Male bias in health care utilization for under-fives in a rural community in western India.

Male bias in health care utilization for under- fives in a rural community in western India B. Ganatra1 & S. Hirve1 A cross-sectional survey, in 1991, of 3100 families in 45 contiguous villages in the Pune district of Maharashtra state showed that 456 under-5-year-olds had suffered an acute respiratory infection and/or diarrhoea during the previous 7 days. Significantly more boys (88.9%) than girls (76.5%) were treated by a registered private medical practitioner (odds ratio (OR) = 2.51). Referrals for further treatment were followed by parents significantly more often in the case of their sons (69.2%) than daughters (25%) (OR = 6.75). An average of Rs 35 (US$ 1.16) was spent on the treatment of a son, compared with Rs 23 (US$ 0.76) for a daughter. In general, parents were willing to travel a greater distance (>2 km) to seek medical treatment for their sons. These differences persisted even after adjusting for severity of illness, parent's income, occupation and education, and the birth order of the child. Intervention programmes directed at under-fives would need to correct the bias against girls if equitable access to health care is to be achieved. Male-female differentials in mortality patterns among under-5-year-olds in developing countries have been well documented, the majority of countries showing how any biological advantage of females is overrid- den by societal norms so that female mortality far exceeds male mortality after the perinatal and early neonatal periods (1-9). This difference in mortality is the clearest sign of overt and covert discriminatory behavioural practices which favour the treatment of sons over daughters. Several studies have suggested that family allocations of resources including food, money and maternal attention tend to deprive female children of health care (2-4, 7, 10-13). This study investigates the extent of sex bias in the utilization of health care facilities for children with common illnesses in a rural area of Maharashtra state. Materials and methods A cross-sectional survey of 45 contiguous villages (population, 120 000) in the Pune district of Maha- rashtra state was carried out from August to Novem- ber 1991. A total of 3100 families with children under 5 years of age were interviewed by medico-social workers. All under-five children with a history (non- fatal of acute respiratory infection (ARI) and/or diar- rhoea during the previous 7 days were included in the study. The duration and severity of the most I Research Scientists, K.E.M. Hospital Research Centre, Sardar Mudliar Road, Rasta Peth, Pune 411 011, India. Requests for reprints should be sent to this address. Reprint No. 5457 recent illness (ARI or diarrhoea) were recorded. Severity was graded, based on a simple classification of symptoms that were most likely to be remembered by the families. Severity of diarrhoea took into account the frequency of loose motions, associated vomiting, and the presence of dehydration (indicated by a depressed fontanelle, sunken eyes, and decrea- sed urine output). Severity of ARI was graded on the basis of fever and a history of breathing difficulty (rapid breathing and stridor). Details on health care utilization variables (e.g., whether medical treatment was sought, from whom, and on what day of the ill- ness ; the type of treatment taken and whether refer- ral was advised and accepted; the distance travelled to seek medical advice; and the amount of money spent on the illness and transport) were collected by a questionnaire. Data were also collected on factors that could potentially influence health care utilization like birth order, socioeconomic status, and parents' education. The sex of the child was recorded but the medicosocial workers who collected the data were not told the aims of the study in order to minimize interviewer bias. All variables were analysed, by sex, and differ- ences in proportions were tested using Student's t- test. Logistic regression analysis helped indicate whether sex remained a significant factor influencing treatment variables after controlling for type and severity of illness and other confounding factors. Results Among the 3100 families (with under-fives) inter- viewed, a total of 456 children had had an episode of ARI or diarrhoea within 7 days, giving a prevalence Bulletin of the World Health Organization, 1994, 72 (1): 101-104 © World Health Organization 1994 101 B. Ganatra & S. Hirve of 9.4% for ARI and 5.3% for diarrhoea. There was no significant difference in the prevalences between boys and girls. Two fatal episodes involving girls who died as a consequence of ARI were excluded from the analysis. Treatment was sought by the families in 77.8% of episodes of illness, with no apparent sex bias. Medical advice, in the case of both sons and daugh- ters, was on average sought on the second day of ill- ness, while treatment was actually started on average on the fifth day after onset of the illness. Further analysis of children for whom treatment was sought showed a significantly higher proportion of girls (18.8%, compared to 6.7% for boys) who were taken for treatment to a paramedical worker (auxiliary nurse-midwife or multipurpose health worker). On the other hand, a significantly higher proportion of boys (88.9%, compared to 76.5% for girls) were treated by a registered private practitioner (Table 1). This difference persisted even after adjust- ing for severity of the illness. A higher proportion of boys were advised refer- ral (7.3%), compared with girls (4.6%); this differ- ence was not statistically significant. But the parents took referral action significantly more in the case of sons (69.2%) than daughters (25%). Families spent more money on the treatment of sons (average, Rs 35 (US$ 1.16)) than daughters (average, Rs 23 (US$ 0.76)). This difference remain- ed statistically significant even after controlling for the type and severity of the illness. In 11% of the female children and in 3.3% of male children, only the free health care facilities and services were uti- lized and no money was spent on medicine, transport or consultations. This difference was significant. Boys were also favoured when one had to travel a greater distance to seek medical treatment. Thus, for a significantly higher proportion of girls medical advice was sought only within a 2-km radius (i.e., in the village itself); the proportion was higher for boys when the distance to travel was more than 2 km. Logistic regression analysis (Table 2) showed that after controlling for severity and type of illness and other factors like parents' education, occupation and income, the chance of seeking medical advice from a private practitioner was 2.5 times greater for boys than for girls. Sons were also favoured, com- pared with daughters, by parents who were twice as likely to travel outside the village for medical care and about 3.7 times as likely to spend more for treat- ment. If referral for further treatment was advised it was 6.8 times more likely to be availed of in the case of a son, compared to a daughter. There was no statistically significant sex bias in factors like resorting to home remedies, the type of treatment (oral medications, injections, etc.), and Table 1: Proportions of males and females for selected health care utilization factors Proportions a (%)p Health care factors Boys Girls Difference value Treated by private doctor 88.9 76.5 12.4 <0.05 Treated by paramedic 6.7 18.8 -12.1 <0.05 Referral advised 7.3 4.6 2.7 N.S.b Referral availed 69.2 25.0 44.2 <0.05 Distance travelled: <2 km 30.0 45.1 -15.1 <0.05 2-5 km 38.3 32.4 5.9 N.S. >5 km 31.6 22.5 9.1 N.S. Expense: Nil 3.3 11.0 - 7.7 <0.05 Rs 1 to Rs 10 14.4 33.7 -19.3 <0.05 Rs 11 to Rs 50 62.8 49.4 13.4 <0.05 Rs 51 to Rs 100 13.9 5.2 8.7 <0.05 >Rs1OO 5.5 0.6 4.9 N.S. a Percentages are those children for whom treatment was sought. b N.S. = not significant. whether this was taken fully or partly. Wherever sig- nificant differences were noted, they persisted after adjusting for parents' education, income and occupa- tion, the birth order of the child, and the number of living brothers and sisters. Discussion Since data collection involved a 7-day recall a cer- tain degree of recall bias was inevitable. The inter- viewers were not told the purpose of the study, so this source of error was minimized. Only the more common childhood illnesses (ARI and diarrhoea) were considered in this study, and it is not possible to say how gender bias would manifest in the case of more severe childhood illnesses. The under-five mortality rate for the study area is 70 per 1000 live births, diarrhoea with dehydration accounting for 20% of all deaths and ARI for 12%. The proportion of under-five mortality is higher for females (54%, compared with 46% in males), which is statistically significant (unpublished data from Table 2: Adjusted odds ratios for risk factors in health care utilization, showing extent of male bias Risk factor Male:female odds ratio Treated by private doctor 2.5 (1.3, 4.7)a Distance travelled 2.0 (1.2, 3.2) (>2 km cf. <2 km) Expense (>Rs 50 cf. <Rs 50) 3.7 (1.6, 8.6) Referral availed 6.8 (0.9, 49.2) a Figures in parentheses are the 95% confidence limits. WHO Bulletin OMS. Vol 72 1994102 Male child bias In health care utilization in India WHO's 'Study on low birth weight and infant mor- bidity and mortality' and the ongoing WHO 'Rural cohort study on child survival'). However, it was outside the scope of this study to determine the degree to which sex bias in health care utilization is linked to these differences in mortality. The preference for a male child is near universal and is seen to varying degrees even in the devel- oped world (14, 15), but is most manifest in societies which are male-oriented (1, 3, 5, 6, 16-18). Differen- tials in utilization of health care is one way in which this preference manifests. Several studies in Punjab (3, 7, 12, 19-21) have shown differences in health care utilization for girls, medical attention being sought much less frequently and only at a much later stage of illness. The patterns of discrimination found in these earlier studies were more overt than in our study, and this may be due to secular changes with time. It is also in keeping with the observation that the preference for sons becomes less marked from northern to southern India as the economy changes from dry land to wet land cultivation and the produc- tive values of daughters increases (1).a Although our study area has a well developed primary health care infrastructure with regular con- tacts between the health care providers and the people, the higher value placed by the community on private physicians must account for the significantly higher proportion of boys being treated privately. Rahaman (11) found decreases in attendance by females at a diarrhoea treatment clinic as distance from the centre increased. In our study, parents took their sons over greater distances to get treatment or for referral, compared with taking their daughters. Birth order and the number of living brothers or sis- ters were not found to influence the money spent on treatment or the distance travelled for treatment, unlike reports from other studies (3). Our findings suggest that the preference for sons and other local customs may directly influence the health status of females, including the utilization of health services and child survival technologies on behalf of girl children. Where fees are charged or referrals to a higher level of care are involved, fami- lies are likely to deny these more often in the case of daughters. Therefore, intervention programmes directed at under-fives would need to correct the bias against girls if equitable access to health care is to be achieved. a Leslie J et al. Weathering economic crisis - the crucial role of women in health. Paper presented at the Second Takemi Symposium in International Health, Harvard School of Public Health, 1986. Acknowledgements We are very grateful to Dr Banoo Coyaji and Dr V.N. Rao, of the K.E.M. Hospital Research Centre for technical gui- dance and support. This study is a result of the ongoing Rural Cohort Study on Child survival, which is funded by the WHO Regional Office for South-East Asia. Resume Biais en faveur du sexe masculin dans l'utilisation des soins de sante chez les moins de cinq ans dans une communaute rurale de l'ouest de l'Inde Une enquete transversale realisee en 1991 sur 3100 families dans 45 villages contigus du district de Poona dans l'Etat de Maharashtra a montre que 456 enfants de moins de cinq ans avaient souffert d'une infection respiratoire aigue (IRA) eVou de diarrhee au cours des sept jours pr6ce- dents. Seuls les episodes non mortels ont ete exa- min6s. La gravit6 des cas a et6 notee au moyen d'une echelle simple bas6e sur les faits retenus par la famille. La prevalence des IRA et de la diar- rh6e 6tait analogue entre les deux sexes. Dans 77,8% des cas, l'enfant a ete vu par un professionnel de sante, en moyenne deux jours apres le d6but des sympt6mes. Aucun biais li6 au sexe n'etait alors apparent. Une proportion signifi- cativement plus grande de garcons (88,9%) que de filles (76,5%) ont ete traites par des praticiens priv6s, ce qui semble compatible avec l'estime dans laquelle la communaut6 tient les praticiens prives, dont elle pense qu'ils fournissent de meil- leurs soins. En moyenne, 35 Rs (US$ 1,16) ont ete d6- pens6s pour le traitement d'un fils, contre 23 Rs (US$ 0,76) pour une fille. Une plus forte propor- tion de filles ont ete trait6es dans le village, tandis que les gar,cons 6taient plus souvent emmen6s dans un centre de soins plus 6loign6 (OR = 2,0). Les families acceptaient plus souvent que 1'enfant soit adress6 a un service specialis6 s'il s'agissait d'un garcon (OR = 6,8). Partout ou des differences significatives ont 6te notees, elles ont persist6 apres correction de la gravite de la maladie, du niveau d'etudes des parents, de leur revenu et de leur profession, et du rang de naissance de l'enfant. Bien que la zone etudiee ait une infrastructure de soins de sant6 primaires bien d6velopp6e, avec des contacts reguliers entre la communaute et les agents de soins de sant6, il semble que la preference pour les gargons se manifeste dans l'utilisation des services de sant6 et des soins de WHO Bulletin OMS. Vol 72 1994 103 B. Ganatra & S. Hirve survie. Lorsque les soins sont payants ou qu'il est question d'envoyer 1'enfant a un service speciali- se, les familles en refuseront le coOt plus souvent s'il s'agit de filles. Par cons6quent, tout program- me d'intervention ax6 sur les moins de cinq ans devra tenir compte de ce biais en faveur des gar- gons, afin d'assurer un acces 6quitable aux ser- vices de soins. References 1. MacCormack CP Health and the social power of women. Social science medicine, 1988, 26: 677-683. 2. Koenig MA, D'Souza S. Sex differences in child- hood mortality in rural Bangladesh. Social science medicine, 1986, 22: 15-22. 3. Das Gupta M. Selective discrimination against female children in rural Punjab, India. Population and development review, 1987, 13: 77-100. 4. D'Souza S, Chen L. Sex differentials in mortality in rural Bangladesh. Population and development review, 1980, 6: 257-270. 5. Levine NE. Differential child care in three Tibetan communities: beyond son preference. Population and development review, 1987, 13: 281-304. 6. Lopez AD. Sex differentials in mortality. WHO Chronicle, 1984, 38: 217-224. 7. Wyon J, Gordon J. The Khanna study: population problems in rural Punjab. Cambridge, MA, Harvard University Press, 1971. 8. Chen L et al. Sex bias in the family allocation of food and health care in rural Bangladesh. Popula- tion and development review, 1981, 7: 55. 9. Kumar G. Gender, differential mortality and devel- opment: the experience of Kerala. Cambridge jour- nal of economics, 1989, 13: 517-539. 10. Basu AM. Is discrimination in food really necessary for explaining sex differentials in childhood mortal- ity? Population studies, 1989, 43: 193-210. 11. Rahaman M. et al. A diarrhea clinic in rural Bangla- desh: influence of distance, age and sex on atten- dance and diarrhoeal mortality. American journal of public health, 1982, 72: 1124-1128. 12. Singh S. et al. Medical care in fatal illness of a rural Punjab population: some social, biological and cultural factors and their ecological implications. Indian journal of medical research, 1962, 50: 865-880. 13. Singh S et al. Immunization practices among chil- dren in an urban area. Indian journal of community medicine, 1986, 11: 94-97. 14. Clealand J et al. Preferences for the sex of children and their influence on reproductive behavior. In: World fertility survey (Comp. series No. 27). Voor- burg, International Statistical Institute, 1983: 10-14. 15. Williamson NE. Boys or girls? Parents' preference and sex control. Population bulletin, 1984, 33: 1-35. 16. Ramanamma A, Bambawale U. The mania for sons: an analysis of social values in South Asia. Social science medicine, 1980, 14: 107-1 10. 17. Dyson T, Moore M. On kinship structure, female autonomy and demographic behavior in India. Popu- lation and development review, 1983, 9: 35-60. 18. Jones M. How daughters suffer. People, 1987, 14: 30-31. 19. Bhatia S. Status and survival. World health, April 1985: 12-14. 20. Freed RS, Freed SA. Beliefs and practices resulting in female deaths and fewer females than males in India. Population environment, 1989, 10: 144-159. 21. Kielmann A et al. Child and maternal health ser- vices in rural India - the Narangwal experiment. Vol. 1. Integrated nutrition and health care (World Bank publication). Baltimore, The Johns Hopkins University Press, 1983. 104 WHO Bulletin OMS. Vol 72 1994