Answers are grounded in this paper and cite their sources.
Causal inference from observational referral records in community clinics is routinely threatened by unmeasured indication confounding: the same operator skill mix that determines whether an ultrasound is clinically indicated also determines whether it is requested. Instrumental variable (IV) designs are frequently proposed as a remedy, but their operating characteristics under realistic clustering and weak instruments remain poorly characterised for this setting. Power was estimated across 2000 replications per design cell for cluster counts of 8, 16, 24 and 32 and effect sizes of 3, 5 and 8 percentage points. The analysis was theoretical and computational; no observational data were used or claimed. Under the declared variance structure, 54 clusters were required to detect a 5 percentage point effect at 80% power, and the minimum detectable effect at 80% power declined monotonically as cluster count rose. The results establish that the BACI instrument is viable only above a cluster threshold that exceeds the size of most community clinic networks, and that IV robustness claims in this literature must be reported against an explicit variance structure rather than asserted.
No takes yet. Share an insight, caveat, or question.
Mwinyikione et al. (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: