Gender-based violence screening cascades in South African antenatal care: a stepped-wedge implementation evaluation of routine enquiry and reproductive health outcomes

Routine enquiry for gender-based violence (GBV) in South African antenatal care is widely mandated, yet the design parameters that determine whether a stepped-wedge implementation evaluation can detect its effects on reproductive health outcomes remain uncharacterised. Power was estimated across 2000 design cells for cluster counts of 8, 16, 24 and 32 and effect sizes of 3, 5 and 8 percentage points.56 clusters are required to detect a 5-percentage-point effect at 80% power, that power is highly sensitive to the number of clusters at small effect sizes, and that the minimum detectable effect at 80% power falls steeply as clusters increase from 8 to 32. Kernel density estimates of power across design cells are bimodal, and residual diagnostics indicate that the BACI interaction is well identified when cluster counts exceed 24. The findings imply that stepped-wedge evaluations of GBV screening cascades in antenatal care are underpowered at the cluster counts typically available to district health services, and that implementation evaluations should be designed around the minimum detectable effect rather than the nominal sample size.