Implications of integrated nature conservation for human reproductive health: a case study from Ranomafana National Park, Madagascar

Abstract

This article explores the impact of national parks, established as integrated conservation and development projects (ICDPs), on human reproductive health, using Ranomafana National Park in Madagascar as a case study. A comparative approach, studying the villages within and beyond the park area, was used. Statistical information, semi‐structured key person and focus group interviews, observation, as well as previous literature, were used. The article illustrates how health programmes of the park project attempted to meet local needs, but how cultural and economic issues affected the success of the project. The study concludes that the impact of the ICDP on human reproductive health was modest. However, the human aspects in conservation have to be emphasised in the future.

1 INTRODUCTION

In recent years, community‐based or other programmes for people‐oriented natural resources management, such as integrated conservation and development projects (ICDPs), have been the main forms for conducting ‘socially’ sound and acceptable conservation (e.g. Adams & Hulme, 2001). They stem from the failure of top‐down projects and incorporate the new understanding of taking into account local people's views, needs and knowledge. The basic idea behind ICDPs is to integrate development, in some form or another, with conservation. The development components of ICDPs can be classified into three categories: compensation, alternatives, and enhancement (Abbot et al., 2001). However, ICDPs and other people‐oriented conservation programmes have been confronted by many problems. The idea of offering compensation or economic alternatives does not lead automatically to nature conservation. The suggestion that poverty alleviation will enhance conservation ‘relies on a number of tenuous assumptions about human behaviour’ (Wilhusen et al., 2002). Many ICDPs have had to face the problem that development activities have not removed the pressure on the natural resources, but on the contrary have accelerated destructive activities due to development that attracts immigrants to the area (Langholz, 1999).

Conservation literature is filled with studies of the failures of these people‐oriented approaches (Brandon et al., 1998; Adams & Hulme, 2001; Brechin et al., 2002; Wilhusen et al., 2002, etc.). However, most studies concentrate on examining the failure to achieve conservation objectives, not the impact of conservation and development on the people's living conditions and wellbeing. Many studies have explored people's perceptions of nature conservation and the national park (e.g. Infield, 1988; Hartrup, 1994; Ite, 1996; Abbot et al., 2001), but not many studies have been conducted on the impact of development projects as such, or on the joint impact of development and conservation on the people's wellbeing, especially using reproductive indicators as ‘signs’ of a project's success. This study is a case study from a ten‐year‐old integrated conservation and development project in Madagascar, one of the conservationists' hot spots.

The main objective of the study is to examine the possible change in reproductive health, as measured by the onset of reproductive life, fertility, modern contraceptive use, birth conditions, prenatal health care and infant mortality. We evaluate whether the villages in the park area, according to reproductive indicators, differ from those outside and if so, whether the reason is the ICDP. Using a comparative approach, the impact of the ICDP can be estimated. We assume that factors other than the ICDP that influence reproductive health have been more or less similar in the villages inside and outside the park area. We also describe the health activities of the ICDP and consider the meaningfulness of the health component in conservation projects.

2 REPRODUCTIVE HEALTH

The study is based on an assumption that nature conservation should be socially sustainable and improve people's wellbeing (Wollenberg & Colfer, 1997). Wellbeing can be defined as consisting of all the aspects of life that provide people with a secure life, such as secure access to resources, economic and decision‐making opportunities, as well as justice, heritage, identity and safety, and health (Wollenberg & Colfer, 1997). Here we have chosen one indicator of wellbeing, namely reproductive health.

Since the Cairo Conference on Population and Development in 1994, reproductive health has been identified broadly to mean not merely the absence of diseases concerning reproduction, but total physical, mental and social wellbeing in reproduction, as well as the right to make decisions about one's own sexual life. Reproductive health includes both male and female perspectives. Female reproductive health concentrates usually on maternal health (Bergestrom, 1994: 307) and a woman's right to control and affect her life on reproductive issues. In this study, we concentrate on reproductive indicators as well as family planning in order to describe the living circumstances and their change near the Ranomafana National Park. One of the reasons why the health component of our case study project was included was to reduce fertility, as a growing population was seen as a threat to conservation (Grenfell, 1995). The aim is to achieve fertility reduction through family planning and improved reproductive conditions, which both are supposed to lead to a decrease in infant mortality.

We believe that, by studying reproductive health, we will focus on one of the most important and comprehensive themes of community development. At the World Summit for Sustainable Development in Johannesburg in 2002, reproductive health was emphasised as one of the main themes of environmentally sound development. The human reproductive perspective on the conservation issue is valuable, as the human aspect in conservation has been recognised and emphasised globally (e.g. Brechin et al., 2002; Killeen & Khan, 2002).

3 RANOMAFANA NATIONAL PARK

The Ranomafana National Park (RNP) Project was established as an ICDP in 1991 to conserve existing mountainous rain forests – the habitats of a new lemur species that was identified in the area in 1986 (Wright, 1997). The RNP area consists of 43 500 hectares of forested land, with an altitudinal gradient varying from 1 374 m to 400 m above sea level (Wright, 1997). The ecological circumstances in the park area differ from mild climate ‘lowlands’ with steep slopes, to flat and cold highlands. The ICDP was founded by the United States Agency for International Development (USAID) and organised by two American universities (Kightlinger et al., 1990). The land which now constitutes the park was considered uninhabited by the park founders, although some farmers lost their farming lands when park boundaries were defined (Peters, 1998b).

The project included six integrated components: health, conservation education, economic development, park management, ecotourism, and biodiversity research (Wright, 1997). The funding from USAID ended in 1997 and the management of the park, as well as the responsibility for economic rural development and ecotourism, was transferred to Association Nationale pour la Gestion des Aires Protégées (ANGAP), the body that manages almost all protected areas in Madagascar. ANGAP receives most of its funding from the World Bank (World Bank, 1996). Health, conservation education and coordination of biodiversity research were continued by an environmental organisation, Madagascar Institute pour la Conservation des Ecosystemes Tropicaux (MICET), which receives its funding from foreign foundations and research projects. Nowadays, development projects are funded by distributing 50 per cent of the entrance fees to local people's projects (ANGAP, 1997). The numbers of visiting tourists have increased steadily every year, being around 12 000 in 2000 (ANGAP, 2000). Entrance fees provide a substantial amount of money to people's projects – around US$50 000 annually (Wright & Andriamihaja, 2002). However, problems with distribution of funds have emerged (Korhonen, 2003).

People around the RNP live in approximately 93 small communities, ranging between 10 and 600 people, with a total population of about 25 000 (Grenfell, 1995). The Malagasy administration consists of a hierarchical system: provinces, fivondronanas (sub‐jurisdictions), firaisanas (municipalities) and fokotanys (groups of two to five villages). A fokotany is the lowest official administrative unit and the villages are led by traditional leaders (apamjakas) or by the village elders' council. The park overlaps seven municipalities that have villages in the peripheral zone. Development projects of the park are carried out in this peripheral zone, which is defined as a 3‐km belt around the park, including those fokotany and firaisana centres that have villages in this belt (Grenfell, 1995). In this article, the villages inside the peripheral zone are called ‘villages in the park area’ and those outside this zone ‘villages outside the park’.

The municipality of Ranomafana has been the centre of interventions and activities because it has the largest number of villages situated in the park area. Tourist access to the park begins near the Ranomafana town, which lies on the national highway (although badly damaged), and has had a history as a tourist attraction for more than 50 years due to the hot springs (Peters, 1998a). In recent years, an increasing number of hotels and other tourist services have started in Ranomafana town. However, the park area also includes various remote villages that are totally inaccessible during the rainy season. Their situation is entirely different from that in Ranomafana town.

People in the park area have been classified into two ethnic groups: Tanala and Betsileo, although ethnicity cannot be considered a fixed characteristic. Especially in Madagascar, the division into certain ‘ethnic’ groups has been defined by the economic activities that the people practise, rather than their ethnic characteristics (Kottak, 1971). These groups of people should be regarded as culturally and economically adapted groups, rather than ethnic tribes (see, for example, Marcus, 2000: 98). People in the lowlands have adapted the farming technique of shifting cultivation (tavy) suitable for local ecological conditions, but they also practise paddy rice and cash crop cultivation, such as coffee and bananas. People living in the western and higher elevations mainly cultivate paddy rice in terraces. Tavy, practised mainly in the lowlands of the park, is considered the most serious threat to primary forests and the park (Oxby, 1985; ANGAP, 1998). Other factors have also contributed to forest destruction, such as the exploitation of hardwood and cash crop cultivation (mainly coffee), introduced by the French in the beginning of the last century. Forestland was converted to other land‐use purposes, which affected the substitute agriculture (Hanson, 1997). The Ranomafana area was seen as a region of endless forests, and due to population growth people moved there just over 200 years ago from the central plateau, looking for easier living conditions (Ferraro & Rakotondranjaona, 1991).

4 METHODS

The research design is comparative. First, the study compares reproductive health indicators in villages and municipalities in the park area and outside it. Second, the situation is compared over time. By using this approach we aim to distinguish the possible impact of the ICDP. We believe that a comparative approach gives strength for the impact assessment, although we do acknowledge that eliminating all the other influencing factors is difficult. In addition, we do not want to consider the reproductive health indicators separate from the context of the ICDP, and we also attempt to address some aspects of conservation that may have influenced the specific reproductive issue under study.

The data were collected from local health statistics, by visiting villages and interviewing key persons, conducting focus group discussions and making systematic observations in the villages. Furthermore, park project documents, follow‐up studies and previous literature were used. Table 1 presents the data sources of the study. The fieldwork was carried out in two phases between September–November 2001 and October–November 2002.

Summary of data sources of the study

Data source	Unit and number	Park/non‐park	Year(s)
Health statistics	Ten municipalities	6/4	1999–2001
School enrolment statistics	26 schools (fokotany)	14/12	1990–2000
Key person interviews	13 villages and ten municipalities	7/6 and 6/4	2001
Focus group interviews	20 groups in five villages	3/2	2002
MICET follow‐up surveys	Six villages	6/0	1998
Initial park surveys (Kightlinger et al., 1992a, 1992b)	18 villages	18/0	1990–1

Health and demographic data were available at the municipality (firaisana) level. We selected four municipalities that do not have any part in the park area, and six municipalities that are partly situated in the park area. Health data were collected from local health centres and from the provincial health office (Pro‐Santé) in Fianarantsoa, the province capital. The statistical information from the health centres included data on births in health centres, prenatal visits and contraceptive use in the study municipalities from 1999 to 2001. No population‐based statistical health data were available from 1990. Educational information was available from the early 1990s, and was obtained from regional school district offices. Previous doctoral studies conducted in the park area villages (Harper, 2002; Kightlinger et al., 1992a; Hardenbergh, 1993) were used for comparison. Furthermore, follow‐up studies conducted by MICET in their target villages were used.

The idea behind the methods employed is to use existing quantitative material instead of traditional surveys. Of course, due to the different levels of official material, summarising the data is more complicated than in village‐level surveys. However, these official records are believed to produce scientifically reliable data, especially as Malagasy administration holds strong traditions of respect for orders due to their own culture and tradition of the French colonial period as well as the centralised communist system. However, statistics provided a good overview on the issue but did not describe the actual situation in the most affected villages near the park border. This is one of the reasons why qualitative data were also included.

For village‐level data, we selected seven villages in the park area, and six outside it (Figure 1). The villages were situated in three highland and three lowland municipalities. Of the six control villages, two were near the park area and the other four villages from municipalities further away from the park. These two control villages near the park were chosen in order to compare them with the villages in the park area. They lie near each other, but one is involved in the park project and the other not. Villages situated further away from the park were selected in order to eliminate the park's influence in all sectors. In village selection, the road and accessibility were considered important factors. Villages both near the main road as well as far from it were included in the study.

Map of study area

In every village, village elders were consulted, and traditional midwives and health animators were interviewed with semi‐structured interviews. Health animators were volunteer villagers who were elected by the villagers or chosen by local health authorities to work in the villages. They had received some education from doctors at local health centres. In addition, doctors and midwives working in local health centres, or CSBs (Centre Santé de Base), were interviewed. In total, 12 health animators, four doctors and ten midwives (seven village midwives and three CSB midwives) were interviewed.

The interviews were carried out with a Malagasy translator who was a university student with experience in conducting rural development interviews. The questions were translated into Malagasy beforehand. The interpreter translated the main points of an answer during the interview so that the researcher could ask additional questions. All the interviews were tape‐recorded and later transcribed and translated into English. Interview techniques and questions were tested in a pilot village in the park area, but because there were hardly any changes in technique the pilot village was included in the study. In every village visited, systematic observations regarding housing and general living conditions were carried out and informal discussions were held with villagers before and after the interviews.

During the second fieldwork period, 20 focus group interviews were conducted in five villages in order to deepen the information and gain the perceptions of ordinary people. Three of those villages were the same as visited during the first phase. One village in the park area was changed and the fifth village was Ranomafana Centre.

5 RESULTS AND COMMENTS

5.1 Health interventions of the RNP Project

A health component was included in the RNP Project, as proposed during the initial discussions by villagers who wanted improvements in their health situation. Villagers had experienced changes in health resources during the past decades (Harper, 2002). In the French colonial period up to 1960 Western medicine and health care had been available for all, and older villagers remembered those times well (Harper, 2002; Harrison, 1992: 86). After independence in 1960, health care in rural areas was reduced considerably. Madagascar implemented a socialist system from the beginning of the 1970s until the mid‐1980s, and health services were reduced further due to national health policies that concentrated health‐care improvements in urban areas. In the mid‐1980s, the World Bank started structural adjustments that further limited the rural health sector improvements. By 1991, access to modern health care was poor and there was a lack of doctors and medicines. As residents had experienced better times before, it was natural for them to ask for improvements to such a situation, especially when foreigners were enquiring about their needs (Harper, 2002).

One of the main aims of the health project was to reduce fertility because an increasing number of people was seen as a threat to the park (Swanson, 1996; Harper, 2002). According to Swanson (1996) the health component was also needed to decrease the need for medical plants obtained from the forest, although the study by Hardenbergh (1993) shows that the majority of plants used for medicinal purposes can be found in the peripheral area of the villages, not deep in the forest.

Initial health surveys were conducted in 1989 in seven villages in the peripheral zone of the future park (Figure 2). In 1990 and 1991, the survey was expanded to 18 villages and a wide range of health‐related data was collected (Kightlinger et al., 1992a). Already from the beginning, the team also provided medical treatments for basic illnesses and conducted health education (Peters, 1993). Malagasy personnel, including a doctor and nurses, continued the health services. The health component of the RNP Project consisted of these health team visits for treatment of basic illnesses and health education for five years (Wright, 1997). Later on, the health team included vaccination campaigns and family planning education. The number of villages visited increased to cover more of the population, but the visits became more infrequent (Peters, 1993). The credibility of the health project suffered from an unexplained exclusion of some of the initial target villages in the mid‐1990s (Harper, 2002).

Timeline of health interventions by the Ranomafana National Park

When USAID funding ended in 1997, health programmes were continued by MICET. They hired the previous personnel, but health‐related activities now concentrated on preventive health care. The objectives of the MICET health programme were to develop a local health‐care system, fight malnutrition and evaluate the status of the health situation in the park area villages. The change in activities was influenced by the decrease of funds, but also by the government's policy to concentrate all health activities in state‐run health centres. MICET had its own family planning office in Ranomafana Centre, from where the health team visited remote villages. Two women's groups for family planning were established in two municipality centres in cooperation with state health centres (Wright, 1997). From 1998 to 2000 family planning activities (education and contraceptives) and sanitation assistance included 12 villages of the 93 available villages, with two annual visits as well as cooperation in five health centres. The size and the composition of the health team decreased in time – in 2001, there was only one doctor in Ranomafana and in 2002, no health activities of the park were realised. This was due to the decrease in MICET's funding and the change in its scope to concentrate on scientific biological research instead of development activities, as was agreed when the USAID‐funded ICDP period ended. Also, the change in environmental policies to a corridor and landscape approach decreased funding from NGOs' activities from already established protected areas to new target areas.

Nowadays, other health organisations work in the national park area as well as in the villages outside it. For example, Secaline has held a successful infant nutrition and health campaign (Marek et al., 1999). During the 1990s, consultations in state‐run health centres were free of charge but medicines and hospital fees had to be paid. However, health centres have increased health and general hygiene‐related activities by building latrines and training volunteer villagers to educate people in the villages. Traditional midwives were given training at health centres in order to improve delivery conditions in rural villages. After 2002, when a new national president was elected, health care was administered free of charge and also medicines were given for free. This, however, resulted in the end of medicine stocks in health centres. The research focus of this study is the period before 2002.

5.1.1 Women's position and education

Women's position and education are one of the most important underlying issues of reproductive health and the reduction of fertility. In the case study villages women's activities included household matters such as cooking and taking care of children, as well as growing vegetables, planting rice and going to the market. Lone mothers and poor women worked in the fields as hired labour. There were no clear differences in women's positions or activities between the villages in the park area and beyond. Conservation restrictions on the use of forest resources seem to have had no major effect on women's daily workload, as fuelwood is collected mainly by men and children.

In general, Malagasy culture has strong support for equity between genders, although traditionally power has been held by men. For example, in one village a woman complained to a group of other women about her husband's violence. The women discussed the issue jointly and made a decision that the husband had to honour. In terms of education, both parents make decisions about children's schooling.

Female primary school enrolment was equal to that of males in the study area, and the ICDP seems to have had no effect on girls' schooling. For example, the percentage of girls of all pupils in primary schools in the park area was almost equal to that outside the park (Table 2). The average share of girls in the third grade from the girls in the first grade between 1990 and 2000 is 46 per cent in the park area, and 51 per cent outside thereof. The early figures varied notably, especially in the park area. If we divide the time period into two (1990–4) and (1995–2000), the mean percentage in the park area increased somewhat (from 40 to 50) but not in the area outside the park, where it decreased from 54 to 49. Dropping out of primary school was not related to gender, as these figures are similar to those of boys. Attending primary school has only recently become obligatory in Madagascar. In rural areas, however, poor families do not send their children to school, as they are needed as agricultural labour. Long distances to school and inability to pay school fees are also an issue (Korhonen & Lappalainen, 2004). A large number of children drop out of school after a few years. On average, every second girl who starts primary school stops before the third grade. Girls' low educational status seems to be the result of a generally poor educational situation and not discrimination against girls' schooling.

Percentage of girls (girls/total number of pupils at school) in primary schools between 1990 and 2000, and the percentage of girls in the third grade from girls in the first grade between 1990 and 2000, in the Ranomafana park area and outside it

Area	1990	1991	1992	1993	1994	1995	1996	1997	1998	1999	2000
% of girls in primary schools
Park area	60	49	56	49	54	49	49	50	49	49	50
Non‐park	49	51	53	52	55	47	49	48	48	48	51
% of girls in the third grade*
Park area	34	32	47	56	31	26	38	49	36	55	97
Non‐park	60	70	58	25	59	43	39	46	55	58	51
Note: *Figures were derived by dividing the number of girls in the third year by the number of girls in the first year in every year. For example, the number of girls in the third grade in 1990 was divided by the number of those who were in the first grade in the same year.

5.1.2 Starting sexual activity and having the first child

Traditional Malagasy marriage was an informal relationship where women were free to leave their partners (Peters, 1992; Harper, 2002). Young girls could have children before marriage. Already in 1990, before the establishment of the park, the first child was usually born in the mother's middle teens – 75 per cent of the women participating in the survey in the park area had their first child before age 20 (Kightlinger et al., 1992b).

Although traditionally girls could have had a rather liberal sexual life, many sources report that nowadays girls start sexual activity and fall pregnant even younger than before. According to DHS surveys, 40 per cent of adolescent girls in Fianarantsoa province in 1997 were in reproductive life, compared with 29 per cent in 1992 (DHS, 1992, 1997). The same trend, of an earlier onset of sexual activity and having the first child, was stated by the midwives and doctors interviewed, as well as in focus group discussions.

In the park area and outside it, sexual activities were said to start at 14 or 15 years, and more girls had their first child at the age of 15–17 years. According to MICET (1998), the average age of the first sexual intercourse in six villages in the park area was 16,9 years. It is higher than the ages reported in interviews. The increased number of teenager pregnancies is not improving the reproductive health situation in the area.

According to the interviews, general social change, the increasing numbers of children in a family, combined with a difficult economic situation, were given as reasons for the decreasing age at which girls' sexual activity started. These conditions make parents incapable of looking after their teenagers who imitate what they have seen on television or on a market day. The tight economy makes young girls more willing to men's suggestions, as men promise them some clothes and food, something which the parents are often not able to provide. The RNP Project could have an accelerative impact on social change, as it was reported that people no longer respect the traditional way of life in Ranomafana town, which is the most affected by ‘external’ influences related to the park, such as ecotourism initiated by the ICDP.

5.2 Contraceptive use

This study concentrated only on modern contraceptives in its evaluation of park activities, although some traditional contraceptives were used in villages. The prevalence and use of modern contraceptives are low in Madagascar. According to DHS surveys in 1997, the national average of contraceptive use was 10 per cent. In 1990, the national prevalence was estimated at only 2–3 per cent (Harrison, 1992:87). Table 3 shows the use of contraceptives in the study villages. At the municipal level the average rate of contraceptive use in the park area was 6 per cent, the same as at the village level. In the municipalities outside the park area, the average rate was only 2 per cent.

Use of modern contraceptives (% of women of reproductive age) in villages in the Ranomafana park area and outside it (n=number of villages)

	Non‐park area in 2001**
	Park area in 1990* (n = 18)	Park area in 2001** (n = 7)	Near the park (n = 2)	Far away (n = 4)	Total (n = 6)
Average %	0	6	7	2	4
Range %	–	0–18	3–11	1–5	1–11
Sources: *Kightlinger et al. (1992b); **interviews with health animators in 2001. The number of contraceptive users was estimated by health animators, which we divided later by the total number of women of reproductive age in each village.

The prevalence of contraceptive use both inside and outside the park area is lower than the national rate of 10 per cent. However, the RNP Project may have had an increasing effect on contraceptive use – in the past ten years the use of contraceptives has increased in the park area. In the municipality of Ranomafana the rate is 10 per cent, but in the other municipalities it is much lower. Also, in a study carried out in 1999, 91 per cent of women in the centre of Ranomafana knew about modern contraceptives and 24 per cent used them (Rakotoarimino, 1999). In the villages outside the park, use of contraceptives seems to be more common in those that are near to the park. It is likely that the park's family planning programmes have affected the prevalence of contraceptives in Ranomafana, as well as in other park municipalities. The MICET health and family planning clinic is situated in Ranomafana, and although the doctors also visit the other municipalities the main working area is Ranomafana. Already ten years ago in Ranomafana town, women had a general interest in family planning which was not reported in other villages (Kightlinger et al., 1992b). In 1997, MICET started to work with an emphasis on family planning. According to health statistics, the number of contraceptive users rose dramatically – in 1995 there had been seven new users; in 1996, 32; in 1997, two; and in 1998 there was 80; in 1999, 80; and in 2000, 84 new users.

The most common forms of contraceptives in all municipalities and villages were first, injections and then oral contraceptives (pills). This had been the case during the extent of the park's health project (Bakoliarisoa, 1998; Rakotoarimino, 1999).

However, the rates inside the park area vary a great deal (see Table 3). The only village without any contraceptive users was in the park area, less than 10 km from a village where 10 per cent of the women were using contraceptives. This latter village is situated on the national highway and was one of the villages visited regularly by the health team, whereas the former was never included in the park's health activities.

According to the interviews, the reasons for the low rates of contraceptive use seem to be the low educational levels of women, difficult access to family planning information, fear of side effects and religious beliefs. A main constraint on the use of contraceptives was women's fear of and uncertainty about them. Health animators, as well as focus group interviews in all villages confirmed that women were afraid of the rumours that contraceptives would cause diseases and infertility. Traditional Malagasy culture does not hold major obstacles for contraceptive use but as Catholicism is quite strong, especially among Betsileo, the number of children a woman has is considered a decision of God. Many people stated that their ancestors did not use contraceptives, so why should they? Harper (2002) suggests that in tight economic times, people rely more on the power of ancestors to solve social problems.

These findings support previous studies about constraints on contraceptive use in developing countries (e.g. Bongaarts & Bruce, 1996; FHI, 2002). Bongaarts & Bruce (1996) found that although geographic access to services remains a problem, the principal reasons for not using contraceptives are lack of knowledge, fear of side effects, and social and familial disapproval.

Even though the fear of side effects seems to prevail among the women, according to the doctors, the only side effects they have encountered have been some weight gain and absence of menstruation. Doctors reported their difficulties in working and educating the villagers to use contraceptives: ‘They eat pills like candy, all in same day […] They never respect the appointment for having an injection as they do not know the run of time.’ However, the problems with side effects of modern contraceptives, especially injectables, are real and must not be glossed over. Hormonal birth control needs accurate information and continuous use, which is very difficult and sometimes impossible in local circumstances. Fear of side effects of hormonal contraceptives cannot be overlooked. Because of difficulties in the use and continued supply of contraceptives, side effects are real (see, for example, Paul et al., 1997; Philips, 2001), and could be terrifying for women who are not aware of them. The study made in Ranomafana Centre also indicated misuse as one of the main problems in the adoption of modern contraceptives (Rakotoarimino, 1999).

5.2.1 Fertility

In Ranomafana, the population has increased during the past decades. In the municipality of Ranomafana, the population increased 25 per cent between 1960 and 1970, 44 per cent between 1970 and 1980 and 18 per cent between 1980 and 1990 (Kightlinger et al., 1992a). During the existence of the park between 1990 and 2000, the population in Ranomafana increased by 83 per cent. In 1996, the eastern fokotanys of Ranomafana were split to form a new municipality, Kelilalina. The high population growth cannot be explained only by high birth rates; immigration has also contributed to it. In 1998, there were 6 723 inhabitants and 197 immigrants in Ranomafana; in 1999, 189; and in 2000, 2 899. Immigration to Ranomafana was noticed already in mid‐1990, when Grenfell (1995) classified it as the second most important indirect human pressure on the park.

Traditionally, Malagasy culture places a high value on having many children. Children are seen as a sign of wealth and vitality, as well as proof of one's existence. Also, children provide social security when people age, and they continue the family line. In particular, boys are needed to support the parents in older age as women will marry and leave the community (Harper, 2002). Children are also seen as a labour resource for working in the fields.

In 1992, the number of children per woman in the peripheral zone of the park was about the same as the national average, which was 5,4 in 2000 (World Bank, 2002). In 1992, the average number of children per woman in the park area was 5,7 (Kightlinger et al., 1992b). In a 1997 MICET survey, the average number of children per woman was 6,3 (MICET, 1998). In the interviews with the health animators and midwives no exact figures for the fertility rate were established, but estimations indicate that the number of children has not changed dramatically during these ten years of the park's existence. We were able to establish some figures by calculating the number of births and the number of fertile‐age women, and by assuming 35 years of similar fertility for each woman (Table 4).

Fertility rates in the Ranomafana National Park area and the area outside thereof

Area	1990‐1*	1998 (MICET survey)	1999**
Park area	5,7	6,3	5,3
Non‐park area	–	–	4,9
Sources: *Kightlinger et al. (1992b); **the number of women of reproductive age (15–49 years) was obtained from health statistics and was estimated at 23 per cent of the total population in a municipality. The number of births was obtained from the census data.

In focus‐group discussions it was noticed that people want to have fewer children, although many felt that limiting the number of one's children was not in the hands of human beings. A study conducted in Ranomafana Centre in 1998 indicated that the majority of women who used contraceptives did so in order to space out the deliveries and not necessarily to limit the number of children (Bakoliarisoa, 1998). However, in the park area villages the average number of children a woman hoped for was around four and in the villages out of it around five. Ten years ago, 65 per cent of the women participating in the survey in the park area wanted to have six or more children (Kightlinger et al., 1992b). Thus, at least for the women in the park area, it seems that there has been some change concerning the desire to have many children, although the actual number has not reduced.

Lower fertility was one of the goals of the park's health project, as a growing population was seen to be a major threat to natural resources (Grenfell, 1995). However, although villagers perceive that population growth causes shortage of land and other resources, a large number of children in a family is seen to maximise the potential support as one ages. Particularly because the high infant mortality makes the number of surviving children uncertain (Harper, 2002), the perception of children being a threat to the forest was not the kind of goal that villagers would understand.

5.3 Prenatal health care and delivery conditions

In 2000, most pregnant women visited the health centre before the birth (Table 5), despite the fact that births usually took place in the village than at the centre. In the park area, the proportion has increased during the past ten years but the same was true for the women outside the park. Midwives in the villages and health centres reported an increase in prenatal visits in both areas. It seems that an increase in prenatal visits is a more general trend and not related to the RNP Project. In fact, the average percentage rate out of the park area was 79. The increase in prenatal visits was noticed as well by the province‐level survey in 1999 (MGHC, 2001). However, conservation has restricted access to forest resources that are also used to supplement the diet (Hardenbergh, 1993). This is particularly important to the poorest women when being pregnant.

Share of deliveries in health centres and women with prenatal visits in the Ranomafana park area and outside thereof

	Park area in 1990*	Park area in 2000**	Non‐park area in 2000**
Number of municipalities and villages	18 villages	Six municipalities	Four municipalities
Deliveries in health centres (%)	29	15	23
Range (%)	0–86	3–22	11–48
Pregnant women having prenatal visits (%)	49	71	79
Range (%)	0–89	35–100	40–98
Average number of prenatal visits per visiting woman	–	2,4	2,3
Range	–	1,3–3,1	1,8–3
Sources: *Kightlinger et al. (1992b) survey in 18 villages; **health centre statistics, average for 1999–2001.

Comparison in time and between areas inside and out of the park indicates that the ICDP has not increased deliveries at health centres. The percentage rates of giving birth in a health centre varied from 3 to 48. The two municipalities with the lowest rates lie partly in the park area. In Ranomafana municipality, the rate of deliveries at health centres was 22 per cent, the second highest in the studied municipalities. Most of the women who gave birth in health centres in Ranomafana were living in Ranomafana town. In 1990‐1 in the park area, 70 per cent of the children were born at home and 29 per cent in a health centre (Kightlinger et al., 1992b). Deliveries in the villages currently seem to continue or increase. The percentage rates of births in health centres has decreased from 29 to 15 in ten years.

In the studied municipalities, the highest rate of giving birth in health centres was in the municipality outside the park area where almost half the deliveries (48 per cent) took place in the health centre. This municipality lies on a high plateau where distances between villages are not very great. In the park area, access to many villages can be very difficult due to the steep slopes and rivers. In addition to long distances to the health centre, high fees for overnight stays, equipment and medicines and the shame of not having extra clothes for the mother and the child were given as reasons for not having a child in the health centre. People around the park live at subsistence level and have the lowest income in the country, making them unable to pay hospital fees (Harper, 2002). The decrease in health‐centre deliveries in the park area might indicate harder economic times, as the costs of deliveries in health centres are currently too high for ordinary people.

5.3.1 Infant mortality

The rates of infant mortality given for the years from 1990 to 1998 were captured from mayors' census books by counting the notified deaths. The average rate from 1999 to 2001 was calculated from health centre statistics, and concerns only the deliveries at health centres (Table 6).

Infant mortality rates (population based), averages of deaths per 1 000 live births (infants under one year old)

Area	1990	1993	1994	1996	1998	1999–2001*	% of births in health centres
Park area	34	40	29	48	34	74	15
Non‐park area	57	31	72	54	45	60	29
Note: *Deaths/children born live in health centres.

The highest and the lowest figures at municipal level are again inside the park area and there are no significant differences between the park area and the areas outside it. The figures in Table 6, however, are generally much lower than the country's average, according to World Bank statistics of 88 per 1 000 live births in 2000 (World Bank, 2002). The average rate of infant mortality between 1990 and 1998 in the park area was 37 per 1 000 live births and in the area outside it 52 per 1 000 live births. The differences between these figures can be explained by variations in reporting deaths and births, with the overall percentage of births taking place in the health centre, as well as the births that occurred in the health centre possibly being the most problematic ones. These factors could explain some of the higher figures in health centres. High figures could also be the result of better communication, as the villagers generally informed the authorities of deaths, which probably explains the low average rate in the park area.

Harrison (1992: 86) estimates that infant mortality in the villages in the park area is higher than the country's average, which is close to 200 deaths per 1 000 live births. In the interviews in the present study, no reliable information for establishing the exact figures could be obtained. However, village midwives reported that infant deaths in deliveries have decreased both inside and outside the park area. By proportioning the number of deaths in deliveries to the total number of births that traditional midwives had assisted in a village, we were able to establish some estimations for a rate of deaths in village deliveries. Midwives witnessed infant deaths in an average of 20 per cent of deliveries in the park area and 22 per cent outside the park. Although midwives reported a general decrease in infant deaths, two midwives in the park area reported that in 2001 there had been many infant deaths for unconfirmed reasons. In general, the reasons for deaths were malaria, diarrhoea, convulsions and general infant morbidity.

It is very difficult to estimate the effect of conservation restrictions on infant mortality, particularly when the figures do not show any clear trends. However, it was noticed that deliveries in hospital have decreased in the park area because of the weakened economic situation. This could also mean an increase in infant mortality.

6 CONCLUSIONS

To summarise, the impact of the ICDP on reproductive issues has been modest. It seems that the ICDP has had no positive effect on decreasing the onset of sexual activity and having the first child; on the contrary, it might have been accelerating teenager pregnancies. The RNP Project has increased contraceptive use although the scale has been very minimal. Prenatal health care has increased in Ranomafana, but the ICDP does not seem to be the reason for this. Births in health centres have decreased, possibly due to the weakened economic situation.

Harper (2002) stated that people's health had worsened, as the RNP had enhanced the decline of economic resources started by other external factors such as structural adjustment. People's purchasing power had seriously declined and this was especially true concerning the costs of medicines for ordinary people. Marcus (2000) reported that participation in ICDP activities in Ranomafana did not play any role in the wealth of villagers. The factors influencing economic success were not related to the ICDP in any way. Our reproductive indicators showed the poor state of reproductive health in the park area, although the objective of the project had been to improve the situation. It seems that this goal was not achieved.

However, there were hardly any differences in the indicators inside and outside the park area. Conditions in the park area should be at least at the same level as in the other areas, and that was shown in this study. However, there were great differences between the villages and municipalities, which are not shown in the average figures calculated for the whole park area. In Ranomafana Centre and in the roadside lowland villages, the reproductive health situation seems to be better than in remotely situated municipalities and villages. However, remote villages also faced problems caused by nature conservation, including limited expansion of agricultural land and access to forest resources that are used in various ways to supplement livelihoods (e.g. collection of honey, crayfish and tubers).

Even though ICDPs aim at both development and nature conservation, in many cases the development activities have been considered subordinate to conservation, and biodiversity conservation is a real primary goal (Hughes & Flintan, 2001). This is why, when budget constraints of the park's health intervention became apparent after ICDP funding ended, the focus of health activities shifted to family planning, which was seen to have a clearer connection with conservation objectives. It is true that family planning aims to improve people's living conditions, and serves both the development and conservation goals. However, it must be conducted in a culturally sensible manner. For example, the health project's Western working methods might not have been applicable to the local context. Many educated Malagasy project workers' attitude towards their rural uneducated counterparts is that people are sufficiently informed but they ‘stubbornly’ cling to their old habits. On the contrary, the research findings show women's fear of side effects of contraceptives, which is also a sign of inadequate information. It would have been crucial to establish why women refused to take contraceptives, in order to increase education and adjust the health programmes to be more suitable for the local people. Also, the organisation that had the responsibility of continuing the health activities was neither a health nor a development organisation but an environmental organisation, led by conservationists who consider human health problems minor to the protection of nature.

A health project was also originally included because it was requested by the local people. As the ICDP aims to follow a participatory approach this was the least it could do, as the idea of the park did not come from the local people. The original idea of improving the health situation was a good one, and villagers were aware of the link between the health team activities and the park. However, the unexplained cessation of visits to certain villages as well as sporadic health team visits, which focused later on family planning and finally stopped altogether, were not perceived by the villagers as a trade‐off against their lands (Peters, 1998b; ANGAP, 1999). The idea of an ICDP, namely development as a tool for enhancing conservation, was gradually abandoned.

Nowadays, the inhabitants of Ranomafana are beginning to realise that their fields cannot feed an increasing number of people, and that conservation is a restricting factor for the expansion of their fields. Among the other policy and development reforms, family planning and health services in a conservation project are essential to solving this problem.

To conclude, experience from this case study can be applied to any other conservation project, particularly those implementing the ICDP approach. Regarding the results of this article, equal commitment to both development and conservation is the key issue in ensuring the success of socially sustainable conservation. Understanding the cultural aspects of reproductive health and enhancing the empowerment and education of women are also issues that a successful conservation project should take into account. The health and education components of conservation projects are crucial to socially sound development, which conservation policies should aim to enhance.

Additional information

Notes on contributors

Elina Hemminki 1

Respectively, Researcher, and Senior Research Fellow, Department of Social Policy, University of Helsinki; and Research Professor, National Research and Development Centre for Welfare and Health, Helsinki, Finland. The authors thank their field research assistants, Andry Rakotoarivao and Chantal Sololiana, as well as MICET and ANGAP personnel in Ranomafana and Antananarivo. They also wish to thank Jari Niemelä, Juhani Koponen and Janice Harper for valuable comments on this article. Finally, the authors want to thank all the villagers, health animators, midwives and doctors who enthusiastically shared their information with the authors. The research was supported financially by the Academy of Finland (projects 02‐530‐001 and 45664).

Notes

Respectively, Researcher, and Senior Research Fellow, Department of Social Policy, University of Helsinki; and Research Professor, National Research and Development Centre for Welfare and Health, Helsinki, Finland. The authors thank their field research assistants, Andry Rakotoarivao and Chantal Sololiana, as well as MICET and ANGAP personnel in Ranomafana and Antananarivo. They also wish to thank Jari Niemelä, Juhani Koponen and Janice Harper for valuable comments on this article. Finally, the authors want to thank all the villagers, health animators, midwives and doctors who enthusiastically shared their information with the authors. The research was supported financially by the Academy of Finland (projects 02‐530‐001 and 45664).