CoChat

Applications and implementation considerations for stepped-wedge designs in sub-Saharan Africa: a systematic review

Open access

The DOI matches its Crossref record; 122 of the 122 references it cites were found in a registry or an index; 340 of the 364 sources its claims cite were checked for support.

Abstract

Systematic review of 85 stepped-wedge design (SWD) studies in sub-Saharan Africa from 2000 to 2024 evaluating public health, clinical, and social interventions, highlighting methodological rigor, logistical phasing advantages in resource-limited health facilities, and the necessity of accounting for underlying secular time trends and cluster correlation.

Overview

CoChat Summary

TL;DR

This systematic review synthesizes 85 stepped-wedge design studies conducted in sub-Saharan Africa from 2000–2024, covering 1,895,788 participants across HIV/AIDS, maternal and child health, tuberculosis, malaria, and other domains. It finds SWD is used mainly for ethical and logistical phased implementation, but reporting of design and implementation details is inconsistent.

Key findings

  • 85 studies were included, involving 1,895,788 participants; sample sizes ranged from 17 to 780,000. Most studies were facility-based (72, 84.7%), while 13 (15.3%) were community-based.
  • HIV/AIDS was the largest health domain (34 studies, 40%), followed by maternal and child health (17, 20%), tuberculosis (11, 12.9%), malaria (7, 8.2%), and general health (7, 8.2%).
  • Studies were concentrated in Uganda (16, 18.8%), Kenya (14, 15.3%), and South Africa (11, 12.9%), with fewer from West and Central Africa.
  • Design parameters varied widely: the abstract reports 2–12 steps and 4–54 clusters, while the narrative reports 1–27 steps and 3–142 clusters; study duration ranged from 2 to 60 months.

Methods

  • Systematic review following PRISMA and MOOSE, registered with PROSPERO (CRD42024530774); searched PubMed, Web of Science, CINAHL, PsycINFO, Cochrane Library, Google Scholar, and citation tracking from January 2000 to March 2024.
  • Included original research trials using SWD in SSA with clusters, randomization, and controls; excluded non-SWD, non-SSA, individual-randomized, retrospective SWD, editorials, commentaries, and non-peer-reviewed articles. Screening yielded 873 titles/abstracts, 93 full texts, and 85 included studies.
  • Data were extracted on study characteristics, SWD implementation, statistical methods, and sample size calculations; no meta-analysis was performed.

Limitations

  • Stated: geographic concentration in a few countries, variability in methodological reporting, limited representation of non-communicable diseases, inconsistent reporting of study durations and intervals, and absence of a meta-analysis limiting quantitative synthesis.
  • Evident: the PRISMA flow numbers conflict with the text (2,562 vs 1,565 records; 93 vs 132 full texts), and cluster/step ranges differ between the abstract and narrative.
  • Evident: the search was restricted to English-language publications for feasibility, despite the eligibility criteria mentioning English or French.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.3389/fepid.2026.1529289

Cites

122 of 122 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

Figure 1 Diagrammatic illustration of a stepped-wedge design. This figure illustrates a typical stepped-wedge trial with five clusters observed over six time periods (T1–T6). All clusters begin in the control condition and transition sequentially to the intervention condition at predefined time points until all clusters receive the intervention. White cells indicate control periods, while shaded cells represent intervention periods. This design allows within-cluster and between-cluster comparisons over time.
Figure 1 Diagrammatic illustration of a stepped-wedge design. This figure illustrates a typical stepped-wedge trial with five clusters observed over six time periods (T1–T6). All clusters begin in the control condition and transition sequentially to the intervention condition at predefined time points until all clusters receive the intervention. White cells indicate control periods, while shaded cells represent intervention periods. This design allows within-cluster and between-cluster comparisons over time.CC BY · PubMed Central
Figure 2 PRISMA flow diagram of study selection. The PRISMA flow diagram summarises the study identification, screening, eligibility assessment, and inclusion process. A total of 1,565 records were identified from databases and supplementary sources. After removal of duplicates, 873 records were screened, 132 full texts were assessed for eligibility, and 85 studies met the inclusion criteria for the systematic review.
Figure 2 PRISMA flow diagram of study selection. The PRISMA flow diagram summarises the study identification, screening, eligibility assessment, and inclusion process. A total of 1,565 records were identified from databases and supplementary sources. After removal of duplicates, 873 records were screened, 132 full texts were assessed for eligibility, and 85 studies met the inclusion criteria for the systematic review.CC BY · PubMed Central
Figure 3 Heatmap showing distribution of stepped-wedge studies by country and health domain. This heatmap presents the distribution of included studies across sub-Saharan African countries and major health domains (e.g., HIV, TB, maternal and child health, malaria, NCDs). Darker shading indicates a higher number of studies within each country–domain combination. The figure highlights the concentration of stepped-wedge studies in Uganda, Kenya, and South Africa, with HIV and maternal and child health being the most frequently studied domains.
Figure 3 Heatmap showing distribution of stepped-wedge studies by country and health domain. This heatmap presents the distribution of included studies across sub-Saharan African countries and major health domains (e.g., HIV, TB, maternal and child health, malaria, NCDs). Darker shading indicates a higher number of studies within each country–domain combination. The figure highlights the concentration of stepped-wedge studies in Uganda, Kenya, and South Africa, with HIV and maternal and child health being the most frequently studied domains.CC BY · PubMed Central
Figure 4 Types of cluster units used in stepped-wedge studies. This schematic summarises the types of clusters used in included studies. Cluster units commonly included health facilities/clinics, communities or districts, schools, faith-based units, and household/village clusters, reflecting the flexibility of stepped-wedge designs across diverse implementation settings.
Figure 4 Types of cluster units used in stepped-wedge studies. This schematic summarises the types of clusters used in included studies. Cluster units commonly included health facilities/clinics, communities or districts, schools, faith-based units, and household/village clusters, reflecting the flexibility of stepped-wedge designs across diverse implementation settings.CC BY · PubMed Central
Figure 5 Categories of interventions evaluated using stepped-wedge designs. This figure categorises interventions evaluated in the included studies into five broad groups: biomedical and medicines; diagnostics and digital tools; health system strengthening; training and capacity building; and community and environmental interventions. The wide range of intervention types demonstrates the adaptability of stepped-wedge designs across clinical, health system, and community-based interventions.
Figure 5 Categories of interventions evaluated using stepped-wedge designs. This figure categorises interventions evaluated in the included studies into five broad groups: biomedical and medicines; diagnostics and digital tools; health system strengthening; training and capacity building; and community and environmental interventions. The wide range of intervention types demonstrates the adaptability of stepped-wedge designs across clinical, health system, and community-based interventions.CC BY · PubMed Central
Figure 6 Strategies used to balance clusters prior to randomisation. This diagram shows balancing strategies applied before randomisation, including stratification (by region/province, facility volume/caseload, urban–rural status) and matching/pairing (by geographic proximity and catchment size). These approaches were used to improve baseline comparability and reduce imbalance across clusters.
Figure 6 Strategies used to balance clusters prior to randomisation. This diagram shows balancing strategies applied before randomisation, including stratification (by region/province, facility volume/caseload, urban–rural status) and matching/pairing (by geographic proximity and catchment size). These approaches were used to improve baseline comparability and reduce imbalance across clusters.CC BY · PubMed Central
Figure 7 Randomisation methods used in stepped-wedge studies. This figure summarises the randomisation approaches used across studies, including computer-generated sequences (e.g., Stata/SAS, Excel random number generators), public or stakeholder-led methods (e.g., lotteries, drawing names), two-stage or block randomisation, and quasi-random or systematic allocation. The diversity reflects varying methodological and contextual constraints.
Figure 7 Randomisation methods used in stepped-wedge studies. This figure summarises the randomisation approaches used across studies, including computer-generated sequences (e.g., Stata/SAS, Excel random number generators), public or stakeholder-led methods (e.g., lotteries, drawing names), two-stage or block randomisation, and quasi-random or systematic allocation. The diversity reflects varying methodological and contextual constraints.CC BY · PubMed Central
Figure 8 Allocation concealment and masking practices. This schematic presents how allocation concealment was handled in included studies, categorised as no concealment, partial concealment (e.g., allocation revealed shortly before each step), or not reported. Logistical constraints commonly limited full concealment, reflecting real-world implementation challenges in stepped-wedge trials.
Figure 8 Allocation concealment and masking practices. This schematic presents how allocation concealment was handled in included studies, categorised as no concealment, partial concealment (e.g., allocation revealed shortly before each step), or not reported. Logistical constraints commonly limited full concealment, reflecting real-world implementation challenges in stepped-wedge trials.CC BY · PubMed Central
Figure 9 Distribution of number of steps across included studies. This histogram shows the frequency distribution of the number of steps used in stepped-wedge designs. Most studies used between 3 and 7 steps, with fewer studies employing very short (1–2 steps) or highly granular designs (≥10 steps), reflecting trade-offs between feasibility and temporal resolution.
Figure 9 Distribution of number of steps across included studies. This histogram shows the frequency distribution of the number of steps used in stepped-wedge designs. Most studies used between 3 and 7 steps, with fewer studies employing very short (1–2 steps) or highly granular designs (≥10 steps), reflecting trade-offs between feasibility and temporal resolution.CC BY · PubMed Central
Figure 10 Distribution of the number of clusters per study. This plot displays the frequency distribution of the number of clusters used across included studies. Most studies involved fewer than 30 clusters, while a small number of large-scale studies used substantially higher numbers of clusters, indicating variability in scale and design complexity across stepped-wedge trials in SSA.
Figure 10 Distribution of the number of clusters per study. This plot displays the frequency distribution of the number of clusters used across included studies. Most studies involved fewer than 30 clusters, while a small number of large-scale studies used substantially higher numbers of clusters, indicating variability in scale and design complexity across stepped-wedge trials in SSA.CC BY · PubMed Central
Figure 11 Rollout patterns used in stepped-wedge designs. This schematic illustrates common stepped-wedge rollout patterns, including wave-based rollouts with regular time intervals, paired or grouped transitions of clusters, and modified stepped-wedge designs retaining parallel controls or incorporating observation–intervention phases. These variations reflect adaptations of the stepped-wedge design to operational and logistical constraints.
Figure 11 Rollout patterns used in stepped-wedge designs. This schematic illustrates common stepped-wedge rollout patterns, including wave-based rollouts with regular time intervals, paired or grouped transitions of clusters, and modified stepped-wedge designs retaining parallel controls or incorporating observation–intervention phases. These variations reflect adaptations of the stepped-wedge design to operational and logistical constraints.CC BY · PubMed Central

Verification

Claims the paper makes

209 Valid · 35 Not valid · 41 Not checked · 166 Own results and interpretations

167 more claims · 54 with their cited sources checked

Create a free account to see all 451 claims and to verify the 8 references still open.

See all claims — free account
  • Probably not a substantial claim. 225
  • It may state two separate claims as one. 122
  • The quote only partly states this claim. 4

Create a free account to see these claims.

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

5 of 10 · the rest aren’t loaded

5 more findings

Create a free account to see every finding of this check.

See all findings — free account

Citations

Cites

Show all 122 references

Metadata

Authors
Musa, Z. A., Akinsolu, F. T., Salako, A. O., Abodunrin, O. R., Ola, O. M., & Ezechi, O. C.
Affiliations
  • Nigerian Institute of Medical Research (NG)
  • Lead City University (NG)
Corresponding author
Akinsolu, Folahanmi Tomiwa
Venue
Frontiers in Epidemiology
Published
2026-03-25
Volume
6
Pages
1529289-1529289
ISSN
2674-1199 (ISSN-L)
OpenAlex
W7140307959
Cited by
0
References
120
FWCI
0.00
Type
journal article
Language
en
Open access
Locations
  • Frontiers in Epidemiology (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Topics
  • Health Sciences › Health Professions › General Health Professions › Health Policy Implementation Science
  • Physical Sciences › Mathematics › Statistics and Probability › Statistical Methods in Clinical Trials
  • Physical Sciences › Mathematics › Statistics and Probability › Advanced Causal Inference Techniques
Keywords
stepped wedge design, sub-Saharan Africa, implementation science, public health interventions, sample size determination, intra-cluster correlation, cluster randomization, HIV/AIDS, maternal and child health, tuberculosis, malaria, research ethics, primary outcomes, intervention effectiveness evaluation, research capacity building, health professions education, health system strengthening, health care financing, health policy evaluation

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

The claim verdicts this report counts are experimental: their accuracy is not yet measured.

A language model drew the claims from the paper on 2026-10-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-10-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Cited sources are checked with a language model’s help (gpt-6-sol).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Applications and implementation considerations for stepped-wedge designs in sub-Saharan Africa: a systematic review. https://app.cochat.ai/papers/10.3389/fepid.2026.1529289
The paper
Musa, Z. A., Akinsolu, F. T., Salako, A. O., Abodunrin, O. R., Ola, O. M., & Ezechi, O. C. (2026). Applications and implementation considerations for stepped-wedge designs in sub-Saharan Africa: a systematic review. Frontiers in Epidemiology. https://doi.org/10.3389/fepid.2026.1529289

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.