Table A2.

Study rigor assessment

Citation	Study design includes pre/post intervention data	Study design includes control or comparison group	Study design includes cohort	Comparison groups equivalent at baseline on socio‐demographics	Comparison groups equivalent at baseline on outcome measures	Selection process minimizes bias a	Participants randomly allocated to the intervention	Control for potential confounders	Follow‐up rate ≥75%
Baumgartner et al. 2014	Yes	No	No	N/A	N/A	Yes	No	Yes	N/A
Chabikuli et al. 2009	Yes	No	No	N/A	N/A	Yes	No	No	N/A
Chibwesha et al. 2011	No	No	Yes	N/A	N/A	No	No	No	NR
Church et al. 2012; Church et al. 2015	No	Yes	No	N/A	N/A	Yes	No	Yes	N/A
Coyne, Hawkins, and Desmond 2007	Yes	Yes	No	N/A	N/A	Yes	No	No	N/A
Grossman et al. 2013; also Shade et al. 2013 and Onono et al. 2015	Yes	Yes	No	Yes	No	Yes	Yes	Yes	N/A
Hoke et al. 2014	Yes	No	No	N/A	N/A	Yes	No	No	N/A
Kosgei et al. 2011	No	Yes	Yes	Yes	No	No	No	Yes	Yes
McCarraher et al. 2011	Yes	Yes	Yes	No	No	No	No	Yes	Yes
Phiri et al. 2016	No	Yes	No	NR	NR	No	No	No	N/A
Sarnquist et al. 2014	No	Yes	Yes	Yes	Yes	Yes	Yes	No	Yes
Thyda et al. 2015	Yes	No	No	N/A	N/A	NR	No	No	N/A
Wanyenze et al. 2015	No	Yes	No	No	N/A	Yes	No	Yes	N/A
Wielding and Flynn 2015	Yes	No	No	NR	NR	Yes	No	No	N/A

N/A = Not applicable. NR = Not reported.

^aFor most studies, integration is a facility‐level intervention, and the target populations for the included studies were women living with HIV who were attending HIV treatment. Therefore randomization was not superior to other approaches to reduce selection bias, including using a census of all visits (e.g., the studies using electronic medical data) or eligibility screening of all or a randomly selected subset of women attending the sites during an enrollment period. Studies using either of these approaches for selection of the study sample were assessed as having minimized selection bias.