patients. S Afr Med J. 1997;87(11):1540–3.

Benedict AO. The perception of illness in traditional Africa and the development of traditional medical practice. Int J Nurs. 2014;1(1):51–9.

Afaya A, Ramazanu S, Bolarinwa OA, Yakong VN, Afaya RA, Aboagye RG, Daniels-Donkor SS, Yahaya AR, Shin J, Dzomeku VM, et al. Health system barriers influencing timely breast cancer diagnosis and treatment among women in low and middle-income Asian countries: evidence from a mixed-methods systematic review. BMC Health Serv Res. 2022;22(1):1601.

Elewonibi B, BeLue R. The influence of socio-cultural factors on breast cancer screening behaviors in Lagos, Nigeria. Ethn Health. 2019;24(5):544–59.

Saeed S, Asim M, Sohail MM. Fears and barriers: problems in breast cancer diagnosis and treatment in Pakistan. BMC Womens Health. 2021;21(1):151.

Anyigba CA, Awandare GA, Paemka L. Breast cancer in sub-saharan Africa: the current state and uncertain future. Exp Biol Med (Maywood). 2021;246(12):1377–87.

Mensah S, Dogbe J, Kyei I, Addofoh N, Paintsil V, Osei T. Determinants of late presentation and histologic types of breast cancer in women presenting at a Teaching Hospital in Kumasi. Ghana J Cancer Prev Curr Res. 2015;3(4):00089.

Download references

Acknowledgements

The author did not receive any funding for this study.

Author information

Authors and affiliations.

College of Nursing, Yonsei University, 50-1, Yonsei-ro, Seodaemun-gu, Seoul, 03722, South Korea

Agani Afaya

Department of Nursing, School of Nursing and Midwifery, University of Health and Allied Sciences, Ho, Ghana

Department of Population, Family and Reproductive Health, University of Ghana School of Public Health, Accra, Ghana

Emmanuel Anongeba Anaba

Department of Nursing, Faculty of Allied Health Sciences, College of Health Sciences, Kwame Nkrumah University of Science and Technology, Kumasi, Ghana

Victoria Bam

School of Nursing and Midwifery, Griffith University, Queensland, Australia

Richard Adongo Afaya

Department of Internal Medicine, Tamale Teaching Hospital, Tamale, Ghana

Ahmed-Rufai Yahaya

College of Public Health, Medical and Veterinary Sciences, James Cook University, Townsville, Australia

Abdul-Aziz Seidu

School of Clinical Medicine, University of New South Wales, Sydney, Australia

Bright Opoku Ahinkorah

You can also search for this author in PubMed   Google Scholar

Contributions

AA, and EAA conceived the study, analyzed and wrote the methods section. AA, VB and RAA conducted the literature search and wrote the background. AA, RAA, and RY screened the included articles and extracted the data. AA, AS and BOA conducted literature search and discussed the results. All the authors reviewed and provided intellectual content and modification. All the authors reviewed and approved the final draft of the manuscript.

Corresponding author

Correspondence to Emmanuel Anongeba Anaba .

Ethics declarations

Ethical approval and consent to participate.

This study did not require ethical approval.

Consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

Additional information

Publisher’s note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Supplementary Material 1

Rights and permissions.

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ . The Creative Commons Public Domain Dedication waiver ( http://creativecommons.org/publicdomain/zero/1.0/ ) applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and permissions

About this article

Cite this article.

Afaya, A., Anaba, E.A., Bam, V. et al. Socio-cultural beliefs and perceptions influencing diagnosis and treatment of breast cancer among women in Ghana: a systematic review. BMC Women's Health 24 , 288 (2024). https://doi.org/10.1186/s12905-024-03106-y

Download citation

Received : 02 March 2023

Accepted : 22 April 2024

Published : 14 May 2024

DOI : https://doi.org/10.1186/s12905-024-03106-y

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Breast cancer
• Religious and spiritual beliefs
• Systematic review

BMC Women's Health

ISSN: 1472-6874

• Submission enquiries: [email protected]
• General enquiries: [email protected]

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

• View all journals
• My Account Login
• Explore content
• About the journal
• Publish with us
• Sign up for alerts
• Systematic Review
• Open access
• Published: 17 May 2024

Risk factors and incidence of central venous access device-related thrombosis in hospitalized children: a systematic review and meta-analysis

• Maoling Fu 1 , 2 ,
• Quan Yuan 2 ,
• Qiaoyue Yang 1 , 2 ,
• Yaqi Yu 1 , 2 ,
• Wenshuai Song 1 , 2 ,
• Xiuli Qin 1 ,
• Ying Luo 1 ,
• Xiaoju Xiong 1 &
• Genzhen Yu 1  

Pediatric Research ( 2024 ) Cite this article

66 Accesses

Metrics details

The risk factors for central venous access device-related thrombosis (CRT) in children are not fully understood. We used evidence-based medicine to find the risk factors for CRT by pooling current studies reporting risk factors of CRT, aiming to guide clinical diagnosis and treatment.

A systematic search of PubMed, Web of Science, Embase, Cochrane Library, Scopus, CNKI, Sinomed, and Wanfang databases was conducted. RevMan 5.4 was employed for data analysis.

The review included 47 studies evaluating 262,587 children with CVAD placement. Qualitative synthesis and quantitative meta-analysis identified D-dimer, location of insertion, type of catheter, number of lumens, catheter indwelling time, and central line-associated bloodstream infection as the most critical risk factors for CRT. Primarily due to observational design, the quality of evidence was regarded as low certainty for these risk factors according to the GRADE approach.

Because fewer high-quality studies are available, larger sample sizes and well-designed prospective studies are still needed to clarify the risk factors affecting CRT. In the future, developing pediatric-specific CRT risk assessment tools is important. Appropriate stratified preventive strategies for CRT according to risk assessment level will help improve clinical efficiency, avoid the occurrence of CRT, and alleviate unnecessary suffering of children.

This is the latest systematic review of risk factors and incidence of CRT in children.

A total of 47 studies involving 262,587 patients were included in our meta-analysis, according to which the pooled prevalence of CRT was 9.1%.

This study identified several of the most critical risk factors affecting CRT in children, including D-dimer, insertion location, type of catheter, number of lumens, catheter indwelling time, and central line-associated bloodstream infection (CLABSI).

Similar content being viewed by others

Acute kidney injury

Using National Early Warning Score (NEWS) 2 to help manage medical emergencies in the dental practice

Management of hypertensive crisis: British and Irish Hypertension Society Position document

Introduction.

Central venous access device (CVAD) is an infusion device inserted through different parts to make the tip of the catheter to the vena cava. In the clinic, CVAD is mainly divided into the following four categories: tunneled central venous catheter (CVC), nontunneled CVC, peripherally inserted central catheter (PICC), and totally implantable venous access port (TIVAP). 1 Pediatric patients often require stable, multifunctional, and comfortable long-term vascular access due to factors such as poor puncture cooperation, small vessel diameter, poor peripheral venous visibility and tolerance, high water content in the body leading to easy dehydration, and easy changes in condition after diseases. 2 The application of CVAD can significantly reduce the frequency of venipuncture, relieve the stimulation of drugs on the venous blood vessels, alleviate the pain and fear of the children, improve their medication compliance, ensure the effectiveness of intravenous infusion, and improve the quality of disease treatment. 3 , 4 , 5 Therefore, CVAD is widely used in pediatric clinics and has become an indispensable aspect of complex medical care for children with severe and chronic diseases.

Although CVAD has become an important tool in the pediatric treatment and nursing process, there are also risks of complications related to it, including CVAD-related thrombosis (CRT), phlebitis, fluid and blood leakage at the puncture point, catheter displacement, catheter obstruction, central line-associated bloodstream infection (CLABSI) and so on. 6 , 7 Among these, CRT is one of the most common and serious complications. The prevalence of CRT in children varies significantly by country, age, disease, and medical institution, ranging from 2 to 81%, 4 , 8 , 9 , 10 while in Chinese children without prophylactic treatment ranges from 20 to 66%. 11 , 12 CRT has no obvious clinical symptoms in the early stage, but it may still cause serious side effects, not only increasing the patient pain and medical costs but also delaying treatment timing, affecting prognosis and quality of life, and in severe cases, may even lead to thromboembolism, endangering life. 13 , 14 , 15

Identifying risk factors and incidence of CRT facilitates clinical practitioners in the early identification of high-risk patients, designing specific preventive strategies, treatment regimens, and management plans, thereby effectively reducing the incidence of CRT in hospitalized children and alleviating unnecessary patient suffering. However, most current research on CRT involves only small-scale groups in isolated nursing units or specific disease types. To date, no up-to-date systematic review provides pooled estimates of the risk factors and prevalence of CRT in children. Therefore, this study had a dual purpose: 1. to explore potential risk factors for CRT in children and to determine a pooled level of CRT prevalence; and 2. to provide evidence-based recommendations to improve the recognition, control, and treatment of CRT in children, as well as better nursing management for CRT.

This review was conducted following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. 16 The detailed research protocol can be accessed on the PROSPERO website (registration number: CRD42023421353).

Search strategy

Eight electronic databases were utilized to conduct a thorough literature search: PubMed, Web of Science, Embase, Cochrane Library, Scopus, China National Knowledge Infrastructure (CNKI), Sinomed, and Wanfang. The search in these databases was conducted from the earliest records available up to January 31st, 2024. The search strategy used a combination of Mesh terms and free words. The following Mesh terms and free words were mainly used: “child,” “children,” “adolescent,” “infant,” “pediatrics,” “central venous access device-related thrombosis,” “CRT,” “catheter-related thrombosis,” “catheter-related venous thrombosis,” “CVC-related thrombosis,” “risk factors,” “protective factors,” “predictors,” “causality,” “influencing factors”. The full search strategy for each database is available in the Supplementary Materials. In addition, we screened the reference lists of all included studies for relevant studies that met the criteria. Grey literature was searched as well. Some authors were contacted through email to gather more information or clarify any uncertainties.

Inclusion criteria

The study population was hospitalized children aged ≤18 years.

The primary research objective was to explore the risk factors for CRT.

The study results have at least one statistically significant predictor.

Case-control studies or cohort studies.

Published in English or Chinese.

Exclusion criteria

Catheter-related infection, catheter dysfunction, or other catheter complications as the primary outcome indicators.

Repeated published research.

Case reports, study designs, or clinical trials.

Reviews, editorials, letters, and conference abstracts.

In vitro or animal research.

Data were incomplete and could not be extracted.

Unable to find the original article.

Data extraction

Data from each eligible study were independently extracted by two reviewers using a pre-designed data collection form. Any disagreements were resolved by discussions among all authors. Data on the following characteristics were obtained from all included studies (see Supplementary Table S 1 for details):

Basic information: first author, country, year of publication, study duration, and study design.

Demographic characteristics: study population, sample size, number of CRT, and CRT rate.

Catheter-related features: catheter type, CRT type, and diagnostic method.

Potential risk factors for CRT: odds ratios (OR) or relative risks (RR) values and 95% confidence interval (CI) were extracted for each risk factor. If the study did not provide specific values, it was calculated by constructing a 2 × 2 contingency table.

Quality assessment

Two reviewers evaluated the quality of each study independently using the Risk of Bias Assessment for Nonrandomized Studies tool, 17 with any differences settled via group discussion. The tool assessed six domains of risk of bias: participant selection, confounding variables, exposure measurement, blinding of outcome assessment, incomplete outcome data, and selective outcome reporting. If all six domains were rated as low risk, the overall risk of bias for the study was low. The overall risk of bias was moderate if at least one domain was rated as unclear risk, and no domain was rated as high risk, and high if one or more domains were rated as high risk.

To ensure the accuracy of the assessment results, a third reviewer randomly selected five studies to check the data extraction and quality assessment.

Qualitative synthesis and quantitative meta-analysis

Qualitatively classify each risk factor as definite, likely, unclear, or not a risk factor based on the total number of studies with low and moderate bias risks and the proportion of studies demonstrating positive association (Box 1 in the supplementary material). If a risk factor was reported by more than two studies with low or moderate risk of bias, and the definition and reference range were sufficiently consistent, a quantitative meta-analysis was performed to estimate the combined OR.

Data were analyzed using Revman 5.4 software. In the meta-analysis of risk factors and CRT rate, the generic inverse variance method was applied, which only required effect estimate and standard error (SE). 18 The SE was obtained by inverse transforming the 95% CI applying the standard normal distribution. Heterogeneity tests were performed on the studies included in the Meta-analysis to examine for the combinability of the results of each independent study. P  ≥ 0.05 and I-squared ( I 2 ) < 50% considered less heterogeneity between studies and therefore a fixed-effects model was chosen for the analysis, conversely, P  < 0.05 or I 2  ≥ 50% considered greater heterogeneity, and a random-effects model was chosen.

Certainty of the evidence

The Grading of Recommendations Assessment, Development, and Evaluation (GRADE) method was used to assess the certainty of the evidence. In this method, observational studies were initially classified as low-quality evidence and then downgraded and upgraded according to five downgrading and three upgrading principles. The 5 downgrading factors included risk of bias, inconsistency, indirectness, imprecision, and publication bias, and the 3 upgrading factors included the magnitude of an effect, dose-response gradient, and effect of plausible residual confounding. Based on these considerations, the overall certainty of each piece of evidence was rated as one of four levels: high, moderate, low, or very low.

The initial search of the databases extracted a total of 4193 articles, of which 1656 were duplicates and removed. The titles and abstracts of the remaining 2537 articles were screened according to the inclusion criteria and 142 were selected for full-text search. After a rigorous eligibility review, 45 articles met the inclusion criteria. In addition, two articles were found to meet the eligibility criteria in a search of the reference lists of the selected articles and grey literature. In the end, a total of 47 articles were included in this review, of which 43 contributed to the qualitative synthesis and quantitative meta-analysis (Fig.  1 ).

Demonstrate the screening and inclusion process for systematic literature search.

Of the 47 studies, 19 were prospective 4 , 13 , 19 , 20 , 21 , 22 , 23 , 24 , 25 , 26 , 27 , 28 , 29 , 30 , 31 , 32 , 33 , 34 , 35 and the rest were retrospective, 9 , 12 , 36 , 37 , 38 , 39 , 40 , 41 , 42 , 43 , 44 , 45 , 46 , 47 , 48 , 49 , 50 , 51 , 52 , 53 , 54 , 55 , 56 , 57 , 58 , 59 , 60 , 61 of which 10 were multicenter 4 , 9 , 13 , 21 , 23 , 26 , 27 , 28 , 49 , 59 and 37 were single-center. 12 , 19 , 20 , 22 , 24 , 25 , 29 , 30 , 31 , 32 , 33 , 34 , 35 , 36 , 37 , 38 , 39 , 40 , 41 , 42 , 43 , 44 , 45 , 46 , 47 , 48 , 50 , 51 , 52 , 53 , 54 , 55 , 56 , 57 , 58 , 60 , 61 The sample sizes ranged from 47 to 158,299, with the two largest being 71,782 13 and 158,299, 59 respectively. In addition, three studies constructed clinical prediction models. 22 , 28 , 47 Table  1 lists the summary characteristics of the included studies.

Study populations and CRT rates in included studies

These studies investigated a series of hospitalized children of different ages and departments, of which 12 studies with all hospitalized children as the study population, 12 studies with PICU hospitalized children as the study population, six studies with NICU hospitalized children as the study population, one study with all ICU hospitalized children as the study population, four studies with leukemia children as the study population, two studies with infants under 1-year-old as the study population, and the other ten studies with children with a specific disease as the study population.

The combined CRT rate was 9.1% (95% CI : 5.7–14.5%) with a high degree of heterogeneity ( I 2  = 100%). The combined CRT rate was 11.5% (95% CI : 5.7–23.1%; I 2  = 99%) in both male and female children. The frequency of CRT in PICU and NICU was available from 13 articles with 234,464 children and 7 articles with 6093 infants, which combined CRT rates were 10.7% (95% CI : 3.8–23.7%; I 2  = 100%), 2.9% (95% CI : 1.0–6.5%; I 2  = 96%), respectively. The combined CRT rate of children with leukemia was 13.0% (95% CI : 2.9–38.3%; I 2  = 98%) (Supplementary Material Figs. S 1 – 6 )

Quality of the CRT studies

The methodological quality of the included studies varied (Fig.  2 and Supplementary Material Fig. S 7 ). Nine studies had a low overall risk of bias, as all six domains were categorized as low risk. Four studies had a high overall risk of bias, three of which were associated with confounding variables and one to participant selection. The remaining 34 studies had a moderate overall risk of bias, with at least one of the six domains having an unclear risk.

A summary presentation of the assessment results of risk of bias for the 47 studies.

Risk factors of CRT in included studies

The 47 included studies reported 61 statistically significant risk factors for CRT (Table  1 ). These factors were classified into three categories: patient-related risk factors (37.7%, 23/61); CVAD-related risk factors (34.4%, 21/61), and treatment-related risk factors (27.9%, 17/61).

Based on the qualitative synthesis, six variables were considered to be definite risk factors for CRT, including D-dimer, location of insertion, type of catheter, number of lumens, catheter indwelling time, and CLABSI. Eleven variables were considered likely associated with CRT, including gastrointestinal diseases, history of catheterization, thrombophilia, geographic location of line placement, catheter dysfunction, number of catheters, insertion length (cm), catheter to vein ratio, dialysis, hypertonic liquid, and cardiac catheterization. For 42 variables, the relationship with CRT was deemed unclear due to conflicting results from studies assessed as having low and moderate risk of bias, or because they were positively associated in only one study. Additionally, birth weight and gestational age were considered non-risk factors (Table  2 ).

Meta-analyses were implemented for risk factors that were reported by at least two low or moderate risk of bias studies with a consistent definition and reference range (Table  3 and Figs.  3 – 6 ).

Forest plots of odds ratios (OR) that were included in the quantitative meta-analysis and the associated overall OR. For each OR, the size of the red square region is proportional to the corresponding study weight. Diamond shape intervals represent the overall OR. I 2 represents the fraction of variability among the individual OR that cannot be explained by sampling variability.

Forest plots of odds ratios (OR) that were included in the quantitative meta-analysis and the associated overall OR. For each OR, the size of the red square region is proportional to the corresponding study weight. Diamond shape intervals represent the overall OR. I 2 represents the fraction of variability among the individual OR that cannot be explained by sampling variability.

GRADE assessment of evidence

Supplementary Table S 2 shows GRADE assessments for the certainty of evidence. Due to the design of the observational studies, all evidence was initially rated as low certainty. Based on five downgrading and three upgrading principles, 17 pieces of evidence were still rated as low certainty, and the remaining 44 pieces of evidence were downgraded to very low certainty for serious inconsistency and imprecision.

Our study is the latest systematic review of risk factors and the incidence of CRT in hospitalized children. Based on 47 studies included in the current meta-analysis, which involved a total of 262,587 patients, the pooled prevalence of CRT is 9.1%. We conducted a qualitative synthesis analysis of 61 predictive factors and a quantitative meta-analysis of 38 factors, identifying six definite factors, 11 likely factors, and 42 unclear factors associated with CRT. Definite predictors included being of D-dimer, location of insertion, type of catheter, number of lumens, catheter indwelling time and CLABSI. The findings of our systematic review provide the latest comprehensive evidence summary that can inform the early identification of children at risk for CRT and the development of intervention measures to prevent and reduce CRT.

Implantable and temporary medical devices such as CVAD are exposed to blood for weeks to years depending on the type of CVAD in place. Since CVAD is an artificial surface and lacks an endothelial layer that inhibits platelet coagulation and adhesion, it is thought to potentially activate the contact pathways, ultimately leading to thrombosis. Assembly of artificial surface contact systems might be part of the host defense mechanism against foreign substances, but it can lead to kinin and thrombin generation, and complement activation. 62 This eventually promotes thrombosis and inflammation. The presence of CVAD is the most common risk factor for venous thromboembolism (VTE). CRT accounts for 10% of deep vein thrombosis (DVT) in adults and 50–80% in children. 10 , 55 , 63 The incidence of CRT in hospitalized children has increased significantly by 30–70% over the past 20 years, 64 , 65 which may cause serious medical complications besides increasing healthcare expenditures and length of stay.

We discover that a higher level of D-dimer is an independent risk factor for CRT in hospitalized children, consistent with the results of adult studies. 66 D-dimer is a soluble fibrin degradation product deriving from the plasmin-mediated degradation of cross-linked fibrin that is increased or positive in secondary hyperfibrinolysis, such as hypercoagulable states, disseminated intravascular coagulation, and thrombolytic therapy. 67 , 68 Increased D-dimer suggests an association with thrombotic disorders in the body of various origins and an increase in fibrinolytic activity. D-dimer has been extensively investigated for excluding the diagnosis of VTE and is used routinely for this indication. 67 , 69 Therefore, for early recognition and to reduce the incidence of CRT, D-dimer levels should be closely monitored before and after catheterization. However, the elevated D-dimer test results cannot fully explain the cause and location of CRT formation and must be analyzed in conjunction with clinical and other test results. Inherited thrombophilia, caused by genetic defects leading to a deficiency or abnormality in associated proteins, including protein C, protein S, antithrombin, the coagulation factor V Leiden mutation, and factor II mutation G20210A, 70 is considered a potential risk factor for CRT. The prevalence of thrombophilia varies widely among different populations, with a reported prevalence of 10% to 59% in pediatric VTE patients. 71 Children with gastrointestinal diseases like short bowel syndrome (SBS) and inflammatory bowel disease (IBD) have an increased risk of developing CRT during hospitalization. The precise mechanism behind this association is still uncertain according to current research. It may be attributed to the heightened inflammation levels during catheterization, particularly in patients with active IBD episodes or admissions during surgery, which leads to a period of increased inactivity. 55 This suggests that delaying placement during the most active period of inflammation may reduce the rate of thrombosis.

A narrative review pointed out that age is one of the most significant risk factors for VTE. In children, CRT shows a bimodal distribution, with the highest incidence rate in infancy and adolescence. 10 The higher incidence in infancy may be due in part to the smaller diameter of the vein, making insertion difficult and requiring multiple attempts. However, whether age is a risk factor for CRT is still highly controversial. The study by Chojnacka et al. did not find a statistically significant difference, 39 although a trend toward a similar bimodal distribution was found in the study population. Cancer, cardiovascular disease, sepsis, asphyxia, and neurological diseases are also considered unclear factors for CRT. Pediatric patients diagnosed with leukemia have multiple risk factors for VTE formation, such as the presence of hypercoagulable blast cells, the pro-thrombotic nature of the cancer itself, and treatment with steroids and L-asparaginase. Chen et al. 38 and Jaffray et al. 4 concluded that children with leukemia are more likely to develop CRT. Sepsis causes the coagulation mechanism to become fragile, which in turn activates the coagulation system and creates thrombosis. 72 However, a study by Onyeama et al. 52 showed that sepsis was significantly associated with a reduced incidence of CRT, and the exact mechanism is currently unknown.

The location of insertion and type of catheter are critical risk factors for CRT. The incidence of CRT is higher in femoral vein catheterizations compared to subclavian and jugular vein catheterizations in children, which is contrary to findings in adult patients. 73 The femoral location is a larger vessel and allows placement of a larger size catheter. Femoral CVAD is prioritized in urgent and emergency situations. In such cases, the patients tend to be more critically ill and often immobilized, further exacerbating the low-flow state. In addition, there may be vein compression and kinking beneath the inguinal ligament with leg movement, which may increase the risk of CRT. 27 PICC catheters provide a reliable medium to long-term route to intravenous therapy for children, but compared with other types of catheters, the risk of CRT is higher. We speculate that the long tunnel length and relatively large lumen size of the PICC, compared to the diameter of the vessel at the insertion site, may lead to increased blood flow obstruction. 52 Additionally, patients with PICC may be more likely to be diagnosed with symptomatic VTE than tunneled lines (TLs) because PICC is often placed in smaller vessels and journeys through the arm or leg causing limb pain and swelling, whereas TLs are located in the chest.

The risk of CRT increases with the number of lumens. A possible explanation for this finding is that multilumen catheters tend to have larger catheter sizes and thus occupy more area within the vessel lumen, leading to obstruction of normal blood flow within the veins. The relationship between CRT and CLABSI is bidirectional. Following catheter insertion, a fibrin sheath forms around the catheter. Microorganisms, especially staphylococcus aureus, easily adhere to the fibrin sheaths, and may lead to CLABSI. 74 Conversely, CLABSI can trigger inflammatory reactions, leading to further progression of thrombosis. CVAD duration is positively associated with the risk of CRT. Catheter placement may cause mechanical injury to the vein. As the indwelling duration increases, many damaged smooth muscle and endothelial cells become embedded within the fibrin, resulting in thrombus formation. In addition, prolonged indwelling increases the chance of platelet contact with the vessel lining, activating coagulation factors and thrombin, increasing the risk of thrombosis. 22 Therefore, nurses should perform routine maintenance of the catheter in children who require long-term CVAD indwelling. The duration of CVAD should be monitored, the necessity of its indwelling should be assessed daily, and the catheter should be removed as early as possible while ensuring treatment.

As obstruction of venous blood flow from the CVAD is considered an essential causative mechanism for the development of VTE, a high ratio between catheter size and vein diameter could be a risk factor for CRT. The 2012 international guidelines on pediatric CVC insertion recommend that the ratio between the catheter’s external diameter and the cannulated vein’s diameter should not exceed 0.33. 75 However, this suggestion is only based on expert opinions and currently lacks relevant clinical data support. Therefore, further research is still needed to verify it. Catheter dysfunction is mainly caused by small clots or fibrous sheaths wrapping around the tip of the catheter. Prolonged accumulation may lead to incomplete or complete blockage of blood vessels, becoming a gathering point for thrombosis. 74 Journeycake et al. observed that the risk of VTE was highest in pediatric cancer patients with multiple episodes of catheter dysfunction. 76 A study of pediatric brain tumor patients reported that VTE was more common in patients with catheter dysfunction. 77 Thus, these studies and the current data support the need to consider catheter dysfunction as a possible risk factor for CRT and to design further screening and intervention studies for early identification and prevention of catheter dysfunction.

The rationale for studying the relationship between the insertion side of CVAD and the risk of CRT is based on the anatomy of the upper body venous system. The left brachiocephalic vein is longer and courses more horizontally than the right side, thus entering the superior vena cava at a sharper angle. The right jugular vein is the most direct and shortest route for the CVAD to enter the heart. By contrast, the CVAD located in the left jugular vein has a greater distance to the heart and passes through 2 angles in the venous system, which may cause endothelial damage and increase the likelihood of blood flow obstruction and venous wall adhesion. 26 However, our meta-analysis did not find a statistically significant increase in the risk of CRT with left-sided placement compared to right-sided placement. The ideal location for the catheter tip is the junction of the superior vena cava and the right atrium. This location is preferred because of the higher blood flow rate, which may be protective against thrombosis. 43 Currently, the pediatric literature on the effect of optimal tip position on CRT is scarce and inconclusive. In addition, catheter tips do not always remain in that position after initial placement. Therefore, tip movement should be a significant concern in pediatric patients, especially active, growing, and requiring long-term catheter use.

Providing renal replacement therapy is a lifelong task for pediatric end-stage renal disease (ESRD) patients. Although successful transplantation can be achieved even in young patients, the lifespan of the graft is limited. Consequently, many transplant recipients may be put back on dialysis as part of their ESRD treatment. 78 CVC remains the main vascular access for hemodialysis in children. Long-term reliance on CVC is related to a high incidence of catheter dysfunction and failure. The frequent need for recurrent CVC placement in such patients leads to an elevated risk of central vein stenosis and CRT. Cardiac catheterization is also a possible risk factor for CRT. Appropriate anticoagulation is required during catheterization, without which the risk of thrombosis is up to 40%. However, the use of unfractionated heparin in pediatric patients is challenging because the coagulation system and heparin response are different from that of adults. 79 There’s a need for further research to determine if children are receiving adequate doses of heparin during cardiac catheterization to prevent thrombosis without increasing the risk of bleeding complications. The incidence of VTE in adult patients who are chronically bedridden and braked is 3.59 times higher than in patients with normal activity levels. 80 In critically ill or surgical children, mechanical ventilation is often performed in the early stages, requiring continuous use of multiple sedative or inotropic drugs to reduce cardiac load and protect pulmonary function. During sedation, the child is in a braked state, limb activity is reduced or even inactive, blood flow slows down, and blood stagnates in the veins, increasing the chance of platelet adhesion to the endothelium, which may increase the risk of CRT. Therefore, passive movements such as limb abduction, internal rotation, elbow flexion and elbow extension should be performed appropriately when the child’s condition permits.

Nutritional support is an important part of critical illness treatment, including enteral and parenteral nutrition (PN). CVAD is the supply channel for total parenteral nutrition (TPN), and some children may even need this method to provide calories for a long time. High glucose and calcium concentrations in PN are both possible triggers of CRT, and PN has been shown to upregulate the extrinsic coagulation cascade, especially with long-term use. 60 Diamanti et al. reported that the incidence rate of TPN complicated with CRT was 20%. 81 Mannitol or glycerol fructose are widely used as hypertonic drugs in clinical practice, which can increase plasma osmolality to dehydrate tissues after entering the body. At the same time, it may cause a cellular stress response, induce apoptosis, and can activate inflammatory cytokines and coagulation pathways to induce thrombosis. Jiang et al. 22 found vasoactive drugs to be a risk factor for CRT. The possible reason is that vasoactive drugs can cause strong vasoconstriction, endothelial function damage or impairment, and promote fibrinogen synthesis. However, this is contrary to the findings of Marquez et al. 28 and Faustino et al. 21 Therefore, larger prospective studies are still needed to assess this risk factor more precisely.

The strengths of this study include the systematic identification of all relevant studies of risk factors for CRT in hospitalized children and the classification of risk factors into three categories, patient-related risk factors, CVAD-related risk factors, and treatment-related risk factors, to offer a logical progression of the possible causes of CRT in children. However, several limitations of this systematic review should be stated. Firstly, as most of the studies originate from Western countries, extrapolating these results to Eastern populations is questionable. Second, significant heterogeneity was encountered in our analysis, potentially stemming from variations in regimen, duration, population enrolled, and center setting, among other factors. This diversity necessitates a cautious interpretation of the results. In addition, only a few high-quality studies with a low risk of bias, and many of the studies suffer from significant sources of bias. Furthermore, the effect in many occasions was assessed by very few studies. Therefore, the evidence to support it is low, which needs to be validated in future studies. Finally, risk factors for CRT could not be made causal assertions since the majority of studies were retrospective.

Conclusions

In conclusion, we have identified several critical factors that affect CRT, including D-dimer, location of insertion, type of catheter, number of lumens, catheter indwelling time, and CLABSI. Nevertheless, none of the included studies considered the impact of socio-demographic factors on CRT, such as parental education level, occupation, and family economic status. Therefore, larger sample sizes and well-designed prospective studies are still needed to clarify the predictors affecting CRT in the future. In addition, there is a lack of pediatric-specific CRT risk assessment tools, which need to be further developed and validated. Machine learning (ML), as a method for designing risk assessment models that help to efficiently explore and mine useful information, has been widely used in recent years to solve a variety of challenging medical problems. Likewise, the application of ML in CRT risk diagnosis may contribute to a more precise assessment. In clinical practice, it is necessary to take appropriate stratified preventive measures according to the level of CRT risk assessment of children, to improve the efficiency of clinical work, reduce the burden of clinical work, and minimize the occurrence of CRT under the premise of ensuring the safety of children.

Data availability

The datasets used and/or analyzed during the current study are available from the corresponding author upon reasonable request.

Yeow, M. et al. A systematic review and network meta-analysis of randomized controlled trials on choice of central venous access device for delivery of chemotherapy. J. Vasc. Surg. Venous Lymphat. Disord. 10 , 1184–91.e8 (2022).

Article   PubMed   Google Scholar  

Cellini, M. et al. Guidelines of the Italian Association of Pediatric Hematology and Oncology for the management of the central venous access devices in pediatric patients with onco-hematological disease. J. Vasc. Access 23 , 3–17 (2022).

Ares, G. & Hunter, C. J. Central venous access in children: indications, devices, and risks. Curr. Opin. Pediatr. 29 , 340–346 (2017).

Jaffray, J. et al. Peripherally inserted central catheters lead to a high risk of venous thromboembolism in children. Blood 135 , 220–226 (2020).

Article   CAS   PubMed   Google Scholar  

Zhang, J. J. et al. Factors affecting mechanical complications of central venous access devices in children. Pediatr. Surg. Int. 38 , 1067–1073 (2022).

Article   PubMed   PubMed Central   Google Scholar  

Akhtar, N. & Lee, L. Utilization and Complications of Central Venous Access Devices in Oncology Patients. Curr. Oncol. 28 , 367–377 (2021).

Ullman, A. J., Marsh, N., Mihala, G., Cooke, M. & Rickard, C. M. Complications of Central Venous Access Devices: A Systematic Review. Pediatrics 136 , e1331–e1344 (2015).

Östlund, Å. et al. Erratum to ‘Incidence of and risk factors for venous thrombosis in children with percutaneous non-tunnelled central venous catheters’ (Br J Anaesth 2019; 123: 316-24). Br. J. Anaesth. 123 , 918 (2019).

McLaughlin, C. M. et al. Symptomatic catheter-associated thrombosis in pediatric trauma patients: Choose your access wisely. Surgery 166 , 1117–1121 (2019).

Citla Sridhar, D., Abou-Ismail, M. Y. & Ahuja, S. P. Central venous catheter-related thrombosis in children and adults. Thromb. Res. 187 , 103–112 (2020).

Zhou, X. et al. A retrospective analysis of risk factors associated with catheter-related thrombosis: a single-center study. Perfusion 35 , 806–813 (2020).

Li, S. et al. Risk factors for central venous catheter-related thrombosis in hospitalized children: a single-center a retrospective cohort study. Transl. Pediatr. 11 , 1840–1851 (2022).

Patel, N., Petersen, T. L., Simpson, P. M., Feng, M. & Hanson, S. J. Rates of Venous Thromboembolism and Central Line-Associated Bloodstream Infections Among Types of Central Venous Access Devices in Critically Ill Children. Crit. Care Med. 48 , 1340–1348 (2020).

Timsit, J. F. et al. A state of the art review on optimal practices to prevent, recognize, and manage complications associated with intravascular devices in the critically ill. Intensive Care Med. 44 , 742–759 (2018).

Ullman, A. J. et al. Pediatric central venous access devices: practice, performance, and costs. Pediatr. Res. 92 , 1381–1390 (2022).

Hutton, B. et al. The PRISMA extension statement for reporting of systematic reviews incorporating network meta-analyses of health care interventions: checklist and explanations. Ann. Intern Med. 162 , 777–784 (2015).

Kim, S. Y. et al. Testing a tool for assessing the risk of bias for nonrandomized studies showed moderate reliability and promising validity. J. Clin. Epidemiol. 66 , 408–414 (2013).

Borenstein, M., Hedges, L. V., Higgins, J. P. & Rothstein, H. R. A basic introduction to fixed-effect and random-effects models for meta-analysis. Res. Synth. Methods 1 , 97–111 (2010).

Beck, C., Dubois, J., Grignon, A., Lacroix, J. & David, M. Incidence and risk factors of catheter-related deep vein thrombosis in a pediatric intensive care unit: a prospective study. J. Pediatr. 133 , 237–241 (1998).

Dubois, J. et al. Incidence of deep vein thrombosis related to peripherally inserted central catheters in children and adolescents. Cmaj 177 , 1185–1190 (2007).

Faustino, E. V. et al. Incidence and acute complications of asymptomatic central venous catheter-related deep venous thrombosis in critically ill children. J. Pediatr. 162 , 387–391 (2013).

Jiang, W. et al. Construction and validation of a risk prediction model for central venous catheter-associated deep venous thromboses in children with congenital heart disease after surgery. Chin. J. Nurs. 57 , 2217–2224 (2022).

Google Scholar  

Faustino, E. V. et al. Factor VIII May Predict Catheter-Related Thrombosis in Critically Ill Children: A Preliminary Study. Pediatr. Crit. Care Med. 16 , 497–504 (2015).

Jones, S., Butt, W., Monagle, P., Cain, T. & Newall, F. The natural history of asymptomatic central venous catheter-related thrombosis in critically ill children. Blood 133 , 857–866 (2019).

Kim, E. H. et al. Central venous catheter-related thrombosis in pediatric surgical patients: A prospective observational study. Paediatr. Anaesth. 32 , 563–571 (2022).

Male, C. et al. Central venous line-related thrombosis in children: association with central venous line location and insertion technique. Blood 101 , 4273–4278 (2003).

Male, C., Julian, J. A., Massicotte, P., Gent, M. & Mitchell, L. Significant association with location of central venous line placement and risk of venous thrombosis in children. Thromb. Haemost. 94 , 516–521 (2005).

Marquez, A., Shabanova, V. & Faustino, E. V. Prediction of Catheter-Associated Thrombosis in Critically Ill Children. Pediatr. Crit. Care Med. 17 , e521–e528 (2016).

Menéndez, J. J. et al. Incidence and risk factors of superficial and deep vein thrombosis associated with peripherally inserted central catheters in children. J. Thromb. Haemost. 14 , 2158–2168 (2016).

Rubio Longo, M. C. et al. Catheter-related deep vein thrombosis in newborn infants. Arch. Argent. Pediatr. 119 , 32–38 (2021).

PubMed   Google Scholar  

Östlund, Å. et al. Incidence of and risk factors for venous thrombosis in children with percutaneous non-tunnelled central venous catheters. Br. J. Anaesth. 123 , 316–324 (2019).

Sol, J. J. et al. Chronic Complications After Femoral Central Venous Catheter-related Thrombosis in Critically Ill Children. J. Pediatr. Hematol. Oncol. 37 , 462–467 (2015).

van Rooden, C. J. et al. Infectious complications of central venous catheters increase the risk of catheter-related thrombosis in hematology patients: a prospective study. J. Clin. Oncol. 23 , 2655–2660 (2005).

Zeng, X., Zhang, C. & Shi, Y. Analysis of risk factors for complicated catheter-related thrombosis in children. Chin. J. Emerg. Med. 29 , 719–723 (2020).

Wei, Y. et al. The incidence and risk factors of catheter-related-thrombosis during induction chemotherapy in acute lymphocytic leukemia children. Chin. J. Hematol. 38 , 313–317 (2017).

CAS   Google Scholar  

Deng, G. & Liao, Q. Analysis of risk factors for venous thrombosis after PICC placement in critically ill children. Int. I Nurs. 39 , 775–777 (2020).

Badheka, A. V. et al. Catheter related thrombosis in hospitalized infants: A neural network approach to predict risk factors. Thromb. Res. 200 , 34–40 (2021).

Chen, K. et al. Risk factors for central venous catheter-related thrombosis in children: a retrospective analysis. Blood Coagul. Fibrinol. 27 , 384–388 (2016).

Article   Google Scholar  

Chojnacka, K., Krasiński, Z., Wróblewska-Seniuk, K. & Mazela, J. Catheter-related venous thrombosis in NICU: A case-control retrospective study. J. Vasc. Access 23 , 88–93 (2022).

Derderian, S. C., Good, R., Vuille-Dit-Bille, R. N., Carpenter, T. & Bensard, D. D. Central venous lines in critically ill children: Thrombosis but not infection is site dependent. J. Pediatr. Surg. 54 , 1740–1743 (2019).

Diamond, C. E. et al. Catheter-Related Venous Thrombosis in Hospitalized Pediatric Patients with Inflammatory Bowel Disease: Incidence, Characteristics, and Role of Anticoagulant Thromboprophylaxis with Enoxaparin. J. Pediatr. 198 , 53–59 (2018).

Noailly Charny, P. A. et al. Increased Risk of Thrombosis Associated with Peripherally Inserted Central Catheters Compared with Conventional Central Venous Catheters in Children with Leukemia. J. Pediatr. 198 , 46–52 (2018).

Gnannt, R. et al. Increased risk of symptomatic upper-extremity venous thrombosis with multiple peripherally inserted central catheter insertions in pediatric patients. Pediatr. Radio. 48 , 1013–1020 (2018).

Gray, B. W. et al. Characterization of central venous catheter-associated deep venous thrombosis in infants. J. Pediatr. Surg. 47 , 1159–1166 (2012).

Haddad, H. et al. Routine surveillance ultrasound for the management of central venous catheters in neonates. J. Pediatr. 164 , 118–122 (2014).

Lambert, I., Tarima, S., Uhing, M. & Cohen, S. S. Risk Factors Linked to Central Catheter-Associated Thrombosis in Critically Ill Infants in the Neonatal Intensive Care Unit. Am. J. Perinatol. 36 , 291–295 (2019).

Li, H. et al. Prediction of central venous catheter-associated deep venous thrombosis in pediatric critical care settings. BMC Med. Inf. Decis. Mak. 21 , 332 (2021).

Article   CAS   Google Scholar  

Lovett, M. E. et al. Catheter-associated deep vein thrombosis in children with severe traumatic brain injury: A single-center experience. Pediatr. Blood Cancer 70 , e30044 (2023).

MacLean, J. et al. Need for tissue plasminogen activator for central venous catheter dysfunction is significantly associated with thrombosis in pediatric cancer patients. Pediatr. Blood Cancer 65 , e27015 (2018).

Noonan, P. J., Hanson, S. J., Simpson, P. M., Dasgupta, M. & Petersen, T. L. Comparison of Complication Rates of Central Venous Catheters Versus Peripherally Inserted Central Venous Catheters in Pediatric Patients. Pediatr. Crit. Care Med. 19 , 1097–1105 (2018).

Pei, L. et al. Clinical characteristics and risk factors of symptomatic central venous catheter-related deep vein thrombosis in children. Chin. Pediatr. Emerg. Med. 23 , 450–454 (2016).

Onyeama, S. N. et al. Central Venous Catheter-associated Venous Thromboembolism in Children With Hematologic Malignancy. J. Pediatr. Hematol. Oncol. 40 , e519–e524 (2018).

Shah, S. H. et al. Clinical risk factors for central line-associated venous thrombosis in children. Front. Pediatr. 3 , 35 (2015).

Shin, H. S., Towbin, A. J., Zhang, B., Johnson, N. D. & Goldstein, S. L. Venous thrombosis and stenosis after peripherally inserted central catheter placement in children. Pediatr. Radio. 47 , 1670–1675 (2017).

Smitherman, A. B. et al. The incidence of catheter-associated venous thrombosis in noncritically ill children. Hosp. Pediatr. 5 , 59–66 (2015).

Steen, E. H. et al. Central Venous Catheter-Related Deep Vein Thrombosis in the Pediatric Cardiac Intensive Care Unit. J. Surg. Res 241 , 149–159 (2019).

Wang, J. & Ren, G. Peripherally inserted central catheter related venous thromboembolism in children with acute leukemia: a factorial analysis. Chin. J. Biomed. Eng. 27 , 288–293 (2021).

Dubbink-Verheij, G. H. et al. Femoral Vein Catheter is an Important Risk Factor for Catheter-related Thrombosis in (Near-)term Neonates. J. Pediatr. Hematol. Oncol. 40 , e64–e68 (2018).

Tran, M., Shein, S. L., Ji, X. & Ahuja, S. P. Identification of a “VTE-rich” population in pediatrics - Critically ill children with central venous catheters. Thromb. Res 161 , 73–77 (2018).

Wisecup, S., Eades, S. & Turiy, Y. Characterizing the Risk Factors Associated With Venous Thromboembolism in Pediatric Patients After Central Venous Line Placement. J. Pediatr. Pharm. Ther. 20 , 358–366 (2015).

Zhu, W., Zhang, H., Xing, Y. Clinical Characteristics of Venous Thrombosis Associated with Peripherally Inserted Central Venous Catheter in Premature Infants. Children 9 , https://doi.org/10.3390/children9081126 (2022).

Ekdahl, K. N. et al. Innate immunity activation on biomaterial surfaces: a mechanistic model and coping strategies. Adv. Drug Deliv. Rev. 63 , 1042–1050 (2011).

Article   CAS   PubMed   PubMed Central   Google Scholar  

Takemoto, C. M. et al. Hospital-associated venous thromboembolism in children: incidence and clinical characteristics. J. Pediatr. 164 , 332–338 (2014).

Boulet, S. L. et al. Trends in venous thromboembolism-related hospitalizations, 1994-2009. Pediatrics 130 , e812–e820 (2012).

Raffini, L., Huang, Y. S., Witmer, C. & Feudtner, C. Dramatic increase in venous thromboembolism in children’s hospitals in the United States from 2001 to 2007. Pediatrics 124 , 1001–1008 (2009).

Lin, S., Zhu, N., YihanZhang, Du, L. & Zhang, S. Development and validation of a prediction model of catheter-related thrombosis in patients with cancer undergoing chemotherapy based on ultrasonography results and clinical information. J. Thromb. Thrombolysis 54 , 480–491 (2022).

Johnson, E. D., Schell, J. C. & Rodgers, G. M. The D-dimer assay. Am. J. Hematol. 94 , 833–839 (2019).

Favresse, J. et al. D-dimer: Preanalytical, analytical, postanalytical variables, and clinical applications. Crit. Rev. Clin. Lab Sci. 55 , 548–577 (2018).

Weitz, J. I., Fredenburgh, J. C. & Eikelboom, J. W. A Test in Context: D-Dimer. J. Am. Coll. Cardiol. 70 , 2411–2420 (2017).

Darlow, J. & Mould, H. Thrombophilia testing in the era of direct oral anticoagulants. Clin. Med. 21 , e487–e491 (2021).

Monagle, P. et al. American Society of Hematology 2018 Guidelines for management of venous thromboembolism: treatment of pediatric venous thromboembolism. Blood Adv. 2 , 3292–3316 (2018).

Meziani, F., Gando, S. & Vincent, J. L. Should all patients with sepsis receive anticoagulation? Yes. Intensive Care Med 43 , 452–454 (2017).

Saber, W. et al. Risk factors for catheter-related thrombosis (CRT) in cancer patients: a patient-level data (IPD) meta-analysis of clinical trials and prospective studies. J. Thromb. Haemost. 9 , 312–319 (2011).

Journeycake, J. M. & Buchanan, G. R. Thrombotic complications of central venous catheters in children. Curr. Opin. Hematol. 10 , 369–374 (2003).

Lamperti, M. et al. International evidence-based recommendations on ultrasound-guided vascular access. Intensive Care Med 38 , 1105–1117 (2012).

Journeycake, J. M. & Buchanan, G. R. Catheter-related deep venous thrombosis and other catheter complications in children with cancer. J. Clin. Oncol. 24 , 4575–4580 (2006).

Deitcher, S. R., Gajjar, A., Kun, L. & Heideman, R. L. Clinically evident venous thromboembolic events in children with brain tumors. J. Pediatr. 145 , 848–850 (2004).

Mandel-Shorer, N., Tzvi-Behr, S., Harvey, E. & Revel-Vilk, S. Central venous catheter-related venous thrombosis in children with end-stage renal disease undergoing hemodialysis. Thromb. Res 172 , 150–157 (2018).

Chen, D., Långström, S., Petäjä, J., Heikinheimo, M. & Pihkala, J. Thrombin formation and effect of unfractionated heparin during pediatric cardiac catheterization. Catheter Cardiovasc Inter. 81 , 1174–1179 (2013).

Reynolds, P. M. et al. Evaluation of Prophylactic Heparin Dosage Strategies and Risk Factors for Venous Thromboembolism in the Critically Ill Patient. Pharmacotherapy 39 , 232–241 (2019).

Diamanti, A. et al. Prevalence of life-threatening complications in pediatric patients affected by intestinal failure. Transpl. Proc. 39 , 1632–1633 (2007).

Download references

This study was supported by the Fundamental Research Funds for the Central Universities [grant numbers YCJJ20230244] and Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology Research Fund [grant numbers 2022C09].

Author information

Authors and affiliations.

Department of Nursing, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, China

Maoling Fu, Qiaoyue Yang, Yaqi Yu, Wenshuai Song, Xiuli Qin, Ying Luo, Xiaoju Xiong & Genzhen Yu

School of Nursing, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, China

Maoling Fu, Quan Yuan, Qiaoyue Yang, Yaqi Yu & Wenshuai Song

You can also search for this author in PubMed   Google Scholar

Contributions

GY and YL framed the review questions on the basis of input from MF and QY. YY and XQ conducted the literature search. MF, WS, and QY screened and evaluated the identified papers. GY and YY performed data extraction and analysis. MF, WS, XQ and QY prepared the initial manuscript with revisions and comments from GY, YL, and XX. All authors approved the final manuscript as submitted and agreed to be accountable for all aspects of the work.

Corresponding author

Correspondence to Genzhen Yu .

Ethics declarations

Competing interests.

The authors declare no competing interests.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary checklist, supplemental digital tables1, supplemental digital tables2, supplemental digital, rights and permissions.

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ .

Reprints and permissions

About this article

Cite this article.

Fu, M., Yuan, Q., Yang, Q. et al. Risk factors and incidence of central venous access device-related thrombosis in hospitalized children: a systematic review and meta-analysis. Pediatr Res (2024). https://doi.org/10.1038/s41390-024-03225-0

Download citation

Received : 06 October 2023

Revised : 18 March 2024

Accepted : 25 March 2024

Published : 17 May 2024

DOI : https://doi.org/10.1038/s41390-024-03225-0

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

Quick links

• Explore articles by subject
• Guide to authors
• Editorial policies

An official website of the United States government

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

• Publications
• Account settings

Preview improvements coming to the PMC website in October 2024. Learn More or Try it out now .

• Advanced Search
• Journal List
• HHS Author Manuscripts

Systematic reviews and meta-analyses in the health sciences: Best practice methods for research syntheses

The journal Social Science & Medicine recently adopted the Preferred Reporting Items for Systematic Reviews and Meta-Analyses ( PRISMA; Moher et al., 2009 ) as guidelines for authors to use when disseminating their systematic reviews (SRs).

After providing a brief history of evidence synthesis, this article describes why reporting standards are important, summarizes the sequential steps involved in conducting SRs and meta-analyses, and outlines additional methodological issues that researchers should address when conducting and reporting results from their SRs.

Results and conclusions:

Successful SRs result when teams of reviewers with appropriate expertise use the highest scientific rigor in all steps of the SR process. Thus, SRs that lack foresight are unlikely to prove successful. We advocate that SR teams consider potential moderators (M) when defining their research problem, along with Time, Outcomes, Population, Intervention, Context, and Study design (i.e., TOPICS + M). We also show that, because the PRISMA reporting standards only partially overlap dimensions of methodological quality, it is possible for SRs to satisfy PRISMA standards yet still have poor methodological quality. As well, we discuss limitations of such standards and instruments in the face of the assumptions of the SR process, including meta-analysis spanning the other SR steps, which are highly synergistic: Study search and selection, coding of study characteristics and effects, analysis, interpretation, reporting, and finally, re-analysis and criticism. When a SR targets an important question with the best possible SR methods, its results can become a definitive statement that guides future research and policy decisions for years to come.

1. Introduction

Because they organize information from multiple studies on a subject, systematic reviews (SRs) have become an increasing important form of scientific communication: As a proportion of reports in PubMed, SRs increased over 14000% since 1987, over 500% since 2000, and over 200% since 2010. Social Science & Medicine recently adopted the Preferred Reporting Items for Systematic Reviews and Meta-Analyses ( PRISMA; Moher et al., 2009 ) reporting standards and guidelines for authors to use when developing their review manuscripts for publication. This article describes in detail why such reporting standards are important and outlines methodological issues that researchers should address when conducting and reporting results from their SRs.

To set the stage, we first briefly review the history of the practice of research synthesis. Then, we offer a series of recommendations for future SRs based on our experience conducting, peer reviewing, and editing SRs. We also (a) recommend improvements for some conventions, such as a priori consideration of potential moderators that may be associated with different results; and (b) compare the PRISMA reporting standards with a popular measure of the quality of meta-analysis, the AMSTAR 2 ( Shea et al., 2017 ), concluding that simply reporting methods according to the PRISMA checklist is no guarantee of a high quality SR. Although we (c) make best practice suggestions for methods involved in systematic reviewing across all disciplines, in this article, our focus is on the health sciences. We place special emphasis on the notion that a SR should take a team approach – to be rigorous and useful to the field, reviews must have authors with substantive and methodological expertise and many tasks should be performed in duplicate to reduce errors. In writing this article, our hopes are to empower readers to determine whether a particular SR “embodies ‘megaenlightenment’ or ‘mega-mistake’” ( Nakagawa et al., 2017 ), to enable research syntheses of the highest standards to appear in publications, and to spur better scientific knowledge, along with meaningful improvements in practice. Although the examples we use are drawn from health-related literature, the methods are drawn comprehensively from across science.

2. Meta-analysis is the ‘original big data’

In essence, SRs compile the results of two or more independent studies on the same subject. SRs may or may not have quantitative component to summarize the outcomes of the studies reviewed (viz. meta-analysis). In conventional practice, the term meta-analysis is often used to presume that the evidence has first been systematically retrieved and reviewed. (Similarly, meta-analytic methods may pool any two or more studies, without otherwise being systematic about gathering comparable other studies.) As another form, meta-syntheses integrate qualitative information gathered from multiple studies of the same phenomenon. In turn, meta-reviews (viz. scoping reviews, overviews) are reviews of reviews. Whether a SR, meta-analysis, meta-synthesis, or meta-review, all SRs are a form of evidence or research syntheses.

Although many narratives on the history of systematic reviewing call attention to Karl Pearson’s ( Simpson and Pearson, 1904 ) integration of correlations pertaining to the effectiveness of a typhoid vaccine as the original SR with meta-analysis, this practice has a much longer history. In fact, pooling data is a central concept of Bayes’ ( Bayes et al., 1763 ) theorem, such that a prior based on previous observations improves the prediction of a future outcome. Unfortunately, this strategy was not routinely applied until the early 1800s and was not applied rigorously or widely until the computer age, when modern software made it more feasible.

In terms of actual pooling of data from independent studies, historian of statistics Stephen M. Stigler (1986) documented that the practice appeared as early as 1805; astronomers, for example, gained additional precision by pooling their separate observations of the same cosmic events. This practice of quantitatively pooling study results was not termed meta-analysis until Gene V. Glass (1976) coined it. Given (a) that the practice of data pooling has existed for far longer than even primitive electronic computers have existed, and (b), that the essence of SRs has been to pool all available data on the subject (i.e., to be the biggest database yet available), it is clear that meta-analysis is the original big data.

Although one might suspect that the health sciences first recognized the importance of the SR process, ironically, as Chalmers et al. (2002) documented, the first may well have been the 19th century physicist, John William Strutt, a Nobel Prize winner who wrote:

“If, as is sometimes supposed, science consisted in nothing but the laborious accumulation of facts, it would soon come to a standstill, crushed, as it were, under its own weight … [what ] deserves, but I am afraid does not always receive, the most credit is that in which discovery and explanation go hand in hand, in which not only are new facts presented, but their relation to old ones is pointed out” ( Strutt, 1884 , p. 20)

Whereas some conceptual and computational strategies were available in the 19th century for reviews, methodological training of the time was starkly unaware of the problems of poor SR methods. Thus, Strutt had a good idea but lacked the methods to generate high quality SRs, and the practice of reviewing scientific literature remained a dubious one that permitted personal biases to interfere with accurate conclusions. Consumers of literature reviews tended to trust those presented by scholars who had conducted many studies in a domain, and, who, not surprisingly, usually confirmed not only the trends in their own studies but also their own pet theories. Thus, these were reviews, but they were rarely if ever systematic, at least not until late in the 20th century, when following Glass’s (1976) lead, scholars increasingly became aware that reviewing of scientific studies is itself a scientific method that must be carefully applied and made transparent enough that independent scholars can judge the validity of the conclusions; indeed, transparency is necessary in order to replicate these results. In short, accurate, comprehensive pooling of results offers the potential for increasingly better understanding of a given phenomenon. It means, for example, that the conditions when treatments can improve health outcomes will become better known, or that the plausible causes of a health condition or outcome will be better known.

As the 20th century advanced, the statistics behind pooling evidence became increasingly sophisticated, and, with computers to aid them, pooling large numbers of studies became a routine activity, helping to ignite what Shadish and Lecy (2015) dubbed the “meta-analytic big bang.” At the same time, the numbers of studies on seemingly every subject grew exponentially, such that science often seemed to be crushed under its own weight, to borrow Strutt’s words. In essence, some scientific phenomena may be too popular subjects among scientists for their own good. And, specific phenomena may comprise such a large mass of data that they are humanly impossible to cumulate into a comprehensive SR without using quantitative strategies ( Johnson and Eagly, 2014 ).

This narrative should not be taken to imply that SRs’ success as a scientific strategy has been linear and without controversy. To be sure, controversies emerged, although it is intriguing that the most forcible resistance emerged from scholars who found their own pet hypotheses challenged by upstart SRs ( Chalmers et al., 2002 ). In recent decades, scholars have jumped on the SR bandwagon under the recognition that, by pooling data from all relevant independent studies, the promise is that human welfare and activities will improve. Unfortunately, some commercial entities have also attempted to use the popularity of SRs as a tool to improve their own economic bottom lines; for example, a recent meta-review documented that meta-analyses with industry involvement (e.g., authors on the staff of big pharmaceutical companies) are massively published: 185 meta-analyses of antidepressant trials appeared in a seven-year stretch ( Ebrahim et al., 2016 ) and they seldom reported caveats about the drugs’ efficacy ( Ebrahim et al., 2016 ; Ioannidis, 2016 ).

Aided in large part by the exponential growth in SR and meta-analytic software ( Polanin et al., 2017 ; Viechtbauer, 2010 ), it is now possible for authors to generate meta-analytic statistics quickly and efficiently. Yet, it is clear that doing so is no guarantee of a rigorous and trustworthy SR. Indeed, conducting and publishing a SR with less than optimal methods may actually worsen human welfare ( Ioannidis, 2010 ). Importantly, rigorous SRs are in the best position to identify subjects for which the evidence is the thinnest. Thus—at their best—SRs help to focus resources on the most needed new research. In short, science crucially needs strong SRs.

3. Assumptions involved in systematic reviews

The foregoing history did not detail specifically how it is that poor rigor may undermine SRs. To put these assumptions into perspective, we describe the basics of systematic reviewing, characterized as seven main steps, which Fig. 1 briefly summarizes. (For more nuanced discussions or tutorials see, e.g., Borenstein et al., 2011 , Card, 2015 , Higgins, 2008 , Siddaway et al., 2019 ). This section outlines the assumptions involved in systematic reviews by organizing them by step of the SR process. First, the steps are highly synergistic: As Fig. 1 depicts, rigorous work in the earlier steps facilitates better progress on the later steps as well as better, more transparent reporting. Importantly, if early steps of the process are undertaken with low rigor, then a systematic review’s conclusions ought to be regarded with considerable suspicion. Similarly, SR teams often refine their methods as the process continues, which entails returning to repeat earlier steps of the process until the SR is completed with sufficient quality. Table 1 summarizes the advice that we provide in the remaining sections of this article.

The meta-analysis process depicted in seven steps that build on each other and that sometimes must be repeated as feedback learned during the process emerges.

Methodological steps necessary to conduct systematic reviews (SRs), along with best-practice recommendations (the text expands on these points).

3.1. Formulating the research problem

In Step 1, the SR team formulates the research problem, a step that relies on the SR team members’ understanding of the literature from both a substantive and methods (including statistical assumptions) perspective. Importantly, if the team lacks this strong conception, then the SR will not be worth doing. Note that Step 1 is crucial from a practical standpoint: The broader the research problem, the more resources will be needed to complete the review in a reasonable time-frame. Thus, a poorly framed SR may amount to a waste of valuable resources. Such a SR may become so cumbersome that it may not even be completed at all–or may take so long to complete that the literature is irrelevant by publication. Reviewing prior SRs may help to guide the team by identifying the more interesting questions that a new SR could address. Doing so also ensures that the new SR is worth the massive resources that are often needed for it to be rigorous and comprehensive. Importantly, such an initial scoping search of the existing synthesis literature may demonstrate whether a new SR should even be conducted (e.g., because past SRs lacked rigor or are out of date) or may highlight how a different focus could enhance the existing literature base. The SR team should search online registries (e.g., Prospero, Cochrane Collaboration, Campbell Collaboration), evidence gap maps, and individual journals relevant to their discipline to see if similar SRs are already in progress or have been conducted and whether these address the same questions. If a team judges extant or ongoing SRs as rigorous and as having addressed the most important questions, then there is no reason to continue. In contrast, if there is a strong need for a new SR and the team has the necessary conceptualization, then the remaining steps of the process may also be formalized in advance and with high rigor. As part of this process, authors of the proposed review should describe existing review literature and how this new review would build upon those previous reviews prior to embarking on a new review, a section that could then be incorporated in the final review manuscript.

When the SR focuses on a particular treatment, formulating a research problem routinely utilizes a form of the PICOT, PICOS acronyms ( Haynes, 2006 ), or, as we introduce, the TOPICS + M acronym:

T: Time may concern the period when studies are of interest (e.g., if the SR team is conducting a rapid review, typically only focusing 12 or 18 months’ worth of studies) or may involve questions around duration of effects (e.g., interventions immediate effect versus longer-term outcomes). It also may concern how an effect is changing over time for a particular population (e.g., that perceived racism’s linkage to depression is worsening with time).

O: The outcome is the measure or measures used to evaluate the impact of the intervention (e.g., depressive symptoms). To prevent reporting bias, it is ideal to pre-specify eligible outcomes with as much detail as possible: Depending on the focus and scope of the review, it may also be necessary to detail the type of measure of the outcome that is relevant for the review (e.g., only standardized measures of the outcome, such as objective measures of blood pressure).

P: SRs must address the population to which the chosen topic is most relevant, which may vary from clinical diagnoses (e.g., patients with high levels of depressive symptoms) to ethnic and/or racial groups. We recommend also specifying the geographic locale(s) of interest (e.g., Latinx who have recently immigrated to the U.S.) as it will affect the choice of literature sources, especially the location of grey literature.

I: If the SR concerns treatments, then interventions refers to the particular type of treatment provided to samples from the population (e.g., mindfulness training to reduce stress).

C: The comparison references the standard to which the intervention is compared. Treatment studies routinely include control groups such as standard of care or placebo. Alternatively, a treatment study may compare multiple treatments or may instead focus on one treatment and use baseline levels of a condition to examine how much the condition changes over time.

S: Study design refers to the method used to evaluate the phenomenon in question, such as randomised controlled trials (RCTs), uncontrolled trials, and cohort or cross-sectional studies. SRs often focus on study design types that are considered high quality (e.g., only RCTs), but there is a strong caveat to this focus. Because it is usually not clear what makes for the best possible study for a given problem, and the most ecologically valid studies may in fact suffer flaws, methodologists ( Johnson et al., 2015 ) have recommended broader samples of designs, so that it can be determined whether, in fact, results differ between contrasting study design types. (We return to this issue in Step 3.) If the goal is a meta-review (a review of reviews), then the type of review targeted should be stated (e.g., meta-analyses, meta-syntheses).

M: Moderators (also known as effect modifiers ) are factors that the SR team evaluates for their potential associations with outcomes. We add “M” because it is nearly always the case, even in medical literature ( Higgins et al., 2003 ), that studies exhibit significant levels of heterogeneity; that is, treatment studies’ effects routinely vary more than sampling error expects them to vary. The existence of heterogeneity implies the presence of one or more factors that make study effects vary and that may be present in some of the included studies or differ across included studies. Consequently, it nearly always benefits a SR team to consider what factors may make their effects larger or smaller, more positive or more negative. Of course, to the extent that there is variation in the TOPICS elements, these too can serve as moderators. If moderators are included in a SR, then they are often added as secondary questions to be answered, but we recommend pre-specifying them at the protocol stage and thus we have incorporated the “M” element into the standard PICOS framework. Finally, we recommend that, even if the SR team has no a priori expectations about moderators, they should still specify “M” elements; instead these can act as sensitivity analyses (e.g., to show that an intervention is equally efficacious across men and women or across racial categories). For example, Lennon et al.’s (2012) SR focused on HIV prevention interventions for heterosexual women; this SR team theorized that intervention success hinges on levels of depression the women have at the outset of the trial (we return to this example, below). Notably, this factor had never been considered in prior meta-analyses of HIV prevention trials, making it a hidden moderator ( Van Bavel et al., 2016 ).

TOPICS + M affords a linguistic advantage as well as a methodological one, as literally SRs must define their topic of focus (plus moderators), although original sources on PICO imply theorizing about moderators (e.g., Haynes, 2006 ). Note that, as we elaborate below, some methodological quality scales also give credit to SRs that make a priori attempt to determine causes of heterogeneity (e.g., Shea et al., 2017 ). Numerous other definitions similar to TOPICS + M exist, and we summarize some prominent ones in Table 2 ; some of these are useful for non-intervention study designs. Whether a formal TOPICS + M statement is used or an alternative takes its place, Step 1 should lead the SR team to as nuanced a statement of the research problem as possible. The work in this step sets the stage for the remaining steps. For example, it helps the team to identify the best literature search terms and literature sources (e.g., especially grey literature sources that involve using sources other than electronic databases), to use more systematic searching to determine which studies will be included (and excluded) in the review, and to determine which features of the studies most need coding.

Selected popular structures for formulating questions systematic reviews as well as the currently proposed one, TOPICS + M.

NA = Not applicable to structure.

As part of completing Step 1, the SR protocol should be drafted for pre-registration, which details planned aims and analyses, before proceeding further (e.g., with PROSPERO or the Open Science Framework [OSF]). Doing so conveys a certain degree of trustworthiness to the SR, especially because systematic reviewing is essentially retrospective in nature ( Ioannidis, 2010 ): Review authors are often already aware of the individual findings of many of the studies they plan to include. Thoroughness may also help ward off competition from other teams: It is up to the SR team to be thorough and thoughtful in specifying primary and secondary research questions, inclusion/exclusion criteria, and planned analyses. Thus, the SR team should engage in most of the background work and piloting of screening and coding forms prior to beginning the review process, following contemporary protocol guidelines ( Lasserson et al., 2016 ; Moher et al., 2015 ) prior to registration. Accordingly, peer review prior to publication should also compare the submitted SR with any available pre-registered protocol.

3.2. Finding and selecting studies

Searches and selection criteria..

In Step 2, SR teams conduct systematic literature searches to find as many qualifying studies as possible ( Nakagawa et al., 2017 ). If the research statement is well developed, then, as noted, it economizes the search for studies and the process of determining which located studies match inclusion criteria and qualify for the review. Of course, there are exceptions to this rule: In the case of Lennon et al.’s (2012) HIV prevention SR, it was necessary to retrieve far more full-text reports for inspection, as these scholars realized that trials may not report mental health dimensions in their titles and abstracts.

Similarly, the rationale for each selection criterion should be clear from the background to the review and should also be carefully justified when reporting the SR. For example, review authors often decide to base inclusion criteria around the study design of included studies (e.g., including only RCTs), an approach that may not always be warranted: In every situation, the SR team must consider the review question and use knowledge of the existing body of evidence when making decisions about how to include and combine different study designs and make these assumptions clear to the reader.

Grey literature.

A number of disciplines have well-documented reporting biases such that studies with significant and positive effects are more likely to be submitted for publication and eventually published, and, as a result, the synthesis outcome is somewhat inflated ( Dwan et al., 2008 ; Easterbrook et al., 1991 ; Fanelli, 2012 ; Franco et al., 2014 ; Munafo et al., 2009 ; Polanin et al., 2017 ). Consequently, it has become a standard of high quality SRs to search for grey literature, unpublished work relevant to the defined TOPICS + M. Although some authors argue that grey literature is unpublished because it has lower quality, unless unpublished studies are gathered, it cannot be known whether in fact unpublished studies have low rigor; as well, in many disciplines, much published research has low rigor and, often, unpublished studies (e.g., theses, dissertations) have high quality ( Hanna, 2015 ). Thus, authors should attempt to be as comprehensive in their searches as possible and to search for and include grey literature, although how exhaustive the search is will depend largely upon available resources, and even when using comprehensive search practices, it is likely reviews still lack some relevant studies ( Delaney and Tamás, 2018 ). The scope of the literature search should be addressed in the conclusions and limitations section of the SR, especially if given decisions made by the review team, it is likely there is missing and relevant literature to the review; only review teams with incoming substantive expertise of the field can estimate how likely bias would be, and teams should always critically reflect on potential implications of their decisions.

Mono-language bias.

It may come as a surprise to native-English speakers, but much scientific literature appears in other languages: To the extent that studies appear in multiple languages, mono-language searches or inclusion criteria based on language of publication are a weakness, as there is no guarantee that results replicate across the cultures where these studies were conducted. For example, a meta-analysis found that a particular gene association reversed when Chinese literature was retrieved and compared to studies reported in English ( Pan et al., 2005 ). Similarly, a meta-review of 82 meta-analyses revealed that behavioural studies conducted in the U.S. achieve larger effect sizes than in other countries, whereas there was no difference for non-behavioural studies ( Fanelli and Ioannidis, 2013 ). SR teams should consider whether their research question is likely to have publications in other languages and may need to adapt their scope if they decide not to include non-English literature. In general, as with the collection (or not) of grey literature, SR teams should reflect on this decision and detail whether it influences the findings.

Necessary teamwork.

No matter how refined a research statement is, literature searches are nearly always highly time- and energy-dependent. Because the number and diversity of databases quickly changes, it has increasingly become a hallmark of high quality SRs that their teams include librarians and information retrieval specialists who are trained in systematic literature retrieval; doing so routinely improves and economizes searches. Team members should screen abstracts independently and in duplicate. If reliability is high on a randomly selected portion of the potential studies, then it is permissible to do single screening, which saves resources: The AMSTAR 2 tool suggests that there should be a kappa score between two independent reviewers of 0.80 or greater to ensure trustworthiness of decisions ( Shea et al., 2017 ). Although there has been no set standard for how large of a randomly selected portion to use, based on our experience, we would suggest examining the first 10% and then every 10% thereafter until high agreement (i.e., benchmark suggested by AMSTAR) is reached. Nonetheless, in some cases, this process may simply lead to the realization that all tasks must be done in duplicate in order to ensure trustworthiness of the findings. Another option is to use a liberal screening approach such that one person screens all potentially eligible studies and a second screener only reviews those that the first person excluded, to ensure the search does not exclude anything that should be included. We have emphasized the teamwork involved in successful SRs; committed, honest team members help to correct each other’s mistakes. Practically speaking, without duplicate effort, there is no way to report reliability.

The future of report screening.

In recent years, computer scientists are have developed automated search engines and screening tools ( Carter, 2018 ; Marshall et al., 2018 ; Przybyla et al., 2018 ; Rathbone et al., 2015 ). Following human interaction to train them, which again should utilize duplication of effort, these widgets then make fast work of abstract selection. One problem with this method is its technical demands ( Paynter et al., 2017 ). Another cost- and time-efficient method is crowdsourcing selection of literature using online workers ( Mortensen et al., 2017 ).

3.3. Coding studies for substantive and methodological features

A strong formulation of the research problem leads to coding methods that capture the most interesting aspects of the studies, those that the SR team expects will moderate effects. For example, SRs of interventions commonly examine which behaviour change techniques were tapped in an effort to improve participants’ health ( Michie et al., 2013 ); the dosage of treatment is another common dimension. A coding formulation should not overlook (a) items that tap risk of bias and/or methodological quality; and (b), descriptive features of studies that will accurately depict the underlying characteristics of included studies and may be helpful for post hoc observations. As much as possible, the authors should specify items for data extraction prior to reviewing the studies; then, they should pilot data extraction forms to ensure coders understand the coding dimensions and requirements. Some frameworks that could be useful tools to create a detailed data extraction form include PROGRESS-Plus for variables related to equity issues between/among included populations ( O’Neill et al., 2014 ), TIDieR for key intervention components ( Cotterill et al., 2018 ; Hoffmann et al., 2014 ) or the Behaviour Change Taxonomy and its associated coding system ( Michie et al., 2013 ), and the clinicaltrials.gov checklist for a fully-reported outcome measure (e.g., see https://prsinfo.clinicaltrials.gov/data-prep-checklist-om-sa.pdf ).

It is helpful to include a codebook along with the standardized form that details key definitions or operationalisations the coders will need to ensure accurate coding; software facilitates the process, especially with large SRs. As with screening, it is best to double-code independently and in duplicate and resolve discrepancies as they arise. Review authors should also consider what items may need to be coded qualitatively (e.g., using direct quotations from reports) and what content should be coded in a summary manner (e.g., using categorical response options). These strategies should be evaluated ahead of time by trying the dimensions on small (but diverse) samples of studies. Meeting often throughout this part of the review process will help ensure that data extraction mistakes are caught and addressed quickly. Finally, advances in artificial intelligence are beginning to economize these operations ( Sumner et al., 2019 ).

In science, generalizations of new knowledge depend on having valid observations in the first place, and it is well known that people routinely use misleading or false information (cf. De Barra, 2017 ; Ioannidis, 2017 ). Thus, if the methodological quality of studies is assessed, then the SR team has a means to determine if conclusions of the SR hinge on the inclusion of studies that lack rigor. Unfortunately, other than psychometric artefacts, which attenuate effect sizes ( Schmidt and Hunter, 2015 ), and bias introduced with p -hacking or selective reporting ( Kirkham et al., 2010a , b ; Rubin, 2017 ; Williamson et al., 2005 ), which tend to exaggerate effects, little is known about the effects of other methodological problems, such as confounds between groups randomly assigned to condition ( Higgins, 2008 ; Johnson et al., 2015 ; Valentine, 2009 ). Thus, it should be evident that even if a SR team limits a review to RCTs, there can still be flaws in individual RCTs due to individual research decisions or mistakes (e.g., poor blinding or allocation concealment, inadequate randomisation, other biases introduced during the process).

The SR team should use the most appropriate methodological quality instrument for the type of study designs that appear in their SR; multiple scales may be needed if the study design inclusion criteria are broad. It is best not to design new scales for this purpose, as there are many standardized and appropriate tools for various study designs and for treatments in particular disciplines. Even as early as 2003, there were nearly 200 tools available to evaluate non-randomized designs ( Deeks et al., 2003 ), let alone quality inventories for RCTs. Tools recommended for SRs focused on outcomes from intervention studies include the Cochrane Risk of Bias for RCTs ( Higgins et al., 2011 ) or the ROBINS-I tool for non-randomized studies ( Sterne et al., 2016 ) as these have been created to address the major issues relating to risk of bias in primary studies and, as such, are considered the gold-standard of risk of bias tools and are widely used.

Many meta-analyses create summative scales from the individual items assessing methodological quality, but this practice makes the operational specificity assumption, which is that the items have units that bear with equal weight, when it is possible that even one crucial flaw could invalidate a study ( Valentine, 2009 ). Thus, it is best to examine whether the specific, individual defects undercut conclusions ( Higgins, 2008 ; Johnson et al., 2015 ). If an intervention effect appears only in the weakest studies, then it undermines any conclusion that the intervention is efficacious. Thus, results from this analysis of risk of bias must be incorporated into the discussion of evidence of the effects. A tool to assist in this regard is the GRADE approach ( Balshem et al., 2011 , Canfield and Dahm, 2011 ), including the GRADE CerQUAL for qualitative meta-syntheses ( Lewin et al., 2015 ), which provide guidance for systematic assessment of the outcome in light of the quality of the evidence in the review and the potential areas of bias among individual studies. The GRADE approach emphasizes ratings based on studies’ limitations, consistency, and precision of outcome findings (i.e., directness of how well a study addressed questions of interest, and publication bias) and results in high, moderate, low or very low ratings of the quality of evidence for a particular outcome. Obviously, GRADE emphasizes exactly the factors that SR teams should be integrating in their SRs. Yet, without standardized tools such as GRADE, there is little systematic guidance for the presentation of conclusions and the result is that stakeholders could easily misread reports due to lack of precision (or even in some cases subjective reporting) on the part of the researcher ( Knottnerus and Tugwell, 2016 , Lai et al., 2011 ; Yavchitz et al., 2016 ). Thus, while the rigor of GRADE depends on how well it is applied, without such a standardized system, review authors can allow bias (e.g., significant results) to interfere with nuanced and accurate reporting.

3.4. Calculating effect sizes

SRs qualitatively or quantitatively pool results. A meta-analysis puts study outcomes on the same metric to pool results; effect sizes may examine associations between variables, mean levels of a phenomenon, or both. In a SR without meta-analysis that focuses on outcomes, qualitative descriptions of results replace the pooling of effect sizes, but authors should also present single effect sizes for each study or the available quantitative results from the reports.

Calculating effect sizes is a central problem when examining treatments with continuous outcomes (e.g., quality of life, depression scores, substance use), which generally requires use of the standardized mean difference ( SMD ) due to multiple continuous outcomes and a variety of measurement tools. In contrast, this issue is much less problematic in studies using dichotomous outcomes (e.g., abstinent, mortality rate) or those with correlational data. As we noted, the SMD especially becomes more complicated because of the myriad of details that studies report (or fail to report). One initial step is to reach out to authors to ask them to provide the team with the necessary data. Yet, when primary study authors are unavailable or unresponsive to requests for data, SR teams may not think creatively enough about what to do when the standard statistics, such as means and standard deviations, seem to be missing in study reports. There are many ways to get around poor primary study reporting: (a) Use figures, other results or test statistics (e.g., results from an F test; see the online effect size calculator ( Wilson n.d. )). (b) Extrapolate a matrix of correlations from structural equation results ( Kenny, 1979 ). (c) Estimate the mean and variance from the median, range, and sample size ( Hozo et al., 2005 ). (d) In studies that use analysis of variance (ANOVA), even if the report omits the critical comparison of conditions, it is still possible to determine a pooled standard deviation if other F tests are provided, along with relevant means; thus, if the two means to compare are available, an effect size may be calculated ( Cohen, 2002 , Johnson and Eagly, 2014 ). When in doubt, it is always advisable to consult with a statistician.

The foregoing strategies reveal that there are many ways to calculate effect sizes. When studies report multiple means of calculating effect sizes, all that have the same level of inferential information should be used (e.g., means and standard deviations and t or F tests), to triangulate on the best estimate. Continuous information is more accurate than categorical information; thus, means (e.g., levels of life satisfaction, depression) are better than count information (e.g., proportions recovering from illness). It is also preferable to re-calculate the effect sizes, rather than rely on the original study authors’ calculations, to make sure that the effect sizes and their signs are accurately coded. Note that including studies with weaker statistics is still considered preferable to dropping studies for being relatively inaccurate. Again, consulting with a statistician about problematic cases is always advisable. Then, sensitivity analyses evaluate whether these particular estimates are outliers or unduly influence the model.

After the calculation of effect sizes, the next consideration prior to analysis is whether the effect sizes need to be adjusted for small sample bias (which is conventionally the case for the SMD ), or to a metric that addresses coefficients with undesirable statistical properties (e.g., in the case of correlations transforming to Fisher’s z ), as well as how to identify and handle outliers. Additionally, once the chosen effect size is calculated, attention to the calculation of its standard error is the last important consideration before analysis; the standard error for each effect is an estimate of the degree of sampling error present and a variation of it is used as weights in meta-analytic statistics ( Borenstein et al., 2011 , Hedges and Olkin, 1985 ; Lipsey and Wilson, 2001 ). Note that the standard error is a gauge only of the potential for sampling error and not of other types of error (e.g., low methodological rigor; for more detail on such matters, see sources cited in Table 1 ).

3.5. Analysing the SR database

Analytic assumptions..

In classical meta-analysis, non-independence among studies in a review is another problem that might need to be handled at multiple stages in the review process. The first occurrence is when the effect sizes are calculated because non-independence, if ignored, can result in inaccurate study weighting. Non-independence may result because (a) studies have more than one relevant measure of an outcome; (b) because a comparison group is used more than once in calculating effect sizes (e.g., multiple treatment arms compared with a single control group); or (c) when primary studies do not appropriately adjust for clustering within their own sample (i.e., in the case of a cluster-RCT). Thus, it is important to attend to dependency by first examining how clustering was handled in any of the primary reports utilizing cluster-RCTs: Adjustments to the effect size standard error should be made if primary study authors did not attend to this issue ( Hedges, 2007 ). Regarding non-independence among studies in the sample due to multiple effect sizes per study, if there are sufficient studies, it is best to model the non-independence (e.g., using robust variance estimation (see De Vibe et al., 2017 ; Hedges et al., 2010 ); another approach is structural equation meta-analysis ( Cheung, 2014 )), rather than selecting one effect size or averaging similar effect sizes for each study. These strategies are particularly valuable when there are large numbers of studies available and are preferable to simply averaging effect sizes within studies, a solution that can lose valuable information. Whichever method is chosen to address non-independence among samples in the review, it should be documented.

Analyses should be conducted appropriate to the questions and the literature base. In the health sciences, research questions often address diverse samples, treatments, or environments, all of which may contribute to inconsistencies in results across studies. Consequently, models that follow random-effects assumptions are relatively conservative but typically the most appropriate modelling choice given the likely variability in health sciences questions, interventions, and populations, although there are cases where fixed-effect assumptions are better ( Borenstein et al., 2011 ). Others have argued for the use of unrestricted weighted least squares meta-regression to account for heterogeneity ( Schmid, 2017 , Stanley and Doucouliagos, 2015 ). Given the diversity of options available to synthesis authors, modelling choices should be explicitly stated, and justification should be provided. Yet, despite an increase in power due to pooling studies, many meta-analyses using random-effects models are in fact underpowered to detect effects because of the parameters needed to estimate between-study variance. Thus, power analysis should be considered in advance of undertaking a review of diverse literature. Additionally, any primary analysis must incorporate an assessment of heterogeneity and should consider publication or other reporting biases that could exist alongside substantive moderators.

In a SR without meta-analysis, qualitative descriptions of results replace analyses. SR teams may choose to omit a meta-analysis for several reasons: (a) if the literature is new and small (one cannot do a meta-analysis when there is only one study); (b) studies in the literature lack rigor; (c) the included studies are extremely different and the team has no hypothesized potential moderators; (d) if the review question centres on processes, theory development, research of a qualitative nature; or (e) the SR team lacks the statistical expertise to conduct a meta-analysis. It is important to note that simply lacking the expertise to conduct a meta-analysis is not adequate justification for conducting an outcomes-focused narrative review of a field that has sufficient primary study evidence for a quantitative analysis. Teams without such expertise should identify this need at the outset and set the scope of their SR accordingly, or, enlist the help of an expert meta-analyst early in the process. Authors must also carefully reflect on the potential study designs and areas for heterogeneity at the start of the review process (i.e., during protocol development) and attempt as much as possible to determine ahead of time whether a meta-analysis (versus a narrative review) will be possible. If, during the course of the review, the authors realize that a pooled synthesis is not possible or would not provide clear answers, authors may decide that it is better to map the existing research eligible for the review, rather than focus on outcome (effectiveness) data; however, this post hoc decision should be transparently reported with reasonable justification and all collected outcome data should still be reported in a standardized format (e.g., as effect sizes). That is, the judgement that “the studies were too clinically heterogeneous to combine” is an insufficient rationale for not conducting a meta-analysis. As with all research designs, the best design for the question of interest should be utilized in a synthesis endeavour, given the resources available.

The decision to conduct a meta-analysis–or a qualitative analysis–is one that review authors must consider and justify given the research question and scope. The larger the literature is, the more the risk rises that a reviewer may take shortcuts that reduce the accuracy of the conclusions reached; thus, a SR may fall prey to the same “cherry-picking” problems that vexed reviews before meta-analysis became commonplace: Selecting studies whose results support the SR author’s biases and views. Thus, SR teams need to erect barriers that prevent selectively presenting study findings.

Moderators of effects.

The fact that, in nearly every literature, study results are routinely highly variable led us to advocate in TOPICS + M a prion specification of moderators for substantive dimensions (or as sensitivity analyses). Meta-analyses have the best possibility of locating cross-study inconsistencies, because quantitative indexes have been developed to gauge it (see next sub-section, Heterogeneity ). Literally, these examine whether there is more variability in effect sizes than would be expected by sampling variance alone; that is, the null hypothesis is that there is homogeneity and, when the hypothesis is rejected, the statistical inference is that heterogeneity exists ( Borenstein et al., 2011 ; Hedges and Olkin, 1985 ). Heterogeneity, in turn, implies that there is more than one population effect at work in the literature, or that a range of population effects exists. In turn, a mean effect size for such a literature is not very precise and instead, moderators should be evaluated.

Reviewers can systematically examine heterogeneity by testing moderators driven by applicable theory and relevant research. In Step 1, SR teams specify moderators chosen for the analysis; unexpected moderators tested post hoc should be so identified for readers. It is important to remember what the moderator means in the meta-analytic context and interpret results appropriately. For example, a variable such as percentage of male participants is analysed at the study, not individual level; if gender is identified as a significant moderator of trial outcomes, then, outcomes improved to the extent that studies sampled larger portions of a certain gender (e.g., females). The SR team is not justified to infer that “this intervention worked better for females” only that it worked better for studies with larger proportions of female participants. If they have sufficient numbers of studies, SR teams should evaluate moderators that are significant in bivariate models in simultaneous, multiple-moderator models. Locating patterns whereby some moderators uniquely predict effect sizes potentially reduces confounding between predictors and thus yields a clearer picture of results ( Tipton et al., 2019 ). Similarly, SR teams should examine whether results hinge, either overall or by interaction with moderators, on the inclusion of lower quality studies; findings that hinge on studies with low rigor should be interpreted with appropriate caution ( Johnson et al., 2015 ; Valentine, 2009 ).

When reporting moderators, it is helpful to depict results in either graphical or tabled form. Johnson and Huedo-Medina (2011) introduced the moving constant technique, with which analysts use meta-regression models to create graphs or tables of estimated mean effect sizes plotted against moderator values, including confidence intervals, or confidence bands around the meta-regression line. This technique can also be used to estimate mean effect size values and confidence intervals at moderator values of interest for moderators that reached statistical significance. Specifically, analysts may move the intercept to reflect interesting points along or beyond a range of independent variable values. Returning to Lennon et al.’s (2012) meta-analysis of HIV prevention interventions for women, these scholars showed that trials succeeded better for samples with higher baseline depression: As Fig. 2 shows, on average, risk reduction was large and significant for samples with the highest mean levels of depression, whereas for samples with lower levels of depression, interventions failed to change risk. (Separate analyses also revealed that reducing depression was associated with greater risk reduction.) Results presented in this form help show at what levels of a moderator an effect exists. Returning to the significant gender moderator example above, now, we can answer the question of whether, on average, trials significantly succeeded for male vs. female samples. Such estimates, in turn, can be highly informative when interpreting the nature of the phenomenon being studied in the meta-analysis, especially when a comparison to an absolute or a practical criterion is important.

Empirical demonstration of the moving constant technique: Sexual risk reduction following a behavioural intervention as a function of each sample’s baseline depression. Sexual risk behaviour declined following the intervention at the last available follow-up to the extent that samples had higher levels of baseline depression (treatment [control] group effects appear as darker [white] triangles and the size of each plotted value reflects its weight in the analysis). The solid regression line indicates trends across initial levels of depression; dashed lines provide 95% confidence bands for these trends. Reproduced from Lennon et al. (2012) .

The moving constant technique also permits analysts to estimate confidence intervals for an effect size at particular values of one or more independent variables (and thus to avoid artificially dichotomizing continuous predictor variables). In multiple moderator models, an extension of the moving constant technique is to show what average effect appears for combinations of moderators. For example, MacDonald et al. (2016) showed that resistance exercise creates large to very large beneficial effects for the systolic blood pressure of (a) non-White samples (b) with hypertension at baseline, (c) who are not taking medications, and (d) who perform eight or more exercises per session. (Similar effects emerged for diastolic blood pressure.) Another reason analysts should emphasize the confidence intervals around such point estimates is because they are more conservative. In the MacDonald example just provided, their model estimated a 1.02 standard deviation average blood pressure improvement for resistance exercise, but the 95% confidence interval on this estimate was 0.67 to 1.36, so there is a considerable range.

When literature have sufficient size and spatial variability, spatiotemporal meta-analyses can incorporate ecological- and/or community-level variables in their models ( Johnson et al., 2017 , Marrouch and Johnson, 2019 ). For example, Reid et al. (2014) examined whether the success of HIV prevention interventions for African Americans depended on levels of prejudice these people experienced. In fact, interventions failed, on average, for trials conducted in relatively prejudiced communities and succeeded better to the extent that Whites were not prejudiced. Using spatiotemporal factors along with study-level factors can help diverse disciplines to converge, a practice that ought to be encouraged for a journal like Social Science & Medicine, which prides itself on its interdisciplinary nature. Of course, as Kaufman et al. (2014) recommended, SR teams may seek guidance from experts in multiple disciplines and also pursue ecological and/or spatiotemporal models of the effects in their reviews ( Johnson et al., 2017 ), though of course these matters should be considered at the outset, in Step 1.

Heterogeneity.

Heterogeneity in a review is one of the most important areas to assess and an area that is often ignored or, even worse, misinterpreted. We have already noted that its existence sharply challenges the interpretation of an overall mean effect size (in the case of a random-effects model, it literally implies a mean of means). The real question for meta-analyses is whether there is still large or significant heterogeneity remaining after applying moderators. Conventional practice is to use and report multiple assessments including I 2 , τ 2 , and Q (and its associated p -value), as there are a number of documented limitations with individual tests of heterogeneity. For example, although in theory I 2 does not depend on the number of studies and is easily interpretable on a scale of ~0–100% ( Higgins et al., 2002 , Higgins et al., 2003 ), I 2 is a relative rather than absolute measure of heterogeneity; it also increases with the inclusion of larger samples of studies and thus may artificially increase in a particular literature over time ( Borenstein et al., 2017 ; Rucker et al., 2008 ). Additionally, research has demonstrated that in meta-analyses with smaller k, I 2 can have substantial bias ( Von Hippel, 2015 ); for this reason, reporting of the random-effects variance, τ 2 , is also recommended ( Schwarzer et al., 2017 ). Similarly, Q values also vary by choice of measure of effect; thus, directly calculating I 2 from Q yields results that deviate from the theoretically intended values ( Hoaglin, 2017 ). I 2 has a limited maximum value (100%), but is literally based on Q , which has an infinite maximum value; consequently, I 2 is not quite linear, although a convention has emerged with 25% being small, 50% medium, and 75% large heterogeneity ( Higgins et al., 2003 ). Nonetheless, analysts should keep in mind that the inference of heterogeneity is a yes or no inference based on a significant test statistic.

Another consideration with heterogeneity is whether the heterogeneity may be attributed to one or more outliers included in the SR ( Viechtbauer and Cheung, 2010 ). If removal of one or two outlying effects markedly decreases heterogeneity, then the SR team can examine these as a post hoc effort to determine whether they differ in important respects from other studies. Alternatively, it may be that some moderator should have been coded but was missed when conceptualizing the review. Finally, outlying effects can be winsorized (e.g., reduced to a reasonable magnitude compared to other effects) so that they do not unduly influence results.

Publication bias.

Publication bias (viz. small-studies or reporting bias) refers to a tendency for certain findings to be published, generally those that reach statistical significance, rather than null, findings. There are multiple tests SR teams can use to assess the potential of this bias in a review; each has limitations and therefore SR teams are advised to triangulate data from multiple assessments. Although many sources recommend examining visual plots, such as funnel plots, this practice can be quite subjective and thus more critical sources have recommended quantitative approaches, such as regression-based assessments: Tests such as Begg’s ( Begg and Mazumdar, 1994 ), Egger’s ( Egger et al., 1997 ), and PET-PEESE ( Stanley and Doucouliagos, 2014 ) examine whether there is a skew to effects (such that smaller studies exhibit larger effects). It is important to realize that such tests generally assume a single population effect size; therefore, in the face of heterogeneity, inferences of publication bias are perilous. In parallel, the existence of publication bias can artificially restrict the range of known effect sizes, perhaps even to the point of (artificial) homogeneity. It is important to note that a number of reasons may drive the appearance of an asymmetrical funnel plot, including heterogeneity due to moderators ( Lau et al., 2006 ). Thus, SR teams must carefully consider how likely it is that reporting and publication bias exists in their particular field and the likely impact it has on the review. As well, we must caution against the use of Failsafe N, as it relies on arbitrary assumptions ( Becker, 2005 ).

Many sources advocate direct tests of publication bias, such as comparing effects garnered from journal articles versus those from theses and dissertations (e.g., Card, 2015 , Johnson and Eagly, 2014 ), and we advocate the very practical strategy of examining whether null effects appear in peer-reviewed publications and whether grey literature studies routinely achieve significant effects. Contour-enhanced funnel plots can be a useful tool for this purpose, plotting effect sizes with differing symbols for publication status. Fig. 3 , Panel a, shows a literature in which study effects routinely reached statistical significance no matter whether published (solid dark blue markers) or not (grey markers). Panel b shows a literature in which the grey literature routinely does not reach significance; here is one with strong evidence of publication bias. Finally, Panel c shows a literature with heterogeneity, no matter whether grey or published studies are considered, and it is the goal of the SR team to find moderators of effects, if possible.

Contour-enhanced funnel plots showing effect sizes from three literature (a) one with no clear evidence of selective (e.g., publication) bias, as even published studies (solid triangles) commonly achieve null results and unpublished studies (hollow triangles) achieve statistically significant outcomes (this distribution is also homogeneous, τ 2 = 0.00047, I 2 = 0%); (b) one with marked evidence of selection bias, with only published studies routinely finding a significant effect and unpublished studies routinely finding non-significant effects ( τ 2 = 0.00047, I 2 = 0%); and (c), a literature with marked heterogeneity ( τ 2 = 0.0145, I 2 = 61%). The contours surrounding the null value show at which points individual effects reach significance. Effects in the white zone are statistically non-significant, where the significance level is set at p > . 05.

3.6. Interpretation and dissemination

In elaborating Steps 1 through 5, we have already offered much advice that is relevant to Step 6, interpretation and dissemination. Broadly speaking, the SR’s methodological operations need to be clearly stated and its key assumptions carefully defended. As Fig. 1 implies, performing Steps 1–5 with rigor logically eases matters of interpretation and dissemination. The review report should be considered as though it were a primary experiment: (a) What do readers need to know so that they could reproduce the results of this SR? (b) What about the methods should be related so that readers understand potential biases introduced by the review team during the review process?

Given the rising frequencies of meta-reviews, published SRs will likely be reviewed for best practice SR standards using standard review criteria (e.g., SAMSTAR 2: hea et al., 2017 ; risk of bias in systematic reviews [ROBIS]: Whiting et al., 2016 ). Because of potential feedback loops in the process ( Fig. 1 ) whereby the SR team serendipitously learns better strategies for reviewing the studies in question, it is incumbent on the team to report when, what, and why the process returned to an earlier step (e.g., because it was discovered that search terms omitted studies that would have qualified). The team must identify post hoc adjustments that occurred following registration and provide rationale for these deviations because it is not safe to assume that readers will compare a protocol to a published review. Transparency is important to reduce potential bias in reviews. For example, recent meta-reviews demonstrate that as many as 20–30% of systematic reviews are prone to selective inclusion and reporting of outcomes ( Kirkham et al., 2010a , b ; Mckenzie, 2011 ; Page et al., 2013 ; Tricco et al., 2016 ).

Although such information partially mirrors the PRISMA checklist, the checklist is formally a list of reporting standards. It does not directly necessarily address methodological quality. Indeed, Table 3 overlays the PRISMA checklist ( Moher et al., 2009 ) with an important, recently developed assessment of methodological quality for systematic reviews ( AMSTAR 2; Shea et al., 2017 ). There is only partial overlap: Thus, while authors may think they are adequately following SR standards when they “follow PRISMA standards,” they may miss some key areas for conducting a methodologically rigorous SR. Thus, we assert that merely following PRISMA reporting guidelines does not guarantee high quality.

Comparison of AMSTAR 2 with the PRISMA checklist for systematic reviews (SRs), ordered using AMSTAR 2’s items, and using the AMSTAR 2’s “yes” categories as high quality; this comparison also indicates the AMSTAR 2 items deemed critical by its authors (marked with ✪ ), as well as two that the current authors believe are also critical ( ✰ ).

Note. Omitted PRISMA items are stylistic and do not reflect on methodological features. NRSI = non-randomized studies of interventions. RoB = Risk of bias. SR = systematic review.

As Table 3 details, conclusions must take into account whether studies exhibit deficits in methodological quality. Limitations about the primary studies involved should primarily be discussed in the results section (e.g., when presenting risk of bias information) and how these influence conclusions to be made from the review or when discussing how primary study authors should do a better job of measuring and reporting certain variables (e.g., if evidence for an effect is only found in the weakest studies, then its existence is in considerable doubt). However, SR authors must also focus on the limitations introduced by the decisions they made or methods used: For example, perhaps there was not enough funding to employ a double-screener or perhaps the review team could use only one language for searches. Thus, review authors must thoughtfully consider how these decisions may (or may not) influence the implications from the review and discuss them for the reader.

Finally, in the discussion and concluding sections, audiences will benefit most if findings have been transformed into a meaningful metric as pooled effect sizes of a literature are not always intuitively meaningful in and of themselves. Thus, a SR should answer the question: What do these findings mean? For example, if the review assessed whether brief alcohol interventions were effective in reducing adolescent substance use, the outcome could be transformed into a clinically meaningful metric such as number of days’ reduction in use, rather than a standardized mean difference ( Hennessy and Tanner-Smith, 2015 ). Similarly, Macdonald et al. (2016) converted their standardized mean difference estimates for the impact of exercise into millimetres of mercury (Hg), the standard interpretive guide for clinicians. Other sources provide similar tips for presentations and help SR teams ponder their findings and what they imply for the domain being reviewed (e.g., Borenstein et al., 2011 ; Card, 2015 , Johnson and Eagly, 2014 ).

3.7. Re-analysis, development, or criticism

Some literature develop rapidly, outdating extant SRs, and making updated SRs valuable sooner. Alternatively, a SR team might theorize that dimensions that were not considered in a published SR might help explain observed heterogeneity; as long as the original SR’s methods were of high quality (see Tables 1 and ​ and3), 3 ), then the previous SR’s database, if available, may be reanalysed to evaluate these hypotheses, although of course, if the SR is dated, a new literature search should be performed along with Steps 3–7 as necessary. This recommendation takes advantage of data archiving; sharing databases is a way to save work and accelerate progress on the newest SR, but it assumes that the previous SR team has done an adequate job on their own journey through all the steps we have documented. Critics may target a SR that reaches conclusions they do not trust, but optimally, such critiques should take quantitative form, focusing on the degree to which SRs had rigor and pointing to areas where new, original research should be conducted, or, alternatively, new SRs.

To stay relevant to current conditions, many reviews should be updated regularly, depending on how rapidly new studies are emerging, what research question is addressed, and the nature of research in that scientific discipline. When to update a SR will hinge on the nature of the research question and the discipline. The Cochrane Collaboration even engages in “living systematic reviews” that are updated monthly ( Elliott et al., 2017 ), but this may not be appropriate for many types of research questions (e.g., in disciplines where treatment practices rarely change and where there have already been high quality SRs). However, evidence indicates that in some faster-moving fields (cardiovascular treatment research), 7% of reviews are out-of-date by publication and another 23% are out of date within two years of publication ( Shojania et al., 2007 ). Thus, it is up to individual review teams to have a sense of when an update will be appropriate.

4. Conclusions

In this article, we have endeavoured to provide best practice recommendations for research synthesis, which Table 1 summarizes as brief “dos and don’ts”. We highlighted a number of tools to guide researchers in the practice of research synthesis, although it is worth noting that these quality inventories have imperfections, but also represent a best-of-science approach at the present time. There is no doubt that we will see improvements emerge in the coming years and decades. As Shadish and Lecy (2015) concluded, meta-analysis is one of the central methodological developments in science, or, in their more dramatic terms, the spark of its own big bang. Nonetheless, the entire scientific community must continually take steps to ensure the highest scientific rigor in the SR process and in reporting and consuming results from systematic reviews. As we have argued in this article, when done with great rigor, SRs can yield great insights, relevant both to science and practice. They can also point the way to future studies that are optimized to fill gaps in the evidence base; in this way, well-done SRs ought to improve the efficiency of science. When a SR targets an important literature with the best known methods, the results can become a definitive statement that guides future research and policy decisions for years to come.

Acknowledgments

The U.S. National Institutes of Health (NIH) Science of Behavior Change Common Fund Program facilitated preparation of this article through an award administered by the National Institute on Aging (a sub-contract from U.S. PHS grant 5U24AG052175). We thank Rebecca L. Acabchuk, Julie M. Brisson, Chris S. Elphick, Eliza Grames, and Catherine Panter-Brick for constructive feedback on earlier drafts; we also thank Ilinca C. Johnson for generating the acronym TOPICS + M (see text).

• Balshem H, Helfand M, Schünemann HJ, Oxman AD, Kunz R, Brozek J, Vist GE, Falck-Ytter Y, Meerpohl J, Norris S, 2011. Grade guidelines: 3. Rating the quality of evidence . J. Clin. Epidemiol . 64 ( 4 ), 401–406. [ PubMed ] [ Google Scholar ]
• Bayes T, Price R, Canton J, 1763. An Essay towards Solving a Problem in the Doctrine of Chances . [ PubMed ] [ Google Scholar ]
• Becker BJ, 2005. Failsafe N or file-drawer number. In: Rothstein HR, Sutton AJ, Borenstein M (Eds.), Publication Bias in Meta-Analysis Prevention, Assessment and Adjustments . Wiley Chichester, England, pp. 111–125. [ Google Scholar ]
• Begg CB, Mazumdar M, 1994. Operating characteristics of a rank correlation test for publication bias . Biometrics 50 ( 4 ), 1088–1101. [ PubMed ] [ Google Scholar ]
• Booth A, 2006. Clear and present questions: formulating questions for evidence based practice . Libr. Hi Tech 24 ( 3 ), 355–368. [ Google Scholar ]
• Borenstein M, Hedges LV, Higgins JP, Rothstein HR, 2011. Introduction to Meta-Analysis . John Wiley & Sons. [ Google Scholar ]
• Borenstein M, Higgins JPT, Hedges LV, Rothstein HR, 2017. Basics of meta-analysis: 12 is not an absolute measure of heterogeneity . Res. Synth. Methods 8 ( 1 ), 5–18. [ PubMed ] [ Google Scholar ]
• Canfield SE, Dahm P, 2011. Rating the quality of evidence and the strength of recommendations using GRADE . World J. Urol . 29 ( 3 ), 311–317. [ PubMed ] [ Google Scholar ]
• Card NA, 2015. Applied Meta-Analysis for Social Science Research . Guilford Publications. [ Google Scholar ]
• Carter BU, 2018. Single screen of citations with excluded terms: an approach to citation screening in systematic reviews . Syst. Rev . 7 ( 1 ) 111-018-0782-x. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Chalmers I, Hedges LV, Cooper H, 2002. A brief history of research synthesis . Eval. Health Prof . 25 ( 1 ), 12–37. [ PubMed ] [ Google Scholar ]
• Cheung MW, 2014. Modeling dependent effect sizes with three-level meta-analyses: a structural equation modeling approach . Psychol. Methods 19 ( 2 ), 211–229. [ PubMed ] [ Google Scholar ]
• Cohen BH, 2002. Calculating a factorial ANOVA from means and standard deviations . Understand. Stat.: Statistical Issues in Psychology, Education, and the Social Sciences 1 ( 3 ), 191–203. [ Google Scholar ]
• Cooke A, Smith D, Booth A, 2012. Beyond PICO: the SPIDER tool for qualitative evidence synthesis . Qual. Health Res . 22 ( 10 ), 1435–1443. [ PubMed ] [ Google Scholar ]
• Cotterill S, Knowles S, Martindale A, Elvey R, Howard S, Coupe N, Wilson P, Spence M, 2018. Getting Messier with TIDieR: Embracing Context and Complexity in Intervention Reporting . [ PMC free article ] [ PubMed ] [ Google Scholar ]
• De Barra M, 2017. Reporting bias inflates the reputation of medical treatments: a comparison of outcomes in clinical trials and online product reviews . Soc. Sci. Med . 177 , 248–255. [ PubMed ] [ Google Scholar ]
• De Vibe M, Bjørndal A, Tipton E, Hammerstrøm KT, Kowalski K, 2017. Mindfulness based stress reduction (MBSR) for improving health, quality of life, and social functioning in adults . Campbell Syst. Rev . 11 . 10.4073/csr.2017.11. [ CrossRef ] [ Google Scholar ]
• Deeks J, Dinnes J, D’amico R, Sowden A, Sakarovitch C, Song F, Petticrew M, Altman D, 2003. Evaluating non-randomised intervention studies . Health Technol. Assess . 7 ( 27 ) iii–x, 1-173. [ PubMed ] [ Google Scholar ]
• Delaney A, Tamás PA, 2018. Searching for evidence or approval? A commentary on database search in systematic reviews and alternative information retrieval methodologies . Res. Synth. Methods 9 ( 1 ), 124–131. [ PubMed ] [ Google Scholar ]
• Denyer D, Tranfield D, Van Aken JE, 2008. Developing design propositions through research synthesis . Organ. Stud . 29 ( 3 ), 393–413. [ Google Scholar ]
• Dwan K, Altman DG, Arnaiz JA, Bloom J, Chan AW, Cronin E, Decullier E, Easterbrook PJ, Von Elm E, Gamble C, Ghersi D, Ioannidis JP, Simes J, Williamson PR, 2008. Systematic review of the empirical evidence of study publication bias and outcome reporting bias . PLoS One 3 ( 8 ), e3081. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Easterbrook PJ, Berlin JA, Gopalan R, Matthews DR, 1991. Publication bias in clinical research . Lancet 337 ( 8746 ), 867–872. [ PubMed ] [ Google Scholar ]
• Ebrahim S, Bance S, Athale A, Malachowski C, Ioannidis JP, 2016. Meta-analyses with industry involvement are massively published and report no caveats for antidepressants . J. Clin. Epidemiol . 70 , 155–163. [ PubMed ] [ Google Scholar ]
• Egger M, Smith G, Schneider M, Minder C, 1997. Bias in meta-analysis detected by a simple, graphical test . Br. Med. J . 315 ( 7109 ), 629–634. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Elliott JH, Synnot A, Turner T, Simmonds M, Akl EA, Mcdonald S, Salanti G, Meerpohl J, Maclehose H, Hilton J, Tovey D, Shemilt I, Thomas J, Living Systematic Review Network, 2017. Living systematic review: 1. Introduction-the why, what, when, and how . J. Clin. Epidemiol . 91 , 23–30. [ PubMed ] [ Google Scholar ]
• European Food SA, 2010. Application of systematic review methodology to food and feed safety assessments to support decision making. Efsa guidance for those carrying out systematic reviews . Eur. Food Saf. Auth. J . 8 ( 6 ), 1–89. [ Google Scholar ]
• Fanelli D, 2012. Negative results are disappearing from most disciplines and countries . Scientometrics 90 ( 3 ), 891–904. [ Google Scholar ]
• Fanelli D, Ioannidis JP, 2013. US studies may overestimate effect sizes in softer research . Proc. Natl. Acad. Sci. U.S.A . 110 ( 37 ), 15031–15036. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Franco A, Malhotra N, Simonovits G, 2014. Social science. Publication bias in the social sciences: unlocking the file drawer . Science 345 ( 6203 ), 1502–1505. [ PubMed ] [ Google Scholar ]
• Garner P, Hopewell S, Chandler J, Maclehose H, Schünemann HJ, Akl EA, Beyene J, Chang S, Churchill R, Dearness K, Guyatt G, Lefebvre C, Liles B, Marshall R, Martínez García L, Mavergames C, Nasser M, Qaseem A, Sampson M, Soares-Weiser K, Takwoingi Y, Thabane L, Trivella M, Tugwell P, Welsh E, Wilson EC, 2016. When and How to Update Systematic Reviews: Consensus and Checklist . British Medical Journal Publishing Group . [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Glass GV, 1976. Primary, secondary, and meta-analysis of research . Educ. Res . 5 ( 10 ), 3–8. [ Google Scholar ]
• Grant MJ, Booth A, 2009. A typology of reviews: an analysis of 14 review types and associated methodologies . Health Inf. Libr. J . 26 ( 2 ), 91–108. [ PubMed ] [ Google Scholar ]
• Hanna JM, 2015. The Quality of Education Leadership Doctoral Dissertations in the United States: an Empirical Review . Doctoral Dissertation edn. Marshall University. [ Google Scholar ]
• Haynes RB, 2006. Forming research questions . J. Clin. Epidemiol . 59 ( 9 ), 881–886. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Hedges LV, 2007. Effect sizes in cluster-randomized designs . J. Educ. Behav. Stat . 32 ( 4 ), 341–370. [ Google Scholar ]
• Hedges LV, Olkin I, 1985. Statistical Methods for Meta-Analysis . Academic Press Inc, Orlando FL, USA. [ Google Scholar ]
• Hedges LV, Tipton E, Johnson MC, 2010. Robust variance estimation in meta-regression with dependent effect size estimates . Res. Synth. Methods 1 ( 1 ), 39–65. [ PubMed ] [ Google Scholar ]
• Hennessy EA, Johnson BT, Keenan C, 2019. Best practice guidelines and essential steps to conduct rigorous and systematic meta-reviews . Appl. Psychol. Health Well-Being (in press). [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Hennessy EA, Tanner-Smith EE, 2015. Effectiveness of brief school-based interventions for adolescents: a meta-analysis of alcohol use prevention programs . Prev. Sci. Official J. Soc. Prev. Res . 16 ( 3 ), 463–474. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Higgins JP (Ed.), 2008. Cochrane Handbook for Systematic Reviews of Interventions , 5 edn. Wiley-Blackwell, Chichester. [ Google Scholar ]
• Higgins J, Thompson S, Deeks J, Altman D, 2002. Statistical heterogeneity in systematic reviews of clinical trials: a critical appraisal of guidelines and practice . J. Health Serv. Res. Policy 7 ( 1 ), 51–61. [ PubMed ] [ Google Scholar ]
• Higgins JP, Thompson SG, Deeks JJ, Altman DG, 2003. Measuring inconsistency in meta-analyses . BMJ Br. Med. J. (Clin. Res. Ed.) 327 ( 7414 ), 557–560. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Higgins JP, Altman DG, Gotzsche PC, Juni P, Moher D, Oxman AD, Savovic J, Schulz KF, Weeks L, Sterne JA, Cochrane Bias Methods Group and Cochrane Statistical Methods Group, 2011. The Cochrane Collaboration’s tool for assessing risk of bias in randomised trials . BMJ (Clin. Res. Ed.) 343 , d5928. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Hoaglin DC, 2017. Practical challenges of I(2) as a measure of heterogeneity . Res. Synth. Methods 8 ( 3 ), 254. [ PubMed ] [ Google Scholar ]
• Hoffmann T, Glasziou P, Boutron I, Milne R, Perera R, Moher D, Altman D, Barbour V, Macdonald H, Johnston M, Lamb S, DIXON - Woods M, Mcculloch P, Wyatt J, Chan A, Michie S, 2014. Better reporting of interventions: template for intervention description and replication (TIDieR) checklist and guide . Br. Med. J . [ PubMed ] [ Google Scholar ]
• Hozo SP, Djulbegovic B, Hozo I, 2005. Estimating the mean and variance from the median, range, and the size of a sample . BMC Med. Res. Methodol . 5 13-2288-5-13. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Ioannidis JP, 2010. Meta-research: the art of getting it wrong . Res. Synth. Methods 1 ( 3–4 ), 169–184. [ PubMed ] [ Google Scholar ]
• Ioannidis JP, 2016. The mass production of redundant, misleading, and conflicted systematic reviews and meta-analyses . Milbank Q . 94 ( 3 ), 485–514. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Ioannidis JP, 2017. Does evidence-based hearsay determine the use of medical treatments? Soc. Sci. Med . 177 , 256–258. [ PubMed ] [ Google Scholar ]
• Joanna Briggs Institute, 2011. In: Joanna Briggs Institute Reviewers’ Manual: 2011 Edition . University of Adelaide, Adelaide, South Australia. [ Google Scholar ]
• Johnson BT, Eagly AH, 2014. Meta-analysis of social-personality psychological research. In: Reis HT (Ed.), Handbook of Research Methods in Social and Personality Psychology, second ed . Cambridge University Press, London, pp. 675–707. [ Google Scholar ]
• Johnson BT, Huedo-Medina TB, 2011. Depicting estimates using the intercept in meta-regression models: the moving constant technique . Res. Synth. Methods 2 ( 3 ), 204–220. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Johnson BT, Low RE, Macdonald HV, 2015. Panning for the gold in health research: incorporating studies’ methodological quality in meta-analysis . Psychol. Health 30 ( 1 ), 135–152. [ PubMed ] [ Google Scholar ]
• Johnson BT, Cromley EK, Marrouch N, 2017. Spatiotemporal meta-analysis: reviewing health psychology phenomena over space and time . Health Psychol. Rev . 11 ( 3 ), 280–291. [ PubMed ] [ Google Scholar ]
• Kaufman MR, Cornish F, Zimmerman RS, Johnson BT, 2014. Health behavior change models for HIV prevention and AIDS care: practical recommendations for a multi-level approach . J. Acquir. Immune Defic. Syndr . 66 ( Suppl. 3 ), S250–S258. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Kenny DA, 1979. Correlation and Causality . Wiley, New York: 1979. [ Google Scholar ]
• Kirkham JJ, Altman DG, Williamson PR, 2010a. Bias due to changes in specified outcomes during the systematic review process . PLoS One 5 ( 3 ), e9810. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Kirkham JJ, Dwan KM, Altman DG, Gamble C, Dodd S, Smyth R, Williamson PR, 2010b. The impact of outcome reporting bias in randomised controlled trials on a cohort of systematic reviews . BMJ 340 , c365. [ PubMed ] [ Google Scholar ]
• Knottnerus JA, Tugwell P, 2016. The way in which effects are analyzed and communicated can make a difference for decision making . J. Clin. Epidemiol . 72 , 1–3. [ PubMed ] [ Google Scholar ]
• Kugley S, Wade A, Thomas J, Mahood Q, Jorgensen AMK, Hammerstrom K, Sathe N, 2016. Searching for Studies: A Guide to Information Retrieval for Campbell . The Campbell Collaboration, Oslo. [ Google Scholar ]
• Lai NM, Teng CL, Lee ML, 2011. Interpreting systematic reviews: are we ready to make our own conclusions? A cross-sectional study . BMC Med . 9 ( 1 ) 30–30. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Lasserson T, Churchill R, Chandler J, Tovey D, Higgins JPT, 2016. Standards for the reporting of protocols of new Cochrane Intervention reviews. In: Higgins JPT, Lasserson T, Chandler J, Tovey D, Churchill R (Eds.), Methodological Expectations of Cochrane Intervention Reviews . Cochrane, London. [ Google Scholar ]
• Lau J, Ioannidis JP, Terrin N, Schmid CH, Olkin I, 2006. The case of the misleading funnel plot . BMJ 333 ( 7568 ), 597–600. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Lennon CA, Huedo-Medina TB, Gerwien DP, Johnson BT, 2012. A role for depression in sexual risk reduction for women? A meta-analysis of HIV prevention trials with depression outcomes . Soc. Sci. Med . 75 ( 4 ), 688–698. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Lewin S, Glenton C, Munthe-Kaas H, Carlsen B, Colvin CJ, Gulmezoglu M, Noyes J, Booth A, Garside R, Rashidian A, 2015. Using qualitative evidence in decision making for health and social interventions: an approach to assess confidence in findings from qualitative evidence syntheses (GRADE-CERQual) . PLoS Med . 12 ( 10 ), e1001895. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Lipsey M, Wilson DB, 2001. Practical Meta-Analysis . Sage Publications, Thousand Oaks Ca, USA. [ Google Scholar ]
• Macdonald HV, Johnson BT, Huedo-Medina TB, Livingston J, Forsyth KC, Kraemer WJ, Farinatti PT, Peseatello LS, 2016. Dynamic resistance training as stand-alone antihypertensive lifestyle therapy: a meta-analysis . Journal of the American Heart Association 5 ( 10 ). 10.1161/JAHA.116.003231. [ PMC free article ] [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Marrouch N, Johnson BT, 2019. Geocoding: using space to enhance social psychological research. In: Blanton H, Lacroix LM, Webster GD (Eds.), Measurement in Social Psychology . Psychology Press, New York, pp. 201–227. [ Google Scholar ]
• Marshall IJ, Noel-Storr A, Kuiper J, Thomas J, Wallace BC, 2018. Machine learning for identifying randomized controlled trials: an evaluation and practitioner’s guide . Res. Synth. Methods 9 ( 4 ), 602–614. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Mckenzie, 2011. Methodological Issues in Meta-Analysis of Randomized Controlled Trials with Continuous Outcomes , PhD edn. Monash University, Melbourne, Australia. [ Google Scholar ]
• Michie S, Richardson M, Johnston M, Abraham C, Francis J, Hardeman W, Eccles MP, Cane J, Wood CE, 2013. The behavior change technique taxonomy (v1) of 93 hierarchically clustered techniques: building an international consensus for the reporting of behavior change interventions . Ann. Behav. Med . 46 ( 1 ), 81–95. [ PubMed ] [ Google Scholar ]
• Moher D, Liberati A, Tetzlaff J, Altman DG, PRISMA Group, 2009. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement . J. Clin. Epidemiol . 62 ( 10 ), 1006–1012. [ PubMed ] [ Google Scholar ]
• Moher D, Shamseer L, Clarke M, Ghersi D, Liberati A, Petticrew M, Shekelle P, Stewart LA, 2015. Preferred reporting items for systematic review and meta-analysis protocols (PRISMA-P) 2015 statement . Syst. Rev . 4 ( 1 ), 1. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Mortensen ML, Adam GP, Trikalinos TA, Kraska T, Wallace BC, 2017. An exploration of crowdsourcing citation screening for systematic reviews . Res. Synth. Methods 8 ( 3 ), 366–386. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Munafo MR, Stothart G, Flint J, 2009. Bias in genetic association studies and impact factor . Mol. Psychiatry 14 ( 2 ), 119–120. [ PubMed ] [ Google Scholar ]
• Nakagawa S, Noble DW, Senior AM, Lagisz M, 2017. Meta-evaluation of meta-analysis: ten appraisal questions for biologists . BMC Biol . 15 ( 1 ), 18. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• O’Neill J, Tabish H, Welch V, Petticrew M, Pottie K, Clarke M, Evans T, PARDO Pardo J, Waters E, White H, Tugwell P, 2014. Applying an equity lens to interventions: using PROGRESS ensures consideration of socially stratifying factors to illuminate inequities in health . J. Clin. Epidemiol . 67 ( 1 ), 56–64. [ PubMed ] [ Google Scholar ]
• Page MJ, Mckenzie JE, Forbes A, 2013. Many scenarios exist for selective inclusion and reporting of results in randomized trials and systematic reviews . J. Clin. Epidemiol . 66 ( 5 ), 524–537. [ PubMed ] [ Google Scholar ]
• Pan Z, Trikalinos TA, Kavvoura FK, Lau J, Ioannidis JP, 2005. Local literature bias in genetic epidemiology: an empirical evaluation of the Chinese literature . PLoS Med . 2 ( 12 ), e334. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Paynter R, Banez LL, Erinoff E, Lege-Matsuura J, Potter S, 2017. Commentary on EPC methods: an exploration of the use of text-mining software in systematic reviews . J. Clin. Epidemiol . 84 , 33–36. [ PubMed ] [ Google Scholar ]
• Polanin JR, Hennessy EA, Tanner-Smith EE, 2017. A review of meta-analysis packages in R . J. Educ. Behav. Stat . 42 ( 2 ), 206–242. [ Google Scholar ]
• Przybyla P, Brockmeier AJ, Kontonatsios G, LE Pogam MA, Mcnaught J, Von Elm E, Nolan K, Ananiadou S, 2018. Prioritising references for systematic reviews with RobotAnalyst: a user study . Res. Synth. Methods 9 ( 3 ), 470–488. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Rathbone J, Hoffmann T, Glasziou P, 2015. Faster title and abstract screening? Evaluating Abstrackr, a semi-automated online screening program for systematic reviewers . Syst. Rev . 4 80-8015-0067-6. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Reid AE, Dovidio JF, Ballester E, Johnson BT, 2014. HIV prevention interventions to reduce sexual risk for African Americans: the influence of community-level stigma and psychological processes . Soc. Sci. Med . ( 103 ), 118–125 1982. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Rubin M, 2017. When does HARKing hurt? Identifying when different types of undisclosed post hoc hypothesizing harm scientific progress . Rev. Gen. Psychol . 21 ( 4 ), 308. [ Google Scholar ]
• Rucker G, Schwarzer G, Carpenter JR, Schumacher M, 2008. Undue reliance on I(2) in assessing heterogeneity may mislead . BMC Med. Res. Methodol . 8 79-2288-8-79. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Schmid CH, 2017. Heterogeneity: multiplicative, additive or both? Res. Synth. Methods 8 ( 1 ), 119–120. [ PubMed ] [ Google Scholar ]
• Schmidt FL, Hunter JE, 2015. Methods of Meta-Analysis: Correcting Error and Bias in Research Findings, third ed . Sage publications, Thousand Oaks, CA. [ Google Scholar ]
• Schwarzer G, Schumacher M, Rücker G, 2017. Sole Reliance on I 2 May Mislead . BMJ Publishing Group Ltd and British Cardiovascular Society. [ PubMed ] [ Google Scholar ]
• Shadish WR, Lecy JD, 2015. The meta-analytic big bang . Res. Synth. Methods 6 ( 3 ), 246–264. [ PubMed ] [ Google Scholar ]
• Shea BJ, Reeves BC, Wells G, Thuku M, Hamel C, Moran J, Moher D, Tugwell P, Welch V, Kristjansson E, 2017. AMSTAR 2: a critical appraisal tool for systematic reviews that include randomised or non-randomised studies of healthcare interventions, or both . BMJ 358 , j4008. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Shojania KG, Sampson M, Ansari MT, Ji J, Doucette S, Moher D, 2007. How quickly do systematic reviews go out of date? A survival analysis . Ann. Intern. Med . 147 ( 4 ), 224–233. [ PubMed ] [ Google Scholar ]
• Siddaway AP, Wood AM, Hedges LV, 2019. How to do a systematic review: a best practice guide for conducting and reporting narrative reviews, meta-analyses, and meta-syntheses . Annu. Rev. Psychol . 70 , 747–770. [ PubMed ] [ Google Scholar ]
• Simpson R, Pearson K, 1904. Report on certain enteric fever inoculation statistics . Br. Med. J . 1243–1246. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Stanley TD, Doucouliagos H, 2014. Meta-regression approximations to reduce publication selection bias . Res. Synth. Methods 5 ( 1 ), 60–78. [ PubMed ] [ Google Scholar ]
• Stanley TD, Doucouliagos H, 2015. Neither fixed nor random: weighted least squares meta-analysis . Stat. Med . 34 ( 13 ), 2116–2127. [ PubMed ] [ Google Scholar ]
• Sterne JAC, Higgins JPT, Reeves BC, ON BEHALF OF THE DEVELOPMENT GROUP FOR ROBINS-I, 2016. A Tool for Assessing Risk of Bias in Non-randomized Studies of Interventions, 7 edn . . [ Google Scholar ]
• Stigler SM, 1986. The History of Statistics: the Measurement of Uncertainty before 1900 . Harvard University Press. [ Google Scholar ]
• Strutt JW, 1884. Presidential Address at the 54th Meeting of the British Association for the Advancement of Science. John Murray, London 1884. [ Google Scholar ]
• Sumner JA, Carey RN, Michie S, Johnston M, Edmondson D, Davidson KW, 2019. Using rigorous methods to advance behaviour change science (vol 2, pg 797, 2018) . Nature Human Behaviour 3 ( 1 ) 101–101. [ PubMed ] [ Google Scholar ]
• Tipton E, Pustejovsky JE, Ahmadi H, 2019. Current Practices in Meta-regression in Psychology, Education, and Medicine . Research synthesis methods . [ PubMed ] [ Google Scholar ]
• Tricco AC, Lillie E, Zarin W, O’brien K, Colquhoun H, Kastner M, Levac D, Ng C, Sharpe JP, Wilson K, Kenny M, Warren R, Wilson C, Stelfox HT, Straus SE, 2016. A scoping review on the conduct and reporting of scoping reviews . BMC Med. Res. Methodol . 16 15-016-0116-4. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Valentine JC, 2009. Judging the Quality of Primary Research . The Handbook of Research Synthesis and Meta-Analysis , pp. 129–146. [ Google Scholar ]
• Van Bavel JJ, Mende-Siedlecki P, Brady WJ, Reinero DA, 2016. Contextual sensitivity in scientific reproducibility . Proc. Natl. Acad. Sci. Unit. States Am . 113 ( 23 ), 6454–6459. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Viechtbauer W, 2010. Conducting meta-analyses in R with the metafor package . J. Stat. Softw . 36 ( 3 ). [ Google Scholar ]
• Viechtbauer W, Cheung MW, 2010. Outlier and influence diagnostics for meta-analysis . Res. Synth. Methods 1 ( 2 ), 112–125. [ PubMed ] [ Google Scholar ]
• Von Hippel P,T, 2015. The heterogeneity statistic I 2 can be biased in small meta-analyses . BMC Med. Res. Methodol . 15 ( 1 ). [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Whiting P, Savovic J, Higgins JP, Caldwell DM, Reeves BC, Shea B, Davies P, Kleijnen J, Churchill R, ROBIS Group, 2016. ROBIS: a new tool to assess risk of bias in systematic reviews was developed . J. Clin. Epidemiol . 69 , 225–234. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Wildridge V, Bell L, 2002. How CLIP became ECLIPSE: a mnemonic to assist in searching for health policy/management information . Health Inf. Libr. J . 19 ( 2 ), 113–115. [ PubMed ] [ Google Scholar ]
• Williamson PR, Gamble C, Altman DG, Hutton JL, 2005. Outcome selection bias in meta-analysis . Stat. Methods Med. Res . 14 ( 5 ), 515–524. [ PubMed ] [ Google Scholar ]
• Wilson DB Practical meta-analysis effect size calculator [online calculator] , n.d.-last update, Available: https://www.campbellcollaboration.org/effect-size-calculato.html , Accessed date: 4 August 2018.
• Yavchitz A, Ravaud P, Altman DG, Moher D, Hrobjartsson A, Lasserson T, Boutron I, 2016. A new classification of spin in systematic reviews and meta-analyses was developed and ranked according to the severity . J. Clin. Epidemiol . 75 , 56–65. [ PubMed ] [ Google Scholar ]

Exploring the perception of parents on children's screentime: a systematic review and meta-synthesis of qualitative studies

Affiliations.

• 1 Child Development Unit, Khoo Teck Puat, National University Health System, Singapore, Singapore.
• 2 University Children's Medical Institute, National University Health System, Singapore, Singapore.
• 3 Paediatrics, Yong Loo Lin School of Medicine, National University of Singapore, Singapore, Singapore.
• 4 Alice Lee Center for Nursing Studies, Yong Loo Lin School of Medicine, National University of Singapore, Singapore, Singapore.
• 5 Alice Lee Center for Nursing Studies, Yong Loo Lin School of Medicine, National University of Singapore, Singapore, Singapore. [email protected].
• PMID: 36966270
• PMCID: PMC10039437
• DOI: 10.1038/s41390-023-02555-9

Background: Screentime (ST) has been increasingly pervasive in young children. Evidence suggests positive and negative effects of ST on children's development. Parents play a crucial role in influencing their children's ST. There is limited research consolidating the parental perceptions related to children's ST. This review consolidates the evidence on the perception of parents on their children's ST.

Methods: Six electronic databases (PubMed, Embase, CINAHL, PsycINFO, Scopus, and ProQuest) were searched from their inception to September 2022. Critical appraisal was conducted using the Critical Appraisal Skills Program (CASP) checklist. Data synthesis was performed using Sandelowski and Barroso's approach.

Results: Twenty studies were included in this review, encapsulating the experiences of 1,311 parents. Three main themes with corresponding subthemes were identified: (1) Varied reasons behind ST; (2) Attitudes toward ST; and (3) Strategies and approaches to managing ST.

Conclusion: The findings highlighted the varied reasons parents promote ST. Parents reported their concerns and confusion about ST and shared some strategies and approaches for better ST. Future research is needed to develop and evaluate educational programs to ensure that parents understand the risks and benefits of ST and in turn, ensure the appropriate adoption of ST for their children.

Impact: In the digital era, parents' attitudes towards screentime (ST) for their children are mixed. This review presents the dilemma parents face and their struggles with using optimal screen time for their children. This review provides up-to-date evidence on the parents' confusion and concerns about ST as well as strategies and approaches used by the parents for better ST. There is an urgent need for evidence-based educational programs to enhance parental knowledge about ST so that they can ensure appropriate ST among children.

Clinical trial registration: PROSPERO Registration No. CRD42022356083.

© 2023. The Author(s), under exclusive licence to the International Pediatric Research Foundation, Inc.

Publication types

• Systematic Review
• Child, Preschool
• Perception*
• Qualitative Research