Written and maintained by CASRAI Editorial Board
Last updated
A mixed-methods systematic review (MMSR) synthesizes quantitative and qualitative primary studies within a single review to answer a question neither evidence type could answer completely alone — for example, whether an intervention works (the quantitative question) and how patients or clinicians actually experience it, and why it does or doesn’t get used (the qualitative question). The methodological difficulty is not finding both kinds of studies; it’s combining findings that were never designed to be compared on the same terms. The Joanna Briggs Institute (JBI) publishes the most widely adopted, step-by-step methodology for doing this rigorously, and its terminology — convergent integrated versus convergent segregated — is what most MMSR protocols and published reviews now reference.
The typology is smaller than it used to be
JBI’s methodology historically described three MMSR designs: convergent (synthesis happens at roughly the same stage, then results are combined), sequential exploratory (qualitative synthesis first, used to inform a subsequent quantitative synthesis), and sequential explanatory (the reverse — quantitative first, qualitative used to explain the findings). The current JBI Manual for Evidence Synthesis narrows its guidance to the convergent design specifically: JBI reports that fewer than 5% of published MMSRs used a sequential design, and updated methodological guidance now concentrates on the convergent approach rather than maintaining full guidance across all three (Stern C, Lizarondo L, Carrier J, et al. “Methodological guidance for the conduct of mixed methods systematic reviews.” JBI Evidence Synthesis, 2020;18(10):2108-2118).
Within the convergent design, the decision that actually shapes a review’s workplan is which of two synthesis approaches to use: convergent segregated or convergent integrated. Both are supported directly in JBI SUMARI, the software JBI provides for running a review through its own methodology, from screening to data extraction to synthesis.
Convergent segregated: synthesize separately, then integrate the results
The segregated approach keeps quantitative and qualitative evidence apart through synthesis and only brings them together at the end, at the level of findings rather than raw data:
- Quantitative studies are synthesized on their own terms — statistical pooling (meta-analysis) where studies are similar enough, or a structured narrative synthesis where they aren’t.
- Qualitative studies are synthesized on their own terms — thematic synthesis, meta-aggregation, or a comparable qualitative synthesis method, independently of the quantitative strand.
- The two sets of findings — a quantitative evidence profile and a qualitative set of synthesized findings — are then integrated in a final configuration step, most often using a joint display that lines the two strands up side by side against a shared set of questions or outcomes.
Segregated is the right choice when the review question has genuinely distinct quantitative and qualitative components — for example, “does the intervention reduce readmission, and separately, what do patients say made it easier or harder to use?” The two strands answer different sub-questions rather than the same one from different angles, so there’s nothing to transform; they just need to be lined up and read together.
Convergent integrated: transform one data type so it can be pooled with the other
The integrated approach is used when quantitative and qualitative studies are actually addressing the same underlying question, and the reviewer wants a single, combined body of evidence rather than two parallel ones. That requires data transformation before synthesis, in one of two directions:
- Qualitizing — converting quantitative findings into textual form (a written description of a statistical result) so they can be pooled alongside qualitative findings in a single qualitative-style synthesis.
- Quantitizing — converting qualitative findings into a numerical or categorical form (for example, counting how many included studies report a given theme) so they can be pooled alongside quantitative data.
Qualitizing is the more common direction in JBI-guided reviews. Once transformed, the combined data set is assembled, findings are grouped by similarity of meaning, and the reviewer produces one set of “integrated findings” that draws on both original evidence types — not two separate findings sets stitched together afterward. This is more labor-intensive and interpretively riskier than the segregated approach (the transformation step is a judgment call the review team has to defend and document), but it produces a genuinely unified answer rather than a side-by-side comparison.
Which one does your question actually need?
The deciding question is not “do I have both quantitative and qualitative studies” — most MMSR candidates do. It’s whether those studies are answering the same question or complementary but distinct questions:
- Same question, different methods used to study it → convergent integrated (transform and pool into one synthesis).
- Different but related questions/dimensions of one phenomenon → convergent segregated (synthesize separately, integrate the findings).
Illustrative example, not a real published review: a team reviewing telehealth follow-up after hip replacement surgery has RCTs measuring 90-day complication rates and qualitative interview studies exploring why patients skip virtual visits. Because the RCTs and the interview studies are addressing different sub-questions (does it work vs. why do people disengage from it), this is a segregated case: synthesize the RCTs, synthesize the interviews, then build a joint display showing where the two strands agree, diverge, or fill each other’s gaps — rather than attempting to force complication-rate data and interview themes into one pooled synthesis.
Reporting and where this connects to CASRAI’s broader synthesis coverage
Whichever approach is used, report it against PRISMA 2020 for the review-conduct reporting basics, plus the GRAMMS (Good Reporting of A Mixed Methods Study) criteria specifically for the integration step — GRAMMS requires authors to state explicitly how the strands were integrated, not simply that both quantitative and qualitative evidence were used. A joint display is the most common way to satisfy that requirement visually as well as in text; see CASRAI’s dedicated guide on building a joint display for the mechanics.
This page sits alongside CASRAI’s other systematic review and evidence-synthesis content: the PICO framework for formulating the review question, CASP checklists and JBI’s critical appraisal checklists for assessing individual study quality within each strand, the SPIDER framework for structuring the qualitative-strand search, and narrative synthesis and SWiM for the quantitative strand when meta-analysis isn’t possible. The underlying concept is also indexed as a Dictionary term on mixed methods research for readers who want the primary-research definition before applying it at the review level.
Frequently asked questions
Can a mixed-methods systematic review still use meta-analysis?
Yes, within the quantitative strand — a convergent segregated review can pool comparable quantitative studies via meta-analysis as part of its independent quantitative synthesis, before integrating those pooled results with the qualitative synthesis. A convergent integrated review is less likely to run a separate meta-analysis, since the quantitative findings are being transformed (qualitized) into the combined synthesis rather than pooled statistically on their own.
Do I need JBI SUMARI to run a JBI-methodology MMSR?
No — SUMARI is JBI’s own software for managing screening, extraction and synthesis through its methodology, and using it is convenient but not mandatory. Reviews following JBI’s convergent segregated or convergent integrated methodology have been published using other reference-management and synthesis workflows, as long as the methodological steps and reporting match JBI’s published guidance.
What’s the difference between qualitizing and quantitizing?
Qualitizing converts quantitative results into narrative/textual form so they can sit inside a qualitative-style synthesis; quantitizing converts qualitative findings into numeric or categorical form (such as counts of how many studies reported a theme) so they can sit inside a quantitative-style synthesis. JBI’s convergent integrated approach more commonly qualitizes.
