Table 1.

General characteristics of included studies

Title	Study Design	Setting	Scale2	Description of Intervention
Ande-michael et al. (2009)	Hospital-based before and after study with qualitative component	Eritrea, remote areas of two coastal regions of the Red sea	655,279 people 11 MWH	11 health facilities with MWH for women living at least 10 km distance from facility. MWHs had an ambulance for referal to higher level facilities for complications. During admission, consumables were provided to women. Community support provided through supplies. Equity considerations made for women residing more than 10 km from health facility. Staff at MWH were trained. Part of a strategy implemented by MOH
Chandra-mohan et al. (1994)	Hospital cohort (childbirth outcomes over time)	Zimbabwe, Rural	208,000 people 1 MWH	Free self-catering temporary accommodations 150 m from labour ward. Women advised to stay at MWH from 36 weeks gestation. Target population was women identified as risk in ANC. MWH offered ANC and health education.
Chandra-mohan et al. (1995)	Cohort analytic (two group pre + post)	Zimbabwe, Rural	208,000 people 1 MWH	See Chandramohan et al., 1994
Danel et al. (2003)	World bank report	Honduras, National	Population nr 5 MWH	Attached to rural hospitals.
Ecker-mann et al. (2008)	Case study with qualitative components	Lao People’s Democratic Republic (PDR), Remote-rural	27,539 people No MWH, 17 to be built	Improve maternal outcomes in remote communities with a high proportion of ethnic minorities and disadvantaged groups economically and in health indices. Women provided with nutrition and baby care training, handicraft training and have opportunity to earn an income while staying at MWH. All given information and opportunities for micro-credit initiatives. MWHs designed for privacy before, during and after birth (for uncomplicated births conducted in MWH in traditional birthing position)
Feresu et al. (2003)	World bank report	Zimbabwe National	Overview of 255 MWH
Fraser (2008)	Case study	Peru, Rural and urban	Population nr 2 MWH (390 available nationally)	Reported outcomes of key interventions to address MMR in Peru. MWH near health centres that refer cases to hospitals. MWH are part of a strategy implemented by MOH
Garcia Prado et al. (2012)	Cross-sectional survey and qualitative components	Nicaragua, Rural	Population nr 18 MWH	Women spend 2 weeks before and 1 week after childbirth at MWH, where food and lodging is provided. Most homes extend their services beyond medical visits and education on SHR, offering advice and counselling on diverse issues (domestic violence, selling handmade prouducts, and obtaining identify cards or land titles). Women referred from mobile health teams and TBAs. Situated near health centres. MOH has a strategy to promote MWHs.
Gaym et al. (2012)	Hospital based cohort with a qualitative component	Ethiopia, Rural	Population nr 9 MWH	Faith based organizations pioneered the construction of MWHs in Ethiopia since the late 1980’s, then adopted by NGOs as well as public health facilities. Conditions within each varied, activities included outreach to increase community awareness of MWHs. Women referred by staff at peripheral health facilities, and outreach teams. Women also came based on recommendations from other women who had used facility. Situated within compound of health facility.
Gorry (2011)	Case study	Cuba, Rural and urban	Population nr 327 MWH	15 MWHs were introduced in 1962 and grew to 327. Existing houses are reconditioned to create a home-like environment for monitoring health and wellbeing of woman and fetus. Concept has been further developed to emphasize nutrition and diet, and provision of ambulatory services so women can take meals and classes at MWH, but return home in the evenings. MWHs follow guidelines designed by Ministry of Public Health’s maternal child health program in collaboration with UNICEF, describing criteria for admission, diagnostic and clinical guidelines for identifying risk factors and protocols for treatment in MWHs.
Kelly et al. (2010)	Hospital cohort (childbirth outcomes over time)	Ethiopia, Rural	800,000 people 1 MWH	40 bed MWH, located within hospital grounds. Original facility built in 1973 in local style with thatched roof, which caught fire in 1999; replaced by corrugated roof. A companion resides at MWH, finds firewood and food, and cooks for her. High-risk women spend last few weeks of pregnancy in MWH.
Knowles et al. (1988)	Case study	Malawi	Population nr 1 MWH	Women referred from other medical facilities and can self-refer. Situated in hospital ground.
Larsen et al. (1978)	Hospital cross sectional survey	South Africa, Rural	nr	Nr
Lori et al. (2013a)	Qualitative study	Liberia, Rural, post conflict	78,446 people 5 MWH	Served women affected by conflict. Women self-refer. Situated near health facilities.
Lori et al. (2013b)	Cohort analytic (two group pre + post)	Liberia, Rural post conflict	>50,000 people 4 MWH	Served women affected by conflict.
Martey et al. (1995)		Ghana, Rural	131,229 people 5 MWH	Nr
Millard et al. (1991)	Hospital cohort study	Zimbabwe, Rural	Population nr 1 MWH	Women self-referred themselves to the facility. 2 min walk from hospital. MOH policy exists supporting MWHs.
Mramba et al. (2010)	Cross sectional survey, qualitative components	Kenya	Population nr 1 MWH	50 m from the maternity unit at a District Hospital. It has a capacity of 40 people: 20 pregnant women and 20 healthcare workers. Referrals mostly by health workers. Referrals from health workers.
Poovan et al. (1990)	Hospital cross-sectional survey	Ethiopia, Rural	300,000 people 1 MWH	Women referred during outreach ANC conducted by nurse midwives and TBAs. Situated close to the hospital.
Ruiz et al. (2013)	Qualitative study	Guatemala, Urban	Population nr 2 MWH	Focus on attracting indigenous women. Women referred from TBAs and health centre physicians. Women could also self-refer. 3 km from the hospital. Part of a MOH strategy to increase utilisation in this region.
Schooley et al. (2009)	Qualitative inquiry (focus groups and in-depth key informant interviews, unstructured, focused observations)	Guatemala	Population nr 1 MWH	Focus on increasing utilisation of health services by indigenous women. Situated adjacent to a local hospital.
Shrestha et al. (2007)	Cross-sectional survey and qualitative component	Nepal, Lowland conflict	Population nr Study not linked to existing MWH (27 MWH available)	Working in a context of conflict. MOH supported MWH to increase health facility utilisation.
Spaans et al. (1998)	Household-level cross-section	Zimbabwe	Population nr 4 MWH	In the hospital grounds.
Tumwine et al. (1996)	Cohort analytic (two group pre + post)	Zimbabwe	100,000 people Number of MWH nr	Women referred by health centre staff, TBAs and could refer themselves. 100 m from hospital.
van Lonkhuij-zen et al. (2003)	Hospital cross-section	Zambia, Rural	60,000 people 1 MWH	Women referred during monthly outreach clinics conducted by midwives. Situated next to hospital.
Wessel(1990)	Case study	Nicaragua, Rural	Population nr 1 MWH	Aimed at supporting refugees from the civil war. Self-referral.
Wild et al. (2012)	Interrupted time series	Timor-Leste Remote-rural	>100,000 people 2 MWH	Connected by a walkway to the hospital, and near a health centre. MOH run as part of their maternal health strategy.
Wilson et al. (1997)	Qualitative study, with MWH utilisation rates	Ghana, Rural	126,000 people 1 MWH	Referrals from private midwives and health posts. Situated in an unused ward in the hospital.

1 Year of study or report; 2 Catchment population reportedly covered by MWH and number of MWH included in article; 3 Health indices reported as background levels in the article only, pertinent to locality, population of interest and time period where available. Health indices as a result of the MWH intervention not included

Abbreviations: MMR = maternal mortality ratio/100000, PMR = perinatal mortality/1000, SBA = skilled birth attendance, IDR = institutional delivery rate, HB = home births, ANC = antenatal care, PHC = primary health centres, TBA = traditional birth attendants, MOH = ministry of health nr = not reported