What can we learn from studying control arms of randomised VAW prevention intervention evaluations: reflections on expected measurement error, meaningful change and the utility of RCTs

Figures & data

Table 1. Summary of the methods of the trials

Evaluation	Study design	Face to face/ ACASI	Participants	Control arm sample	Control arm intervention exposure	Papers
Stepping Stones Creating Futures, South Africa	2-arm RCT: treatment and delayed intervention control. Assessments: baseline, 12 and 24 months	ACASI	Cohort of men and women. Not recruited as couples	Baseline: 336 men, 338 women (17 clusters); Endline: 268 men, 285 women	None	[13,14]
Indashyikirwa couples, Rwanda	2-arm RCT: treatment and control. Assessments: baseline, 12 and 24 months	ACASI	Cohort of men and of women who were married/cohabiting couples	Baseline: 832 women, 830 men; endline: 818 women, 803 men	None	[15,16]
Indashyikirwa community, Rwanda	2-arm RCT: treatment and control. Assessments: baseline and 24 months	Both used (choice provided)	Residents in households in the districts of the study. Repeat cross-sectional data	Baseline and endline: 700 currently-partnered women, 700 men (14 clusters)	None	[15,16]
Change Starts at Home, Nepal	2-arm RCT: treatment and control. Assessments: baseline, 12 and 28 months	F to F	Household residents in the study districts. Repeat cross-sectional data	Baseline: 720 women; Endline: 717 women (18 clusters)	None	[17,18]
Women Empowerment Programme, Afghanistan	2-arm RCT: treatment and delayed intervention control. Assessments: baseline, 12 (data not presented) and 22 months’	F to F	Cohort of women	Baseline: 714 women (453 currently married); Endline: 537 (441 married)	None	[19,20]
Rural Response System, Ghana	2-arm quasi-experimental study: treatment and control. Assessments: baseline and 24 months	F to F	Households recruitment in 2 control districts. 10 male eas, 10 female. Repeat cross-sectional data	Baseline: 1075 men and 1047 women (20 eas); Endline: 1177 men and 1170 women	None	[21,22]
VATU, Zambia	2-arm RCT: treatment and delayed intervention control. Assessments: baseline and 12 months	ACASI	Cohort of couples (men and women). known problems of IPV and the man harmfully using alcohol	Baseline: 125 couples; endline: 114 women and 108 men	Weekly phone calls (or home visits) by study assessors asking questions on suicidal and homicidal ideation, current risk of IPV and child abuse. With safety plans for any identified risk	[23]

Table 2. Prevalence of IPV in the control arm across studies, by IPV type, and quantum of change between time points

	Gender	IPV measure	Baseline	12-mths	24-mths	Change in percentage points baseline – endline	Largest change in percentage points between two time points	Proportionate change (baseline to endline)	Largest change as a proportion of baseline
Stepping Stones Creating Futures, South Africa	Women	Physical	60.1	49.3	54.4	−5.7	−10.8	−0.095	−0.180
		Sexual	30.2	32.1	35.4	5.2	5.2	0.172	0.172
		Severe	56.2	45.1	50.9	−5.3	−11.1	−0.094	−0.198
	Men	Physical	50.4	53.7	44	−6.4	−6.4	−0.127	−0.127
		Sexual	30.4	44	27.2	−3.2	13.6	−0.105	0.447
		Severe	46.6	49.5	40.7	−5.9	−5.9	−0.127	−0.127
Indashyikirwa couples, Rwanda	Women	Physical	33.1	34.4	32.2	−0.9	1.3	−0.027	−0.027
		Sexual	37.6	38.5	38.1	0.5	0.9	0.013	0.024
		Severe	41.1	43.5	42	0.9	2.4	0.022	0.058
	Men	Physical	22.8	19.4	16	−6.8	−6.8	−0.298	−0.298
		Sexual	19.6	17	15.6	−4	−4	−0.204	−0.204
		Severe	22.4	19.4	16.7	−5.7	−5.7	−0.254	−0.254
Change Starts At Home, Nepal	Women	Physical	20.4	12.9	16.0	−4.4	−7.5	−0.216	−0.368
		Sexual	23.9	15.8	23.0	−0.9	−8.1	−0.038	−0.339
		Severe	28.4	19.6	25.8	−2.6	−8.8	−0.092	−0.310
Women’s Empowerment Programme, Afghanistan	Women	Physical	23.9		26.6	2.7	2.7	0.113	0.113
		Severe	16.8		19.6	2.8	2.8	0.167	0.167
RRS, Ghana	Women	Physical	15.3		10.9	−4.4	−4.4	−0.288	−0.288
		Sexual	8.5		6.8	−1.7	−1.7	−0.200	−0.200
		Severe	17.6		11	−6.6	−6.6	−0.375	−0.375
	Men	Physical	10.7		12.6	1.9	−1.9	0.178	−0.178
		Sexual	16		19.8	3.8	3.8	0.238	0.238
		Severe	17		19.4	2.4	2.4	0.141	0.141
Indashyikirwa community Rwanda	Women	Physical	40.9		41.9	1	1	0.024	0.024
		Sexual	46		43.8	−2.2	−2.2	−0.048	−0.048
		Severe	50.9		49.7	−1.2	−1.2	−0.024	−0.024
	Men	Physical	22.4		24.1	1.7	1.7	0.076	0.076
		Sexual	16.7		17.3	0.6	0.6	0.036	0.036
		Severe	19.7		21.4	1.7	1.7	0.086	0.086
VATU couples study, Zambia	Women	Physical	75.2	47.4		−27.8	−27.8	−0.37	−0.37
		Sexual	69.6	46.5		−23.1	−23.1	−0.332	−0.332
		Severe	81.4	44.9		−36.5	−36.5	−0.552	−0.552
	Men	Physical	74.4	49.4		−25	−25	−0.336	−0.336
		Sexual	43.2	38.3		−5	−5	−0.115	−0.115
		Severe	73.7	39.0		−34.7	−34.7	−0.529	−0.529

Table 3. Mean change in the control arms of across studies, all types of IPV and genders

	Mean change in percentage points baseline – endline (95%CI)	Mean largest change in percentage points between two time points (95%CI)	Mean proportionate change (baseline to endline)	Mean largest change as a proportion of baseline
All studies (12 datasets)	7.01 (3.68, 10.33)	8.39 (5.08, 11.70)	0.175	0.216
All studies with no active control arms (10 datasets)	3.21 (1.59,4.83)	4.88 (0.94, 8.83)	0.134	0.177
Cohort studies with no active control arms (5 datasets)	4.00 (0.38, 7.62)	6.29 (0, 17.04)	0.13	0.171
Repeat cross-sectional (5 datasets)	2.47 (0, 5.56)	3.57 (0,10.22)	0.137	0.182

Table 4. Change in the control arms of different study types, by type of IPV and gender

	Gender	Change in % points baseline – endline (mean)	Range	Largest change in % points between two time points (mean)	Range	Proportionate change baseline to endline (mean)	Range	Largest change as a proportion of baseline (mean)	Range
Physical IPV
All studies*	Female	6.70	0.9,27.8	8.06	1.0,27.8	0.162	0.024,0.37	0.201	0.024,0.37
	Male	8.36	1.7,25.0	9.02	1.7,25.0	0.203	0.076, 0.336	0.216	0.076, 0.336
All studies with no active controls ^@	Female	3.18	0.9, 5.7	4.77	1,10.8	0.127	0.024, 0.289	0.173	0.024, 0.368
	Male	4.20	1.7, 6.8	5.03	1.7,9.7	0.17	0.077, 0.298	0.186	0.077,0.298
Cohorts with no active controls ***	Female	3.10	0.9, 5.7	5.23	2.2, 10.8	0.078	0.027, 0.113	0.12	0.066, 0.180
	Male	6.60	6.4, 6.8	8.25	6.8,9.7	0.213	0.127, 0.298	0.245	0.192, 0.298
Repeat cross-sectional***	Female	3.27	1, 4.4	4.30	1.0, 7.5	0.176	0.024, 0.288	0.227	0.024, 0.368
	Male	1.80	1.7, 1.9	1.80	1.7, 1.9	0.127	0.076, 0.178	0.127	0.076, 0.178
Sexual IPV
All studies^$	Female	5.60	0.5, 23.1	6.55	0.9, 23.1	0.134	0.013, 0.332	0.175	0.024, 0.339
	Male	3.32	0.6,5.0	6.04	0.6, 16.8	0.14	0.036, 0.238	0.229	0.034, 0.553
All studies with no active controls **	Female	2.10	0.5, 5.2	3.24	0.9, 8.1	0.094	0.013, 0.172	0.144	0.024, 0.339
	Male	2.9	0.6, 4.0	6.3	0.6, 16.8	0.146	0.036, 0.238	0.258	0.034, 0.553
Cohorts with no active controls ^#	Female	2.85	0.5, 5.2	2.1	0.9, 3.3	0.093	0.013, 0.172	0.067	0.024, 0.109
	Male	3.6	3.2, 4.0	10.4	4.0, 16.8	0.155	0.105, 0.204	0.378	0.204, 0.553
Repeat cross-sectional ***	Female	1.6	0.9, 2.2	4	1.7, 8.1	0.095	0.038, 0.200	0.137	0.048, 0.339
	Male	2.2	0.6, 3.8	2.2	0.6, 3.8	0.137	0.036, 0.238	0.196	0.036, 0.238
Severe IPV
All studies*	Female	7.99	0.9, 36.5	9.91	1.2, 36.5	0.189	0.022, 0.552	0.240	0.024, 0.552
	Male	10.08	1.7, 34.7	10.66	1.7, 34.7	0.228	0.086, 0.529	0.240	0.086, 0.529
All studies with no active controls **	Female	3.23	0.9, 6.6	5.48	1.2, 11.1	0.129	0.022, 0.167	0.189	0.024, 0.375
	Male	3.93	1.7, 5.9	4.65	1.7, 8.8	0.152	0.086, 0.254	0.168	0.086, 0.254
Cohorts with no active controls ***	Female	3	0.9, 5.3	5.43	2.4, 11.1	0.094	0.022, 0.167	0.141	0.058, 0.198
	Male	5.8	5.7, 5.9	7.25	5.7, 8.8	0.191	0.127, 0.254	0.222	0.189, 0.254
Repeat cross-sectional***	Female	3.47	1.2, 6.6	5.53	1.2, 8.8	0.163	0.024, 0.375	0.236	0.024, 0.375
	Male	2.05	1.7, 2.4	2.05	1.7, 2.4	0.114	0.086, 0.141	0.114	0.086, 0.141

* 12 datasets **9 datasets ***5 datasets ^$ 11 datasets ^# 4 datasets ^@ 10 datasets.

Gibbs A, Washington L, Willan S, et al. The Stepping Stones and Creating Futures intervention to prevent intimate partner violence and HIV-risk behaviours in Durban, South Africa: study protocol for a cluster randomized control trial, and baseline characteristics. BMC Public Health. 2017;17:336.

 PubMed Web of Science ®Google Scholar

Gibbs A, Washington L, Abdelatif N, et al. Impact of the stepping stones and creating futures intervention on intimate partner violence and livelihoods in urban informal settlements in South Africa: cluster randomized control trial. J Adolesc Health. 2019;1–13. DOI:10.1016/j.jadohealth.2019.10.004.

 Web of Science ®Google Scholar

Chatterji S, Stern E, Dunkle K, et al. Mixed methods evaluation: community activism to reduce intimate partner violence in Rwanda. J Glob Health. In press;10(1):010406. DOI:10.7189/jogh.10.010406.

 Web of Science ®Google Scholar

Dunkle K, Stern E, Chatterji S, et al. Indashyikirwa program to reduce intimate partner violence in Rwanda: report of findings from a cluster randomized control trial. Pretoria: South African Medical Research Council; 2019.

 Google Scholar

Clark CJ, Spencer RA, Shrestha B, et al. Evaluating a multicomponent social behaviour change communication strategy to reduce intimate partner violence among married couples: study protocol for a cluster randomized trial in Nepal. BMC Public Health. 2017;17:75.

 Web of Science ®Google Scholar

Clark C, Shrestha B, Gerguson G, et al. Impact of the change starts at home on women’s experience of IPV in Nepal. Soc Sci Med - Popul Health. 2019;13:100530. eCollection 2020 Apr.

 Google Scholar

Gibbs A, Corboz J, Shafiq M, et al. An individually randomized controlled trial to determine the effectiveness of the women for women international programme in reducing intimate partner violence and strengthening livelihoods amongst women in Afghanistan: trial design, methods and baseline findings. BMC Public Health. 2018;18:164.

 Web of Science ®Google Scholar

Gibbs A, Corboz J, Chirwa E, et al. Evaluating the impacts of a combined social and economic empowerment training programme on intimate partner violence, gender norms, and livelihoods amongst women in Afghanistan: an individually randomized control trial and qualitative study. BMJ Glob Health. 2020;5(3):e001946. DOI:10.1136/bmjgh-2019-001946. eCollection 2020.

 Web of Science ®Google Scholar

Addo-Lartey AA, Ogum Alangea D, Sikweyiya Y, et al. Rural response system to prevent violence against women: methodology for a community randomised controlled trial in the central region of Ghana. Glob Health Action. 2019;12:1612604. PubMed PMID: 31134866; PubMed Central PMCID: PMC6440026. eng.

 PubMed Web of Science ®Google Scholar

Ogum Alangea D, Addo-Lartey AA, Chirwa D, et al. Evaluation of the rural response system intervention to prevent violence against women: findings from a community-randomised controlled trial in the Central Region of Ghana. Glob Health Action. 2020;13:1711336.

 PubMed Web of Science ®Google Scholar

Kane JC, Van Wyk S, Murray S, et al. Testing the effectiveness of a transdiagnostic treatment approach in reducing violence and alcohol abuse among families in Zambia: study protocol of the Violence and Alcohol Treatment (VATU) trial. Global Mental Health. 2017;4. DOI:10.1017/gmh.2017.10.

 Web of Science ®Google Scholar