Waiswa 2015.

Methods	Uganda Newborn Study (UNEST) 2 armed cluster‐randomised trial.
Participants	Sample size: 63 clusters randomised; baseline survey n = 194 intervention and n = 201 controls; endline survey n = 894 intervention and n = 893 control. Clusters: Uganda, Iganga and Mayuge districts in eastern Uganda, predominantly rural with 65 villages and total population of 70,000 at time of the study. Local health services include 1 100 bed hospital and 19 health centres that provide delivery services. 63 villages randomised (31 intervention and 32 control. Individuals: all consenting pregnant women and their newborns residing in the study area between September 2009 and August 2011. women were eligible for the baseline survey if they had a live birth in the last 4 months; women were eligible for the end line survey if they had a live birth in the last 12 months.
Interventions	Target: community and health systems. Arm 1: CHWs made 5 home visits (2 prenatal and 3 postnatal) with extra visits for sick or small newborns. Health facility strengthening in all facilities (both arms) to improve quality of care. Facility strengthening included: "6‐day in‐service training, provision of a once‐off catalytic supply of equipment and medicines, as well as collaboration with the district health team to continuously improve the quality of care provided to mothers and newborns" (p. 4). Arm 2: the control arm received standard care as well as facility strengthening.
Outcomes	Trial primary outcomes: ANC coverage and services. Birth preparedness, skilled attendance at delivery, and postnatal care, as well as increases in healthy practices including breastfeeding, thermal care, and hygiene. Review outcomes reported: Primary: ANC coverage 4 or more visits; Mortality data were collected but not reported in this publication (authors emailed with no reply 5/15). Secondary: deliveries in a health facility, ANC coverage one visit Follow‐up: baseline survey March‐Aug 2008; Study began 01/09/2009 and ended 01/08/2011; end line survey Sept‐Nov 2011. Home visits were: 2 prenatal (first and third trimesters) and 3 postnatal (days 1, 3 and 7).
Notes	Funders: this study was supported by the Sida/SAREC‐Makerere University‐ Karolinska Institutet Research collaboration as well as by funds provided by Save the Children through a grant from the Bill & Melinda Gates Foundation.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated restricted randomisation was done in a 1‐to‐1 ratio by an independent epidemiologist from the London School of Hygiene and Tropical Medicine.
Allocation concealment (selection bias)	Low risk	Allocation done remotely as above.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Women and staff not blind.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Outcomes assessors not blind.
Recruitment bias (for cluster RCTs)	Low risk	No problems with recruitment reported.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Different populations for pre and post‐intervention surveys.
Selective reporting (reporting bias)	High risk	Mortality data not reported yet in published reports; analysis of a household survey to document mortality outcomes is ongoing; low birthweight not reported according to intervention arm; preterm birth mentioned in discussion but unclear if these data were available.
Analysis bias	Low risk	ITT analysis. Authors report adjustments made for cluster design.
Other bias	Unclear risk	Demographic variables comparable at baseline (age, parity and household wealth); only women with a live birth were eligible; mortality data are not reported; there were no buffer zones between control and intervention villages.
Overall risk assessment	Unclear risk	Conduct of the trial is of low risk of bias, though reporting bias is unclear.