Data extraction and synthesis
Publisert
Thematic synthesis and best fit framework synthesis are both methods that can be used on thick/rich and thin/poor data and provide a good approach for descriptive QES. They will be introduced in this section.
At the NIPH we mostly receive commissions for descriptive QES. This means, for example, QES that describe people’s preferences for or experiences of an intervention. Thematic synthesis and best fit framework synthesis are both methods that can be used on thick/rich and thin/poor data and provide a good approach for descriptive QES. They will be introduced in this section. If you are aiming to conduct a more theoretical or conceptual QES, then a different synthesis approach might be better suited. For help in deciding which synthesis approach is best suited to your QES see the RETREAT framework (7).
About this phase
|
What happens before this step |
Have full texts for all included studies/ sampling complete and are familiar with the content. |
|
Description |
To extract and then synthesize relevant data from the included studies |
|
Why is this step important? |
The synthesis allows us to formulate findings that will be the results of the QES |
|
Responsible |
Project leader and at least one other team member |
|
Tasks |
Create a data extraction form, extract relevant data, confirm your synthesis approach (if different from protocol describe why in changes from protocol), conduct synthesis |
|
Tools |
Some teams choose to use software such as NVIVO (8). Other teams conduct the synthesis in available word processing tools. |
|
Deliverable |
Completed synthesis with the identification of study findings |
By data synthesis we mean going from looking at data from a single study or data collection to focusing on analysing and synthesising data from across many primary studies. We do this by identifying patterns across the data including contradictory findings and data that does not fit, often using an iterative cyclus of analysis. Data synthesis aims to explain patterns of similarity, contradictions and differences (9).
It is important to consider what type of synthesis strategy could be appropriate at the protocol stage and provide a description of this. However, once the included studies are identified it may become apparent that a different synthesis strategy might work better. This is fine if an argumentation for why the synthesis approach has been changed is provided and is transparently described in the section of the final report entitled Changes from the protocol. Decisions around deciding what type of data synthesis to conduct need to be informed by:
- The experience and knowledge of the different team members
- The topic of the synthesis
- Whether the QES has a descriptive or theoretical focus
- If there are any existing theories or frameworks on the topic of interest that could be of use
In this part of the chapter, the two most commonly used synthesis strategies for descriptive QES; thematic synthesis and best fit framework synthesis are presented.
Relevant resources:
Undertaking a qualitative evidence synthesis to support decision-making in a Cochrane context.
- Part 1: Qualitative research and how it fits into systematic reviews
- Part 2: How to choose a method for qualitative evidence synthesis
- Part 3: Framework synthesis, thematic synthesis, meta-ethnography
- Part 4: Confidence in qualitative evidence and reporting a qualitative evidence synthesis
Booth, A., Noyes J, Flemming K, Gerhardus, A., Wahlster, P., Van Der Wilt, G.J., et al. (2016) Guidance on choosing qualitative evidence synthesis methods for use in health technology assessments of complex interventions [Online]. Available from: http://www.integrate-hta.eu/downloads/
Cochrane Webinar series on conducting QES on Meta-ethnography
Thematic synthesis
Thematic synthesis for QES grew out of the thematic analysis approach for primary qualitative research. It was developed and adapted for QES in the early 2000s (10). Thematic synthesis can be used with both thick/rich and thin/poor findings. Broadly, thick and thin refer to levels of contextual description and rich and poor refer to conceptual or theoretical depth of the included studies. See the Cochrane Handbook chapter for more information. It can also be done using software such as NVivo.
Thematic synthesis is a structured, inductive approach to identifying and developing themes from the data. It can produce both descriptive and analytical themes. Descriptive themes stay close to the findings of the original studies. They describe participant preferences, feelings or experiences. Analytical themes go beyond descriptive findings to generate new understandings.
Thematic synthesis consists of three overarching steps:
- Coding findings
- Developing descriptive themes
- Generating analytical themes
There are two types of thematic analysis that can be used in a QES. General thematic analysis where coding is across all data to start the synthesis (10) or constant comparison thematic analysis (3) where the article that comes closest to your review objective is identified and is the codes from this article are used to build the code set.
Relevant resources:
Cochrane Webinar series on conducting QES on Thematic synthesis
Thomas J, Harden A. Methods for the thematic synthesis of qualitative research in systematic reviews. BMC medical research methodology. 2008;8(1):1-10.
Noyes J, Booth A, Cargo M, Flemming K, Harden A, Harris J, Garside R, Hannes K, Pantoja T, Thomas J. Qualitative evidence. Cochrane Handbook for Systematic Reviews of Interventions. 2019 Sep 23:525-45.
Best fit framework synthesis
A framework synthesis uses an a priori framework or theory to predetermine data analysis themes and categories. In 2011, Booth and Carroll extended the framework synthesis method to include data that did not fit in the framework and called it the best fit framework synthesis (11). This approach formally separates the deductive framework coding phase from the inductive theme generation phase for the data that does not fit into the framework. The research team identifies a framework from the literature that is “good enough” to explain more than 50% of the data.
A best fit framework synthesis is a good fit for a specific question with issues and a framework that can be identified a priori and a limited time frame. This synthesis approach easily allows the review team to see what is generalizable across studies and settings. It also fulfils theory testing and theory generation as the framework is adjusted (12).
A best fit framework synthesis consists of 7 steps:
- Define review question
- Systematically identify relevant primary research studies and identify relevant (Best Fit) publications of frameworks and conceptual models/theories
- Extract data on study characteristics from included studies and conduct study quality appraisal
- Code evidence from included studies into a priori framework identified in step 2
- Create new themes by performing secondary thematic analysis on any evidence that cannot be coded into the a priori framework
- Produce a new framework composed of a priori and new themes supported by the evidence
- Revisit evidence to explore relationships between themes or concepts in order to create a model
Relevant resources:
Cochrane Webinar series on conducting QES on Making sense of framework and 'best fit' framework synthesis
Carroll C, Booth A, Cooper K. A worked example of" best fit" framework synthesis: a systematic review of views concerning the taking of some potential chemopreventive agents. BMC medical research methodology. 2011;11(1):1-9.
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 quality & safety. 2015;24(11):700-8.