Table 9. Community actions to address GBV.

	Baseline		Follow-up				Odds-ratio of follow-up to baseline prevalence among control clusters (time effect)			Odds-ratio of intervention to control clusters prevalence at follow-up (intervention effect)
	Interv’n	Control	Interv’n		Control		Est. OR	95% CI	p valuea	Est. OR	95% CI	p valuea
	Percent	Percent	Percent		Percent
	Freq/N	Freq/N	Freq/N		Freq/N
Prevalence of reported actions in the past 12 months
Witnessed an act of GBV or VAC	38.6	43.8	31.3	31.3			0.60	0.48–0.75	<0.001	1.14	0.84–1.55	0.386
	251/650	281/641	192/613	193/616
Took action to stop GBV or help a survivor (among those who witnessed an act)	39.8	36.6	50.5	47.2			1.49	1.06–2.10	0.022	1.09	0.71–1.69	0.690
	100/251	102/279	97/192	91/193
Started a conversation about GBV or VAC	18.8	21.3	21.6	15.8			0.73	0.56–0.95	0.020	1.56	1.12–2.18	0.008
	123/656	137/643	135/624	98/622
Prevalence of reported assessments and awareness of the community’s response to GBV in the past 12 months
Community leaders have ever spoken out or acted to address GBV or VAC	30.2	30.8	38.5		22.8	1.00		0.81–1.24	0.978	2.02	1.51–2.71	<0.001
	198/656	198/643	239/621		142/624
Community has done a good (or very good) job of responding to IPV and VAC	34.8	28.9	38.3		27.4	0.89		0.70–1.13	0.334	1.35	0.98–1.87	0.064
	227/653	185/641	237/619		171/623
										Est. OR	95% CI	p valueb
Aware of a Community Action Group on GBV	Intervention		25.8%			X				2.45	1.50–4.00	<0.001
			161/625
	Control		12.5%			X				Ref	-	-
			78/624
Aware of community volunteers who help GBV survivors get to services	Intervention		31.0%			X				2.49	1.57–3.96	<0.001
			193/624
	Control		15.4%			X				Ref	-	-
			96/624

^a p values are based on a GLMM with cluster-specific baseline prevalence equal to the true baseline prevalence plus a random effect for all clusters including those randomized to the intervention.

^b p values are based on a GLMM with cluster treated as a random effect.