Table 7.

Rwanda difference–in–differences (DiD) regional results: IcSCI and QC groups

	IcSCI*						QC*
Group	IcSCI		Non–intervention		DiD†	N	QC		Non–intervention		DiD†	N
Time	BL	FU	BL	FU			BL	FU	BL	FU
Percent of CHWs who manage all 5 products, in stock on DOV	53	46	58	37	14	351	35	62	58	36	49‡	346

CHW – community health worker, DiD – Difference in Differences, DOV – day of visit, IcSCI – Incentives for Community Supply Chain Improvement, QC – Quality Collaboratives,

*Controls: CHW has formal training on how to manage medicines and health products, CHW has training in pneumonia, malaria, or diarrhea, CHW has obstacles to transport

†DiD is calculated as DID = (Intervention Follow up% – Intervention Baseline%) – (Comparison Follow up% – Comparison Baseline%). Results displayed represent two steps in the analysis of the data: the significance, denoted by the stars, represents the results from the multivariate logistic regression on the time–group interaction variable, which is the key independent variable of a DiD regression. Since the interaction coefficient is non–intuitive, we have instead depicted the difference over time between the intervention and non–intervention groups using the predicted probabilities resulting from the regression. Essentially, this is the net percentage point change in the intervention region once the comparison group change is subtracted.

‡P < 0.001.