Key result
Detecting a ~5-point increase in hypertension treatment cessation from CHW attrition requires 54 clusters.
Why the study?
The consequences of community health worker attrition for chronic-disease treatment are typically assumed rather than estimated, and the ability of a before-after-control-impact design to recover its effect on hypertension treatment cessation without randomisation was unknown.
What can a before-after-control-impact (BACI) design recover about the effect of community health worker attrition exposure on hypertension treatment cessation in Malawian districts when randomisation is unavailable?
What can a before-after-control-impact (BACI) design recover about the effect of community health worker attrition exposure on hypertension treatment cessation in Malawian districts when randomisation is unavailable?
This simulation study demonstrates that evaluating the impact of community health worker attrition on hypertension treatment cessation using a before-after-control-impact design requires 54 clusters to detect a 5-percentage-point effect at 80% power, emphasizing the need to conceptualize attrition as a program-level exposure.
BACI designs may empirically estimate CHW attrition effects on hypertension cessation; leaves open causal inference and generalizability without randomization.
Community health worker (CHW) attrition is treated in much medical-sociological writing as a workforce problem whose consequences for chronic-disease treatment are assumed rather than estimated. the present analysis reframes CHW attrition as an exposure that is assigned by programme managers rather than by patients, and it asks what a before-after-control-impact (BACI) design can recover about the effect of that exposure on hypertension treatment cessation when randomisation is unavailable. No observational data are used or claimed. The analysis characterises the operating characteristics of the BACI estimator under a declared variance structure: between-cluster SD = 6.0, residual SD = 4.5, intraclass correlation = 0.64. Power was estimated across 2000 design cells defined by cluster counts of 8, 16, 24 and 32 and effect sizes of 3, 5 and 8 percentage points. Under these parameters, 54 clusters are required to detect a 5-percentage-point effect at 80% power. The credibility of any causal claim about CHW attrition and treatment cessation depends less on the sophistication of the estimator than on whether the exposure is conceptualised as a programme-level assignment, and it specifies the design conditions under which that conceptualisation is defensible.
No takes yet. Share an insight, caveat, or question.
Banda et al. (2026) studied Hypertension. Community health worker attrition was evaluated on Hypertension treatment cessation. Community health worker attrition in a BACI simulation requires 54 clusters to detect a 5-percentage-point effect on hypertension treatment cessation at 80% power (ICC=0.64).
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: