Health Policy and Planning Advance Access originally published online on March 24, 2006
Health Policy and Planning 2006 21(3):217-230; doi:10.1093/heapol/czl005
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Reducing under-five mortality through Hôpital Albert Schweitzer's integrated system in Haiti
1Former Director General and CEO, Hôpital Albert Schweitzer, 2Director, Haitian Institute for Child Health, Port-au-Prince, 3Former Special Assistant to the Director General and CEO, Hôpital Albert Schweitzer, 4Former Medical Director, Hôpital Albert Schweitzer, 5Former Director of the Community Health Division, Hôpital Albert Schweitzer and 6Co-Founder, Community Health Division, Hôpital Albert Schweitzer, Haiti
Correspondence: Henry B Perry, MD, Future Generations, HC 73 Box 100, Franklin, WV 26807, USA. Tel: + 1 304–358–2000; Fax: + 1 304–358–3008; E-mail: henry{at}future.org
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionConclusionBiographiesEndnotesReferences |
|---|
Background: The degree to which local health systems contribute to reductions in under-five mortality in severely impoverished settings has not been well documented. The current study compares the under-five mortality in the Hôpital Albert Schweitzer (HAS) Primary Health Care Service Area with that for Haiti in general. HAS provides an integrated system of community-based primary health care services, hospital care and community development.
Methods: A sample of 10% of the women of reproductive age in the HAS service area was interviewed, and 2390 live births and 149 child deaths were documented for the period 1995–99. Under-five mortality rates were computed and compared with rates for Haiti. In addition, available data regarding inputs, processes and outputs for the HAS service area and for Haiti were assembled and compared.
Results: Under-five mortality was 58% less in the HAS service area, and mortality for children 12–59 months of age was 76% less. These results were achieved with an input of fewer physicians and hospital beds per capita than is available for Haiti nationwide, but with twice as many graduate nurses and auxiliary nurses per capita than are available nationwide, and with three cadres of health workers that do not exist nationwide: Physician Extenders, Health Agents and Community Health Volunteers. The population coverage of targeted child survival services was generally 1.5–2 times higher in the HAS service area than in rural Haiti.
Discussion: These findings support the conclusion that a well-developed system of primary health care, with outreach services to the household level, integrated with hospital referral care and community development programmes, can make a strong contribution to reducing infant and child mortality in severely impoverished settings.
Key Words: child mortality, local health systems, integrated systems, community, severely impoverished settings
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionBiographiesEndnotesReferences |
|---|
The United Nations General Assembly and the international community agreed in 2001 upon a set of goals for improving, by the year 2015, the lives of the 2 billion people of the world who live in severely impoverished settings (United Nations 2001
). Among these goals is the reduction of child mortality by two-thirds. Although the most effective strategies for achieving this continue to be debated, the record of the past three decades has been extraordinary: for developing countries, under-five mortality has declined by 47%, from 167 to 89 deaths per 1000 live births (UNDP 2003). The number of children throughout the world dying during their first 5 years of life has declined from an estimated 18.9 million per year in 1960 to 10.8 million at present, in spite of an increase in the number of births from 96 million in 1960 to 132 million in 2000 (Ahmad et al. 2000; Black et al. 2003; UNICEF 2003). Much of the progress in reducing under-five mortality in children can be attributed to national programmes to: increase population coverage of immunizations for mothers and children and of family planning services; expand prenatal care and safe delivery practices; promote the prevention and treatment of childhood diarrhoea; promote early diagnosis and proper antibiotic treatment of childhood pneumonia; promote more appropriate infant and young-child feeding practices; and to provide high-dose vitamin A to children. In addition to the individual effects of the vertical, targeted, selective primary health care interventions mentioned above, progress in reducing under-five childhood mortality can also be attributed to the effects of general socioeconomic improvements, such as better overall nutrition in the population, better housing, as well as cleaner drinking water, better sanitation and roads, and access to electricity.
International donor support for health programmes in severely impoverished settings during the past several decades has been primarily for highly selective interventions. Because of the ‘top-down’ and highly selective approach that international donors have chosen, new funds have not been available for strengthening the capacity of local health services to deliver an integrated package of services, and international donor support for strengthening hospital referral services has been almost non-existent.
The evidence that local health care systems can reduce childhood mortality within their service population has been and still remains limited, and therefore the advisability of investing in local health care systems with a strong primary health care component, as a means to reduce childhood mortality, has been debated for the past several decades (Walsh and Warren 1979; Mosley 1988; Abed 1996). Although numerous reports exist concerning the effect of a single programme intervention (such as vitamin A or treatment of childhood pneumonia) on childhood mortality, only a small number of studies have documented an impact of a more comprehensive local primary health care programme on childhood mortality (Gwatkin et al. 1980; Berggren et al. 1981; Chen et al. 1983; Lamb et al. 1984; Tandon et al. 1984; Becker et al. 1993; Bryant et al. 1993; Ewbank 1993; Husein et al. 1993; Pison et al. 1993; Taylor et al. 1993; Arole and Arole 1994; Fauveau 1994; Koenig and Strong 1995; Aziz and Mosley 1997; Taylor and De Sweemer 1997; Perry et al. 2003). In most cases, these studies have been carried out in small populations. Furthermore, many of these studies lack adequate comparison groups and do not provide supporting evidence regarding inputs, processes and population coverage of services to support the assertion that the observed mortality difference can be reasonably attributed to the health programme.
Other studies have found a lack of evidence that declines in mortality can be attributed to specific health programmes (Ewbank and Gribble 1993). Furthermore, a recent global review of programmes to reduce childhood mortality reported that ‘... little is known about the characteristics of delivery strategies capable of achieving and maintaining high population coverage for specific interventions in various epidemiological, health system, and cultural contexts’ (Bryce et al. 2003).
The questions which we address here are the following. (1) Is there a statistically significant difference in the under-five mortality in the Hôpital Albert Schweitzer (HAS) Primary Health Care Service Area compared with Haiti in general? (2) Is the rest of Haiti an appropriate comparison group? (3) If so, can one make a convincing case that the HAS programme is responsible for an observed mortality benefit? These are important questions, because little evidence exists in the scientific literature regarding the actual or potential mortality impact of local programmes which provide a comprehensive array of services in a non-research environment. Rather, the scientific literature concentrates on the mortality impact of single interventions in research settings, and these findings have fuelled the dominance of vertical programming to the detriment of broader, community-oriented approaches. The questions which we address here are complex ones, and attribution of any observed mortality differences to the HAS programme requires ruling out potential confounders which could provide an alternative (and perhaps true) explanation of any observed difference.
We report here the impact of an integrated system of comprehensive community-based primary health care, hospital care and community development services on under-five mortality in a severely impoverished setting. We also describe the inputs, processes and outputs (in terms of population coverage of services) and outcomes (in terms of child mortality and fertility) that contributed to the achievement of these results.
Hôpital Albert Schweitzer in Haiti
Hôpital Albert Schweitzer (HAS), based in Deschapelles, Haiti, is a comprehensive set of programmes of health care and development serving a rural population in the Artibonite Valley since 1956 (see Figure 1). Haiti is the poorest country in the Western hemisphere, and one of the poorest countries in the world. The Human Development Index (based on life expectancy, educational attainment and adjusted real income) ranks Haiti as 150th out of the 175 countries of the world (UNDP 2003).
|
HAS is operated by the Grant Foundation, a not-for-profit organization registered in the United States which has a long-term contract with the Government of Haiti (Berggren 1974
; Berggren et al. 1981, 1995, 1997; Greco 2000; Nicholas 2003; HAS 2005). Dr and Mrs Larimer Mellon, following in the footsteps of their mentor Dr Albert Schweitzer and seeking to implement Schweitzer's philosophy of ‘Reverence for Life’, established HAS to provide compassionate and competent services to a poor rural population (Schweitzer 1933; Paris 2000). At present, HAS provides community-based primary health care services to a population of 148 000 persons in seven Functional Units of the Health District (officially referred to as the Unité Communale Santé, or UCS) of Petite Riviere, Verrettes and LaChapelle. These seven Functional Units comprise the HAS Primary Health Care Service Area (hereafter, HAS PHC Service Area) (Figure 2). The UCS Health District has a total of 258 000 residents in 14 Functional Units. Primary health care services in the other seven Functional Units are provided by other non-governmental organizations (NGOs) and the Ministry of Health.
|
The UCS is an area of 610 square miles, with a central valley plain surrounded by mountains. Approximately one-third of the district population lives in isolated mountain villages, where few roads are passable even with four-wheel-drive vehicles, and where travel time to a Dispensary, Health Center or the Hospital in Deschapelles can be as much as 12 hours walking or riding a horse or donkey.
HAS serves as a referral hospital for the entire UCS. It is the largest private health and development organization in Haiti, with 850 employees and an annual operating budget in Haiti of US$4.9 million. HAS's landmark achievements in the 1960s and 1970s in virtually eliminating neonatal tetanus as a cause of death and in lowering under-five mortality (Berggren 1974; Berggren et al. 1981) helped to lay the foundation for the child survival strategy of the 1980s, which contributed to the marked reductions in under-five mortality around the world.
The HAS community-based primary health care programme consists of the following:
- 1500 volunteer community health workers (Animatrices), one for every 15 households, who provide peer-to-peer health education, assist with Mobile Clinics and Rally Posts (see below), assist with referral to higher levels of care, and promote community involvement in planning, implementation and evaluation of services;
- 80 paid Health Agents, one for approximately every 400–500 households, who make regular home visits and direct monthly Rally Posts for immunizations, growth monitoring/nutritional counselling and referral;
- eight Monitrices (Monitors) who provide liaison with and training of lay midwives and Animatrices and who supervise the community-based nutritional rehabilitation programme;
- seven tuberculosis Accompagnateurs and nine tuberculosis Agents who follow up on tuberculosis contacts and provide directly observed therapy for tuberculosis patients in the home;
- Mobile Clinics at which an auxiliary nurse visits isolated communities every 1–2 months to provide basic curative and family planning services and to refer patients when indicated; and,
- seven Dispensaries/Health Centers, where curative care, immunizations and family planning services are provided.
HAS's programmes are, by their very nature, designed to promote equity by ensuring that those most in need have ready access to essential services, and also ensuring that health services reach every home. The following services are provided free of charge: health education regarding prevention and treatment of the common causes of childhood death, immunizations, detection and rehabilitation of malnutrition, voluntary counselling and testing for HIV infection, detection and treatment of tuberculosis, prenatal care, and family planning. Mobile Clinics are held in the most isolated areas. Almost all primary health care services (except HIV voluntary counselling and testing) are accessible within a 1–2 hour walk. No patient presenting at the Hospital with a life-threatening illness is denied treatment because of inability to pay. Thus, HAS's programmes are geared to limiting inequities, though they certainly have not eliminated them.
HAS operates a comprehensive programme of tuberculosis detection and treatment, with over 400 patients entering treatment annually and only 6.9% failing to complete their entire course of treatment. It also operates a Women's Health Program, including high-risk maternity services and comprehensive family planning services. HAS also provides a comprehensive AIDS prevention and treatment programme, including voluntary counselling and testing for HIV infection and prevention of mother-to-child transmission through the provision of anti-retroviral medication to HIV-positive pregnant women. In 2004, HAS began treatment of patients with AIDS using anti-retroviral medication paid for by the UN Global Fund for AIDS, Tuberculosis and Malaria. The prevalence of HIV infection among pregnant women receiving prenatal care in the UCS Health District was 3.9% in 2001, unchanged since the previous study in 1996. HAS also has strengthened the role of lay midwives and traditional healers by providing training and involving them as integral members of the health system.
The Hospital, with 190 beds and an active outpatient clinic, provides specialty referral care in adult medicine, paediatrics, obstetrics-gynaecology and surgery. The health care system facilitates referral of patients to a higher level of care and counter-referral down to a lower level as appropriate. Care is decentralized as much as possible so that mothers do not have to walk further than 30 minutes to a Rally Post or more than one hour to a Mobile Clinic or a Dispensary/Health Center.
HAS's Division of Community Development works throughout the UCS Health District. It operates programmes for improving water and sanitation (drilling wells, building cisterns, tapping springs, building and promoting the use of water filters, and building latrines), promoting vegetable gardens and reforestation (by selling seeds and seedlings and providing technical assistance), providing opportunities for micro-credit for women by establishing savings and loan groups, promoting the production of handicrafts and purchasing them for resale at HAS outlets, providing literacy training and support of primary education, and operating programmes to promote animal husbandry (including the training and support of veterinary technicians) and to improve agricultural production (including soil conservation).
Health and development activities in other areas of Haiti
Although Haiti has a number of outstanding health and development programmes, they are scattered unevenly throughout the population. Furthermore, health and development activities are rarely combined into single comprehensive systems for populations with clearly defined geographic boundaries. Ministry of Health resources are extremely limited, particularly in rural areas. Finally, the financial and geographic barriers that Haitians face in obtaining curative care are significantly higher than those within the HAS service area. Approximately one-third of the children of Haiti receive more intensive child survival services from NGOs that employ Health Agents and utilize Rally Posts. The remaining portion of the childhood population receives more limited child survival services from the Ministry of Health, such as immunizations, vitamin A supplementation, basic health education, and treatment of children with pneumonia and diarrhoea.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionBiographiesEndnotesReferences |
|---|
Our study design is an ecological one and compares data from two geographically defined populations: the HAS PHC Service Area and the Haitian population as a whole. When possible, we have used the rural Haitian population as the comparison group, but for certain indicators data were not available for the national rural population and, consequently, data for the entire Haitian population were used instead.
The line of reasoning that we implement here builds on the simple typology developed by Donabedian (1988):
Consequently, we sought out information that would provide us with comparable findings for the HAS PHC Service Area and for Haiti nationally.
In order to determine the levels of inputs (health manpower resources and hospital facilities) at HAS, we carried out interviews with HAS staff. We obtained similar national data from a national review published in 1996 (PAHO/WHO 1996).
We obtained process data (utilization of services) for HAS from internal programme reports. Unfortunately, we could not find comparable national data. When calculating hospital inputs on a per capita basis, we made the assumption that 57% of the Hospital's beds and staff were devoted to the care of residents in the HAS PHC Service Area, since that geographic area contains 57% of the population served by the Hospital. The remainder of the population served by the Hospital resides in other Functional Units of the UCS.
We obtained data regarding the coverage of child survival services at HAS from two sources: (1) a 2000 survey of the HAS PHC Service Area carried out by the Institute of Child Health (IHE 2000) and (2) a 2001 survey of 224 women who had given birth during the previous 6 months. For the first, a cluster sample approach was used to identify 1295 women of reproductive age and 945 mothers/guardians of at least one child 0–59 months of age. For the second, 224 women were selected at random from a list of households maintained by HAS. For both surveys, interviewers explained the nature of the survey to respondents, explained to them their right of refusal, and assured them that their responses would remain confidential. The 2000 survey provided data for all of the HAS coverage indicators except for exclusive breastfeeding and prenatal care, which were obtained from the 2001 survey.
The 2000 national demographic and health survey (Cayemittes et al. 2001) provided similar coverage data for Haiti nationally. This cluster sample survey consisted of interview findings from a total of 10 159 women of reproductive age.
Information about socioeconomic characteristics for the HAS PHC Service Area was obtained from the same 2000 survey described above (IHE 2000), based on the responses of 780 household heads. Similar information for the national population was obtained from the 2000 demographic and health survey (Cayemittes et al. 2001), based on the responses of 9595 household heads.
We calculated an odds ratio for each of the child survival interventions. This is the ratio of the odds of an individual in the HAS PHC Service Area receiving a specific intervention to the odds of an individual elsewhere in Haiti receiving the same intervention. In all cases except two, data from rural Haiti were used. For the estimate of exclusive breastfeeding and prenatal care utilization, data were available only for Haiti as a whole.
Mortality and fertility data in the HAS PHC Service Area were obtained from a household survey carried out in late 1999 and early 2000. HAS staff interviewed 3427 women of reproductive age in their homes regarding their reproductive history and the subsequent mortality of live-born children. These women were selected at random from a list of households maintained by HAS and constitute a 10% sample of the women of reproductive age in the HAS PHC Service Area. The respondents interviewed reported 149 deaths of children younger than 5 years of age during the previous 5 years. The 2000 national demographic and health survey (Cayemittes et al. 2001) provided comparable national mortality and fertility data.
Exact Fisher 95% confidence intervals were calculated according to the methods developed by Fleiss (1981) using Computer Programs for Epidemiologists (PEPI) version 4.0 (Abramson and Gahlinger 2001). The statistical significance of the odds ratios was calculated using the likelihood-ratio chi square as performed by PEPI version 4.0 as well. The funding sources for this research had no role in the work that is reported here.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionBiographiesEndnotesReferences |
|---|
Inputs: health manpower resources and hospital facilities
The per capita number of health personnel and hospital beds for the HAS PHC Service Area and Haiti nationwide are shown in Table 1. The number of physicians per capita and the number of hospital beds per capita are lower in the HAS PHC Service Area than for the rest of Haiti. However, the number of graduate nurses and auxiliary nurses is twice as great as for Haiti nationwide. In addition, the Community Health Program and the Hospital make use of three categories of health providers not generally available nationwide: Physician Extenders, Health Agents and Animatrices.
|
The health manpower data shown in Table 1 are for the year 2000 at HAS and for the year 1996 for the national data. Was this level of health manpower at HAS present during the period when the mortality impact was being achieved? The total number of employees at HAS and the numbers of staff working in the Community Health Program and in the Hospital increased from 334 full-time employees in 1993 to 568 in 1999, although the number of physicians working in the Hospital stayed the same during this period (HAS 1993
and 1999). Thus, if anything, the numbers of staff at HAS in the year 2000 represent a modest overstatement of the number present during the previous decade. Processes: utilization of services
In the year 2000, HAS's Health Agents made 97 865 home visits. They also gave 91 131 immunizations and carried out 81 423 weighings of children at Rally Post sessions. The Mobile Clinics provided services for 37 734 patient visits, and the Health Centers provided services for 114 305 patient visits. In addition, 49 006 ambulatory patient visits occurred at the Hospital's Outpatient Clinics. The Hospital admitted 12 306 patients for inpatient care and carried out 2400 surgical procedures, approximately half of which were for life-threatening conditions. The most common diagnoses among children requiring hospital admission in 2000 were malnutrition, pneumonia and diarrhoea. The most common diagnoses among adults requiring hospital admission were tuberculosis, trauma and obstetrical complications. As reported in Table 2, these statistics indicate that there are on average 1.0 ambulatory care visits per person per year, 3.2 home visits made by Health Agents per home, and 39.1 HAS hospital admissions per 1000 population. The data are reflective of the quantity of services which HAS provided during the previous decade. Unfortunately, no comparable national data are available.
|
Some of the activities that the Community Development Division has carried out are shown in Table 3. Many additional activities not shown in Table 3 also took place, including training farmers in improved agricultural techniques and soil conservation, formation and training of local community development committees, production and promotion of briquettes for fuel (using grass, waste paper and other waste products rather than charcoal), literacy training, support of local primary schools, and promotion of local artisans through purchase and sale of their products. Although many of the data in Table 3 are for the year 2000, they are representative of the types of activities that were taking place annually during the previous decade.
|
Outputs: coverage of key child survival services in the service population
The IHE household survey results in 2000 for the HAS PHC Service Area documented the coverage of key child survival interventions in the population, as shown in Table 4 and Figure 3, where they are compared with findings for rural Haiti and Haiti nationwide. For all indicators, the coverage of the service in the HAS PHC Service Area is substantially greater compared with rural Haiti and with Haiti nationwide.
|
|
The percentage of children receiving a high-dose vitamin A capsule during the previous 6 months is 2.8 times greater in the HAS PHC Service Area than in Haiti. Similarly, the percentage of children who received the total recommended series of immunizations is 2.4 times greater, the prevalence of exclusive breastfeeding is 2.3 times greater, the contraceptive prevalence rate is 1.8 times greater, the percentage of most recent births attended by a trained health care provider is 1.5 times greater, the percentage of children with symptoms of serious acute respiratory infection who obtained medical treatment is 2.6 times greater, and the percentage of children with diarrhoea receiving oral rehydration therapy is 1.5 times greater. The percentage of women obtaining three or more prenatal visits during their most recent pregnancy and the percentage of women receiving a tetanus toxoid immunization during their most recent pregnancy are both greater in the HAS PHC Service Area than in Haiti, but the differences are less striking than for the other indicators. Again, though these data for the HAS PHC Service Area are for the year 2000, they are in our judgement representative of the coverage during the previous decade.
The ratio of the odds of an individual (mother or child) receiving a child survival intervention in the HAS PHC Service Area compared with that for an individual in Haiti in general is at least 1.7, and as high as 20.7. In every case, the odds ratio is statistically significant at p<0.001.
Outcomes/impacts: comparison of under-five mortality and fertility rates in the HAS PHC Service Area with those for Haiti
Mortality and fertility rates for the HAS PHC Service Area, for rural Haiti and for Haiti nationwide (as shown in Table 5 and Figure 4) indicate that the risk of death before age five among live-born children is 58% less in the HAS PHC Service Area compared with rural Haiti. The infant mortality rate is 48% less, while the 1–4-year mortality rate is 76% less. The neonatal mortality rate is 28% less (though the difference is not statistically significant) and the post-neonatal mortality rate is 62% less. The total fertility rate in the HAS PHC Service Area is 29% less than in rural Haiti.
|
|
Is the comparison population otherwise similar to the HAS PHC Service Area? Comparison of socioeconomic and nutritional characteristics
One might argue that the HAS under-five mortality rates are more favourable than for the rest of Haiti because the population served by HAS has a higher standard of living or a higher level of childhood nutrition. As shown in Table 6, the HAS PHC Service Area is similar to the rest of rural Haiti in terms of its socioeconomic characteristics. While the level of education is virtually the same for the HAS PHC Service Area as for rural Haiti, the percentage of homes with no electricity and with no sanitation is slightly less (but statistically significant). However, the percentage of homes in the HAS PHC Service Area with no access to a protected source of water is substantially higher (75.2% versus 47.4%) as well as statistically significant. Therefore, the difference in under-five mortality rates cannot be explained on the basis of the socioeconomic characteristics that we were able to measure.
|
Levels of childhood malnutrition in the HAS PHC Service Area are essentially the same as those in the UCS Health District, in rural Haiti and in Haiti nationwide (Table 7). Therefore, the improvements in the under-five mortality rates cannot be explained on the basis of differences in levels of childhood malnutrition.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionBiographiesEndnotesReferences |
|---|
The findings from this study support the conclusion that, in a setting of severe impoverishment and under-development, a system of readily accessible community-based primary health care integrated with a referral hospital and with basic community development activities can reduce 0–59-month mortality by more than half. This conclusion is based on the following:
- The Hôpital Albert Schweitzer serves a population essentially identical in socioeconomic and nutritional characteristics to that for the rest of rural Haiti;
- HAS has established an infrastructure of staff, programmes, facilities and logistical support that addresses the major causes of infant and child death in the population;
- A high volume of health care and development services is being provided which, in turn, has achieved a high level of health service utilization and, compared with the rest of Haiti, a higher coverage of key child survival services; and
- Infant and child mortality rates in the HAS service area are substantially lower than those for the rest of Haiti.
Likewise, these findings are consistent with previous studies that have also documented substantial reductions of infant and child mortality in the HAS service area compared with Haiti in general (Berggren et al. 1981, 1995, 1997).
There are limitations to the ecological design that we have employed here. In assessing the whole-population effects of a complex structural intervention, such as the health and development system at HAS, we are obligated to search for confounders which might have been responsible for the mortality difference. The lack of evidence that the population within the HAS service area is substantially better-off in socioeconomic terms than is the rest of rural Haiti reduces but does not entirely eliminate the possibility of socioeconomic status as a confounder.
If the level of childhood malnutrition in the HAS PHC Service Area were lower than that in the rest of Haiti, then sorting out whether the difference is a result of HAS's programmes or other influences would have been critical. However, this is not the case. There is no evidence that the level of childhood malnutrition is lower in the HAS PHC Service Area, so it does not appear that the level of childhood malnutrition is a confounder which explains the observed mortality difference. In fact, these findings are consistent with (but not necessarily confirming of) the argument that, were it not for HAS's nutrition programmes, the level of childhood malnutrition would have been even higher than observed.
Is the population in the HAS service area different from the rest of Haiti in some other way that might explain the observed difference in under-five mortality? Are there cultural or behavioural patterns that are different, or some difference in the ecology of the area? We are unable to conceive of any such potential confounder. However, having additional socioeconomic indicators to compare would have been helpful.
In our view, the entire array of services and their dynamic interaction as a health and development system – not merely the specific child survival and maternal care interventions themselves, but the community-based primary health care programme, the hospital referral services, and the community development activities working together – are responsible for the lowered mortality that HAS achieved among children younger than 5 years of age.
We are not able to determine precisely how much each element within the system contributed to this mortality impact, although certain elements no doubt contributed more than others. Furthermore, the time for each element to exert its influence differs. The effects of curative hospital care are immediate, for instance, while the effects of primary education may take a generation. However, the various elements within the HAS system have been relatively stable for several decades, so we can be confident that there was sufficient time for the various inputs to be able to influence the processes, outputs and outcomes which appear to have produced the mortality impact.
One of the weaknesses of our study is the lack of formal quality assessments of the processes of service provision at HAS. However, various internal quality audits have been carried out over the time period of this study, and all of them have identified significant issues that needed to be addressed (L King, A Aftab, P Simon, personal communication). There have been no comparisons of the quality of services provided at HAS with those provided throughout Haiti as a whole other than assessments of coverage of services, which are shown in Table 4 and which show much higher levels of coverage within the HAS service area.
In our judgment, based on a combined total of more than 60 years of personal experience working at HAS, curative and preventive health care services together, including the hospital care of seriously ill children and mothers with pregnancy complications, explain a major part of the mortality impact. This is particularly the case in light of the high level of childhood malnutrition and associated morbidity that exists in the HAS PHC Service Area.
HAS has been a global leader in strengthening access to maternal screening for syphilis and reducing the number of cases of congenital syphilis in the population, an important cause of perinatal mortality in populations with a high prevalence of syphilis such as HAS's (Fitzgerald et al. 1998; Fitzgerald et al. 2003; Saloojee et al. 2004). Unfortunately, comparable data for Haiti nationally are not available.
Reduced fertility and an associated increase in birth intervals have been consistently observed to be a causal factor in lowering mortality in children younger than 5 years of age (Setty-Venugopal and Upadhyay 2002). Therefore, the higher rates of contraceptive use and lower rates of fertility in the HAS PHC Service Area appear to be part of the reason for the mortality impact achieved at HAS.
Other possible contributions to the observed reduction in child mortality include improving access to clean water, providing opportunities for women to participate in savings and loan groups, promotion of income-generation projects, and promotion of literacy and primary education. The geographic coverage of these and other community development services were far less than the coverage of child survival health interventions, however.
Although we have no data to support the following assertion, our experience working at HAS leads us to conclude that the active participation of the community was an essential element in making this system effective. Trust within the local population toward the local health system has been built up over a long period of time, and the local population has developed confidence in the quality of health care provided within the HAS system. Furthermore, the community has made a major effort to participate in the services provided by HAS, often investing scarce amounts of time and financial resources in order to promote and to access services.
We estimate that, at present, approximately one-fifth of mortality in children younger than 5 years of age in the HAS PHC Service Area can be attributed to HIV/AIDS.1 Therefore, if this epidemic had not presented itself in the HAS PHC Service Area, then the mortality rates presumably would have fallen by an additional 20%. The additional decline in under-five mortality in Haiti if HIV/AIDS had not been present would have been 8%.2
The leading causes of childhood death in the HAS PHC Service Area, based on both the IHE (2000) survey and on a follow-up to the HAS pregnancy history study, are still readily preventable or treatable. Malnutrition, pneumonia and diarrhoea accounted for 47% of the deaths and vaccine-preventable diseases (neonatal tetanus and measles) for 5% (Perry et al. 2005). Neonatal causes related to prematurity and delivery complications accounted for 12% of deaths. Unfortunately, no comparable national data on cause of death are available.
Identification and treatment of children with severe malnutrition, pneumonia and diarrhoea are all activities that receive priority in the community-based primary health care programme as well as in hospital services. Obviously, there is still need within the HAS health system for further improvements in the prevention, detection and early appropriate treatment of malnutrition, pneumonia and diarrhoea.
HAS operates Ti Foyer (Hearth) Programs in the community for the rehabilitation of severe childhood malnutrition. This approach is now being implemented around the world as the Positive-Deviance Hearth Model (Marsh and Schroeder 2002). HAS also operates an in-patient service for the rehabilitation of severely malnourished children.
Childhood tuberculosis remains a ‘silent’ cause of under-five mortality in the HAS PHC Service Area and elsewhere in Haiti, being often difficult to diagnose and often expressing itself only as severe malnutrition refractory to nutritional rehabilitation (Berggren and Berggren, personal communication, 2002; Graham et al. 2004). The actual burden of disease produced by childhood tuberculosis remains unknown, however.
These findings have implications for human resources policy in Haiti and potentially other resource-poor settings, particularly the greater reliance on lower-level personnel and community volunteers and the lesser reliance on physicians than is the case for Haiti as a whole. Physician Extenders, Health Agents and Community Health Volunteers are all integral members of the health team at HAS, but they are not commonly utilized in Haiti as a whole. They make it possible to maintain contact with all households and to keep costs down.
Formal training programmes for Physician Extenders do not exist in Haiti and are uncommon in other resource-poor settings. HAS trained its own. Developing health systems similar to HAS's will require the development of formal training programmes for this mid-level category of workers.
Having effective lower-level staff requires higher-level supervisory staff with the professional expertise to provide appropriate support and training. Thus, a system such as HAS's needs to have the capacity to attract and sustain high-quality senior level professionals. HAS has been able to provide long-term professional leadership, programme staff, infrastructure, logistical support and financial support that have led to long-term consistency of locally accessible and affordable community-based primary health care and referral hospital services. The strong community outreach component has made possible high population coverage of basic services for maternal and child health care in the HAS PHC Service Area.
In short, HAS has extended health services down to the household level and up to the level of a high-quality district hospital with specialist care, including in surgery. Linking these various elements closely together has led to effectiveness in reducing child mortality. Such comprehensive systems of community-based primary health care linked to facility-based primary health care and hospital care will be necessary in other similar low-income settings if optimal benefits of health services are to be achieved in the future. Therefore, global and national policies as well as global and national financial support are needed which can encourage and support the development of such ‘model’ systems, which can serve as the basis for the development of similar systems elsewhere. These are affordable. The entire system at HAS cost US$19 per capita in the year 2000, including the cost of the community development programmes.
We think that the results achieved by HAS are possible to achieve in other low-income settings if the above conditions are met. If sufficient long-term financial resources are available along with strong and sustained political will, programmes should be able to achieve major impacts on under-five mortality if similar programme strategies are employed.
|
Conclusion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionBiographiesEndnotesReferences |
|---|
The risk of death before reaching the age of five for live-born children in the HAS PHC Service Area in Haiti is less than half that for children in the rest of Haiti. Altogether, the evidence is strong that the programmes at HAS are responsible for this difference.
Additionally, these findings support the assertion that strengthening local health programmes and promoting community development activities can lead to major reductions in childhood mortality and in fertility. The experience of HAS suggests that the components of the system which contribute to these demographic impacts include: (1) preventive and curative services for reducing mortality from the most frequent, readily preventable or treatable causes of childhood death; (2) a system for maintaining contact with every household to promote utilization of basic maternal and child health services; (3) hospital care that is readily accessible, affordable and integrated with the primary health care programme; (4) the provision of services of sufficient quality at an affordable price so that local people have trust and confidence in them and will utilize them; (6) the provision of services that are effective in preventing and treating illness; and (5) the promotion of community development activities that address the underlying causes of childhood mortality, including lack of financial resources, education, clean water and sanitation.
In severely impoverished settings, the establishment of similar local systems for health and development could play an important role in achieving the Millennium Development Goals for reducing childhood mortality by two-thirds by the year 2015 (UNDP 2003).
|
Biographies |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionBiographiesEndnotesReferences |
|---|
Henry Perry, MD, PhD, MPH, served as the Director General and CEO of Hôpital Albert Schweitzer from 1999 to 2003. Prior to that, he had worked with community-based primary health care programmes in Bolivia and Bangladesh and founded Curamericas, an NGO working in Latin America. He currently teaches in the masters programme of Future Generations, Franklin, WV, USA, focused on community change and conservation.
Michel Cayemittes, MD, serves as the Executive Director of Institut Haïtien de l’Enfance, which carries out health and demographic research in Haiti. Dr Cayemittes has a long history of involvement in community-based primary health care in Haiti, including providing leadership to the community outreach programmes of Grace Children's Hospital.
Francois Philippe, MA, provided leadership for the collection of mortality data when he worked at Hôpital Albert Schweitzer from 1999 to 2001. He taught in the public school system in New York City prior to that time. Currently he works as a consultant in labour management and organizational development in the health care field in New York and New Jersey, USA.
Duane Dowell, MD, has a long period of involvement with Hôpital Albert Schweitzer covering over 40 years. Between 1968 and 1980, he served in rotation with two others as chief of the paediatric department, during which time he provided inpatient and outpatient paediatric care and developed programmes for the hospital management of severely malnourished children. He has also participated in the community-based provision of preventive services for children. In 1999 he returned to HAS as Medical Director and, over the next 5 years, provided medical leadership for the Hospital Division.
Jean Richard Dortonne served as Director of the Community Health Division at Hôpital Albert Schweitzer from 2000 to 2003. Prior to that time he worked in other community health programmes in Haiti. Since then, he has been working with the University of Montreal in Gabon on community health projects.
Henri Menager served as Director of Community Health Division at Hôpital Albert Schweitzer from 1990 to 1995. Prior to that, he worked with Save the Children in Haiti. He currently works as an epidemiologist with the Kansas Department of Health and Environment, USA.
Erve Bottex, MD, served as Assistant Director of the Community Health Division at Hôspital Albert Schweitzer from 1993 to 1995 and Director from 1995 to 1996. Prior to that, he worked in the Community Health Department of the Christian Mission Bienfaisance Hospital of Pignon, Haiti, from 1989 to 1993. He taught as a faculty member of the Salzburg Seminar in Austria from 1996 to 2000, featuring health care initiatives around the world. He currently works with Columbia University at the Harlem Hospital Center Division of Medicine/Infectious Diseases in New York City.
Warren Berggren, MD, and his wife Gretchen founded the community health programme at Hôpital Albert Schweitzer in 1967 and have worked there off and on since that time. From 1958 until 1964, he and his wife served as medical missionaries in the Belgian Congo. He also served on the faculty of the Harvard School of Public Health from 1968 to 1993. He headed the child survival initiatives of Save the Children USA from 1983 to 1993. Although officially retired, he continues to consult with health programmes around the world.
Gretchen Berggren, MD, with her husband Warren, founded the community health programme at Hôpital Albert Schweitzer in 1967. She has worked there off and on since that time. From 1958 until 1964, she and her husband served as medical missionaries in the Belgian Congo. She served on the faculty of the Harvard School of Public Health from 1968 to 1993. In addition, she worked with Dr Nevin Scrimshaw at the Massachusetts Institute of Technology on child nutrition projects. She is widely known for her contributions to the development of the Hearth/Positive Deviance Approach to rehabilitation of childhood nutrition, now being applied throughout the world.
|
Acknowledgements |
|---|
The authors would like to acknowledge the guidance and critical influence of the late Dr John Wyon on the design of the HAS outreach services almost four decades ago. We are grateful to the people of Haiti for their cooperation in obtaining the information used in this report; to the Board of Directors of the Grant Foundation for their commissioning of this study; to Future Generations for its support during the writing of this paper; to Robert Northrup and Jack Bryant for their assistance in the design of the demographic and health survey carried out by the Institut Haitien de l’Enfance in the HAS service area; to Jack Bryant, Doug Ewbank, Leslie King, Robert Northrup, Jon Rohde, Dory Storms, Carl Taylor and John Wyon (prior to his death) for their helpful comments on earlier drafts of this paper; and to anonymous reviewers for their helpful suggestions. We would like to acknowledge the assistance of Bryan Proffitt and Angela Barton, who computerized the reproductive health history data, and of Baker Perry, who produced the maps. Finally, the authors would like to acknowledge the strong long-term support of the many individual and organizational donors who have made the work at HAS possible. The evaluation activities which made this report possible were funded by the Bill and Melinda Gates Foundation, the United States Agency for International Development, and the general operating funds of HAS.
|
Endnotes |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionBiographiesEndnotesReferences |
|---|
1This conclusion was based on the following assumptions: 4% of mothers are HIV-positive, the rate of mother-to-child transmission is 33%, and 90% of HIV-positive newborns will die before the age of five. According to these assumptions, 11.9 children per 1000 live births will die before the age of five from HIV infection. This is equivalent to 19% of under-five mortality (11.9/62.3).
2If the under-five mortality due to HIV/AIDS in Haiti is also 11.9 deaths per 1000 live births, then the percentage of under-five deaths in Haiti due to HIV/AIDS is 8.0% (11.9/149.4).
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionBiographiesEndnotesReferences |
|---|
Abed FH. 1996. Health and development: lessons from the grassroots. Journal of Diarrheal Disease Research 14:118–24.
Abramson JH and Gahlinger PM. 2001. Computer Programs for Epidemiologists: PEPI Version 4.0 Salt Lake City, UT Sagebrush Press.
Ahmad OB, Lopez AD, Inoue M. 2000. The decline of child mortality: a reappraisal. Bulletin of the World Health Organization 78:1175–91.[ISI][Medline]
Arole M and Arole RS. 1994. Jamkhed: A comprehensive rural health project London Macmillan.
Aziz KMA and Mosley WH. 1997. The history, methodology, and main findings of the Matlab Project in Bangladesh. In Das Gupta M, Aaby P, Garenne M, Pison G (Eds.). Prospective community studies in developing countries Oxford and New York Clarendon Press/Oxford University Press pp. 28–53.
Becker SR, Diop F, Thorton JN. 1993. Infant and child mortality in two counties of Liberia: results of a survey in 1988 and trends since 1984. International Journal of Epidemiology 22:(Suppl. 1), S56–63.[Abstract]
Berggren WL. 1974. Administration and evaluation of rural health services. I. A tetanus control program in Haiti. American Journal of Tropical Medicine and Hygiene 23:936–49.
Berggren WL, Ewbank DC, Berggren GG. 1981. Reduction of mortality in rural Haiti through a primary-health-care program. New England Journal of Medicine 304:1324–30.[Abstract]
Berggren G, Menager H, Genece E, Clerisme C. 1995. A prospective study of community health and nutrition in rural Haiti from 1968 to 1993. In Scrimshaw NS (Ed.). Community-based longitudinal nutrition and health services: classical examples from Guatemala, Haiti and Mexico Boston, MA International Foundation for Developing Countries pp. 143–78.
Berggren G, Berggren W, Menager H, Genece E. 1997. Longitudinal community health research for equity and accountability in primary health care in Haiti. In Das Gupta M, Aaby P, Garenne M, Pison G (Eds.). Prospective community studies in developing countries Oxford and New York Clarendon Press/Oxford University Press pp. 157–88.
Black RE, Morris SS, Bryce J. 2003. Where and why are 10 million children dying every year? The Lancet 361:2226–34.[CrossRef][ISI][Medline]
Bryant J, Marsh D, Khan KS, et al. 1993. A developing country university oriented toward strengthening health systems: challenges and results. American Journal of Public Health 83:1537–43.
Bryce J, el Arifeen SE, Pariyo G, et al. 2003. Reducing child mortality: can public health deliver? The Lancet 362:159–64.[CrossRef][ISI][Medline]
Cayemittes M, Placide MF, Barrère B, Mariko S, Sévère B. 2001. Enquête Mortalité, Morbidité et Utilisation des Services (EMMUS-III) Haiti, 2000 Calverton, MD Ministère de la Santé Publique et de la Population, Institut Haïtien de l’Enfance and ORC Macro.
Chen L, Rahman M, D'Souza S, et al. 1983. Mortality impact of an MCH-FP program in Matlab, Bangladesh. Studies in Family Planning 14:199–209.[Medline]
Donabedian A. 1988. The quality of care: how can it be assessed? Journal of the American Medical Association 269:1743–8.
Ewbank DC. 1993. Impact of health programs on child mortality in Africa: evidence from Zaire and Liberia. International Journal of Epidemiology 22:564–72.
In Ewbank DC and Gribble JN (Eds.). Effects of health programs on child mortality in Sub-Saharan Africa 1993. Washington, DC National Academy Press.
Fauveau V. 1994. Matlab: Women, children and health Dhaka, Bangladesh Pioneer Printing Press.
Fitzgerald DW, Behets FM, Lucet C, Roberfroid D. 1998. Prevalence, burden, and control of syphilis in Haiti's rural Artibonite region. International Journal of Infectious Diseases 2:127–131.[CrossRef][Medline]
Fitzgerald DW, Behets F, Preval J, et al. 2003. Decreased congenital syphilis incidence in Haiti's rural Artibonite region following decentralized prenatal screening. American Journal of Public Health 93:444–6.
Fleiss JL. 1981. Statistical methods for rates and proportions New York John Wiley and Sons.
Graham SM, Gie RP, Schaar HS, et al. 2004. Childhood tuberculosis: clinical research needs. International Journal of Tuberculosis and Lung Diseases 8:648–657.
Greco R. 2000. The cardboard box. Journal of the American Medical Association 284:534.
Gwatkin D, Wray JD, Wilcox JR. 1980. Can health and nutrition interventions make a difference? Overseas Development Council Monograph No. 13 Washington, DC Overseas Development Council.
Hôpital Albert Schweitzer (HAS). 1993. Annual report Sarasota, FL Grant Foundation.
Hôpital Albert Schweitzer (HAS). 1999. Annual report Sarasota, FL Grant Foundation.
Hôpital Albert Schweitzer (HAS). 2005. Website: [http://www.hashaiti.org]..
Husein K, Adeyi O, Bryant J, Cara NB. 1993. Developing a primary health care management information system that supports the pursuit of equity, effectiveness and affordability. Social Science and Medicine 36:585–96.
Institut Haïtien de L’Enfance (IHE). 2000. Enquête de santé sur l’Unité Communale de Santé (UCS) de Petite-Rivière, des Verrettes et de La Chapelle Port-au-Prince, Haiti Institut Haïtien de L’Enfance.
Koenig M and Strong M. 1995. Assessing the mortality impact of an integrated health programme: lessons from Matlab, Bangladesh. In Rashad H, Gray R, Boerma T (Eds.). Evaluation of the impact of health interventions Liège, Belgium International Union for the Scientific Study of Population pp. 361–95.
Lamb WH, Foord FA, Lamb CM, Whitehead RG. 1984. Changes in maternal and child mortality rate in three isolated Gambian villages over ten years. The Lancet 2:912–4.[ISI][Medline]
The positive deviance approach to improve health outcomes: experience and evidence from the field. Food and Nutrition Bulletin 2002. 23:(Suppl.), 5–137.[Medline]
Mosley WH. 1988. Is there a middle way? Categorical programs for PHC. Social Science and Medicine 26:907–8.
Nicholas S. 2003. Haiti's Hospital Albert Schweitzer: the legacy of Larimer and Gwen Mellon (editorial). American Journal of Public Health 93:527–9.
Pan American Health Organization/World Health Organization (PAHO/WHO). 1996. Health Situation Analysis: Haiti 1996 Port-au-Prince, Haiti Pan American Health Organization.
Paris B. 2000. Song of Haiti: The lives of Dr. Larimer and Gwen Mellon at the Albert Schweitzer Hospital of Deschapelles New York Public Affairs.
Perry H, Shanklin D, Schroeder D. 2003. Impact on infant and child mortality of a community-based, comprehensive primary health care program in Bolivia. Journal of Health, Population and Nutrition 21:383–95.[ISI][Medline]
Perry H, Ross A, Ferdinand F. 2005. Cause of under-five mortality in rural Haiti: findings from the Hôpital Albert Schweitzer service area. Pan American Journal of Public Health 18:178–86.
Pison G, Trape JF, Lefebvre M, Enel C. 1993. Rapid decline in child mortality in a rural area of Senegal. International Journal of Epidemiology 22:72–80.
Saloojee H, Sithembiso V, Goga Y, et al. 2004. The prevention and management of congenital syphilis: an overview and recommendations. Bulletin of the World Health Organization 82:424–30.[ISI][Medline]
Schweitzer A. 1933. Out of my life and thought Baltimore, MD The Johns Hopkins University Press (republished in 1998)..
Setty-Venugopal V and Upadhyay UD. 2002. Birth spacing: three to five saves lives. Population Reports, Series L, Number 13 Baltimore, MD Johns Hopkins Bloomberg School of Public Health, Population Information Program.
Tandon B, Sahai NA, Vardhan A. 1984. Impact of integrated development services on infant mortality rate in India. The Lancet 2:157.
Taylor CE and De Sweemer C. 1997. Lessons learned from Narangwal about primary health care, family planning and nutrition. In Das Gupta M, Aaby P, Garenne M, Pison G (Eds.). Prospective community studies in developing countries Oxford and New York Clarendon Press/Oxford University Press pp. 101–29.
Taylor WR, Chahnazarian A, Weinman J, et al. 1993. Mortality and use of health services surveys in rural Zaire. International Journal of Epidemiology 22:(Suppl. 1), S15–9.[Abstract]
UNICEF. 2003. The State of the World's Children Report 2003 New York UNICEF.