The tyranny of distance: maternity waiting homes and access to birthing facilities in rural Timor-Leste
Kayli Wild a, Lesley Barclay b, Paul Kelly c & Nelson Martins d
a. Menzies School of Health Research, Charles Darwin University, PO Box 41096, Casuarina, Northern Territory 0811, Australia.
b. University Centre for Rural Health, University of Sydney, Sydney, Australia.
c. National Centre for Epidemiology and Population Health, Australian National University, Canberra, Australia.
d. Ministry of Health, Dili, Timor-Leste.
Correspondence to Kayli Wild (e-mail: email@example.com).
(Submitted: 04 April 2011 – Revised version received: 04 October 2011 – Accepted: 10 October 2011 – Published online: 02 November 2011.)
Bulletin of the World Health Organization 2012;90:97-103. doi: 10.2471/BLT.11.088955
Maternity waiting homes are residential facilities where women who live remotely can wait before giving birth at a hospital or health centre. The commonly accepted hypothesis is that more women from remote areas would access birthing facilities if they could wait for the onset of labour in a maternity waiting home. Thus, the aim of implementing a maternity waiting home strategy is to reduce maternal and perinatal mortality by improving access to skilled birth attendance and emergency obstetric care, particularly for women in rural and remote areas.1
Maternity waiting homes have been documented in the scientific literature since the 1960s.2 In 1991, the World Health Organization (WHO) highlighted the potential advantages of implementing maternity waiting homes as part of a package of essential obstetric services. WHO stated that: “More attention should be focused on this little publicized, but highly important, approach to obstetric care at first referral level in rural hospitals, and the results should be evaluated and published.”3 Shortly afterwards, WHO released a review of maternity waiting homes containing a list of what was needed when establishing such homes as part of a national safe motherhood programme.1 The Safe Motherhood Action Agenda, published in 1997, reinforced the need for further evaluation of this strategy.4
Since being included as an option in the safe motherhood programme, implementation of maternity waiting homes has become an increasingly popular strategy in health systems in developing countries. More countries are incorporating these homes into ministry of health policy, often backed by United Nations (UN) agencies and other large donors. Organizations such as the United Nations Population Fund (UNFPA), The World Bank and WHO have recently supported maternity waiting home projects in countries as diverse as Afghanistan, Cambodia, Eritrea, the Gambia, the Lao People's Democratic Republic, Lesotho, Maldives, Mongolia, Morocco, Mozambique, Nepal, Sri Lanka and Timor-Leste.
Maternity waiting homes represent a significant investment in terms of construction, running costs and human resources, and their efficacy is often contested among WHO and UN country representatives. Furthermore, no research has assessed whether women from remote areas are more likely to use a hospital once a maternity waiting home is available. Two papers that reported on distance simply compared distance from home for users of maternity waiting homes and for those going directly to hospital.5,6 However, a study design that assesses distance from home between two groups after an intervention provides no indication as to whether or to what extent maternity waiting homes improve access for women in remote areas. A before-and-after study based on place of residence for all women giving birth in a health facility would provide a more accurate assessment of whether the establishment of maternity waiting homes affects the use of birthing services.
The implementation of maternity waiting homes in Timor-Leste provided an opportunity to conduct a before-and-after distance analysis. This analysis was possible because of the unique way in which the homes were incorporated into birthing facilities, given that all women who gave birth in the health facility also used the maternity waiting home. This study reports on the impact of maternity waiting homes on the use of facility-based birthing services by rural women in two districts of Timor-Leste.
Timor-Leste is a small country situated in the Indonesian archipelago. In 1999, after 25 years of occupation by the Indonesian military, Timor-Leste voted for independence. The history of war, ongoing political instability, limited infrastructure and poor social determinants of health have combined to produce one of the highest maternal mortality ratios in the world, recently estimated to be between 557 and 929 per 100 000 live births.7,8 The proportion of women giving birth with a skilled attendant continues to be low (30%), and only 22% of women give birth in a health facility.7
In 2005, the Timor-Leste Ministry of Health and its technical advisers developed the National Maternity Waiting Home Strategy. The aim was to improve access to facility-based delivery for women in remote areas and thus to reduce the high maternal mortality rate.9 Under this strategy, two maternity waiting homes were piloted – one in Lospalos, Lautem district, in the east of the country, and the other in Same, Manufahi district, in the south. Lospalos and Same, being the district capitals, had the only health centres equipped for deliveries within their districts.
The maternity waiting homes were promoted through community meetings held at the beginning of the piloting process, radio broadcasts and a television commercial. A health promotion film on the importance of facility-based delivery – Women’s War – was screened in numerous villages across the country. In addition, women were informed about the maternity waiting homes at mobile clinics and during antenatal visits. Despite this fairly extensive awareness campaign, it was difficult to reach women in more remote areas because they generally lacked access to radio, television or village-based health services.
As outlined in the national strategy, any pregnant woman was eligible to stay at the maternity waiting homes from weeks 36 to 38 of gestation and for three or four days after delivery.9 Priority was given to women with certain risk factors (Table 1) and to those from rural areas. In practice, no systematic referral or data collection mechanism was in place for women with risk factors, and the maternity waiting homes catered to all women regardless of their risk status. Women were more likely to seek facility delivery and use the maternity waiting homes if they were having their first baby, had experienced complications in a previous pregnancy, or had a perceived problem in the current pregnancy.10
Table 1. Population, expected births and data for health centres and maternity waiting homes (MWHs) in Lautem and Manufahi districts, Timor-Leste, 2007
The maternity waiting home in Lospalos was a separate structure or “wing” that was connected to the health centre by a walkway; births took place in the health centre’s delivery room. The maternity waiting home in Same was built across the road from the health centre and, because of the poor conditions in the health centre, delivery facilities were later moved from the health centre to the maternity waiting home. Because of lack of space at both health centres, all women who delivered at the health facility were accommodated in the maternity waiting home before and/or after giving birth.
The new maternity waiting homes dramatically improved the physical condition of the facilities available to women before and after giving birth. Families were happy with the cleanliness and the fact that the home was a separate space specifically for pregnant and postpartum women, away from “sick” patients. This resulted in greater user satisfaction with birthing services at both sites. Originally, maternity waiting homes were intended to provide space for two family members to accompany each woman. However, on implementation, this policy was changed and family members were allowed to visit only at certain hours during the day. This was largely a result of the way in which the buildings were designed, with all women kept together in one large room and no additional space or separate rooms in which to accommodate families. Children were usually looked after by other family members. The need for additional space for postpartum women and families emerged as a common theme in interviews with users.
No data were available on length of stay at the maternity waiting homes, either before or after birth. The only information collected in health-centre records was the date of delivery. Interviews with families and health centre staff indicated that women tended to come to the maternity waiting homes when labour pains began and thus would stay there for up to two days before giving birth. A major advantage of the new maternity waiting homes was that women could stay from three to seven days after delivery to receive postpartum care.
Ethical approval was obtained from the Human Research Ethics Committee at Menzies School of Health Research, Australia. Permission to carry out the study was provided by the Timor-Leste Ministry of Health. This research was conducted as part of a larger study in which 124 interviews were conducted with policy-makers, district health staff and users of maternity waiting homes.10 For the quantitative component reported here, data were collected directly from birth registration books and entered into a data-collection table. Individual records were de-identified and date of delivery, place of birth, type of attendant and village of residence were recorded. Any ambiguous entries were clarified with the midwife. The travelling distance to the village of residence was then added to each record. Precise information on distance was available from UN police offices, which were located in each district. UN staff had mapped the distances by driving the routes and measuring the kilometres travelled on their odometers. It was important to obtain local information on actual distance because roads labelled on maps were often inaccurate or impassable, and detour routes provided a more accurate estimation of the actual distance travelled. Other information (e.g. population figures, maps and service targets) was also collected from district offices, UN agencies and local police stations.
Distance categories were developed based on the divisions present in the data and on biological plausibility. For example, many villages were within 5 km, and a heavily pregnant woman could possibly walk this far. In contrast, 50 km is a long way in the remote context of Timor-Leste, even with transport, and only a few villages were this far away. It was therefore appropriate to divide distance categories into 0–5, 6–25, 26–50 and more than 50 km from the health centre. These categories were also used by Høj et al. to assess the effect of distance to the health facility on maternal mortality.11 Entries with missing village of residence were excluded from the analysis. Statistical analysis was conducted using Stata version 7 (StataCorp. LP, College Station, United States of America). A χ2 distribution test with three degrees of freedom was used to measure the difference in the distribution of births in these distance categories before and after the intervention. A P-value of less than 0.05 was the cut-off for statistically significant differences.
The expected number of births occurring in a health facility for each distance category was also calculated because such data are lacking for Timor-Leste. In contrast to the overall urban or rural estimate that is currently available for the country, this calculation can illustrate the use of facility-based birthing services as distance from the health facility increases. Population figures for each village were obtained from the district offices in Same and Lospalos. Each village was matched against its distance from the health centre to calculate the total population in each distance category. To calculate the expected number of births, the population was multiplied by 0.045 (because 4.55% is the accepted birth rate estimate for the country that district health officials use to calculate service targets). The observed number of facility-based births was then divided by the expected number of births in the population to estimate the percentage of women in rural and remote areas who were accessing facility-based care when a maternity waiting home was in place. Pearson's χ2 test for trends in ordinal data was used to assess the decline in the expected number of births for each distance category. Again, a P-value of less than 0.05 was the cut-off for statistically significant differences.
Same, Manufahi district
Most of the women who delivered in the health centre in Same lived within a distance of 5 km, both before and after the implementation of the maternity waiting home (Table 2). As shown in Table 2, there was no significant difference in the distribution of births between distance categories after the maternity waiting home was implemented. This means that women from remote areas were no more likely to have a facility-based birth once the maternity waiting home was functioning than before it was available.
Table 2. Facility-based births in health centre before and after maternity waiting homes (MWHs) were implemented, by distance from village of residence to MWH, in Same, Manufahi district, and in Lospalos, Lautem district, Timor-Leste
In 2007, only about 9% of all births occurring in Manufahi district took place in a health facility (Table 3). An analysis by distance category showed that for the women who lived within 5 km, 23% of expected births took place at the health facility and the percentage of facility-based births declined with increasing distance from the health centre (Table 3). However, this trend was not statistically significant.
Table 3. Expected births and facility-based births, by distance from village of residence to maternity waiting home (MWH) in Manufahi and Lautem districts, Timor-Leste
Lospalos, Lautem district
In Lospalos, 62% of the facility-based births occurred among women who lived within 5 km of the facility, and 84% among women who lived within 25 km (Table 2). As in Same, there was no significant difference in the distribution of facility births between distance categories after the maternity waiting home was implemented. This suggests that women from remote areas were no more likely to have a facility-based birth once the maternity waiting home was implemented than before it was available.
In 2007, about 17% of all births in Lautem district took place in Lospalos health centre, the only facility in the district equipped for deliveries (Table 3). As in Manufahi district, the number of facility-based births and distance to the health centre were inversely related. In Lautem, this trend was statistically significant.
The maternity waiting home strategy was designed to overcome the tyranny of distance for women in remote areas of Timor-Leste. However, the results showed that in both Same and Lospalos, women who lived within 5 km of the health centres were the group most likely to use the accommodation facilities at the maternity waiting homes. This finding is not consistent with the results of two other studies that also assessed the distance from home among women using maternity waiting homes. A study in Zambia showed an average distance to home of 22 km for users of a maternity waiting home, compared with 13 km for those who went directly to hospital.6 In a study in Zimbabwe, women who used a maternity waiting home were more likely to live more than 40 km away (48%) than those who went directly to hospital (22%).5 However, this comparison provides no basis for knowing whether women using the maternity waiting homes would have gone to hospital anyway.
The analysis of the use of health facilities before and after maternity waiting homes were implemented in Timor-Leste showed the proportion of facility-based births did not increase significantly among women who lived more than 25 km away, in either Manufahi or Lautem. Therefore, the hypothesis that maternity waiting homes improve access to facility-based birth for women in remote areas of Timor-Leste is rejected. This is not surprising, given that more than 50% of Timorese women surveyed in the 2010 Demographic and Health Survey (DHS) claimed that both distance and transport were major barriers to seeking health care.7 Maternity waiting homes did not bring services closer to women or improve transport; thus, major barriers to accessing care persisted despite the logic behind the implementation of the homes.
Problems with record keeping at the health centre led to the exclusion of 18% of the births in Same from the analysis; village of residence was not recorded for three consecutive months in 2006. Therefore, the findings from Same should be interpreted with caution. However, the pattern of use evident in Same was similar to that in Lospalos, where there were more births and all records were complete.
The before-and-after study design without a concurrent control group provides limited evidence from which to conclude that successful outcomes are a result of the intervention. However, the absence of an observed effect on facility-based delivery is a significant finding because it demonstrates that accommodation facilities, on their own, are unlikely to have a major impact on the use of services. Other authors discussing the maternity waiting home strategy state that “a combination of approaches may be most effective to overcoming obstacles to care”.4 The findings from this research suggest that a more holistic approach may be warranted. Use of maternal health services in Timor-Leste depends on a range of individual, social, political and health-system factors.10 Hence, the findings from this study may not be generalizable to other settings. It is important to accurately assess how maternity waiting homes affect access to care; the approach and distance categories used in this study could be replicated in other countries implementing maternity waiting homes.
The percentage of births in the population that occurred in a health facility was low in both Manufahi (9%) and Lautem (17%) and resembled the rates found in the 2010 DHS (11% in Manufahi and 21% in Lautem).7 However, much greater inequities in access were demonstrated by providing a breakdown of the percentage of women who accessed facility-based delivery living at various distances from the health centres. For example, the national estimate that 53% of urban women accessed facility delivery compared with 12% of rural women7 disguises larger gaps in use between those who lived within 5 km of a health centre (23% in Manufahi and 46% in Lautem) and those who lived more than 50 km away (1% in Manufahi and 2% in Lautem).
A before-and-after distance analysis of the use of the first two maternity waiting homes to be implemented in Timor-Leste has demonstrated that, contrary to its objectives, the strategy of implementing such homes did not result in a higher proportion of women from remote areas giving birth in health facilities in Lospalos and Same. Place of delivery is prominent in many of Timor-Leste’s health policies. Given the failure of maternity waiting homes to improve access for remote women, further research should be conducted to determine which models of care increase coverage and produce the best outcomes for women and neonates. The results of such research could then be used to inform national policy. Fruitful areas of research in Timor-Leste include the impact of maternity transport services, skilled attendance for home births, and decentralization of birthing facilities to subdistricts.
Since the early 1990s, WHO has called for studies assessing the impact of maternity waiting homes.1,3,4 Because the success of implementing such homes will depend on the context, efforts should be made to evaluate and publish findings on the use of this strategy in other countries. The methodology for distance analysis presented here provides a framework for further research using data collected routinely in rural health centres.
This research was funded through an Australian Postgraduate Award received by K Wild. The authors have not entered into any agreements with the funding agency that may have affected the integrity of the research. K Wild has had full control of all primary data.
- Maternity waiting homes: a review of experiences. Geneva: World Health Organization, Division of Reproductive Health; 1996 (WHO/RHT/MSM/96:21). Available from: http://whqlibdoc.who.int/hq/1996/WHO_RHT_MSM_96.21.pdf [accessed 27 October 2011].
- Stewart D, Lawson J. The organization of obstetric services. In: Lawson J, Stewart D, editors. Obstetrics and gynaecology in the tropics and developing countries. London: Edward Arnold Publishers Ltd.; 1967. pp. 305–12.
- Essential elements of obstetric care at first referral level. Geneva: World Health Organization; 1991. Available from: http://helid.digicollection.org/en/d/Jh0227e/1.html [accessed 27 October 2011].
- Starrs A. The safe motherhood action agenda: priorities for the next decade. New York: Family Care International; 1997.
- Chandramohan D, Cutts F, Millard P. The effect of stay in a maternity waiting home on perinatal mortality in rural Zimbabwe. J Trop Med Hyg 1995; 98: 261-7 pmid: 7636923.
- van Lonkhuijzen L, Stegeman M, Nyirongo R, van Roosmalen J. Use of maternity waiting homes in Rural Zambia. Afr J Reprod Health 2003; 7: 32-6 doi: 10.2307/3583343 pmid: 12828140.
- National Statistics Directorate. Timor-Leste Demographic and Health Survey 2009–10. Calverton: ICF Macro; 2010.
- Hogan MC, Foreman KJ, Naghavi M, Ahn SY, Wang M, Makela SM, et al., et al. Maternal mortality for 181 countries, 1980-2008: a systematic analysis of progress towards millennium development goal 5. Lancet 2010; 375: 1609-23 doi: 10.1016/S0140-6736(10)60518-1 pmid: 20382417.
- Process of developing the national strategy on maternal waiting houses. Dili: Timor-Leste Ministry of Health; 2005.
- Wild K, Barclay L, Kelly P, Martins N. Birth choices in Timor-Leste: a framework for understanding the use of maternal health services in low resource settings. Soc Sci Med 2010; 71: 2038-45 doi: 10.1016/j.socscimed.2010.09.012 pmid: 20971540.
- Høj L, da Silva D, Hedegaard K, Sandstrom A, Aaby P. Factors associated with maternal mortality in rural Guinea-Bissau. A longitudinal population-based study. BJOG 2002; 109: 792-9 pmid: 12135216.