Integrating Qualitative and Quantitative Evidence in a Review

Many research questions cannot be answered by one type of evidence alone. Trials can show whether an intervention works on average, but not why it works for some people and not others, or how it is experienced by those who receive it. Qualitative studies can illuminate experiences, mechanisms and contexts, but not estimate effects. Mixed-methods reviews bring both kinds of evidence together, aiming for a richer and more useful synthesis than either could provide separately. Doing this well is challenging: the two traditions have different logics, different quality criteria and different ways of presenting findings. This guide explains the main approaches to integrating qualitative and quantitative evidence in a review, how to plan the integration from the start, how to present integrated findings and how a source-linked assistant such as Yila can help you extract comparable information from very different kinds of studies.

Why integrate at all

Integration is worth the effort when your question has several facets that different evidence types address. A review of a community health programme might ask whether the programme improves outcomes, what makes it acceptable to participants and what factors influence implementation. Quantitative studies address the first question; qualitative studies address the second and third. Integrating them can explain variation in effectiveness, identify conditions for success and inform the design of better interventions.

Integration also helps decision-makers, who rarely care only about average effects. They want to know whether an intervention will work in their setting, for their population and with their resources.

Choose an integration design

Several designs exist for mixed-methods reviews. In a convergent design, qualitative and quantitative evidence are synthesised in parallel and then brought together, either by transforming one type into the other or by juxtaposing the two syntheses. In a sequential design, one synthesis informs the other; for example, a qualitative synthesis might generate hypotheses about what makes interventions effective, which are then tested in a quantitative synthesis. In a segregated design, the syntheses are conducted separately and integrated only at the interpretation stage.

The choice depends on your question, the available evidence and your resources. Decide early, because the design affects your search, screening, extraction and analysis.

Plan the review questions

Write separate sub-questions for each type of evidence and an overarching question about how they relate. For example: What is the effect of the programme on outcome X? What are participants' experiences of the programme? How do participants' experiences help explain variation in effects across studies?

Clear sub-questions guide inclusion criteria, extraction and synthesis for each strand, and the overarching question guides integration.

Search and screen for both types

Searches for mixed-methods reviews need to capture both qualitative and quantitative studies. This may require additional search terms or filters for qualitative research, which is often poorly indexed in databases. Consider searching sources where qualitative work is common, such as specialist journals and theses.

Screening criteria should be defined separately for each type of evidence, reflecting what makes a study relevant for each sub-question. Some studies, including mixed-methods primary studies, contribute to both strands.

Appraise each type appropriately

Quality appraisal must respect the different logics of qualitative and quantitative research. Use appraisal tools designed for each study type. Do not judge qualitative studies by criteria designed for trials, or vice versa. For mixed-methods primary studies, consider appraisal tools that address the integration within the study as well as each component.

Record appraisal judgements consistently and decide in advance how they will be used, for example in sensitivity analyses or in weighting confidence in findings.

Extract with integration in mind

Extraction forms should capture the information each strand needs and the information required for integration. For quantitative studies, this includes effect estimates, populations, settings and intervention characteristics. For qualitative studies, it includes themes, participant quotations, contexts and the authors' interpretations. For integration, extract details that allow connections, such as intervention components, implementation features and contextual factors, in comparable terms across both strands.

Yila is useful here because the relevant details are often scattered across sections and reported in different ways. Asking how a qualitative study described the intervention's delivery, or which components a trial's intervention included, returns answers linked to the passages that report them. This helps you build comparable extraction records across very different studies.

Synthesise each strand rigorously

Each strand should be synthesised using methods appropriate to its evidence. Quantitative synthesis might involve meta-analysis or structured narrative synthesis. Qualitative synthesis might use thematic synthesis, meta-ethnography or framework synthesis. Do each well before attempting integration; a weak synthesis in either strand undermines the whole review.

Integrate explicitly

Integration should be an explicit analytic step, not a vague juxtaposition. Common techniques include matrices that map qualitative themes against quantitative findings, logic models that show how intervention components, mechanisms and contexts relate to outcomes and narrative integration that discusses where findings converge, complement or contradict each other.

When qualitative and quantitative findings agree, integration strengthens confidence. When they complement each other, it provides a fuller picture. When they conflict, it raises questions worth investigating: perhaps quantitative outcomes miss what participants value, or qualitative studies come from different contexts than the trials.

Present findings clearly

Integrated findings can be complex. Use tables and diagrams to show how evidence from each strand contributes. A joint display, in which qualitative and quantitative findings appear side by side with an integration column, is a useful format. Clearly label which findings come from which type of evidence and how confident you are in each.

Assess confidence in integrated findings

Frameworks exist for assessing certainty in quantitative findings and confidence in qualitative findings. Apply them to each strand and consider how they bear on integrated conclusions. An integrated finding that rests on low-certainty quantitative evidence and moderate-confidence qualitative evidence should be presented with appropriate caution.

Keep a transparent trail

Mixed-methods reviews involve many judgements, and readers need to see how conclusions were reached. Document your integration methods, decisions and reasoning. Keep notes linking integrated findings to the specific studies and passages that support them. Source-linked notes produced with Yila help maintain this trail across the very different documents in a mixed-methods review.

Integrating qualitative and quantitative evidence is demanding, but it can produce reviews that are far more useful than single-method syntheses. By planning the design and questions from the start, appraising and synthesising each strand appropriately, extracting comparable details, integrating explicitly and documenting every step, you can build a review that explains not just whether something works, but how, why and for whom.