OVERDUE FINES ARE 25¢ PER DAY PER ITEM Return to book drop to remove this checkout from your record. Wilflflzllfllljflllflflflllllflflli’IUUIIJIII 11‘3“" y This is to certify that the thesis entitled FACTORS INFLUENCING WOMEN'S CHOICES 0F OBSTETRICAL CARE IN A NORTHERN DISTRICT IN THE PEOPLE'S REPUBLIC OF BENIN presented by CAROLYN FISHEL SARGENT has been accepted towards fulfillment of the requirements for Ph.D. . Anthropology degree 1n WWW —7 Mag professor 57/2/22 Date 0-7 639 OVERDUE FINES ARE 25¢ PER DAY PER ITEM Return to book drop to remove this checkout from your record. © 1979 CAROLYN F l SHEL SARGENT ALLRIGH'I'S RESHWED FACTORS INFLUENCING WOMEN'S CHOICES OF OBSTETRICAL CARE IN A NORTHERN DISTRICT IN THE PEOPLE'S REPUBLIC OF BENIN By Carolyn Fishel Sargent A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Anthropology 1979 ABSTRACT FACTORS INFLUENCING WOMEN'S CHOICES OF OBSTETRICAL CARE IN A NORTHERN DISTRICT OF THE PEOPLE'S REPUBLIC OF BENIN By Carolyn Fishel Sargent The intent of this study is to determine the factors which influence women's utilization of cosmopolitan and indigenous maternity care options in a Bariba region of the People's Republic of Benin. The study seeks to address issues pertaining to health care decision—making, the status of women, and to provide a thorough description of the indigenous obstetrical system. Research methodology included eighteen months of participant observation in the District of Kouande, administration of interview schedules to 40% of the women of reproductive age in the primary research site of Pehunko and structured and unstructured interviews with eighteen midwives in six Bariba villages. The majority of Bariba women are found to deliver at home and to maintain a preference for the customary ideal of a "solitary delivery," in which the parturient delivers unassisted except in difficult deliveries, when a midwife may be called. Nonetheless, seven years of statistics indicate a slow increase in clinic utilization; clinic patrons are primarily drawn from among wives of civil servants and wealthy traders. The study asserts that in attempting to delineate the factors influencing Bariba obstetrical care preferences, it is necessary to consider a set of salient factors. In general, the set of factors shown to influence obstetrical care choices are conceptualized as monetary and non—monetary constraints and incentives which interact to facilitate or to inhibit choice of a maternity care option. The data indicate that an understanding of the implications of delivering alone is integral to comprehending preferences for birth assistance. Further, the implications of solitary delivery derive from concepts of witchcraft as a causative agent of misfortune and to infanticide practices related to delivery of witch babies. The conceptual dichotomy of the public and domestic domains is utilized to consider sex roles in Bariba society, where males are identified with the public and females with the domestic domain of activities. Reproductive processes are found to comprise the primary arena in which women control decision-making. Moreover, professional specialization in reproductive disorders provides a rare avenue for women to transcend the limitations of sex. By selecting the national health service delivery option, women's customary realm of responsi— bility is diminished and this effect constrains women from utilizing the national health service. The study proposes that clients select a practitioner based on the definition of the problem to be treated, and on the degree of similarity in attributes (homophily) between themselves and the practitioner. Absence of homophily operates as a constraint to clients evaluating alternatives for obstretical care. Choice of practitioner is also suggested to be related to expectations held by clients regarding attributes, techniques and practices characteristic of a competent midwife. In considering clinic utilization in contrast to home delivery, the factors of central importance are categorized as l) role modeling and 2) boundary maintenance behavior. The significance of role modeling derives from the fact that the original clinic clientele consisted of southern civil servants transferred north who were habituated to cosmopolitan health care. This clientele served as an example of innovative behavior to the local population. Boundary maintenance behavior refers to the selection of health services as a means for Bariba and non-Bariba to maintain ethnic distinctions. Increasing clinic utilization by peasants and traders is suggested to be a precursor to the breakdown of boundary maintenance behavior and to represent a facet of the process of national integration. A Yaayi Ganigi, Yaayi Yo, Yaayi Baaba Ngobi Yon Sika, avec mes profonds remerciements et surtout a Kora Zaki Zaliatou, avec affection. ii ACKNOWLEDGMENTS I would like to thank the Social Science Research Council, the Fulbright-Hays Doctoral Dissertation Research Abroad Program, the Ford-Rockefeller Population Policy Program and the National Science Foundation for financing the fieldwork upon which this study is based. Among the many individuals who have assisted me in this research, I want to express particular gratitude to my major professor, Arthur J. Rubel, who provided me with advice, direction and encouragement throughout my graduate career. I also wish to thank the other members of my committee--John M. Hunter, John Hinnant and Brigitte Jordan, for offering guidance and support during the preparation of this dissertation. In addition, I have benefitted from the provocative insights of William Derman who provided comments on my research in particular and anthropology in general. I owe special thanks to my family for their patience and unflagging sense of humor during my graduate studies, and finally I am indebted to the women of Pehunko, who generously endured repeated interviewing and interruptions in their daily schedules and made this study possible. iii TABLE OF CONTENTS List of Tables . . . . . . . . . . . List of Figures . . . . . . . . . . . Chapter I. II. III. IV. PROBLEM STATEMENT AND THEORETICAL Introduction . . . . . . . Problem Statement . . . . . . . The Study Community . . . . . . Data Collection . . . . . . . . Theoretical Perspective . . . . PERSPECTIVE . . . . . . O O O O O O O O I O O O C O C O O O O O O O O O C O O O O O O O O O C O O O C O O O O O O 0 Conclusion. A Model of Factors Influencing Service Utilization . . . . . . . . . OTHER LITERATURE RELEVANT TO THE STUDY OF DECISION-MAKING o o 0 o o o o o 0 Decision Models . . . . . . . . Midwifery Literature . . . . . HEALTH CARE Alternative Utilization of Services . . . . . . . . . Characteristics of Midwives . Techniques and Practices . . Integration of Indigenous Midwives in Health Service . . . . . . Conclusion 0 I O Q o o . Bariba Conceptions of the Order Diagnosis and Treatment . . . . Divination . . . . . . . . . . The Use of Substances . . . . . Conclusion . . . . . . . BELIEFS AND PRACTICES SURROUNDING Menstruation and Clitoridectomy Conception . . . . . . . . . . Development of Fetus . . . . . Contraception . . . . . . . Abortion . . . . . . . . . . . Sterility . . . . . . . . . . . Conclusion . . . . . . . . . . iv National THE CULTURAL CONTEXT OF THERAPEUTIC CHOICE . . . . . . . of the Universe . . . . vii . viii O‘h‘O‘b3P4 ha F‘h‘ 45 45 51 51 56 65 72 78 8O 81 89 97 100 106 110 111 115 118 119 120 121 123 v. VI. VII. VIII. IX. STATUS AMONG THE BARIBA: THE ROLES AND RESPONSIBILITIES OF WOMEN O O C O O O O C O O O O O O O O O O O C O O O 0 Status in Bariba Society . . . . . . . . . . . . . . . Position of women . . . . . . . . . . . . . . . . . . Economic Subsistence . . . . . . . . . . . . . . . . . Political Arena . . . . . . . . . . . . . . . . . . . Domestic Relations . . . . . . . . . . . . . . . . . . Marriage . . . . . . . . . . . . . . . . . . . . . . Household Responsibilities . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . SOCIOLOGICAL AND CAREER ATTRIBUTES OF MIDWIVES . . . Healers: Midwives and Medicine People . . . . . . . Tingi: The Medicine Person . . . . . . . . . . . . . Apprenticeship . . . . . . . . . . . . . . . . . . . The Power of Words . . . . . . . . . . . . . . . . . Midwife as Healer . . . . . . . . . . . . . . . . Midwife as Category . . . . . . . . . . . . . . . Implications of Role Expectations for Birth Assistance Status Characteristics of Midwives . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . THE MEANING OF THERAPY AND EFFICACY IN RELATION TO OBSTETRICAL CARE PREFERENCES . . . . . . . . . . . . . BIRTH ASSISTANCE IN THE RURAL AREA: PATTERNS OF DELIVERY ASSISTANCE AND CLIENT-PRACTITIONER ENCOUNTERS . . . . . Delivery Assistance: Patterns of Selection in the Rural Area . . . . . . . . . . . . . . . . . . Person Present at Last Delivery . . . . . . . . . Clients and Cord- Cutters . . . . . . . . . . The Baby-washer . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . Midwifery as a Therapeutic System . . . . . . . . . . Client-Practitioner Encounters . . . . . . . . . . . . The Case of Adams . . . . . . . . . . . . . . . . . The Case of Sako . . . . . . . . . . . . . . . . . . The Case of the Prolapsed Cord . . . . . . . . . . . The Case of the Terrifying Breech . . . The Case of Bona . . . . . . . . . . . . . . . . . . Pain as a Cultural Phenomenon . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . UTILIZATION OF NATIONAL HEALTH SERVICES FOR MATERNITY CARE IN THE DISTRICT OF KOUANDE . . . . . . . . . . . Clinic Vs. Home Delivery: A Pehunko Sample . . . . . . Utilization of the Pehunko Dispensary . . . . . . . . Pehunko Women at the Kouande Maternity Clinic . . . . The Kouande Maternity Clinic: General Utilization . . Number of Deliveries . . . . . . . . . . . . . . . . Distance . . . . . . . . . . . . . . . . . . . . . . 125 127 131 132 137 138 138 140 144 148 148 148 152 155 156 157 159 160 199 202 218 219 221 225 228 234 235 245 245 252 255 263 264 271 283 287 288 292 298 300 300 304 Prenatal Consultations . . . . . . . . . . Reasons for Utilizing the Clinic . . . . . . . Occupational Status . . . . . . . . . . . . . Morbidity and Mortality . . . . . . . . . . . The Etic Perspective . . . . . . . . . . . . . The Clinic Setting . . . . . . . . . . . . . . The Emic Perspective . Conclusion . O O O 0 O O Q X. SUMMARY AND CONCLUSIONS . . . . . . . . . . . . . Assumptions Features Features Features Features Features of of of of of O O O C O O O O I O O O O O O Q the Disorder . . . . . . . . . . the Decision-Making Process . . . the Practitioner . . . . . . . . the Clients Seeking Obstetrical Care . the Model . . . . . . . The Cross-Cultural Study of Midwifery APPENDICES Appendix A: Demographic Data . . . . . . . . . . Appendix B: Female Circumcision Songs . . . . . . . FOOTNOTES . BIBLIOGRAPHY 0 vi 305 307 310 312 314 322 329 332 341 342 344 346 350 352 356 360 363 363 365 367 375 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. LIST OF TABLES Number of WOmen Interviewed by Neighborhood . . Salient Features of Indigenous Birth Attendants (from selected literature) . . . . . . . . . . . Matrones by Age . . . . . . . . . Birth and Residence Locations of Midwives . . Marital Status of Matrones . . . . . . . . . . . Matrone Recruitment Patterns . . . . . . . . . . Reproductive Histories of Thirteen Matrones . . . . . Birth Attendants by Neighborhood (major categories) . Matrones as Cord- Cutters by Neighborhood . . . . . . . . Birth Attendants and Cord- Cutters by Relationship to Client Baby-washer by Relationship to Client . . . . . . . . . . Neighborhood of Client and Baby-Washer . . . . Baby-Washer by Neighborhood . . . . . . . . . . . . Techniques and Practices of Bariba Matrones . . . . Home and Clinic Deliveries . . . . . . . . . . Characteristics of women Delivering All Children at Clinic Women Delivering First Child at Clinic, Subsequent Children at Home . . . . . . . . . . . . . . . . . . . . . . . . . . Occupations of Husbands of Women Delivering at Pehunko Dispensary in 1976 . . . . . . . . . . . . Occupations of Husbands of Pehunko Parturients . . Distance: Village of Residence-Maternity . . . . Prenatal Consultations and Clinic Deliveries . . . Consultations by Residence . . . . . . . . . . . . . . . . Deliveries of Consultees by Residence . . . . . . Home and En Route Deliveries by Year . . . . . . . . . . . Responses Given for Clinic Utilization . . . . . . . . . Occupations of Fathers of Babies Delivered at the Kouande Clinic . . . . . . . . . . . . . . . . . . . . . . . . . Maternal Mortality per Clinic by Year . . . . . . . . . . . Causes of Death in Cotonou Maternity Clinic . . . . . . . Distribution of Personnel . . . . . . . . . . . . . . . . . Official Rates of Obstetrical Services . . . . . . . . . . Age Structure of Pehunko According to Census . . . . . . Official Census and Research Sample . . . . . . . . . . . . Proportion of Live Births Ending in Infant or Child Death . RatiO‘of Clinic Users to Village Size in Relation to Distance Traveled . . . . . . . . . . . . . . . . . vii 13 64 162 163 164 168 174 224 226 227 228 228 229 243 288 289 291 293 298 301 305 306 306 307 307 311 323 324 328 335 363 364 364 373 O O O O Q O‘UIJ-‘wNH LIST OF FIGURES Map of People's Republic of Benin . . . . . . . . . . . . . . . 9 Map of Site Locations . . . . . . . . . . . . . . . . . lO Concepts of Sickness Causation among Bariba . . . . . . . . . . 91 Women Delivering First in Clinics, then Home . . . . . . . . . 290 Totals by Mbnth . . . . . . . . . . . . . . . . . . . . . . . . 302 Clinic Utilization by Year . . . . . . . . . . . . . . . . . . 303 viii CHAPTER I PROBLEM STATEMENT AND THEORETICAL PERSPECTIVE Introduction Throughout the Third World, a notable trend has been the proliferation of parallel health care systems in which cosmopolitan medical services coexist with a variety of indigenous health care services. Among the most ubiquitous of such services are those which provide maternal and child health care. Within the domain of maternal and child health services, an increasing number of options are available to client populations, including hospitals, private clinics, missionary health services, indigenous herbalists, diviners, midwives, injection dealers; hence decisions must be made regarding the preferred choice among the existing alternatives. The choice is not necessarily conceived as an absolute preference or rejection of one medical system in contrast to another. Rather When new forms of medical care, scientific medicine included, are made available to people whose health problems previously have been met solely or largely by an indigenous system, the basic decision they have to make is not whether to accept the new or adhere to the old. Instead, they have now a variety of options open to them that can be, and almost always are, exercised on a situation basis (i.e., the course of action that seems most appropriate for the particular at hand.) The strategies that underlie these decisionmaking processes have come to be called the 'hierarchy of resort in curative practice'. . . . (Foster and Anderson 1978: 248). Alternative utilization of maternal and child health services, and in particular, obstetrical services, represents one arena within which to study this "hierarchy of resort in curative practice."1 Obstetrical care is broadly defined here to encompass prenatal care, delivery assistance and postpartum care, fertility counseling and infant care, and thus refers to a range of widely sought-after services. Correspondingly, a consideration of the utilization of obstetrical services in an area such as rural Africa highlights a number of significant issues. Among these are the delineation of patterns of utilization of cosmopolitan and indigenous practitioners; the efficacy of different alternatives (including modes of measuring or evaluating 'efficacy'); and the role for indigenous healers in national health plans. All of the above issues are relevant to determining the most feasible methods of delivering Optimum health care to rural African populations. It might be argued, for instance, that integration of indigenous midwives in a national health service would have a twofold effect: it would provide an effective screening and referral system in a context where there is an insufficient number of physicians and nurse-midwives to assure care for the client population; and it would offer an obstetrical care alternative which meets client expectations of the midwife role. An additional topic of significance which is related to the utilization of obstetrical services in rural Africa is the role of women in the household, as professional specialists, and in society in general. In the West African setting, childbearing is fundamental to a woman's status, and reproduction is thus of particular concern to women. Moreover, specialists dealing with reproductive functions tend to be female. Research on factors influencing choices of obstetrical assistance, then, has implications for the study of decision-making, the provision of health care to rural African populations, the cross- cultural study of the status of women, and the study of indigenous midwifery. Problem Statement The intent of the research presented in this dissertation is to determine the factors which influence the utilization of obstetrical services in a Bariba region of the People's Republic of Benin.2 The research focused on that social field which includes a set of alter- native health care, and specifically obstetrical care resources, and comprises all healers, clients, relatives and advisors who impinge on or participate in health-seeking behavior and the healing process. This research project was developed as an observational experiment (Freedman et a1. 1978: 10—22), contrasting the village of Pehunko, where customary birth practices prevail, and the administrative center of Kouande, site of the district maternity clinic, where usage of both indigenous midwives and national health services is more common. Data on health service utilization in the People's Republic of Benin are scarce; also lacking are clinical and analytical studies of the comprehensive role of the indigenous midwife in providing repro- ductive counseling and treatment, and prenatal and neonatal care, both in Benin and in West Africa in general. Such data are potentially valuable both in supplementing information on beliefs and practices surrounding health-seeking behavior in Africa, and in contributing to planning of maternal and child health services which meet the actual and perceived needs of rural African populations. Moreover, this research attempted to develop a thorough description of the indigenous 4 Bariba obstetrical system. The latter was considered valuable both to supplement the relatively scant data on indigenous midwifery and to provide information on the prevailing birth practices. The research is also intended to provide an understanding of the context for change in birth practices by demonstrating that Bariba midwifery is a therapeutic system comprising a shared set of values, beliefs and practices. Viewed from this perspective, Bariba midwifery represents one choice, currently the prevailing alternative, for obstetrical assistance. A detailed presentation of the context of indigenous midwifery is imperative in order to comprehend the weighing of alternative obstetrical care settings and services, the goals and expectations of clients and practitioners, and resulting preferences. The Bariba case is particularly interesting to consider because of the contrast between the apparent attraction of maternity services in other countries and regions, and the reluctance to use maternity clinics and the verbal depreciation of their value by rural Bariba. It is difficult, with the inadequate statistics available, to discuss rates of utilization of maternity clinics. There does seem to be a trend towards increasing use of the Kouande clinic, although only a small percentage of district women regularly rely on its services.3 Dispensary services, on the other hand, are more consistently utilized, especially for child care and treatment of endemic diseases. In comparison, other observers elsewhere have noted that obstetrical services tend to attract clients more quickly than other services, due to the (alleged) observed benefits of improved care (cf. Maclean 1976, Foster, 1962, Landy 1977). It is intriguing, then, to consider the reasons for the slow increase in utilization of the national health 5 obstetrical services in the region in which this research project was undertaken. In subsequent portions of this chapter, the structure of the arguments to be set forth in the discussion of the above issues will be presented, beginning with a description of the study community and methods of data collection. This section will be followed by a discussion of the general theoretical perspective and the major hypotheses to be considered in the ensuing presentation. The hypotheses will be examined in light of the Bariba data but have implications wider than this delimited context, as subsequent arguments will suggest. The hypotheses under consideration fall into several major categories: Assumptions Underlying the Mbdel of Health Services Utilization 1. People usually do have choices to make; they have given wants and values and will select the most effective means to reach a goal or set of goals. 2. A primary consideration affecting a decision will be comparative costs of alternatives; these costs may be conceptualized as both monetary and non-monetary costs. Features of the Mbdel l. A multi-factor model of utilization of health care services is more explanatory than single-factor models previously devised by other analysts of health service utilization. Features of Clients Seeking Obstetrical Care 1. Social characteristics such as education and occupation are related to preferences for obstetrical assistance and setting; thus most rural women, other than wives of civil servants and wealthy traders will prefer home deliveries except in obstetrical emergencies believed to surpass the competence of indigenous midwives. Features of Practitioners (Midwives) 1. Clients will select a practitioner based on the degree of homophily (similarity in attributes) between themselves and the practitioner. Absence of homophily is considered to operate as a constraint to clients evaluating the desirability of alternative obstetrical assistance. 2. Choice of practitioner, moreover, will be related to expectations held by clients regarding attributes, techniques, and practices characteristic of a competent midwife. Features of the Disorder 1. Clients will select a practitioner or service on the basis of the definition of the problem to be treated and correspondingly, on the basis of decisions regarding the type of practitioner most competent to deal with the disorder as diagnosed. 2. Where a shared definition of the problem is absent, a service cannot offer care addressing the perceived needs of the population and will not be a preferred choice to a majority of the population. Features of the Decision-Making Process 1. women's decisions regarding choices of obstetrical assistance are not precipitous but are the outcome of cultural expectations, past experiences and existing alternatives. 2. The domain of reproduction represents one of the primary arenas in which women control decision-making. The Study Community The primary research in Pehunko was selected because of its location as a crossroads market center, through which communications from surrounding villages and towns flow regularly. Because of the regular transmission of information among the populations of the settlements in this region, it was hypothesized that reports concerning obstetrical care alternatives would circulate, and women would be aware of the options available to them and have at least a basis for evalu- ating the various alternatives. Moreover, from the perspective of studying a population's choice of alternative health services, Pehunko was considered appropriate due to its location at the intersection of several settlements housing maternity facilities and dispensaries. In Pehunko itself are a government nurse for curative medicine and a nurse for endemic diseases (i.e., leprosy, TB, meningitis). Pehunko, then, 7 seemed to serve as an illustrative site to investigate the alternatives available to a Bariba woman seeking maternity care and to examine the factors involved in her choice of a midwife. An additional consider- ation was Pehunko's historical significance as a Bariba chiefdom; because of the continuing importance of Bariba traditions in Pehunko at the time the research proposal was developed, it was believed to be a productive site for observing both Bariba obstetrics, and rituals and customary practices which might illuminate facets of Bariba healing. Pehunko is located in the northwestern Atakora province; the Bariba5 are a major ethnic group in this province, accounting for approximately 200,000 of the over 365,000 inhabitants, and occupying a wide area which stretches from the western border of Nigeria, to the eastern edge of the Atakora mountains, and from Randi in the north to Kouande in the south (see Figure 1). Numerous ethnic groups inhabit the Atakora, including the Somba, Yoabu, Pila—Pila, Fulbe, Dompago and Berba. Among these, the Bariba have a lengthy history of prominence. The Bariba, who have been characterized as a feudal society (Lombard 1965), are patrilineal and display a decentralized social organization linking a superior pro- vincial chief with chiefs of other provinces and with lesser chiefs. Incorporated thus in Bariba society were the ruling wasangari nobles, commoners (baatombu), slaves of varying origin, Dendi merchants, Fulbe herders, and autochthonous ethnic groups. ("Dendi" is a term long used by the colonial administration to refer to Mande, Sarakolle, Hausa, and other Muslim settlers who were usually traders.) Islam spread throughout the Bariba region primarily via these "Dendi" traders from Bornu and the Sudan (Lombard 1965: 38); Lombard FIGURE 1.-— Map of People's Republic of Benin* * Taken from The Atlas of Africa, Editions Jeune Afrique, 1973. FIGURE 1 . VOLTA B \l' \/‘—-—N If , e '3’ 10 \ FIGURE 2.——Map of Site Locations 11. points out that whereas elsewhere in West Africa, the upper or ruling class is usually Muslim, among the Bariba, the aristocracy tended to remain animist and the subordinate Fulbe and Dendi to be Muslim. However, Islam has been gaining strength among the aristocracy of the Pehunko region. Pehunko is a stronghold of Islam in the surrounding districts; I estimate approximately 24% of the population in the village is nominally Muslim. The economy of Pehunko and of the entire region is almost completely agricultural, based on staple food crops of yams, sorghum, and millet with increasing dependence on cashcrops of peanuts, cotton, and rice. Additionally, Pehunko is an important cattle market. Currently, the name Pehunko refers both to a Commune, or agglomeration of villages, and to "Pehunko Center," the village which originally was known by that name (derivation: $2352, rock; 32253, black). The Commune of Pehunko comprised twelve villages at the time of the study (the District is now being reorganized). Of these villages, five are Fulbe villages and were dropped from this study because of the necessity for focusing on primarily Bariba healing practices. Data Collection It quickly became apparent that political and logistical difficulties would inhibit extended interviewing of midwives and women of reproductive age in the seven Bariba villages, as originally planned. As a compromise measure, midwives in six villages were interviewed by means of interview schedules and informal conversations and observations over a period of a year (see Figure 2 for site locations). Of the 19 midwives interviewed, ten were selected because they had been identified 12 as part of a government effort to find and train indigenous midwives. The others came to my attention through references from their clients or friends. The primary difficulty in studying midwives, as will become clear in subsequent discussion, involved arranging first-hand obser- vations of the midwives assisting at deliveries. Because the midwives resided in villages located up to 20 kilometers from Pehunko (on roads sometimes passable only on foot) where I was based, and because mid- wives were rarely called before a woman was in advanced labor, it was extremely difficult to arrange to be present at deliveries in villages other than those in Pehunko center. Constraints also deterred interviewing women of reproductive age. These constraints derived both from sampling difficulties related to recordkeeping in the Commune and to political tensions. I began attempts at designating a sample by drawing on Commune census data. The population of the Rural Commune of Pehunko is officially listed as 12,228 inhabitants and the population of "Pehunko Center and Neighbor- hoods" as 2,114, of whom 1,113 are male and 991 are female. However, when I counted the inhabitants enrolled in the Commune census, I found 1409 listed, of whom 780 were women. The Mayor of Pehunko urged me to adopt the latter figures. Originally, I intended to draw a random sample of women of reproductive age in Pehunko in order to determine reproductive histories, including patterns of utilization of birth attendants. However, problems arose in attempting to track down villagers listed in the census. Many names had been falsified, men were recorded as women, the dead had not been eliminated, and new family members over 14 (the eligibility age for taxes) had not been 13 added. Thus I was obliged to abandon efforts at formal sampling based on the census. Instead, I followed networks of my own, in both aristocratic and commoner families, expanding through the networks of those families and adding others whom I was able to find through the census reports. The number of women of reproductive age listed in the census of Pehunko Center (277) was used as the guideline for the size of the sample of women of reproductive age whom I interviewed. I attempted to interview at least 277 women but was only able to arrange interviews with 117 women (42%). These women were drawn from the six major neighborhoods of Pehunko (see Table 1). TABLE l.--Number of Women Interviewed by Neighborhood Neighborhood Estimated # Women Repro. Age # Women Interviewed* Sinawararou 80 25 Pehunko Center 90 66 Gbankerou 10 9 Tance 15 6 Pehunko Gando ll - Zongo 52 17 Unknown 19 - * (See appendix A for age breakdown of women interviewed) Demographic data were extremely difficult to obtain on the Pehunko agglomeration. Census data were clearly not reliable and the political situation did not permit census-taking as part of the research. Since the 1972 military takeover of the national government by Mathieu Kerekou and the subsequent attempts to educate the country's population regarding the prevailing socialist philosophy, the national government has become an increasing presence in northern rural areas. Due to suspicions regarding government roles in tax collection, 14 marketing, ideological formation, and education a systematic effort to collect demographic data was initially deemed unfeasible. This tenta- tive conclusion was intensified when a mercenary invasion of the country during the research project (January 16, 1977) led to a vigorous radio broadcasting campaign to alert Beninois to be wary of speaking with or showing hospitality to foreigners, who might be mercenaries or spies. Some demographic data were obtained, nonetheless, by means of a small survey of household size. These data were obtained towards the end of the research project, by interviewing women who had participated in a survey of women of reproductive age and gearing the questioning towards the size of the group for whom the woman cooked, washed, etc. On occasions when the head of the household was present or appeared during the discussion of family size and composition, however, the woman was ordered to cease providing answers immediately. Such data as were obtained will be dispersed throughout the text where relevant. In addition to interviewing in Pehunko, some interviewing was undertaken in the village of Kouande, the district capital where the maternity clinic for the district is located. In spite of its position as district capital, Kouande was even more difficult to obtain data on than Pehunko, possibly due to political conflicts among district officials at that time. The only information which could be obtained was that the urban portion of the Commune of Kouande had an official population of 17,739; Kouande Center, where the clinic was located, had a population of about 3,100. Population figures for the rural portion of the district were unavailable and there were no available statistics on the breakdown of males and females in the district population. Thus 15 it was not possible to accurately estimate the proportion of the district population which utilized the district clinic. Interviewing in Kouande basically took the form of participant-observation and inter- viewing of women delivering at the maternity clinic over a two-month period. In discussing constraints on data collection, it is also necessary to note that many customary Bariba activities such as drumming in accompaniment to rituals and to horseriding and praise- singing were restricted by the national government as part of the national campaign to end "mystification" and "power of the chiefs." Thus traditional Bariba celebrations such as Gani, (New Year) when horse races were held and tribute delivered to the chiefs, were not observed during the research period. Similarly, spirit possession cults, which rely on drumming to assist the cult members in dancing and entering a trance state, were required to obtain permission from the revolutionary council (and, it is claimed by cult members, to pay the delegates prior to receiving permission to meet.) During most of the research, cult meetings were infrequent and/or held deep in the bush, surrounded by secrecy. This of course restricted observation of cult sessions and interviews with members, who were for the most part already hostile to any interest shown in their affairs and were suspicious of questioning regarding cult activities. Most data on spirit cults were obtained from two generous midwives, one of whom refused to directly discuss issues of ritual import but delighted in taking me to her own cult dance and to funerals so that I could watch; the other midwife was willing to answer questions and pantomime behavior for me, although she did not take me to any cult meetings. 16 This research project was organized as a collaboration between myself and Professor Eusébe Alihonou, Agrégé, Director of the Department of Obstetrics and Gynecology at the National University of Benin, and was designed as a concomitant medical evaluation of maternal and child health and anthropological investigation of indigenous midwifery in the sample villages. Again, due to the political situation, it was not possible to consistently pursue the medical aspects of the study as planned. Dr. Alihonou was obliged to devote an unexpected proportion of his time to political reorganization meetings in the capital and his efforts to find a sixth-year medical student to work with me were thrown awry when the university year was suddenly rescheduled and students dispersed for vacation or national volunteer service. Nonethe- less, Dr. Alihonou collaborated in the preparation of questionnaires on reproduction and obstetrical care and contributed to the evaluation and interpretation of various findings regarding midwifery practices. He also organized a team to conduct prenatal examinations for the women of Pehunko; this was meant to be one facet of a long-term study which did not in fact materialize; the results of the one set of consultations which are pertinent to the research are cited in the text where appropriate. Theoretical Perspective The analysis of health-seeking behavior in the context of multiple health care alternatives has been addressed in the framework of the larger issues of culture contact, social change, and diffusion of innovations, among others. Anthropologists who have considered the utilization of therapeutic options in such contexts have included Foster (1958, 1962), Gould (1957), Erasmus (1952), Colson (1971), and 17 more recently Riley (1974) and Janzen (1978). As Foster and Anderson noted in a recent work, one prevailing interest in the history of medical anthropology has been the "social— aspects-of-technological-change model" (1978: vi) as exemplified by Saunders (1954), Paul (1955), Erasmus (1952) and Foster (1958, 1962). Such studies focused on the introduction of western medical services into traditional societies, and emphasized that health and disease are aspects of total culture patterns and that social factors influence health care decisions. Studies such as Foster (1958, 1962) and Paul (1955) attempted specifically to deal with planned or directed change; for example, Foster's Problems in Intercultural Health Programs treated the ". . . formal, planned interchange of medical and associated ideas and practices across cultural and subcultural boundaries. . ." (1958: 3) and addressed such topics as cultural, social and psycho- logical barriers and motivations to change, cultural aspects of medical systems, social aspects of medical systems, and the implications for intercultural health programs of cultural and social factors. These works were more directed towards illustrating the importance of cultural and social factors in influencing technological change than in developing specific models of the process of choosing among health care alternatives. Reviewing the literature in a search for more directly compar- able data and useful explanatory models of the selection of therapeutic options, I found that much of the existing research (with notable exceptions which will be discussed below) has relied on single—factor models. In a similar review of the literature, Colson classifies the generalizations and hypotheses frequently used to explain the issue of 18 differential utilization of health services in two categories: these are "features of the disorder" as independent variables, and "features of the patient" as the independent variables (Colson 1971: 227). One hypothesis which falls into the category of "features of the disorder" is what Foster termed the "folk dichotomy." The essence of this hypothesis is that the use of modern as opposed to indigenous sources of therapy would correspond to the supposed origin of the disease. A disorder of natural origin would be in the domain of the modern thera- pist, whereas for a disorder of supernatural origin, the client would seek traditional health care (Colson 1971: 227). Another set of hypotheses which Colson includes in the category of "features of the disorder" suggests that choice of a service or therapy is based on previous evidence regarding the efficacy of the alternatives available. E. Fuller Torrey, for instance, suggests that behavioral disorders might most successfully be treated by indigenous practitioners whereas the outcome of treatment for somatic disorders would be more successful with modern therapy. Erasmus postulated the selection of western medical therapy to be "a matter of observed frequencies of successful outcomes" (Colson, 227). Such hypotheses imply that on the basis of differential efficacy, individuals will select a particular therapeutic option, thus lending support to the idea that evaluation based on empirical observation influences the use of health services (cf. Erasmus 1952: 267). Other generalizations which should be mentioned arising from the study of differential utilization of services include the concept of "shotgun" therapy (Colson 1971: 228) which describes individuals simultaneously using a number of types of therapies; and Gould's 19 hypothesis which suggests that indigenous practitioners would be consulted for chronic ailments, whereas doctors would be sought for acute disorders . . . ("folk medical practices . . . were employed whenever the person's complaints were classifiable as a chronic non- incapacitating dysfunction while doctors were being sought for com— plaints classified as critical incapacitating dysfunctions.") (Gould 1957: 508). Among those hypotheses regarding "features of the patient," the most significant to mention here are the concept of relative wealth and of relative acculturation. "Relative wealth" refers to the rather evident fact that insufficient resources might inhibit prospective clients from utilizing a service; it does not, however, deal with situations where wealth is sufficient but the client nonetheless prefers not to use the service in question. From acculturation theory is derived the hypothesis that the selection of medical services depends on the degree of acculturation of the client. Colson notes the shortcomings of this approach, stating that "the basic factor . . . is usually implicitly the relative degree to which one is involved in some forms of behavior that are not traditionally part of one's own culture." The problem is that "use of modern medical facilities is frequently employed as a measure of acculturation, which makes the acculturation argument tautological" (Colson 1971: 228). I intend to show that a multi-factor model is more explanatory than the models proposed by the writers mentioned above; although those models certainly contain pertinent insights into factors influencing utilization of services, they are too limited in scope. Rather, I will suggest that characteristics of the client, service, therapist and the 20 condition interact in affecting the resulting choice. A set of salient factors, including values and beliefs, exists and may change, varying with the situation (for example with residence of the mother or rank of child). In addition, the co-occurrence of variables may result in one factor affecting or altering another. The client may be viewed as weighing such factors in terms of the relative monetary and non-monetary costs; certain combinations of factors carry differential degrees of risk, uncertainty and benefit. The overall approach to be utilized is derived from rational choice theory, as presented in Heath (1976) and incorporates related ideas of Rogers and Shoemaker (1971), Barth (1967, 1969) and Riley (1974). In the perspective of communications, Rogers and Shoemaker (1971) attempted to integrate anthropological literature in developing an interdisciplinary model of the adoption of innovations which provides a pertinent perspective to the topic in question. In their work, Rogers and Shoemaker stated that innovative decision-making differs from other kinds of decision-making situations in that by definition, an innovation is an alternative which is perceived as "new" by the client population (Rogers and Shoemaker 1971: 99; see also Zaltman and Duncan 1977: 13). Elaborating on this concept, Zaltman and Duncan defined change as the relearning by individuals or groups (1) in response to newly perceived requirements of a given situation requiring action and (2) which results in a change in the structure and/or functioning of social systems (1977: 10). Innovation, in turn, is referred to as a subset of change characterized by its being an entirely new situation or phenomenon that the individual, group or organization is encountering (13). "New" can be a problematic concept, but not unduly in the 21 instance of utilization of maternity clinics, which were imposed by external agents and differ radically in structure, functioning, and philosophy than any indigenous institution. Rogers and Shoemaker developed a model of the diffusion of innovations based on the premise that communication is essential for social change. Although this model is intended to present an inter— disciplinary approach, it is not entirely applicable to this research because of its lack of emphasis on the effects of culture and social organization on diffusion; the diffusion of innovations model was also originally based on data from western societies, but since 1960, has been increasingly supported by research in Africa and Asia (1971: 83). However, two components of the model prove useful in regarding the data on choice-making between health care alternatives. Particularly explanatory is the concept of relative advantage, which the authors subsume in a paradigm of the "innovation-decision process" under "perceived characteristics of innovations." According to Rogers and Shoemaker: Relative advantage is the degree to which an innovation is perceived as being better than the idea it supercedes...Re1ative advantage, in one sense, indicates the intensity of the reward or punishment resulting from adoption of an innovation. There are undoubtedly a number of subdimensions or relative advantage: the degree of economic profitability, low initial cost, lower perceived risk, a decrease in discomfort, a savings in time and effort, and the immediacy of the reward (139). The authors then note that of eight investigations of the relationships between perceived attributes of innovations and their rate of adoption, almost all report a positive relationship between relative advantage and rate of adoption (139). The concept of relative advantage presupposes that the potential adopter is acting rationally, in the 22 sense of considering the most effective means to reach a given goal-—an assumption which Rogers and Shoemaker make explicit (1965); however, their concern is primarily with "objective" rationality, rather than with subjective rationality as perceived by the individual. Thus they cite Homans' measure of irrationality: "Behavior is irrational if an outside observer thinks that its reward is not good for a man in the long run" (165, my italics). It seems, then, that Rogers and Shoemaker are most interested in the change agent or "expert's" interpretations of rationality. They do nonetheless argue that it "is the receiver's perceptions of innovations' attributes" which affect decision-making (167). The concept of relative advantage and more specifically of rationality in decision-making will be shown to have explanatory power in the Bariba example. However, relative advantage is not intended here to imply a unidirectional choice; that is, one choice does not necessarily supplant another, and options may be selected alternatively or simultaneously. The approach of the communications writers such as Rogers and Shoemaker, and Zaltman and Duncan regarding rational choice is suggestive but not adequate. A more sophisticated approach, which focuses on choice models, is that of Heath in Rational Choice and Social Exchange (1976). Heath's book is a systematic account and evaluation of the fundamentals of exchange theory (5). Exchange theory, he states, is one of the rational choice theories and in his critique of the theory of exchange (particularly Blau 1964, Homans 1961, and Thibaut and Kelly 1959) he presents a review of the rational choice approach. It is this review which sets forth clarifying assumptions for the study of Bariba 23 decision-making. Heath explains that: Broadly speaking the rational choice approach...begins with the assumption that men have given wants, goals, values or 'utility functions'....It then assumes that these goals cannot all be equally realized. Men live in a world of scarcity and therefore must select between alternative courses of action....Social life no less than economic life is characterized by scarcity. We may be faced by a scarcity of time rather than money, but it is true nonetheless that we cannot have all we want. Accordingly men must choose and make decisions" (3). I According to Heath, people will make decisions "rationally;' rationality here refers to the process of selecting the most effective means to reach a goal, or preferred goal where more than one goal exists. The assumption of rational choice implies (a) that although norms and roles may prescribe behavior appropriate to certain situations, people do have latitude for individual decision-making and (b) that people deliberate and weigh the advantages of options. Rational choice theories include theories of price, power, collective action and coalition formation, derived from such disciplines as economics, political science, social psychology, psychology, and anthropology (cf. Barth 1966; Blau 1964; Homans 1961; Thibaut and Kelly 1959; Olson 1965). Rational choice theories are concerned with behavior in terms of goals and alternatives; theories may vary, however, in the "model of man" which they presuppose. Thus, in the economic theories of Downs and Olson, the model is that of "a man with fixed, given preferences . . . uninfluenced by feelings of envy or justice . . . All that he is concerned to do is to maximize the future stream of benefits that he will receive, and to this end he selects the course of action that will maximize his expected utility" (170). Contrastingly, the "social model of man" of Thibaut and Kelly (1959) views man as a strategizer, pursuing self—interest. However, in addition "he tries to 24 maximize a rather wider range of satisfactions"-- He seeks status and social approval as well as material gains, and he is also concerned to avoid such intangibles as anxiety and guilt....Socia1 man also suffers from feelings of envy or of relative deprivation; he compares himself both with his contempor- aries and with his own past. The level of satisfaction which he derives from a given outcome is thus not invariant over time and place, and his behaviour has to be explained not only in terms of his present situation and expectations but also in terms of his past history of satisfactions" (Heath: 171). A third model of man is "Skinnerian Man" (cf. Homans 1961) who funda- mentally relies on past experience for selecting behavior; thus he avoids what has proved painful in the past and seeks out that which has been rewarding (171). Choice of the particular "model of man" varies, then, with differing rational choice theorists but according to Heath, the model selected should be justified by empirical data. Heath also addresses the issue of the extent to which the rational choice approach can be used successfully outside economics where it originated, stating that providing there is some scope for choice, there is some scope for a rationalist approach" (60). He contrasts this approach with explanations of behavior based on uncon- scious drives, habitual action or societal prescription and concludes that the merit of the rational choice approach is "that it forces us to abandon the notion of man as a 'cultural dope' blindly following the norms and prescriptions of his culture. Instead it forces us to recognize man as a decision-maker who decides whether or not to conform in the light of the options available to him" (105). In general, Heath's synthesizing work lends itself well to providing a model for Bariba health-seeking behavior, and in return, the Bariba data support the utility of the rational choice approach. Overall, this perspective was determined to be more applicable in 25 considering decisions made by Bariba women than, for example, adopting the "garbage can model" of choice (Cohen et a1. 1972), designed to deal with situations where decisions are not taken methodically. According to this model, choice is not based on preference, but evolves out of action; decision-makers are characterized by goal ambiguity and lack of consistent, shared goals, and tend to act capriciously (Cohen et a1. 1972: 1). The Bariba data, as will be shown, indicate that women and others involved in obstetrical decisions have a clear set of goals, H rather than "goal ambiguity, that these goals are widely held among rural Bariba and that decisions, although they may seem precipitous in any obstetrical emergency are not capricious but measured and evaluated in terms of past experiences and current alternatives. The Bariba data, then, do not support utilization of the "garbage can model" of choice. Contrastingly, rational choice theory offers a useful framework for ordering the data and determining the factors influencing Bariba decision-makers engaged in situations of choice. As Heath explains, situations of choice may be categorized as situations of certainty, of risk, and of uncertainty. Situations of certainty are of course ones where each course of action open to the individual has a single, known outcome or consequence, whereas in the case of risk or uncertainty a number of possible outcomes may follow from any given course of action. .12 the case of risk a numerical probability can be assigned to the likelihoods of each of the outcomes occurring whereas in the case of uncertainty no such probability can be assigned (7). Bariba women, faced with the alternatives of dispensary or clinic settings for delivery or with home delivery (assisted or unassisted) are confronted in the current era with situations of uncertainty. 26 Maternity clinics are still relatively new in the Kouande region, dating back to around 1950 and women in outlying villages far from the district seat may have no role models for clinic experience. The typical woman making such a decision is called upon to weigh the benefits of a more certain situation--the home delivery, although of course the outcome of a delivery is never certain-«with a relative unknown, or at best, a very controversial choice. The analysis of alternative obstetrical care services by Bariba women presented in this dissertation is predicated upon certain assumptions which are fundamental to rational choice theory as synthesized by Heath: 1) Rational choice theory takes off from the idea that the normative approach to behavior neglects the possibility that "people usually d3 have choices to make, not least the choice whether to conform to the appropriate norms and role expectations. There is always some possible alternative course of action, and the costs of non-conformity are merely one set of costs to be weighed in the balance along with all the others" (Heath 1976: 176). Thus this approach adopts the basic assumption postulated by Kluckhohn and Strodbeck that variations exist in the value orientations "of whole societies, of subgroups within societies and of the individual persons, who are, in the final analysis, the actual carriers of culture...." (Kluckhohn and Strodbeck 1961: 3). Correspondingly, individuals may vary in their evaluation of choices, acting in relation to certain values in one situation but differentially in another context. Critics of rational choice theory have remarked that the evaluation of choice by the observer is problematic because of the 27 difficulties inherent in determining the factors influencing a choice. Most frequently, it is assumed that it is possible to determine people's values as manifest in their decisions and choices. However, if the analyst observes decisions made, and then merely assumes that people value that which they choose, the argument becomes tautological. However, rather than viewing the rational choice approach as tauto- logical because choices are defined as "preferred" because they are chosen, the issue might more aptly be perceived as a problem in the methodology of studying values (Cancian 1966: 467). It is not, then, a misdirection of theory. In this study it is assumed that the problem can be minimized by considering values and norms in other settings than the decision contexts, in order to acquire information on the range of salient norms and values which may be drawn upon in decision-making situations. Verbal explanations of choices may then be solicited to assist in clarifying the values which have influenced a decision, or which are used to endorse or justify decisions. It is recognized that anthro- pological methods are more designed to analyze behavior, which can be observed, than to evaluate motivations. 2) Individuals have given wants, goals, and values, which comprise alternatives provided in the context of a particular culture. Individuals will then select between alternative courses of action rationally, in other words, selecting the course of action which is the most effective means to a goal or the course leading to the most preferred goal (Heath 1976: 3). Rational choice theory thus sets out to explain man's behavior, focusing on alternatives, rather than on rules (Heath 1976: 173). 28 With regard to this assumption, it should be noted that one of the major criticisms of rational choice theory suggests that ordinary people are not particularly rational, that they do not weigh or evaluate alternatives as the theory postulates, that they more often choose impulsively, that they do not base decisions on complete and accurate information or try to obtain such information (Heath: 75). However, as Heath advocates, it is necessary to distinguish between "the rationality of a man's beliefs and knowledge and the rationality of what he does given those beliefs" (76). Rational choice theory concerns the choices people make given their beliefs and is not dependent on whether people are believed to act out of self-interest, egoism, duty, friendship, or based on any particular motive. Rather, "rationality has nothing to do with the goals which men pursue but only with the means they use to achieve them" (Heath: 79). It is certainly true that complete information may not be available to or utilized by individuals.6 In addition, constraints of which decision—makers are not aware (e.g., class factors, national policies, international economic competition) may impinge on choices made. Nonetheless, it seems desirable to differentiate the issues of social forces impinging on individuals (and beyond their domain of awareness) and the decision-making process in which the individual believes herself/himself to be engaged. That is, people operate on the premise that they do have choices to make, and that they have bases upon which to make decisions. The fact that they may be unaware of all factors comprehensible to an external analyst does not negate the validity of studying the choice-making process given the facts, beliefs, and values in relation to which the individual acts. 29 Correspondingly, Ortiz states that "rationality of behavior does not imply that there is a constant conscious awareness of having made a choice or even the ability to express it verbally in terms of quantities or factors" (1967: 196). It is important to note, though, that the number of alternatives available to an individual depends on the information available to that individual or significant others (Prattis 1973: 48; Ortiz 1967: 220). In the Bariba context, access to informa- tion to a certain extent correlates with social class; thus civil servants and those employed by government agencies or Europeans have greater access to reports about health services, quality of care, procedures for utilization, and so forth. Town-dwellers also have greater access to information than villagers residing off the main roads, due to rumor networks via shops, market communication, politi- cization of health care, and resultant attempts at publicizing and advocating use of national health care clinics. After more than 25 years of existence, it is the rare individual who is unaware of the existence of maternity clinics and of their function. Nonetheless, those who are involved in 'modern' occupations, particularly the civil service sector (including the agricultural service) and who reside in administrative centers have much more constant and fuller access to information regarding health care alternatives. The significance of this differential access to information is that when considering selection of an obstetrical service, whereas everyone may be aware of the actual existence of the alternatives, certain sectors of the population are better informed than others as to such aspects of the clinic alternative as expected behavior for a patient; familiarity with 3O bureaucratic procedures, and probably treatment procedures and staff behavior. 3) "An actor always has many goals; his actions in pursuit of any one affect and are affected by his actions in pursuit of others" (Heath: 180, taken from Cohen 1968, Modern Social Theory). 4) Overall, the analysis adopts the "social model of man" of Thibaut and Kelly (1959) described on page 20. Empirical justification for the selection of this model is that in the Bariba case, the data suggest that people minimize costs and try to assure reaching their goal to the extent that these two attempts can be attained simultaneously. Parenthetically, it is necessary to note that adoption of this model is not intended to suggest that "maximizing" is a universal characteristic of human nature but rather that in the context of Bariba obstetrical care decision-making, the preferred strategy of Bariba seems to be to maximize, in the sense of choosing an alternative most likely to lead to a specific preference. Following Prattis (1973) I would argue that "decision—making can follow a number of alternative rational paths" (47); these may include maximizing; a 'mini—max' strategy (minimizing chances for maximum loss) or satisficing, a strategy of adopting the first alternative that meets a minimal set of payoff conditions, particularly used in the presence of time and information constraints (Prattis 1973: 47). According to Prattis, rationality should be examined in terms of strategizing, and not exclusively in terms of any one type of strategy; the choice of strategy will be related to the situational logic of the actor making decisions (46). The Bariba data suggest that the strategy of priority in selecting an obstetrical setting and/or assistant is to maximize to meet the set of preferred 31 goals, although in some situations one could argue that people are required to adopt a different strategy (to preserve as many goals as possible) especially in obstetrical emergencies. The probable goals of Bariba women seeking a delivery setting are listed below. These goals are intended to represent the client perspective, taking as ideal types a farmer's wife in the village of Pehunko and a trader's wife in the administrative center of Kouande, where the district maternity clinic is located. The goals are listed roughly in order of priority: Farmer's Wife: Home Delivery, Trader's Wife: Clinic Delivery live mother and baby live mother and baby live mother live mother live baby live baby witchcraft control aspirations to upward mobility manifestation of high standards prestige associated with 'civil of female behavior servant' behavior minimize costs (infrastructural) minimize costs (infrastructural) "Strategy," then, refers to selecting a means of meeting goals. Selecting a particular delivery setting and type of birth assistant has implications for the goals which are most likely to be met--for example, witchcraft control is unlikely to be assured in a national health maternity clinic. The preferred strategy seems to be to attempt to meet the entire set of goals if possible. However, as suggested previously, in situations such as obstetrical emergencies, people may adopt the strategy of preserving as many goals as possible. Thus the goals of witchcraft control and high standards of female behavior may be sacrificed in order to assure "live mother" if the life of the mother appears at risk. The maximizing model, then, is intended to represent the preferred strategy, but not necessarily the only mode of 32 decision-making. One issue which requires discussion when attempting to apply rational choice theory is the treatment of norms, values and beliefs. According to Heath, the rational choice approach attempts explain what people will do giggn the existing norms and institutions (61). In developing a model of Bariba choice-making, the data indicate that it is inadequate to take values and beliefs as "given". In this study, it will be shown that contradictory beliefs and values can co—exist, and be called into play situationally. This hypothesis has been examined and supported by Janzen, among others, in his study on pluralistic medical systems in Zaire; Janzen concluded that individuals entertain separate and unique systems of knowledge and sequentially or simul- taneously act on them (Janzen 1978: 40). In addition, as has been noted, it is important to distinguish between the rationality of an individual's beliefs and the rationality of his actions. Thus a person's assumptions may be false but the actions predicated on those assumptions follow from the premises. Complementary to Heath's perspective on rational choice theory, it is useful to consider the contribution of Earth (1967), writing on the study of social change. Although Barth does not explicitly refer to rational choice theory, certain of his assumptions correspond to those cited above and in general, his theory of social organization and social change supplements the approach presented by Heath, with a particular emphasis on anthropological concepts and problems. Barth suggests that whereas anthropologists have tended to rely on a "morphological concept of custom as the minimal element of form," it is more productive to view social behavior in terms of allocations of time 33 and resources (662). The pattern of behavior should be regarded as a set of frequencies of alternatives; thus we can depict the pattern whereby people allocate time and resources. Different frequencies of allocation of time and resources entail different kinds of community life (662). With this approach, the anthropologist can arrive at a means of "isolating the underlying determinants of social forms, so as to see how changes in them generate changing social systems" (662). A question which arises, then, is the nature of the factors responsible for changing the allocations of time and resources. Although Barth does not emphasize this aspect of the model, he does propose that a reward/cost (incentive/constraint) determinant is fundamental to the process. Thus he states that "people make alloca- tions in terms of the pay-offs that they hope to obtain and their most adequate bases for predicting these pay-offs are found in their previous experience or in that of others in their community," (668) and further, that whether a change in allocation is adopted or "institutionalized" depends on the rates and kinds of pay-offs of alternatives; people may make mistakes in evaluation but learn as they judge outcomes of alternatives, through experience. The comparative rates of pay-off must be seen from the point of view of the actors, or other units of management (668). In the preceding discussion of Heath, the desirability of a focus on relative costs of alternatives was underscored. In Barth's terms, decisions regarding allocations are adjusted and adapted in terms of experiences of the benefits and losses observed from behavior outcomes (1967: 665). (Variations on this theme include Foster and Anderson 1978: 244, suggesting that people will modify pre-existing 34 practices if they perceive advantages which are not outweighed by social or economic costs; Alland 1970: 186, on trial and error of strategies of adaptation; Warren 1978 for an application of Barth's premises to an ethnomedical study in Ghana; and Erasmus 1952 as noted). In relation to the assumptions of rational choice theory pre- viously stated, Barth's propositions correspond to the postulates that individuals do have choices to make, that there are always some possible alternative courses of action and that the relative advantages of the courses of action leading to the goal will determine the pattern of behavior which predominates. Thus Barth states that "the most simple and general model available to us is one of an aggregate of people exercising choice while influenced by certain constraints and incentives" (1967: l). (Lingenfelter 1977 utilized a different perspective leading to the same conclusion). In discussing the contribution of rational choice theory to developing a model of Bariba decision-making, the problem of the treat- ment of values was remarked. In this regard, Barth also provides a useful proposition. He notes that to the structuralist, the constraints on choices have been viewed as "moral: society is a moral system," (1967: l) where "social forms were generated from values" (1967: 15). Contrastingly, Barth suggests that a feedback effect exists: ...values become progressively shared by being known through transactions; the principles of evaluation, and their uses, become public and serve as guides in the choices of others...in an on- going system, where patterns of behaviour are generated from a set of shared values, the resolution of individual dilemmas of choice by the construction of over-arching principles of evaluation will have a feed-back effect on the shared values....not only do the cumulated resultant choices produce patterns which again effect and modify choice; ...instances of transaction affect in turn both the canons and distribution of values, and in part compel the 'correction' of these values. Thus actions can have a feed—back 35 effect which makes them logically on a par, and in a certain sense developmentally prior to, values and social arrangements (1967: 15). This perspective will prove useful in considering how a set of beliefs and values, some of which may be inconsistent or contradictory, may co-exist, and their role in decision-making. The feedback effect of actions provides particular explanatory power in comprehending the gradual shift in patterns of birth practices in the town of Kouande (from preference for home delivery to reliance on the national health clinic), which appears to be accompanied by a change in predominant values. A necessary consideration in the study of decision-making is the identification of decision-makers, including the differential influence of individuals based on gender, age, and other status attributes. In this regard, Janzen's work on the choice of therapeutic alternatives in lower Zaire is suggestive in some respects. Janzen organized his study around what he referred to as the therapy managing group, primarily composed of kin in the rural areas, and additionally, of work and church associates in the urban context. According to Janzen, western and traditional medical systems are complementary in course of therapy, but irreconcilably disparate in terms of cultural logic. Each system possesses an internal rationale, linking therapeutic technique to symptoms or problem identification. These disparate systems are often utilized in conjunction with one another to terminate a case satisfactorily via referral among/between the kin therapy management group, African healers, African personnel working in the European system, and European physicians (the latter to a lesser extent and limited to referral to other western medical practitioners). 36 According to Janzen, the Bakongo have a clear concept of the focus and limits of western medicine, and delineate it as a separate therapeutic system, effective, but for a narrow range of purposes. The primary deficiency of western therapy seems to be its inability to deal with problems of wider social etiology. Choice of a therapy then becomes dependent on the therapy management group's diagnosis of the problem. According to the major Kongo diagnostic differentiation, diseases may be categorized as illnesses "of God" (in the natural order of things) or illnesses "of man" (related to human intentions) (8-9). The classification of an illness in one category or another determines appropriate therapy; thus "In Lower Zaire, most individuals with a common or simple problem such as a cold, fever, malaria, or accidental wound go forthwith to the nearest dispensary. But they may act less assuredly when they have suffered a more serious or chronic infection attributed to nonnatural causes" (127). Cosmopolitan medicine is intrinsically unable to effectively diagnose or treat social causes of illness, which can be dealt with by indigenous healers, diviners, prophets, etc. One of Janzen's most interesting suggestions is that in a decision—making process, there may be majority/minority decisions within the kin group. The majority decision may be to consult a diviner, the minority to consult a physician. If the first decision proves unsatisfactory or only partially resolves the problem, the minority decision may increase to first priority. This process will be shown to occur also with decisions regarding obstetrical care among Bariba women. 37 Two significant differences are evident between the Zairien situation and that of Bariba women. According to Janzen, the Bakongo deal with the discrepancies between medical systems by utilizing indigenous healers to treat the issues of wider social etiology of illness, while utilizing western therapy for obtaining medicines to relieve physical symptoms. (For example Janzen states that "when a disease is diagnosed as having a social or mystical cause, the thera- peutic tradition prescribes simultaneous treatment for localized symptoms and wider causesz" 191). This dichotomy has been suggested for other African societies as well (cf. Maclean 1976; Frankenberg 1976; Asuni 1979). In the case of Bariba obstetrics, however, it will be shown that it is not as feasible to utilize a maternity clinic or hospital for a delivery and simultaneously to use indigenous healers to deal with significant issues of etiology which are not perceived as considerations by western medical personnel. A second difference between the Bakongo and Bariba contexts is that the complexity of the therapy management group appears to be much greater (more formal and with greater participation) among the Bakongo than in the situation where Bariba are deciding how to deal with a delivery. Among the Bakongo, the decision-making responsibility is often assumed by matrilineal kin ranging from mother (biological or classificatory) to mother's brothers and siblings (130). Spouses do not usually have an especially significant voice in the proceedings and sufferers "retain decision-making rights only if they are adult, usually male, capable of walking and travelling, and financially able to pay for care" (130). Contrastingly, Bariba obstetrical care decisions primarily involve the patient and the women of the household 38 holding position of greatest authority (e.g., mother-in-law, father's sister, father's mother). Rarely were group discussions held either formally or informally. Analysis of the Bariba data will center around a consideration of the rewards and costs of obstetrical care alternatives to different sectors of the population. In determining the perceived costs of the alternatives, two factors will be emphasized. These are (l) homophily/ heterophily of therapists and clients in affecting decisions and (2) the cultural meaning of efficacy in affecting decisions. The homophily/ heterophily dichotomy is one which was set forth by Rogers and Shoemaker in discussing opinion leadership and its effect on the process of diffusion of innovations. In this context, homophily is defined as "the degree to which pairs of individuals who interact are similar in certain attributes such as beliefs, values, education, social status and the like," (210) whereas heterophily refers to the "degree to which pairs of individuals who interact are different in certain attributes" (210). Rogers and Shoemaker hypothesize that "change agent success is positively related to...homophily with clients" (242). Similarly, Alland (1970) suggests that similarities and differences between role systems in doner and receptor populations affect the process of change. When analogous roles exist in two different behavioral systems, change need only involve a shift in the content of existing roles. When no such analogues exist, change may require the adoption of an entirely new role or set of roles. What I am suggesting is that analogue roles act as templets for behavior and have the effect of facilitating and directing change (157). The implication of this concept is that when "change" requires the adoption of an entirely new role or set of roles, individuals will be 39 more hesitant to adopt and/or consistently utilize that alternative. This proposition formed a portion of the original research hypothesis for this study, which stated that most village women, other than wives of upwardly mobile men, would prefer to utilize the services of indigenous midwives rather than nurse-midwives. The reason for this preference was suggested to be the lack of role congruency between the position of nurse-midwife and indigenous midwife and correspondingly, the lack of fulfillment by nurse-midwives of role expectations held by rural clients. This discrepancy in attributes between client and nurse-midwife was hypothesized to parallel that between indigenous midwife and nurse—midwife. This explanation was found to be relevant in contributing to the understanding of preferences for birth assist- ance. The implications of this finding for the rational choice approach are that one primary consideration in the selection of a birth assistant is "can the midwife/healer accomplish the task at hand?" Whether a particular type of healer will be selected by clients depends on attributes of the person who could successfully achieve the desired goals; the question of successfully accomplishing healing responsi- bilities, then, incorporates expectations about the attributes of the role occupant, such as sociological and career characteristics. According to the hypothesis considered in the present research, in order to fulfill role expectations, the indigenous midwife must demon- strate attributes which fall within an acceptable range to the client population, rendering her reasonably homophilous. Similarity in attributes such as beliefs, values, past experiences with health care practices influence prospective clients in their selection among health care alternatives. Thus a client will 40 tend to prefer the practitioner who is "homophilous" and shares the above attributes and expectations regarding roles of patients and healers. Correspondingly, the absence of homophily as manifest in the relationship between most rural Bariba women and clinic nurse-midwives operates as a constraint on prospective clients evaluating the possi- bility of utilizing a maternity clinic. In subsequent chapters, an attempt will be made to show that indigenous midwives and their clients are homophilous, that the majority of prospective clients and nurse- midwives are not homophilous, and that lack of homophily renders the maternity clinic option less desirable than home delivery to the rural population. The second issue, the cultural meaning of efficacy, is signifi— cant because of the basic assumption that when choosing a health care service or assistant, people seek a type of assistance which they believe will "work" to solve the problem, however that problem is defined. Thus in evaluating a health service, Donabedian urges that the criterion of "shared definition" is of first-order significance. Because illness, and particularly 'health' are to varying degrees socially, rather than technically, defined, that organization is best that has mechanisms for the inclusion of client and social perspectives on 'health' and illness...." (1972: 105). One might infer, then, that where a shared definition of health and illness is absent, a service cannot offer health care which addresses adequately the perceived needs of the population; such a service will predictably not be a preferred choice among a majority of the client population. Correspondingly, Janzen addresses the issue of the cultural meaning of efficacy when he cites as a primary factor among variables 41 determining choice of therapist the perception of the sufferer's problem and associated expectations of the role of the therapist and techniques of therapy. He states that the diagnosis of a disorder will be linked to a therapeutic technique and therapist perceived as appropriate for the condition (222), and selection of a therapeutic system will depend on assumptions regarding an effective means of addressing the sufferer's problem, however it is identified. Similarly, in this research, the problem at hand is the Bariba definition of pregnancy and birth in contrast to that held implicit by cosmopolitan clinic personnel. Critical to the definition is the variable meaning of the concept of "work" or efficacy; the Bariba data indicate that the cultural definition of the meaning of birth in relation to both cosmological and social factors greatly influences the preference for one sort of birth assistant. This topic will be developed, following an explanation of relevant aspects of Bariba culture, with particular reference to witchcraft beliefs. In general, the present study of the factors influencing use of alternative obstetrical assistance is set in a presentation of the social context for decision—making. The discussion of social environ— ment in the Bariba region of Kouande will focus particularly on the interrelationship between religion, morality and healing, and include aspects of stratification and the role of women in Bariba society. Moreover, beliefs and practices surrounding reproductive processes will be detailed. A significant point which the presentation of these data are intended to support is that although decisions may appear to be impulsive or precipitous, especially in an obstetrical crisis, such decisions are based on previously acquired models of appropriate 42 behavior. The moment of decision-making, then, represents a point in time but more accurately reflects a process comprising a range of cultural expectations and past experiences in which context current alternatives are evaluated and selected. Conclusion: A Model of Factors Influencing Service Utilization Due to the nature of the data, it seems impossible to develop a statistical model of factors influencing decisions; the data are insufficient to determine the effect of combinations of factors or to weight variables. Instead, a descriptive model will be developed which attempts to delineate relevant factors, and to note the probability of a choice in relation to the co-occurrence of certain factors. Previously, the intent of this research was described as to develop a model of salient factors influencing the selection of obstetrical services among Bariba women. The model, as devised, conceptualizes these factors in terms of monetary and non-monetary constraints and incentives affecting the utilization of services. This model will be more thoroughly developed in the last chapter. The potential for weighing the significance of the various factors and predicting the choice of obstetrical assistance in any given case will also be discussed. The available data enable only a partial estimation of variables: thus, the co-occurrence of high monetary and non—monetary costs, as well as infrastructural factors may render utilization of one alternative so unlikely as to be impossible, whereas the congruency of low non-monetary costs, no infrastructural inhibitions and low or irrelevant monetary costs may suggest the high likelihood of utili- zation of a particular alternative. 43 Unfortunately, in the middle range, where a mix of factors is present, it is difficult to predict behavior. The problem which arises seems to be that as the comparative difference in rewards and costs narrows between alternatives, it becomes pregressively more difficult to predict which choice will be preferred; thus where there seems to be little to recommend one choice over another, individuals may seem to be choosing at random, by whimsy, or due to the greater visibility of one alternative at a critical moment. It may, then, be inherently impossible to predict behavior in this intermediate range. To some extent the limitation may be due to insufficient data on clinic utilizers-—data on Kouande clinic patronage was based on six years of clinic records and 35 interviews with parturients at the clinic. These 35 women were the consecutive patients over a two-month period of following clinic cases. Due to the slow rate of utilization and logistic difficulties involved in working in Kouande, a larger caseload could not be interviewed, nor could retrospective studies be undertaken on patients. Therefore, the background analysis of clinic utilizers which seems necessary to clearly delineate factors influencing the use of this service, and particularly to comprehend the situational variability of clinic use, is less than would be optimal. The model which will be presented, then, although it is limited with respect to power of predictability in the "intermediate range", does present an explanatory outline of the factors influencing utilization of obstetrical services, and the implications of co-occurrence of certain factors. 44 Conclusions regarding the set of factors influencing obstetrical care decisions will be set forth following discussion of various aspects of the decision-making process, to be presented as follows: 1) additional literature significant in the history of decision theory but peripheral to the perspective of this study together with literature pertinent to the cross-cultural study of midwifery; (Chapter 2) 2) the values and beliefs which comprise the assumptions upon which choices are based; (Chapter 3) 3) particular beliefs and practices surrounding reproductive processes; (Chapter 4) 4) roles and responsibilities of women; (Chapter 5) 5) sociological and career attributes of indigenous midwives and implications of role expectations for selection of birth assistance; (Chapter 6) 6) the meaning of therapy and efficacy in relation to obstetrical care preferences; (Chapter 7) 7) case studies of the use of alternatives for obstetrical assistance and descriptions of indigenous midwifery practices; (Chapter 8) 8) constraints and incentives, or rewards and costs, of available alternatives for obstetrical care; (Chapter 9) In general, I am assuming that to analyze the determinants of a choice, one must know who is involved in the decision-making, what are the alternative choices, and what is the nature of the social environment providing these alternatives. These topics will be addressed, then, in order to delineate the set of factors influencing choices for obstetrical care among Bariba women. CHAPTER II OTHER LITERATURE RELEVANT TO THE STUDY OF HEALTH CARE DECISION-MAKING Decision Models The preceding chapter included a discussion of literature pertaining to the study of decision-making which was considered particularly relevant to the development of the argument of this thesis. However, decision-making as a topic of theoretical interest in anthro- pology has been addressed from various perspectives which are less relevant to the subject of health care decision-making or to the approach taken in this study than are those studies already described in chapter 1. Illustrative examples of some of these additional orientations to the study of decision-making will be briefly described below, in order to provide a more complete historical perspective on the development of decision models. Subsequently, selected literature regarding the cross-cultural study of midwifery, both in the context of decision-making among health care alternatives and as a substantive issue, will be reviewed. One major focus in the development of decision models has been the study of decision-making in relation to social structure (for an extensive review of this literature see Quinn 1975; much of the following discussion is based on that review). In response to a recognition of the inadequacies of the normative (jural, mechanical) model of social organization (for example as manifest in early British 45 46 social anthropology), numerous writers sought to develop alternative, less deterministic models. Much of the early writing on decision-making centered on aspects of descent and residence affiliation (cf. Keesing 1967). Leach advocated attempting to represent variation within a society via a statistical model, suggesting that "social structures are sometimes best regarded as the statistical outcome of multiple individ- ual choices rather than a direct reflection of jural rules" (1960: 124). Buchler and Selby countered that: The problem is that statistics alone will not give us any notion of the underlying constraints upon behavior, nor any idea of the factors extrinsic to decision, motives, or plans in any social net— work. Thus, for the ethnographer, statistical operations, however much they may explicate the complexity of relations on the ground and provide us with measures of central tendency, do not afford any insight into the motivational or value structure of action that would enable us to understand and predict decisions of actors (1968: 92). A more productive approach which seeks to provide explanatory and predictive power has been referred to as that of decision-making, or decision models (Goodenough 1956; Howard 1963; Keesing 1967). As Quinn summarizes, ". . . while early decision models merely identified and enumerated decision criteria, later models attempted to specify how these criteria entered into actual decisions, with a view to predicting those individual decisions and the patterns of kin group affiliation which resulted from them" (1975: 42). Decision models, she concludes, provide a new and convincing way to think about social structural variability (43), and have attempted to develop the orientation in studies of social structure which focus on individual behaviour as responsive to self-interest rather than obedience to societal norms (Quinn 1975: 19). 47 "Decision-making" models and "structural" models might be dichotomized as probabilistic as opposed to deterministic models of social structure; however, it seems more accurate to postulate that both deterministic and optional considerations may operate together in a given situation. Thus sex or economic status may delimit available alternatives for an individual while further narrowing of options may be based on personal preference or evaluation. Quinn, while taking this fact into account, suggests that the terms 'decision' and 'choice' be reserved for "selection among alternatives which require some active evaluation and calculation on the part of the individual" (Quinn 1975: 41). Another variant of the decision model approach is presented by Lingenfelter in his discussion of the formation of action groups in Yap (1977). The Yap case is intended to provide methodological insights for a comparative study of emic models of culture; the approach is based on Goodenough's proposal that semantic categories and propositions furnish a fundamental starting point for cultural analysis. Lingenfelter questions to what extent specific choices are generated from and constrained by sets of symbols and norms (1977: 332). He concludes that decision-making is more individual evaluation and strategy than utilization of ready-made extant models for behavior. "The ultimate choices . . . are made within the constraints of the symbolic and normative frames but vary directly with the goals, alternatives and priorities of specific actors in particular situations" (344). 48 The position adopted in this thesis seems closest to the decision—making models described above, based on a probabilistic model of social organization and viewing decisions as including active calculation and evaluation of options by the decisionumaker(s), within culturally provided guidelines or constraints. It is significant to note, however, that although this approach has been applied to studies of social structure (whether focused on descent, affiliation, or cultural categories), research on health behavior has not generally utilized an explicit decision-making model, except for rare exceptions such as Janzen, mentioned in chapter 1. Moreover, analysis of strategy or evaluation of options by decision-makers in seeking health care has been disappointingly absent; further, insights from other subject areas such as economic anthropology have not been capitalized on adequately. This is particularly true with regard to considerations of the rational choice perspective, which has been amply discussed in the formalist-substantivist debate (cf. Firth 1967 with references to Dalton 1967, 1969; Polanyi 1957; Burling 1962; Cook 1966. See also Ortiz 1967; Prattis 1973). The absence of discussions on rationality in specific studies of health-seeking behavior and health service utilization, despite the fact that a major controversy in economic anthropology deals with the nature of ration- ality, is especially notable. It seems necessary to remark that in this present analysis, the large body of literature dealing with decision-making and risk from a psychological perspective has not been utilized because of the emphasis in psychology either on individual decision-making to the exclusion of social influences on the individual's behavior, or on cognitive 49 processes of judgment in the individual. Correspondingly, facets of decision-making such as the neurophysiology of information-processing (Simon 1971: 157) and other aspects of "programming" of information within the individual towards a goal of problem-solving are deemed to be outside the realm of this study, for reasons which follow. In some instances, this conclusion is based on an arbitrary estimation of the researcher's competence rather than on the applica- bility of the topic; for example, Simon suggests that when making decisions, people try to simplify the process by examining "small, promising regions of the entire space and simply ignoring the rest" (1971: 159). He describes the human problem solver as unwilling and incapable of enduring much trial-and-error. Similarly, Slovic and Lichtenstein generalized that "judges have a very difficult time weighting and combining information-—be it probabilistic or determin- istic in nature. To reduce cognitive strain, they resort to simplified decision strategies, many of which lead them to ignore or misuse relevant information" (1971: 115). Such conclusions as these, generated in experimental situations, would be relevant to test a field study such as this one, focusing on how individuals within a given culture select alternatives; however, due to time constraints and lack of adequate research methodology, this aspect of decision-making could not be examined. Similarly, highly mathematical approaches to decision-making have also been eschewed due to the nature of the data collected in this research project and also due to doubts as to the possibility of reducing complex decisions to equation form. In one approach, Payne suggests a model of decision-making based on "risk dimensions" which 50 he labels: 1) probability of winning (PW) 2) amount to win (SW) 3) probability of losing (PL) 4) amount to lose ($L) This approach views risky decision behavior as a form of information- processing (Payne 1972: 1). PW, PL, $W, and SL are determined to be the actual sources of information used by decision~makers (27); Payne then reviews "rules or processes used to combine information into a ' such as additive models, and nonlinear models (similar to decision,‘ satisficing, i.e., choosing the minimum level of satisfaction on all dimensions but not necessarily the 'best' choice) (Payne 1972: 27 ff.). Payne then represents a model of risky decision—making taken from Slovic and Lichtenstein (1968) in which A(G) = u + W PW + W2$W + W PL + W4$L l 3 where A(G) is attractiveness of a gamble and the W3 are weights reflecting the relative importance of each risk dimension (Payne 1972: 28). It is possible that were sufficient data available on Bariba decision-making, such a formula could be used in analyzing the decision-making process of women selecting obstetrical assistance. However, with the available information, such a procedure would (a) be overly reductionist and misleadingly simplistic and (b) be unwarranted due to the difficulty of accurately weighting decision factors or components. In sum, with regard to the exclusion of psychological and mathematical approaches to decision-making, Gluckman and Devons have aptly stated: . . . there is a duty of abstention, which requires that if we are to solve certain problems we have to abstain from studying 51 other, though apparently related, problems, and leave these to our colleagues, whether in the same or in some other discipline" (1964: 168). Midwifery Literature Alternative Utilization of Services A review of the literature discloses that discussions of mid- wifery in cross—cultural contexts are presented from several perspec- tives. These include: midwifery as an illustration in considerations of alternative utilization of health services; social characteristics of indigenous midwives; techniques and practices of indigenous midwives; and integration of traditional midwives into national health delivery systems. Examples of the first body of literature were mentioned in chapter 1. In the context of this literature, indigenous midwives are discussed in terms of comparative preferences held by a client popu- lation for alternative types of obstetrical assistance. However, most of the literature on health care services does not focus on midwives (cf. references cited in chapter 1 such as Colson 1971; Erasmus 1952; Gould 1957; Frankenburg and Leeson 1976; Janzen 1978) but rather, focuses on alternative utilization of dispensaries, clinics or hospitals and indigenous herbalists and diviners. A few sources do explicitly refer to the role of indigenous midwives, and these will be described briefly below. One relevant discussion is that of Maclean (1976) which comprises part of a more general presentation on Yoruba sickness behavior. Maclean suggests that illness produces some degree of 52 uncertainty with regard both to causation and treatment. There are always choices to be made between some treatment or none, and among alternative treatments, timing of treatments, and selection of a therapist (Maclean 1976: 291). In her research, Maclean found that people held opinions on which diseases were best treated at home or at the hospital (306); traditional treatment methods were held to be most relevant for treatment of the reproductive system--for example, for impotence, sterility, and disorders of menstruation. ...in the case of matters to do with reproduction, there is such a long time interval between biological cause and effect that all kinds of incidental and accidental events can appear to relate to the ultimate issue. Hence the very powerful continuing hold of traditional medicines, magical devices, prescriptions and pro- scriptions in this field (309-310). Nonetheless, it appears in the Yoruba situation described by Maclean that "empiricism triumphs in modern obstetrics" (312)--that the desire for healthy offspring supplants other traditional concerns such as the objections of men to losing customary decision power over their wives regarding pregnancy care. Maclean concludes that "the great success of hospitals in midwifery and pediatrics is generally recog- nized, especially among the mothers who stand to receive more immediate benefit" (314). Foster (1962) and Landy (1977) also noted the preference for modern obstetrical services in contrast to other modern services based on perceived benefits. Similarly, Frankenberg and Leeson, writing on the selection of a healer in Lusaka state that sick people in Lusaka have a wide range of Options and " . . . behaved empirically. They were determined to get the best of all worlds" (1976: 226). These conclusions are particularly interesting in contrast to the Bariba context where modern obstetrical services are not 53 perceived as necessarily more beneficial than indigenous assistance (the reasons for these differences in evaluation of the benefits to be gained from cosmopolitan health care will be examined in chapter 9). An additional study pertaining to the alternative utilization of services is Riley and Sermsri (1974), mentioned in chapter 1. Riley and Sermsri, who studied the Thai medical system as a context for birth control services, remark that in a country such as Thailand, "there are a great many different kinds of medical services, which with their social and cultural contexts, comprise a 'variegated medical system.'" They note that to a striking extent, the Thai population do not choose to utilize existing government medical services. The authors point out that traditional medicine remains an important service alternative and has in fact influenced the practice of cosmo- politan medicine in Thailand (5); thus they describe aspects of traditional Thai medicine, the social context of medical practice in Thailand, and classify types of practitioners in order to "explain the main considerations in selecting among medical service alternatives" (5). Riley and Sermsri contend that patterns of utilization of health services can be explained in terms of the characteristics of the services as well as of the clients. Among the characteristics of the services, monetary and non-monetary costs to the clients must be evaluated (1974: 58). Thus "monetary costs and the amount of time which must be expended to utilize the various medical services are obvious and important determinants of differential patronage. However, the threshholds at which these obstacles deter utilization of services . . . depend in individual cases upon socio-economic status, distance 54 to the service, and desperation. . . . Many of the less obvious deter- minants of choice could be construed as 'non-monetary costs'" (38). Differences in non-monetary costs help explain the vitality of the traditional segment in spite of the tendency to believe that modern medicines are generally more effective and they help explain why the government health stations are 'skipped' as clients patronize others of the multiple alternatives in the medical system (39). The model of the Thai medical system developed by Riley and Sermsri, which includes "factors affecting personal choices among alternative medical systems" (52) provides a useful point of departure for a similar model of use of obstetrical services by Bariba women and will be discussed at greater length in chapter 9. From a different perspective, Hanks' study of childbirth practices in a Thai village provides an ethnography of Thai childbirth, comprising a study of economic, social, religious and cosmological aspects of beliefs and practices surrounding birth. Hanks describes, for example, how the Buddhist view of the individual, and of disease causation as based on acquisition of "merit" affect the attitude and behavior of women during delivery; in addition, she discusses midwifery as a specialty and also considers attitudes of parturients towards modern hospital care. She states that hospital utilization and acceptance of changes in childbirth practices "proved to be a means of expressing partial withdrawal from the old life of a farmer and a step towards the glitter of urban ways" (2). Overall, this study contributed to the literature on alternative utilization of services by concluding that most villagers preferred local midwives; however, teachers and others aspiring to a higher status proved to be more 55 likely to consult western-trained professionals (1963). Moreover, in an obstetrical emergency, such professionals might also be con- sulted. The conclusions set forth by Hanks were instrumental in the formation of some hypotheses in the Bariba study, in particular, the hypothesis that social class would be found to be a significant factor associated with differential utilization of obstetrical services. One study which specifically discusses utilization of obstet- rical services is McClain (1975) who describes patterns of selection from available options for delivery among women in Ajijic, Mexico. According to McClain, pregnant women in Ajijic may choose from such alternatives as small, private hospitals; federally subsidized hospitals; or home deliveries assisted by indigenous parteras or by a local physician. A few women delivered without assistance of a specialist (43). There appears to be an increasing tendency for women to deliver in hospitals, but considerable variability still exists in preferences for delivery assistance. McClain summarizes that "Regardless of individual motivation, and judging by the increasing frequency of hospital use and the decreasing number of parteras available for traditional obstetrical services, childbirth at home will eventually disappear or become comparatively infrequent as an alternative to hospitalization" (45). McClain also presents personal histories of parteras in Ajijic, reports on type of delivery for 41 Ajijican mothers, and on comparative costs of obstetrical care alternatives. Unfortunately this study is very descriptive and does not examine the factors influencing women to choose a particular option for delivery, nor does it extensively correlate sociocultural context with delivery preferences. However, 56 generally, the article is a thoughtful examination of on—going obstetrical beliefs and practices in a Mexican community and many of the topics considered are also addressed in the study of Bariba birth. Characteristics of Midwives Cross-cultural studies focusing particularly on midwifery are few, and existing studies tend to have emphasized different topics and problems, rendering meaningful cross-cultural comparison of the midwife role difficult. Such comparisons, due to the nature of the data, seem limited to superficial contrasts of status characteristics, and mid- wifery techniques and practices. As Jordan noted recently, there is a general paucity of information on childbirth and much of that which exists is biased towards abnormal births, rather than the uncomplicated. According to Jordan, " . . . regarding the availability of data useful for a holistic conception of childbirth, such data are notable pri— marily for their absence" (1978: 5). Thus she reports that Newton's survey of the HRAF files found that almost 2/3 of the cultures cited contain no description of a normal birth; similarly there are no ethnographic films devoted to the study of births and it is strikingly evident that personal observation of births is rare. The deficiencies of the literature derive mainly from a lack of firsthand observation of midwives in practice and a primarily descrip- tive orientation to the topic. It would be preferable instead to utilize a more holistic approach, which might include: a thorough analysis of midwife characteristics in contrast to non-specialist women in the society; accurate tracing of selection of women for training (e.g., is midwifery a skill transmitted from mother to 57 daughter and which daughter is selected . . .). More detail on apprenticeships than is presented in existing studies would also be useful. Another topic for investigation might be the study of client— practitioner relationships--to what extent is kinship the determining factor as opposed to proximity or skill? In general, where status characteristics are the emphasis of a study, the conclusions should report more than the summary of attributes; rather, the implications of the findings should be interpreted, whether in terms of the role of the midwife as a woman and as a specialist, selection among the midwives, or selection among obstetrical care alternatives. These criticisms in mind, some of the literature on midwifery is relevant to the current study. Jordan's recent work Birth in Four Cultures is perhaps the most thorough presentation available addressing birth as both a physiological and sociocultural event. Jordan provides a detailed description of Yucatecan childbirth practices, a comparative analysis of birthing systems in the United States, Holland, Sweden and Yucatan, and seeks to specify the "ways in which birthing systems are socially organized and culturally produced" (1978: 10). She states her intention as to treat parturition in a biosocial framework, recognizing the "universal biological function and the culture-specific social matrix within which human biology is embedded" (1). This book admirably fulfills its function in suggesting the implications of adopting the medical obstetrical model without scrutiny, and in discussing the desirability for accomodation (for example in Third World countries) between indigenous and medical obstetrics; "accomodation," in this argument, includes changes in both systems to accord with the needs of the people whom the system is 58 intended to serve. Jordan also develops strongly the argument that birth, whether performed as a home delivery assisted by an indigenous midwife or in a hospital setting, presided over by a physician, is a cultural production. However, because her argument is structured as a cross-cultural comparison, its weakness lies in the lack of demonstration of birth in any particular social context. Thus, for example, emphasis on cross- cultural comparison sacrifices the development of such topics as birth in the context of Mayan culture, ideology surrounding birth, women as healers, and women's role in general Mayan society. Correspondingly, the discussion of the role of the midwife in Yucatan is limited to a description of one woman's practice, rather than a more extensive investigation of Mayan midwives in comparison to one another, which would provide data on the status of the midwife in this culture. The limitations of the study of the status of the midwife render Jordan's work less useful for comparing Mayan midwives with other studies focusing on status characteristics of midwives; however, her work remains significant for its cross-cultural comparison of obstetrical techniques and practices and its demonstration that birth is "socially organized and culturally produced." The most thorough review of the literature on status character- istics of midwives is presented in Cosminsky (1976) and will not be repeated in depth here. However, several works which have proven to be illustrative with regard to the Bariba data will be detailed below. These include data on Mayan midwives (Paul and Paul 1975); traditional birth attendants in Cebu (Rubel et a1. 1971); midwives in North Carolina (Mbngeau 1960); midwives in Kentucky (Osgood 1966); and 59 traditional birth attendants in the Southeastern State of Nigeria (Ekanem et a1. 1975). In most societies, parturient women are assisted by birth attendants of varying degrees of specialization. According to Ford (1945), elderly women were found to assist at birth in 58 cultures; in two cultures, they did not assist. In general, ethnographic data on status characteristics of midwives suggest that world-wide, indigenous midwives tend to be female, past middle-age, have been married, have had children, and learned by apprenticeship and/or divine calling (Ekanem et a1. 1975; Cosminsky 1976). 1. Midwifery as a Part-time Specialty, In all cases, midwifery was a part-time specialty and midwives did not depend on midwifery for a living. In the Nigerian case, midwives also engaged in farming and petty trading. Similarly, of the Cebuana birth attendants described by Rubel et a1., 80% claimed to work steadily but 74% reported "other sources of income" (180). The Mayan midwives described by Paul seemed to receive the most significant cash income reported, which compared favorably to income from part-time sewing or weaving, but the Mayan midwives claimed that remuneration depended on the capacity to pay and the will of the client. Mayan midwives also seemed to have a heavier work schedule than the West African midwives whose schedules were discussed, partly because the prenatal care involved more time than in West Africa, where prenatal care may not be traditionally indicated. Time inputs were not elaborated in the studies on the United States, except for caseload in Kentucky. Comparing caseloads, one finds that the Mayan midwives 60 again had the heaviest caseloads, averaging 59 cases a year. The Kentucky midwives ranged from 3 to 75 per year, (three/eight midwives assisted at less than 10 deliveries and four at 20 to 30 deliveries). Of course the caseload differs according to the popularity of the midwife and other contingencies. Bariba matrones, for example, averaged from 9 to 18 cases per year but some do much more. Ekanem, who interviewed 145 midwives, defined a midwife as a person who did at least one delivery per month, that is, 12 per year. 2. Midwife as a Type of Healer "Type of Healer" refers to whether a midwife would be cate- gorized as a healer, shaman, diviner, or other having communication with the occult; as a technician who provides medicines, massage, or other treatments; or as a birth "attendant" who does not actively intervene in the delivery process. Only Paul and Paul (1975) thoroughly discuss this issue. These Guatemalan Mayan midwives are described as having characteristics in common with those of shamans, but midwives are not a "subset" of shamans; the roles are described as separate. 3. Sex of Midwife It is often assumed that midwives are female, as in all these studies under consideration. However, as Ekanem points out, the Danfa Project in Ghana found that 48% of the birth attendants were male, and the Fon of southern P.R. Benin also use male birth attendants. Similarly, Rubel et al. report that 18% of the Cebu birth attendants who responded to questioning were male (1971: 178). Therefore being female cannot be assumed to be an obligatory characteristic of the role. 61 4. Apprenticeship Most midwives excepting Mayan midwives, claimed to have learned by apprenticeship. Mongeau found that the apprentice was often a daughter or granddaughter, chosen by the midwife as her replacement. The apprentice usually began by doing unglamorous, menial tasks; she saw her first delivery after delivering her own first child, and took over at the retirement of her mentor. Osgood found that the midwives he interviewed showed no consistent pattern of training; rather, one group was trained by doctors, one by friends, and one was self-trained. The Nigerian midwives studied by Ekanem tended to have been apprenticed to mothers or grandmothers, or to have received a divine call. Contrastingly, Mayan midwives in Guatemala succeeded to their positions via a divine mandate and denied serving an apprenticeship, although genealogical studies demonstrated that most had a mother or grandmother who had been a midwife or a father who had been a shaman. Mayan mid- wives usually suffer a lingering illness before beginning to practice as midwives; this illness is significant of the "call" to practice, which if denied, may lead to death. In general, Mayan midwives demonstrate the most ritualized role of the midwives under consideration (Paul 1975). 5. Personal Attributes This topic was most thoroughly discussed by Paul with references to Mayan midwives. In discussing the Mayan midwife's initiation to the practice of midwifery, Paul describes a process in which the attributes expected of a woman, such as "verguenza" (modesty, shame, bashfulness... L. Paul 1975: 452) must be put aside in order to attain the role of 62 midwife, which demands "that a woman overcome timidity and fear, and be assertive and commanding in unfamiliar situations, that she trespass the boundaries of space and time that ordinary women respect, that she display unusual fortitude and resourcefulness. She must withstand the terrors of night and contact with the supernatural. She must overcome her fear and disgust at the sight and handling of the newborn baby, of polluting blood and other birth substances, as well as her fear that the body organs will become disordered" (452). 6.‘Agg Midwifery has been assumed to be a service provided by meno— pausal and often elderly women. Although this is true in many instances, some exceptions are notable. The mean age of those midwives studied in Nigeria was 57 compared to age 63 for birth attendants in Ghana. Mbngeau found 90% of the North Carolina midwives to be over 40, and clientele increased with age; Osgood found the average age of his interview subjects to be 70 but remarked that the fact that the midwives were old was incidental to the study because a number were not "grannies" when they began practicing. He suggests that the term "granny midwife" is not appropriate due to the probable number of younger midwives in practice (see Rubel 1966 for other data on Mexican— American parteras vis-a-vis age, calling, etc). Mayan midwives tended to be the youngest of those studied; aged 35-40, they were in their prime as wives and mothers. Overall though, it seems that midwives may begin as apprentices when young women (although some do not even apprentice until they have had several children) but most do not conduct unsupervised deliveries until menopausal and respect seems to increase with age. 63 7. Marital Status Regarding marital status and reproductive histories, substantial data is lacking but Osgood's commentary on Kentucky midwives fits most of the cases under consideration here in that "in marital status, family background and household situations, the local midwives fit in fairly well with the general sociocultural patterns for their age-groups within the area" (Osgood 1966: 767). Ekanem did not report on marital status but found that 89% of the Nigerian midwives had delivered living children, with the distribution as follows: 1-4 children 3 5-9 4 10+ 2 I-‘J-‘b \INI—I o\° MDngeau also notes that the midwives in her sample had had vast experience in childbirth; only 23% had fewer than seven children and 33% had more than eleven children but many midwives had experienced numerous infant or child deaths. Of the Kentucky study of midwives, all but one had four or more children and in Cebu, 41% of the birth attendants surveyed had six or more children (Rubel et a1. 1971: 179). These studies, then, support the premise that midwives commonly are expected to be multiparous women, in order to be perceived as appropriate birth attendants. The preceding discussion was intended to signal some salient features of midwives as noted in several of the more detailed and thorough studies of indigenous midwifery (see Table 2). These status characteristics will be examined and contrasted in chapter 6 with reference to the characteristics of Bariba midwives, in order to determine whether there is a range of social characteristics which are typical of a Bariba midwife, and how these attributes relate to role 64 x x x +m «a x x mauowfiz .m.m +< x x x x .HOnma x moflum> hxonudox x x x +n mm x x mnaaoumo .z x w x +0 He x x w x mumonou x x x w w x x coho: oz mo» moan * N coauume mofiuuos .ocoa .oaoa z huHMHoomm twoommlaoz\amem£m Ho>m lumom loam mauuluumm moowuaouae< Hoamom mo ooze :ouvaanu maumum Homans: ow< xom Aououmuoufia vouuoamm aoumv muamvsouu< suuam maoaowflouH mo mouaumom acowammll.~ mqm