1. Workshop: Framework Synthesis, Meta-Ethnography and Realist Synthesis

2. Shared Topic: Adherence to Antiretroviral therapy (ART) for HIV in Zambia
BACKGROUND: Antiretroviral therapy (ART) has significantly improved morbidity and mortality of individuals infected with HIV. However, lack of adherence to highly active antiretroviral therapy (HAART) remains a key challenge to successful management of patients with HIV/AIDS. Adherence rates lower than 95% are associated with development of viral resistance to antiretroviral medications.
‘Efforts to sustain adherence in Africa and elsewhere remain important goals to optimize outcomes for individuals and global HIV treatment.’ (Mills, Nachega, Buchan, Orbinski, Attaran, Singh et al., 2006).

3. Shared Topic: Adherence to Antiretroviral therapy (ART) for HIV in Zambia
Different Emphases
Barriers and Facilitators to ART (Best Fit Framework Synthesis)
Theory Explaining Adherence to ART (Meta-Ethnography)
Under what Contexts and For Whom is Adherence to ART Effective in Zambia (Realist Synthesis)

4. Reading and Data Extraction

5. Data extraction What is it?
An attempt to reduce a mass of material (your included papers) to a much smaller body of text and numbers, amenable to analysis and the interpretation of findings
Why extract these data? Same as for any review? Combining? Sensitivity analysis?

6. Data extraction form Location Setting Sample (n) Age Gender Ethnicity
Socio-economic status
Intervention (if any)
Quality assessment criteria
Results?
Further citations
See Handouts 1-3
Format?????
Why extract these data? Same as for any review? Combining? Sensitivity analysis? Contextualising and interpreting your findings

7. What results do you extract?
What is your question?
Keep the question in mind as you read:
Are the data relevant to this question?
Is the question answered by the data?

8. Data Extraction Framework Synthesis Meta-Ethnography Realist Synthesis
Extracts data against framework. Coding framework with definitions is provided to increase consistency. Data not explained by framework is“parked” for subsequent inductive stage. Distinction typically made between original quotes and author’s analysis.
Key themes and concepts extracted and peer reviewed for inclusiveness. Distinction made between first-order (original quotes) and second-order (author’s analysis) findings. Findings were coded in duplicate. Discrepancy between codes resolved by third person.
Created tool to cover broad range of information
and to consolidate data from across studies. [Pawson argues such tools are too limiting]. Tool flexible to account for diversity. Reviewers returned to papers if new questions arose. Template applied to body of each report (i.e., introduction, methods, results, and discussion sections) by two members of research team. Differences in coding resolved by consensus with other team members.

9. Quality Assessment
Andrew Booth, Reader in Evidence Based Information Practice, Co-Convenor – Cochrane Collaboration Qualitative Methods Group

10. Before You Begin…
Consider how you will use judgements of quality (cp. 50% of published Cochrane Quantitative Reviews performed quality appraisal but did not make it clear how judgements were used!)
To exclude or to moderate?
Will chosen instrument militate against certain types of research?
Quality of reporting or quality of study?

11. Variability in Practice - 1
21 papers did not describe appraisal of candidate studies
6 explicitly mentioned not conducting formal appraisal of studies
5 papers did a critical appraisal, but did not use a formal checklist
7 described modifying existing instruments
1 used an existing instrument without modification
Dixon-Woods, Booth & Sutton (2007)

12. Variability in Current Practice - 1
23 papers did not describe critical appraisal
5 papers explicitly pleaded against quality assessment of papers or provided valid reason for not conducting quality appraisal.
Criteria used varied between detailed descriptions of relevant items in existing or modified checklists to a set of broad criteria evaluating, for example, rich description of data, credibility or relevance of the original study.
Hannes and Macaitis (2012)

13. Variability in Current Practice - 2
One team used overall judgement (Smith et al., 2005).
Five opted for self-developed assessment instrument
Three used previously developed checklists to create own.
Two mentioned critical appraisal, but did not specify tool.
Most used existing instruments/frameworks. 24 different assessment tools identified:
Critical Appraisal Skills Programme (CASP) (n = 18)
Mays and Pope criteria (n = 6)
Popay criteria (n = 6)
Joanna Briggs Institute (n = 4).
Hannes and Macaitis (2012)

14. Appraising research quality
1. Epistemological criteria: Judgement of ‘trustworthiness’ requires criteria tailored to different research ‘paradigms’.
2. Theoretical Criteria: Explicit theoretical framework shaping the design of the study and informing claims for generalisability
3. Prima facie ‘Technical’ criteria: Used to assess ‘quality’ common to all research traditions e.g.:
Sufficient explanation of background;
Method appropriate to question;
Succinct statement of objectives/research questions;
Full description of methods include approach to analysis;
Clear presentation of findings including justification for interpretation of data etc.
Noyes J (2005)

15. Two dimensional approach to appraising qualitative research
Technical markers –CASP
Epistemological and theoretical markers – Popay et al
Technical Quality High
Description – thicker
Privileges Subjective experience and meanings
Use of theory to build explanations
Technical Quality Low
Description - thinner
Imposed pre-determined framework on respondents narratives.
Limited/no/inappropriate use of theory, little explanatory insight
(Noyes, 2005)

16. Available Tools - 1
CASP – 10 questions to help you make sense of qualitative research
Joanna Briggs Institute - Critical Appraisal Checklist for Interpretive & Critical Research
National Centre for Social Research. Quality in Qualitative Evaluation: A Framework for Assessing Research Evidence. London: National Centre for Social Research/UK Cabinet Office,

17. Available Tools - 2
Dixon-Woods M, Shaw RL, Agarwal S & Smith JA (2004) The problem of appraising qualitative research. Quality & Safety in Health Care, 13,
Hannes K, Lockwood C, Pearson A (2010). A comparative analysis of three online appraisal instruments' ability to assess validity in qualitative research. Qualitative Health Research. 20(12):
Popay J, Rogers A & Williams G (1998) Rationale & standards for the systematic review of qualitative literature in health services research. Qualitative Health Research, 8,
Seale C & Silverman D (1997) Ensuring rigour in qualitative research. European Journal of Public Health, 7,

18. CASP – Technical/Procedural Tool

19. JBI – Theoretical Tool

20. Key Issue How are you going to use the quality assessment?
From quantitative assessment we know authors frequently say they do it – but they don’t incorporate it into results
Is it technical proceduralism gone mad?
Or can we use the assessments to improve our synthesis and subsequent interpretation?

21. Quality Assessment Framework Synthesis Meta-Ethnography
Realist Synthesis
Pragmatic so tends to include all studies . Focuses explicitly on quality of reporting. Qualitative sensitivity analysis used to test robustness of synthesis.
Quality Assessment as Hurdle: Studies using qualitative design and analysis method were included. All studies that focused on ART access and adherence (n = 130) continued to full-text review. Fourteen studies
passed quality appraisal (Wallace et al. 2004; Smith et al. 2005; Walsh & Downe 2005, 2006; Rolfe 2006) and were retained.
Rather than appraising and excluding publications prior to review, reviewer mines each publication for evidence to contribute to fuller development of explanatory model. Instead of evaluating entire reports against a priori standards for appraisal, each piece of evidence is appraised for utility and relevance. Value of evidence depends on whether it contributes to better understanding of questions addressed in review.

22. Data Synthesis

23. What is Data Synthesis?
Process of moving from focus on single studies (cp. Data Extraction and Quality Assessment) to focus on cross-study analysis
Requires identification of patterns across data, including contradictory findings and data that does not fit
Iterative and requires ongoing refinement
Acts as prelude to Analysis which seeks to explain patterns, contradictions and differences

24. Methods of qualitative evidence synthesis
Thematic synthesis; Critical Interpretive Synthesis; Meta-ethnography
Only include “good” qualitative studies (?)
Constant comparison; iterative; interpretations generated from the data by reviewers
Create a theory
Inductive (theory-generating)
Examples:
Thomas J, Harden A. Methods for the thematic synthesis of qualitative research in systematic reviews. BMC Medical Research Methodology 2008; 8.
Campbell R et al. Evaluating meta-ethnography: a synthesis of qualitative research on lay experiences of diabetes and diabetes care. Social Science & Medicine 2003; 65:
Other methods are highly time and expertise intensive
FS uses unsystematic literature review and expert consultation (of unspecified nature) to generate the new model

25. Methods of qualitative evidence synthesis
Framework synthesis:
Only include “good” qualitative studies (?)
Map data from included studies onto an existing framework to test the framework/theory (a role for theory)
Build a conceptual model or framework
Deductive (theory-testing)
Examples:
Oliver S et al: A multidimensional conceptual framework for analysing public involvement in health services research. Health Expectations 2008, 11:72-84.
Brunton G, Oliver S, Oliver K, Lorenc T. A Synthesis of Research Addressing Children’s, Young People’s and Parents’ Views of Walking and Cycling for Transport London. London, EPPI-Centre, Social Science Research Unit, Institute of Education, University of London; 2006.
Other methods are highly time and expertise intensive
FS uses unsystematic literature review and expert consultation (of unspecified nature) to generate the new model

26. Methods of qualitative evidence synthesis
“Best-fit” framework synthesis
Identify relevant pre-existing conceptual models or frameworks
Identify and extract all relevant qualitative studies satisfying review’s inclusion criteria
Code data from included studies against framework
Use secondary thematic analysis/synthesis to generate completely new themes to supplement the framework’s themes
Create new framework and conceptual model or theory
Deductive and Inductive
Framework and Thematic synthesis
Carroll C, Booth A, Cooper K. A worked example of “best-fit” framework synthesis: A systematic review of views concerning the taking of potential chemopreventive agents, BMC Medical Research Methodology 2011; 11: 29
Other methods are highly time and expertise intensive
FS uses unsystematic literature review and expert consultation (of unspecified nature) to generate the new model

27. Integration of qualitative and quantitative evidence
Realist Synthesis (Also The Matrix Method);
Examine Provisional Explanatory Model
Extract Data against Provisional Explanatory Model
Assess for Rigour and Relevance
Develop Propositional Statements
Consider the Strength of Evidence for Each Propositional Statement
Refine and Revise Provisional Explanatory Model
Merten, S., Kenter, E., McKenzie, O., Musheke, M., Ntalasha, H., & Martin‐Hilber, A. (2010). Patient‐reported barriers and drivers of adherence to antiretrovirals in sub‐Saharan Africa: a meta‐ethnography. Tropical Medicine & International Health, 15(s1),
[Candy B, King M, Jones L, Oliver S. Using qualitative synthesis to explore heterogeneity of complex interventions. BMC Med Res Methodol Aug 26;11:124. doi: / ].
Other methods are highly time and expertise intensive
FS uses unsystematic literature review and expert consultation (of unspecified nature) to generate the new model

28. Data Synthesis Framework Synthesis Meta-Ethnography Realist Synthesis
Original best fit framework is expanded to include new themes. Relationship between themes is examined and the data is used to reconstitute a new model. Particular attention is directed at discrepant cases.
Second-order constructs pertinent to adherence identified and cross-compared and presented in results section. Key themes consolidated into a line of argument (third-order analysis), presented in the synthesis ⁄ discussion section.
Reviewers synthesize extracted evidence and
apply it to explanatory model. Then iteratively refine model to best
explain existing data. Where evidence does not fit, reviewer looks for other factors with more explanatory value. Synthesis ends with summary of how explanatory model has been revised.