Bolivia - Demographic and Health Survey -1991

Publication date: 1991

Bolivia Maternal and Child Health in Bolivia Report on the In-depth DHS Survey in Bolivia 1989 ®DHS Demographic and Health Surveys Institute for Resource Development/Macro Systems, Inc. Maternal and Child Health in Bolivia: Report on the In-depth DHS Survey in Bolivia 1989 A. Elisabeth Sommerfelt J. Ties Boerma Luis H. Ochoa Shea O. Rutstein Institute for Resource Development/Macro Systems, Inc. Columbia, Maryland USA April 1991 This report presents the results of an analysis of health data from the Bolivia In-depth Survey, Encuesta Nacional de Demografia y Salud 1989 (ENDSA), which was implemented in 1989 by the Instituto National de Estadistica (INE). The survey is part of the worldwide Demographic and Health Surveys (DHS) program which is administered by the Institute for Resource Development (IRD) under a contract with the United States Agency for International Development (Contract No. DPE-3023-C-00-4083-00). The Instituto Nacional de Estadlstica (INE) is part of the Ministry of Planning and Coordination and functions as the administrative and technological arm of the National System of Statistical Information (Sistema Nacional de Informaci6n). The INE has the responsibility of directing, planning, coordinating, and carrying out the statistical activities of the System. Additional information about ENDSA can be obtained from the Instituto Nacional dc Estadfstica, Casilla 20532, La Paz, Bolivia. Additional information about the DHS Program can be obtained from IRD/Macro Systems, 8850 Stanford Boulevard, Suite 4000, Columbia, MD 21045, USA (Telephone 301-290-2800; Telex 87775; Fax 301-290-2999). CONTENTS Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi Map . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xii 1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.1 Geography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.2 Population . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.3 Programs and Priorities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.4 Bolivia Demographic and Health Survey . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 1.5 Background Characteristics of Survey Respondents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 2 Childhood Mortality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 2.1 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 2.2 Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 3 Causes of Death in Childhood . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 3.1 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 3.2 Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 4 Morbidity: Diarrhea and its Treatment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 4.1 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 4.2 Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 5 Morbidity: Respiratory Infections and theirTreatment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 5.1 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 5.2 Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 6 Nutritional Status . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 6.1 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 6.2 Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 7 Infant Feeding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 7.1 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 7.2 Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 8 Childhood Immunizations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 8.1 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 8.2 Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 iii 9 Prenatal Care and Delivery Assistance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 9.1 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 9.2 Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 10 Maternal Mortality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 10.1 Other Sources of Data on Maternal Mortality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 10.2 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 10.3 Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 Appendix A Accuracy of Diarrhea Recall . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91 A.I Heaping of Responses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91 A.2 Underreporting of Diarrhea . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92 Appendix B Survey Staff . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 Appendix C Questionnaires . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101 iv Table 1,1 Table 2,1 Table 2,2 Table 2,3 Table 3,1 Table 3.2 Table 3.3 Table 3.4 Table 3.5 Table 4.1 Table 4.2 Table 4.3 Table 4.4 Table 4.5 Table 4.6 Table 5.1 Table 5.2 Table 5.3 Table 5.4 Table 5.5 Table 6.1 TABLES Educational level of women . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Childhood mortality by calendar period . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 Childhood mortality by socioeconomic characteristics . . . . . . . . . . . . . . . . . . . . . . 8 Childhood mortality by demographic characteristics . . . . . . . . . . . . . . . . . . . . . . . 10 Causes of death in childhood . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 Symptoms associated with the illness that led to child's death . . . . . . . . . . . . . . . . 13 Causes of death derived from reported symptoms . . . . . . . . . . . . . . . . . . . . . . . . . 15 Causes of death from mothers' reports and derived from reported symptoms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 Size at birth and infant mortality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 Diarrhea prevalence by socioeconomic characteristics . . . . . . . . . . . . . . . . . . . . . 21 Diarrhea prevalence by demographic characteristics . . . . . . . . . . . . . . . . . . . . . . . 22 Diarrhea prevalence by feeding pattems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 Treatment practices for children with diarrhea by socioeconomic characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 Treatment practices for children with diarrhea by demographic characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 Treatment of children with diarrhea by type of health facility visited . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 Prevalence of cough and cough with breathing problems by socioeconomic characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 Prevalence of cough and cough with breathing problems by demographic characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 Place of treatment and type of treatment for children with cough and breathing problems by socioeconomic characteristics . . . . . . . . . . . . . . . . . . . 35 Place of treatment and type of treatment for children with cough and breathing problems by demographic characteristics . . . . . . . . . . . . . . . . . . . . 36 Type of treatment received for respiratory illness in children according to the place where treatment was sought . . . . . . . . . . . . . . . . . . . . . . . . 39 Prevalence of undernutrition by socioeconomic characteristics . . . . . . . . . . . . . . . 43 Table 6.2 Table 6.3 Table 7.1 Table 7.2 Table 7.3 Table 7.4 Table 7.5 Table 7.6 Table 8.1 Table 8.2 Table 8.3 Table 8.4 Table 9.1 Table 9.2 Table 9.3 Table 9.4 Table 9.5 Table 9.6 Table 9.7 Table 10.1 Table 10.2 Table 10.3 Prevalence of undemutrition in 1981 and 1989 by residence . . . . . . . . . . . . . . . . . 45 Prevalence of undernutrition by demographic characteristics . . . . . . . . . . . . . . . . 46 Initiation and duration of breastfeeding by socioeconomic characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 Initiation and duration of breastfeeding by demographic characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 Reasons for stopping breastfeeding by survival status of child . . . . . . . . . . . . . . . 53 Feeding practices for children under three years . . . . . . . . . . . . . . . . . . . . . . . . . . 55 Feeding practices for children under three years who were still breastfeeding at the time of the survey . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 Frequency of breast milk supplements for children under five . . . . . . . . . . . . . . . 57 Immunization coverage by source of information . . . . . . . . . . . . . . . . . . . . . . . . . 60 Immunization coverage by age 12 months for children 12-59 months of age . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 Differentials in immunization coverage by socioeconomic characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64 Differentials in immunization coverage by demographic characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 Prenatal care and tetanus toxoid immunization among currently pregnant women . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68 Prenatal care and tetanus toxoid immunization by socioeconomic characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 Prenatal care and tetanus toxoid immunization by demographic characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70 Place of delivery for births by socioeconomic characteristics . . . . . . . . . . . . . . . . 71 Place of delivery for births by demographic characteristics . . . . . . . . . . . . . . . . . . 73 Maternity care and use of contraception by socioeconomic characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75 Prevalence of caesarean section by socioeconomic characteristics . . . . . . . . . . . . 77 lndirect estimates of matemal mortality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81 Direct estimates of maternal mortality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82 Relative risk of maternal death by age and parity . . . . . . . . . . . . . . . . . . . . . . . . . 83 vi Table A- 1 Table A-2 Length of time since the last episode of diarrhea . . . . . . . . . . . . . . . . . . . . . . . . . . 91 Incidence and prevalence of diarrhea in the two weeks preceding the survey . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93 vii Figure 2.1 Figure 2.2 Figure 2.3 Figure 3.1 Figure 3.2 Figure 4.1 Figure 4.2 Figure 4.3 Figure 4.4 Figure 4.5 Figure4.6 Figure4.7 Figure5.l Figum5.2 Figure5.3 Figure 5.4 Figure 5.5 Figure 5.6 Figure 6.1 Figure 6.2 FIGURES Trends in under-five mortality based on direct and indirect estimates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 Under-five mortality 1979-1988 by socioeconomic characteristics . . . . . . . . . . . . . 9 Under-five mortality 1979-1988 by demographic characteristics . . . . . . . . . . . . . 10 Causes of death based on mother's reports and symptoms . . . . . . . . . . . . . . . . . . . 16 Medical assistance sought before child's death, and place of death . . . . . . . . . . . . 17 Diarrhea prevalence in the two weeks preceding the surveys . . . . . . . . . . . . . . . . 23 Feeding and treatment patterns for children with diarrhea in the two weeks preceding the survey . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 Changes in the amount of liquids and solids given to children with diarrhea in the two weeks preceding the surveys . . . . . . . . . . . . . . . . . . . . . . . . . . 25 Reasons for not taking children with diarrhea to a health facility . . . . . . . . . . . . . 28 Knowledge and use of ORS packets among womcn with children under five . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 Use of ORS packets and cost of packets among women who know about ORS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 Sources of ORS packets among women who know about ORS . . . . . . . . . . . . . . 30 Prevalence of cough with breathing problems among children under five . . . . . . 34 Health care received by children with cough by type of cough . . . . . . . . . . . . . . . 37 Health care received by children with cough and breathing problems by characteristics of mother . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 Treatment received by children with cough and breathing problems by characteristics of mother . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 Specific treatment received by children with cough by type of cough . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 Specific treatment received by children with cough by type of cough and type of facility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 Nutritional status of children 3-36 months . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 Percentage of children 3-36 months classified as undernourished by age . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 ix Figure 6.3 Figure 7.1 Figure 7.2 Figure 7.3 Figure 7.4 Figure 7.5 Figure 8.1 Figure 8.2 Figure 8.3 Figure 8.4 Figure 9.1 Figure 9.2 Figure 9.3 Figure 9,4 Figure A.I Figure A.2 Percentage of children 3-36 months classified as stunted by age and mother's education . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47 Reasons for never breastfeeding or for stopping breastfeeding . . . . . . . . . . . . . . . 52 Reasons for stopping breastfeeding by survival status of children . . . . . . . . . . . . . 53 Frequency of breastfeeding over a 24-hour period by mother's education and length of prior birth interval . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 Percentage of children still breastfeeding by age of child and type of supplement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 Feeding practices among young children by current age . . . . . . . . . . . . . . . . . . . . 56 Immunization coverage for children 12-23 months by source of data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61 Trends in immunization coverage among children immunized by 12 months . . . . 63 DPT3 coverage differentials for children 12-23 months . . . . . . . . . . . . . . . . . . . . 65 Knowledge of immunization sources among women with children under five . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66 Type of prenatal care received by women . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68 Place of delivery and person attending the delivery . . . . . . . . . . . . . . . . . . . . . . . . 72 Mother/child contact the first day after delivery by type of health facility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 Maternity care and ever use of a modem family planning method . . . . . . . . . . . . 75 Daily prevalence of diarrhea among children under five by number of days preceding the survey . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93 Daily incidence of diarrhea among children under five by number of days preceding the survey . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94 X Preface The Bolivia In-depth Survey, Encuesta Nacional de Demografia y Salud 1989 (ENDSA), was implemented by the Instituto Nacional de Estadfstica (INE) with technical and financial assistance from the Institute for Resource Development (IRD). The survey is part of the worldwide Demographic and Health Surveys (DHS) program supported by the United States Agency for International Development. UNICEF/Bolivia provided substantial financial support to cover local costs, and OMSS/OPS and UNFPA (through PRONIMA III) also supported the survey. A nationally representative sample of 7923 women age 15-49 were interviewed in the ENDSA between February and July 1989. Data on women and their children under five were collected using the standard DHS questionnaire with additional health modules. The survey report published by INE and IRD in January 1990 presented the main findings on fertility, family planning, and maternal and child health. This report is an analysis of the health data from the in-depth survey. Successful implementation of the ENDSA required the effort and cooperation of a large number of people. The names of these persons are listed in Appendix B. Special thanks go to the current Executive Director of INE, Ing. Waldo Cerruto, and the former Executive Director, Lic. Marcelo Mercado. Thanks are also extended to Paul Hartenberger and his staff at the USAID Mission in Bolivia, without whose help the survey could not have been carried out. xi Bolivia • COBIJA PERU PANDO BRASIL BENI TRINIDAD • LA PAZ LA PAZ COCHABAMEiA SANTA CRUZ ORURO ORURO POTOSl • • SUCRE • SANTA CRUZ CHILE POTOSI CHUQUISACA TARIJA • TARIJA ARGENTINA PARAGUAY N xii CHAPTER 1 BACKGROUND 1.1 Geography Bolivia is located in central South America and is bordered on the north and east by Brazil, on the south by Paraguay and Argentina, and on the west by Peru and Chile. There are three major ecological regions in its 1,098,591 square kilometer area: the Altiplano, or high plateau, in the west (16 percent); the VaUe, or valleys, in the central region, (19 percent); and, the Llanos, or plains, in the north and east (65 percent). The climate differs greatly in each region, with the result that a diversity of crops are grown throughout the country. In the Altiplano the principal crops are potatoes, oats, beam, beets, and barley. In the Valle, com is the major crop, but wheat, oats, and other cereals are also grown, along with a variety of fruits and vegetables. In the sub-Andean part of the Llanos the most important crop is coca; citrus and other semi-tropical fruits (bananas, papaya, custard apples) are also grown. In the rest of the Llanos, the predominant crops are yucca, com, peanuts, cotton, soya, sugar cane, and tobacco. Bolivia is politically and administratively divided into nine departments: Beni, Chuquisaca, Cochabamba, La Paz, Oruro, Pando, Potosf, Santa Cruz, and Tarija. 1.2 Population The last national census, which took place in 1976, showed a population of 4.9 million. The estimated population in 1988, according to data from the National Population and Housing Survey, was 6.4 million (lnstituto Nacional de Estadfstica, 1989). The population is concentrated in the Altiplano and Valle regions (49 and 28 percent, respectively). Although far more extensive in size, the Llanos region includes only 23 percent of the population. According to the 1988 survey, 49 percent of the population was in settlements of less than 2,000 inhabitants; 32 percent were in cities of 200,000 or more inhabitants. In addition to ecological diversity, Bolivia is characterized by ethnic and linguistic diversity. A large segment of the population continues to speak the indigenous pre-Columbian languages and retains much of the traditional Indian culture. The official language is Spanish and, according to data from the present survey, more than three-quarters of women age 15-49 commonly speak Spanish. However, a significant number of persons use Indian languages (Aymara or Quechua) regularly, especially in the Altiplano and Valle regions. In the Altiplano, 19 percent speak Aymara, and 11 percent speak Quechua. 1.3 Programs and Priorities The strategy of the government regarding maternal and infant health, as expressed in the current National Plan for Child Survival and Maternal Health, focuses on three major areas: social management, primary health care, and the development of local health systems (Sistemas Locales de Salud~SILOS) (Ministerio de Previsi6n Social y Salud Publica, 1989). In this context, integrated care for eligible women (including reproductive health services, assistance at delivery, prenatal, postnatal and newborn care, and encouragement of breasffeeding) has the highest priority. Likewise, integrated care for children under five gives priority to the promotion of breastfeeding, detection and treatment of nutritional deficiencies and control of diarrheal and respiratory illnesses. The immunization program has also received increasing attention, and in recent years fifteen national immunization campaigns have been conducted (Rance et al., 1989). This report is an analysis of these topics. An official population policy has not yet been adopted in Bolivia; however, the National Population Council (CONAPO) has presented guidelines for the formulation of such policies (Consejo Nacional de Poblaci6n, 1988). The fight of couples and individuals to freely decide the number and spacing of children is recognized, and CONAPO proposes the introduction of family planning--more for health than demographic reasons--to women for whom a pregnancy could represent a health risk. 1.4 Bolivia Demographic and Health Survey The Bolivia Demographic and Health Survey (Encuesta Nacional de Demograffa y Salud 1989~ENDSA) was carried out by the Instituto Nacional de Estadfstica (INE) with technical assistance from the Institute for Resource Development (IRD). The survey was part of the worldwide Demographic and Health Surveys Program supported by the United States Agency for lntemational Development. The purpose of the survey was to collect data on a national sample of women (and their children under five), from which estimates could be made regarding fertility levels and trends, fertility preferences, knowledge and use of family planning, indicators of matemal and child health, and infant and child mortality levels. A nationally representative sample of 7923 women 15-49 years of age were interviewed in the Bolivia DHS survey (ENDSA) between February and July 1989. The information about child health is based on live births to these women during the preceding five years. For purposes of the survey, the sample was designed to provide representative estimates for the nine departments, with the exception of Beni and Pando which, due to their low population density, were treated as one entity. For this report, the departments were grouped according to region: La Paz, Oruro and Potosf in the Altiplano; Cochabamba, Tarija and Chuquisaca in the Valle; and Santa Cruz, Beni and Pando in the Llanos. The First Country report for the Bolivia DHS survey presented findings on fertility, family planning, and maternal and child health (INE and IRD, 1990). The present report is a more extensive analysis of issues related to maternal and child health. The analysis is possible because the Bolivia survey was specially designed to include many health questions not asked in previous DHS surveys (see Appendix C). 1.5 Background Characteristics of Survey Respondents Women's level of education is a factor that greatly influences their attitudes and practices regarding their own and their children's health, their reproductive behavior, their attitudes towards ideal family size, and their practice of family planning. However, the number of years of schooling also reflects their socioeconomic status: the higher the level of education, the more favorable the economic situation. Socioeconomic factors, in tum, determine access to health services and the quality of those services. Therefore, the survey respondents' level of education and other characteristics are summarized here. For the purposes of this study, the level of education is divided into four categories: 1) no education; 2) basic education (1 to 5 years of schooling); 3) intermediate education (6 to 8 years); and 4) secondary or higher education (9 or more years). Table 1.1 presents the distribution of women 15-49 years by level of education according to background characteristics. Eighteen percent of the women never attended school and 31 percent reached or surpassed the secondary school level. Figures on education by age group show that women's access to schools has improved in recent decades. Among women 45-49 years, almost half never attended school: in contrast, only 4 percent of women 15-19 years had no education. Likewise, only one in ten women age 45-49 reached secondary school, while among those 15-19 years the proportion is four in ten. As expected, the level of education of rural women is markedly lower than that of urban women. In urban areas, women without education constitute 8 percent of the total, and 46 percent of women have 2 reached secondary school. Among rural respondents, 32 percent have no education, and only 7 percent reached secondary school. With respect to regional differences, women from the Llanos have higher levels of education than women from the Altiplano or Valle. Only 9 percent of the women from the Llanos never attended school, compared with about 20 percent of women in the other two regions. Table 1.1 Educational level of women: Percent distribution of women 15-49 years by level of education, according to selected characteristics, Bolivia ENDSA, 1989 Secondary Number No school or of Characteristic education Bmie Intermediate higher Total women AGE 15-19 4.3 31.4 26.3 38.1 100.0 1682 20-24 6.8 36.0 16.0 41.2 100.0 1311 25-29 10.5 43.1 13.6 32.8 100.0 1341 30-34 19.1 37.6 12.5 30.9 100.0 1117 35-39 28.3 36.8 12.0 22.8 100.0 1073 40-44 33.7 36.5 10.8 19.1 100.0 740 45-49 47.7 32.9 9.4 10.0 100.0 659 RESIDENCE Urban 8.1 27.3 18.6 46.0 100.0 4753 Rural 31.5 49.8 11.3 7.4 100.0 3170 REGION Altiplmao 20.3 35.6 14.8 29.3 100.0 4104 Valle 19.1 38.2 12.3 30.5 100.0 2129 Llanos 8.5 35.7 22.3 33.6 100.0 1691 TOTAL 17.5 36.3 15.7 30.5 100.0 7923 3 CHAPTER 2 CHILDHOOD MORTALITY Based on information from the birth histories, infant mortality was estimated at 96 deaths per 1,000 live births for the period 1979-88. Neonatal mortality was 41 per 1,000 live births, while mortality among children under five years of age was 142 per 1,000 live births. These rates are among the highest in Latin America. The direct and indirect estimates of infant and child mortality show that mortality declined by about 20 percent over the fifteen-year period preceding the survey. Large mortality differentials persist in Bolivia. Infant mortality in urban areas was 79 per 1,000 live births, compared with 112 in the rural areas. The Llanos region appeared to have somewhat lower postneonatal and child (age 1-4 years) mortality rates than the other two regions. Mortality among children of mothers with no formal education was almost three times higher than among children of mothers with at least nine years of education. Children in households where the mother did not watch television or radio daily also had much higher mortality rates than other households. Fertility-related variables had a substantial effect on child mortality. In particular, short birth intervals (less than two years) were associated with a mortality rate three times higher than that for birth intervals of 48 months or longer. Children of birth order 6 and over had higher infant mortality rates than children of lower birth order;first-born children had the lowest mortality rates of all. 2.1 Methodology All respondents, that is, women 15-49 years of age, in the ENDSA were asked to provide a complete birth history, including the sex, date of birth, survival status, and current age, or age at death, for each live birth. Mothers could report the age at death in days, months, or years, depending on the child's age at death. The reliability of mortality estimates calculated from retrospective birth histories depends on how completely deaths (and births) are reported and how accurately birth dates and ages at death are recalled. Generally, underreporting is most severe for deaths which occur very early in infancy. In the ENDSA, the age distribution of deaths within the neonatal and postneonatal periods 1 does not suggest severe underreporting of deaths early in infancy. For example, more than 60 percent of deaths within the neonatal period occurred during the first week of life, and more than 40 percent of all infant deaths took place in the first month of life. However, there was substantial heaping at 12 months of age. Since this heaping would affect the estimates of infant mortality, some of these deaths reported at 12 months were redistributed to ages 10-14 months (INE and IRD, 1990). The direct method of estimation calculates infant and childhood mortality rates from the birth history data (Rutstein, 1984), while indirect estimates of child mortality use data on children ever bom and children still alive, by age of the mother (Sullivan and Wilson, 1982; United Nations, 1983). Unless otherwise specified, the mortality rates reported below are direct estimates. Overall mortality rates based on the indirect method are included, however, for purposes of comparison with other data. Neonatal deaths are defined as deaths among infants who have not yet reached one month of age. The posmeonatal mortality rate is calculated as the difference between the infant mortality rate and the neonatal mortality rate. 2.2 F indings Mortality Levels and Trends Table 2.1 presents child mortality by age, calculated by the direct method of estimation, for various periods before the survey. The estimates for the period 1979-88 are considered the most reliable for recent mortality levels. Infant mortality during this period was 95.8 per 1,000 live births, and 43 percent of the infant deaths occurred in the neonatal period. The probability of dying between birth and the fifth birthday was 141.8 per 1,000 live births. This level of child mortality is one of the highest in Latin America. Table 2.1 Childhood mortality by calendar period: Neonatal, post:neonatal, infant, child and under-five mortality by calendar period, Bolivia ENDSA, 1989 Post- Under- neonatal Infant Child five Neonatal (1-11 (0-11 (l~I (0-4 Calendar period (0 months) months) months) years) years) Ten-year.,period 1979-88" 41.0 54.8 95.8 50.9 141.8 Five year period 1984-88 a 37.6 51.5 89.1 46.3 131.3 1979-83 44.4 58.1 102.5 58.1 152.8 1974-78 44.6 52.7 97.3 69.4 160.0 Note: Rates are determined by the direct method (deaths per 1,000 births). alncludes exposure up to the month of interview The birth history data suggest that under-five mortality has declined from 160 to 131 per 1,000 live births between 1974-78 and 1984-88, with the drop due mainly to a decrease in mortality between ages 1 and 4. Figure 2.1 shows the trends in under-five mortality derived from birth history data (direct method of estimation) and from child survivorship data (indirect method of estimation)) The direct and indirect mortality estimates are generally in agreement, although the indirect method suggests a larger decline in the late 1970s. The indirect estimate of infant mortality, using data from women age 20-34 years, is 102 per 1,000 live births for 1985 (data not shown). The only other source of national child mortality data in the 1980s is the 1983 Contraceptive Prevalence Survey (Coloma and de Ormachea, 1985). Infant mortality was indirectly estimated from these data to be 119 per 1,000 live births, while under-five mortality was 184 per 1,000 live births? Several infant mortality surveys were carried out in urban areas during the first half of the 1980s. Those infant mortality estimates ranged from 89 to 140 per 1,000 (Pedersen et al., 1987). z The indirect estimates were made using the Trussell variant of the indirect child mortality estimation procedure developed by Brass (United Nations, 1983). Model life tables of the South family were used. 3 The sample consisted of three urban ureas and three rural ureas (Coloma and de Ormachea, 1985). 6 250 200 160 100 Figure 2,1 Trends in Under-five Mortality Based on Direct and Indirect Estimates Deaths per 1,000 l i ve blrthe " - _ . 60 0 i p i F , , , r r i , r , , , , i , r J i i i r r i i , p 1974 75 76 77 78 79 80 81 82 83 84 88 86 87 88 lg8g Year - - Oi ,e¢t Eat lmate | - - . Ind)r,©t Egt imat , , Bol iv ia ENDBA-Bg Mortality Differentials Table 2.2 and Figure 2.2 show socioeconomic differentials in infant and child mortality rates. The rates, calculated by the direct method of estimation, cover a ten-year period (1979-1988) to ensure sufficient cases to allow mortality estimates for various population sub-groups. Mortality is higher, at all ages, in rural areas. For infant mortality, rural rates are 1.4 times higher than urban rates. With the exception of the neonatal period, where differences are small, the Llanos has the lowest mortality rates of the three regions. Mortality levels are highest in the Valle. Large differentials can be observed according to the mother's level of education. For all age groups, mortality among children whose mothers have no education is three to four times higher than mortality among children of mothers with at least nine years of schooling. Mortality differentials by the father's level of education generally showed the same pattem, but the differences were less marked (data not shown). Mortality is also substantially lower in families where the father has a white collar job. Mortality is almost twice as high among children of blue collar workers, and more than twice as high among children whose fathers work in agriculture. The socioeconomic variables are strongly con'elated. For example, women who speak an Indian language generally have no formal education. Children of mothers who speak an Indian language have higher mortality rates than children of mothers who speak Spanish. This difference is largest at ages 1-4 years. Exposure to mass media is also related to socioeconomic variables. Media exposure is divided into three categories: viewing television on a daily basis, listening to radio---but not watching television---everyday, and daily exposure to neither medium. Mortality rates are highest for children whose mothers neither watch television nor listen to the radio regularly. If the mother watches television daily, mortality rates are lower. Children of mothers who listen to the radio dally display intermediate mortality rates. Table 2.2 Childhood mortality by socioeconomic characteristics: Neonatal, posmeonatal, Infant, child and under-five mortafity for the calendar period 1979-1988 by socioeconomic characteristics, Bolivia ENDSA, 1989 Post- Under- neonatal Infant Child five Neonatal ( 1 - 11 (0-11 ( 1-4 (0-4 Characteristic (0 months) months) months) years) years) RESIDENCE Urban 35.6 42.9 78.6 38.6 114.1 Rural 46.0 66.0 112.0 63.4 168.4 REGION Altipl ano 38.6 57.1 95.7 51.2 142.0 Valles 45.0 60.6 105.6 59.9 159.1 Llanos 41.3 42.2 83.5 39.7 119.9 MOTHER'S EDUCATION None 57.4 66.2 123.6 65.9 181.3 1-5 years 43.8 64.3 108.1 60.6 162.1 6-8 years 28.8 36.0 64.8 37.6 100.0 9+ years 19.6 26.8 46.4 14.8 60.5 FATHER'S OCCUPATION White collar 26.7 35.4 62.1 25.4 85.9 Blue collar 44.4 58.5 102.8 53.7 151.0 Agriculture 49.3 66.5 115.8 69.1 176.9 Other 33.3 44.2 77.5 34.2 109.1 MEDIA EXPOSURE TV daily 28.7 44.1 72.8 36.7 106.8 Radio daily 40.1 59.5 99.6 51.0 145.6 Neither 62.2 66.3 128.4 75.4 194.1 LANGUAGE Spanish 37.0 49.5 86.5 39.2 122.3 Indian 50.0 66.6 116.6 78.3 185.8 TOTAL 41.0 54.8 95.8 50.9 141.8 Note: Rates are determined by the direct method (deaths per 1,0(30 live births); includes exposure up to the month of interview. RESIDENCE Urban Rural ARi EGION ti lind ~llllam Llanos EDUCATION Nona l-E yearl E-E yelrl E* yelrl LANGUAGE Bpanilh Indian MEDIA TV RIdiO Niithlr Figure 2.2 Under-five Mortality 1979-1988 by Socioeconomic Characteristics . / ~ / 1 1 f / / / / / A i 60 100 180 200 Deaths par 1.000 Live Births I l a I Neormtsl ~Poatneonalal ~Ch l ld (1-4 yra) i Bolivia ENDDA-89 Mortality differentials according to demographic variables are examined in Table 2.3 and Figure 2.3. As expected, neonatal mortality is higher among boys than girls. However, the disadvantage persists into the postoeonatal period, which is less commonly observed. There are no sex differentials at ages 1-4 years. There are only small mortality differentials according to the mother's age at delivery. Children whose mother is 35 years and older have the highest neonatal mortality rates, but the lowest mortality rates at 1-4 years of age. Children of mothers less than 20 years of age have somewhat higher mortality rates after infancy, but the differences are small. At all ages, first births have the lowest mortality rates. The mortality rates for children of birth order 6 and above are 60-70 percent higher, at all ages, than rates for first-born children. Children whose birth order is 2 through 5 have only slightly higher mortality rates than first-born children during infancy; between one and four years of age, however, the rates are 60-70 percent higher, that is, similar to the rates for children of the highest birth order. The length of the preceding birth interval has a large effect on mortality. The longer the interval, the lower the mortality. For children born after a short birth interval (less than 24 months), mortality rates during the first year of life are three to four times higher than for children born after an interval of 48 months or longer. At ages 1-4, the mortality rate is twice as high for children born after short intervals than for children born after long intervals. The effect is large in the neonatal and postneonatal periods, which indicates a biological component, but is remains substantial for children age 1-4, suggesting a behavioral component as well. Finally, infant mortality rates triple in cases of multiple births. The increase is fivefold in the neonatal period and twofold in the postneonatal period. When a child of a multiple birth reaches 12 months of age, no increased mortality risks are observed thereafter. Table 2.3 Childhood mortality by demographic characteristics: Neonatal, postneonatal, infant, child and under-five mortality for the calendar period 1979-1988 by demographic characteristics, Bolivia ENDSA, 1989 Post- Under- neonatal Infant Child five Neonatal (1-11 (0-11 (1-4 (0-4 Characteristic (0 months) months) months) years) years) SEX Male 45.8 59.6 105.5 50.7 150.9 Female 35.8 49.6 85.5 51.1 132.2 MOTHER'S AGE 15-19 years 43.6 57.8 101.4 55.6 151.4 20-34 years 38.5 53.4 91.9 51.0 138.2 35+ years 50.5 58.5 109.0 45.8 149.8 BIRTH ORDER First 34.4 45.9 80.3 33.5 111.1 2-3 35.3 52.8 88.1 52.6 136.1 4-5 41.2 50.7 91.9 59.l 145.6 6+ 56.0 70.6 126.6 58.0 177.3 BIRTH INTERVAL a <24 months 61.4 92.4 153.9 81.6 222.9 24-35 months 36.7 46.9 83.7 45.3 125.1 36-47 months 28.9 34.2 62.9 38.9 99.3 48+ months 19.2 23.6 42.7 40.6 81.6 MULTIPLE BIRTH 211.3 107.2 318.5 41.3 346.6 TOTAL 41.0 54.8 95.8 50.9 141.8 Note: Rates are determined by the direct method (deaths per 1,000 births); rates include exposure up to the month of interview SEX Mi le Femmle MOTHER'8 AGE ,20 20-34 3E* BIRTH ORDER First 2 -3 4-B 6* BIRTH INTERVAL c24 month0 24-38 3E-4T 48* Figure 2.3 Under-five Mortality 1979-1988 by Demographic Characteristics I I I I I l i~ l ~ ) EO 100 150 gOD Deaths per 1.000 Live Blrthi 260 I~ l l l Neonatal ~ Poztnlonztol [~ Child (1-4 yrz) i i m BOliVIa ENDSA'EE 10 CHAPTER 3 CAUSES OF DEATH IN CHILDHOOD Seventy-six percent of the deaths among children who were born in the five years preceding the survey took place at home. Half of the children who died were not taken to any health facility during the illness that led to death. In order to ascertain the major causes of death, the ENDSA respondents were questioned about the illness leading to death among children born during the past five years. During the neonatal period, birth problems were the leading cause, accounting for one-third of neonatal mortality. Low birth weight, as reported by the mother, increased neonatal mortality rates fourfold compared with normal birth weight. Neonatal tetanus was estimated to have caused 5-10 percent of neonatal deaths. During the postneonatal period and early childhood, diarrheal disease was prominent as a cause of death. For almost half the children who died at age 1-11 months, the probable main cause of death was diarrheal disease. Acute respiratory infection was associated with about I in 5 deaths according to mothers' reports of the main cause of death, but was associated with only I in 9 deaths judging by reported symptoms. Measles was a probable cause of death for 3 percent of children 1-11 months (6 percent at ages 1-4 years). In addition, 19 percent of the children 4 months and over hadprobably had measles (rash and fever) in the three months before the survey, according to the reported symptoms. 3.1 Methodo logy Analyzing the medical causes of death in childhnod can be useful in identifying priority areas for health programs. In the ENDSA the probable causes of death were ascertained for deceased children born during the five years preceding the survey. This approach to determining the causes of death is often called the verbal autopsy or postmortem interview technique. Two types of information were used to assess the likely cause of death. First, the respondent was asked to give the main disease or accident causing the death, which was entered by the interviewer in the questionnaire and later coded, using a list of causes of death (Minlsterio de Previsi6n Social y Salud Ptiblica, 1983). Second, for deaths not caused by an accident, inquiries were made into the presence and duration of several specific symptoms and signs during the two-week period preceding the death. These symptoms ~ included diarrhea, diarrhea with blood, difficult breathing, common cold/cough, rash and fever. The mother was also asked whether the baby had been sucking normally during the first days of life, in order to distinguish between neonatal deaths due to tetanus and death due to other causes. The loss of the ability to suckle a few days after birth is typical of neonatal tetanus deaths. The mother was asked whether the child had died at home or in a health facility, and whether medical care was sought for the illness preceding the death. If diarrhea was one of the signs and symptoms present before death, the use of oral rehydration therapy was determined. Twenty-six percent of the respondents said they had a death certificate for the deceased child, but in only 2 percent of the cases was the certificate actually shown to the interviewer. Therefore, the death certificates could not be used to assess the causes of death. Several studies have described the long-term effects of measles on child mortality: children who have had measles are more likely to die from other causes, such as pneumonia, diarrhea, or tuberculosis, in the months afterwards due to reduced resistance to these diseases (Koenig et al., 1990). Therefore, In this report the term "symptom(s)" includes both signs and symptoms observed and reported by the mother. 11 mothers were asked whether the dead child had had rash and fever in the six months before death. I f so, they were asked the duration of the episode and the time elapsed between this illness and the child's death. In addition, the mother's assessment of the neonate's size at birth--available for all live births during the five years preceding the survey- -was used to estimate the impact of low birth weight on mortality, z 3.2 Findings Main Cause of Death Reported by Mothers Table 3.1 summarizes the data on the main cause of death as reported by the mother, according to the age of the child at death. The deaths are divided into three age groups: neonatal (less than 1 month), posmeonatal (1-11 months), and 12 months and over. The latter group consists mainly of deaths among children 1 and 2 years old. Mothers reported a cause of death for 91 percent of the deceased children. Table 3.1 Causes of death in childhood: Percent distribution of deaths among children born in the five years preceding the survey by age at death, according to main causes of death (mothers' reports), Bolivia ENDSA, 1989 Age at death Main caHse ofdeath 1-11 12+ All (mother's report) < 1 month months months deaths Birth problems 32.9 3.8 0.5 13.3 Prematurity 7.7 0.0 0.0 2.7 Tetanus 5.7 2.0 1.9 3.3 Congenital anomaly 1.0 1.1 0.0 0.8 Diarrhea l 3.1 39.1 63.8 35.7 Respiratory illness 17.2 25.7 15.9 20.5 Measles 0.2 2.1 1.2 1.2 Other infections 3.1 1.3 1.7 2.0 Other diseases 4.0 4.4 3. t 4.0 Accidents 7.9 8.4 4.8 7.4 No cause given 7.4 12.1 6.8 9.2 TOTAL 100.0 100.0 100.0 100.0 NUMBER OF DEATHS 199 234 134 567 2 Subjective assessments of the relative size of the infant at birth have been shown to be reasonable indicators of actual size at birth (Moreno and Goldman, 1990). 12 In the neonatal period, problems associated with the delivery, primarily traumatic delivery and neonatal asphyxia, were mentioned by almost one-third of the respondents. Respiratory illness was the next most frequently cited cause of death (17 percent), followed by diarrhea (13 percent). Eight percent mentioned prematurity, and 6 percent tetanus. An unusually high proportion (8 percent) also mentioned accidents as a cause of neonatal death. This is likely to be due to the structure of the questionnaire and its interpretation by both interviewers and respondents. In several cases, deaths due to birth trauma were classified as accidental deaths, and there was not always sufficient detail to distinguish perinatal causes of death (mainly birth problems) from accidents. During the posmeonatal period, diarrhea was reported to be the leading cause of death, cited in 39 percent of the 234 deaths, followed by respiratory infection (26 percent). For children 12 months of age and older, gastrointestinal infections were most common, reportedly causing 64 percent of all deaths. Eight percent of deaths in the postneonatal period and 5 percent of deaths among children 12 months and older were attributed to an accident) According to the mothers' reports, measles did not appear to be a leading cause of death. Cause of Death Derived from Symptoms Table 3.2 presents the symptoms and signs, as reported by the mother, associated with the illness that led to the death of the child. Fever, diarrhea, and breathing difficulties were the most common symptoms. On average, mothers responded that three out of the list of eight symptoms were present. For example, of the children with diarrhea, 73 percent also had fever and 25 percent had breathing difficulties. Table 3.2 Symptoms associated with the illness that led to child's death: Among children born in the last five years who died of disease, the percentage with specific symptoms and combinations of symptoms, Bolivia ENDSA, 1989 Mean Combinations of symptoms Percent number of with other With With With With symptom a symptoms diarrhea dyspnea rash fever Symptom Fever 54.2 1.7 54.6 35.1 13.0 -- Diarrhea 40.7 1.7 -- 24.6 12.6 72.7 Diarrhea with blood 10.4 2.1 -- 40.2 16.8 75.6 Dyspnea 33.8 2.1 29.7 -~ 14.4 56.3 Common cold/cough 27.0 2.1 43.4 56.1 16.8 72.0 Rash 10.9 2.2 47.0 44.8 -- 65.0 Convulsions 3.6 1.9 18.7 37.9 21.4 58.8 Other 33.0 1.2 39.5 15.8 5.6 43.2 aBased on 505 children 3 This represents a total of 26 deaths (unweighted) for these two age groups. The number of deaths reportedly caused by specific accidents were as follows: 6 from falls, 4 from birth trauma causing death in the postneonatal period, l from bums, 1 from a motor vehicle accident, and 3 from other causes. Information was missing for l l cases. 13 This short list of symptoms can be used to assess the probable cause of death. Verbal autopsy validation studies have been carried out in other countries, and their results are used to determine the most probable causes of child death in Bolivia (Garenne and Fontaine, 1986; Kalter et al., 1990). Since a limited number of symptoms were included in the ENDSA, only a few leading causes of death can be identified. These include diarrheal diseases, acute lower respiratory tract infections (primarily pneu- monia), measles, and neonatal tetanus. The first three illnesses are of particular importance after the neonatal period and will be analyzed for all postneonatal and early childhood deaths combined. Initially, three criteria were used to identify deaths due to neonatal tetanus: death occurring between 2 and 30 days after birth, normal sucking during the first days after birth, and the presence of convulsions. The ENDSA question on the presence of convulsions was not very clear, however, and a number of positive responses may have been missed. No deaths met all three criteria. Hence, alternative criteria for identifying babies with possible neonatal tetanus are used in this report. Omitting the question on convulsions and restricting the age range at death to 4-14 days, the period during which most neonatal tetanus deaths occur, 40 deaths were found, which is 20 percent of all neonatal deaths. In five of these cases (2.5 percent of neonatal mortality), the mother had also mentioned tetanus as the cause of her child's death. Using the mothers' diagnoses, 5.7 percent of all neonatal deaths were due to tetanus (see Table 3.1). All deaths occurring among children age 1 month and over were grouped to examine the importance of diarrheal diseases, acute respiratory infections, and measles as causes of death. Diarrhea was considered a cause if the child had had diarrhea (with or without blood in the stool) for at least two days. If the child had had cough for at least four days and breathing difficulties for at least two days before death, pneumonia was listed as a cause (Kalter et al., 1990). Measles was considered a cause if the child's age at death was at least 4 months, a rash was present for at least three days, and the child had fever for at least three days. Table 3.3 shows the percentage of deaths among children born in the five years preceding the survey (excluding neonatal deaths) due to the three causes (derived from the reported symptoms). Diarrheal diseases were considered a probable cause of death in 48 percent of the cases. Chronic diarrhea appeared to play an important role: 38 percent of the children who had diarrhea before death had diarrhea for at least two weeks. In most cases, diarrhea was not associated either with pneumonia or measles. In only 6 percent of all deaths did diarrhea cause or contribute to the child's death in combination with pneumonia or measles. Based on mothers' reports of their children's symptoms, lower respiratory tract infections caused fewer deaths than diarrheal diseases: pneumonia was the probable cause of 12 percent of all deaths. Measles, as defined above and occurring during the last two weeks before death, was a probable cause of 3 percent of the deaths and was associated with 6 percent of all deaths at age 12 months and over. The proportion of deaths preceded by a presumed measles infection was estimated using questions about an illness consisting of rash and fever during the six months preceding the interview (data not shown). Using the same criteria for the diagnosis of measles as above, that is, an illness with a rash lasting three days or longer and with a fever for at least three days, 14 percent of children had a history of measles in the four weeks preceding death, 19 percent in the three months preceding death, and 21 percent in the six months preceding death. For children who died of respiratory diseases, according to the mother's report, 22 percent had had measles in the three months before death. The corresponding figure for diarrheal deaths was 15 percent. This suggests that measles may be more important as a cause of death than the data in Table 3.1 show. 14 Table 3.3 Causes of death derived from reported symptoms: Among children loom in the last five years, the percentage of deaths (excluding neonatal deaths) associated with one or more cause derived from symptoms, Bolivia ENDSA, 1989 Acute lower Cause of death based Diarrheal respiratory on symptom algorithm disease tract infection Measles Death associated with each cause 48.4 l 1.7 3.2 "Cause" was the only cause 42.7 6.1 0.5 "Cause" plus diarrheal diseases -- 3.8 0.8 "Cause" plus acute respiratory infection 3.8 - - 0.7 "Cause" plus measles 0.8 0.7 All three causes t .2 1.2 1.2 Note: "Cause" refers to diarrhea, acute lower respiratory Ixact infection and measles, respectively. Forty-four percent of deaths were not associated with diarrheal disease, ARI, or measles according to the reported symptom algorithm. Causes of Death: A Comparison of the Findings Both analyses of the causes of death highlight the importance of diarrheal diseases after the neonatal period. For about half of the deaths among children older than 4 weeks, diarrhea was either the only cause of death or was an associated cause. In 39 percent of the deaths, diarrhea was listed as the main cause of death by the mother, and the child was reported to have had the symptoms and signs of the illness (Table 3.4 and Figure 3.1). Acute respiratory infection was associated with about 1 in 5 deaths according to the mother's report of the main cause of death, but with only 1 in 9 deaths based on the reported symptoms. In part, this difference may be related to the greater difficulty in identifying pneumonia, as has been observed in other studies. Measles was mentioned as the main cause of death by mothers for only 2 percent of deaths among children age 1 month and over. However, based on the symptoms, measles was a probable cause for 5 percent of the deaths. Size at Birth Low birth weight has been identified as an important risk factor for neonatal and, to a lesser extent, posmeonatal mortality. In Bolivia, birth weights are generally not available, partly because most deliveries occur at home and partly because there is no adequate registration of birth weights on the records kept by mothers. Building upon the positive reports from the DHS experimental survey in Peru (Moreno and Goldman, 1990), mothers were asked to give their subjective assessment of each baby's size at birth for all infants bom in the last five years. The results are given in Table 3.5. 15 Table3.4 Causes of death from mothers' reports of the main cause and derived from reported symptoms: Among children born in the five years preceding the survey, the percentage of deaths (excluding neonatal deaths), according to symptom algorithm and the main cause reported by mothers, Bolivia, ENDSA, 1990 Acute lower Dian'heal respiratory Source for cause of death diseases tract infection Measles CAUSE ACCORDING TO: Symptom algorithm only Mother's report only Both symptom algorithm and mother's report 9.2 4.6 3.2 8.8 15.1 1.7 39.2 7.1 Symptom algorithm and/or mother's report 57.3 26.8 4.9 Not~: Multiple causes are possible (N=367). Figure 3.1 Causes of Death Based on Mothers' Reports and Symptoms Percentage of Deaths Mothers' Reports | [~ Reports and Symptom= I J Symptoms Diarrhea Respiratory Measlaa IIInaii Note: Includel daathl among ohi!;:~n under five. except for neonatal delths. Bolivia ENDSA-89 16 Table 3.5 Size at birth and infant mortality: Percent dls~bution of births in the five years preceding the survey according to size reported by mother, and neonatal and postneonatal mortality rates, Bolivia ENDSA, 1989 Neonatal Postneona[al Number Size at birth Births mortality a mortality ~ of births Very small 9.5 108.9 46.6 551 Smaller than average 20.3 31.5 43.8 1174 Average 56.0 25.5 38.9 3235 Larger than average/ very large C 13.2 16,9 37.7 763 Don't know 0.4 21 TOTAL 100.0 34.0 43.3 5780 Note: Rates are for deaths per 1,000 live births. aExcludes births in 1989 bExcludes births in 1988 and I989 CTwenty-four births were in the very large category (0.4 percent of the total) Almost 10 percent of all neonates were considered "very small" by the respondents, and 20 percent were "smaller than average." Neonatal mortality rates varied by size at birth. Neonatal mortality was 109 per 1,000 live births for "very small" babies, which was four times higher than for babies of "average" size. Postneonatal mortality rates decreased moderately with increasing size at birth, but the differences were small. These mortality rates are consistent with the higher mortality rates reported elsewhere for low birth weight babies and suggest that the mother is able to give a valid estimate of the baby's size at birth even if the exact birth weight is not known. Medica l Care Figure 3.2 shows the types of medical assistance sought for children during their terminal illness. For almost half the children, no medical assist- ance was sought. The public hos- pital was by far the most fre- quently visited source of medical assistance (29 percent). It is no- table that 40 percent of the chil- dren who died from diarrhea (ei- ther as a primary or associated cause) were reported to have re- ceived oral rehydration solution before death. The majority of children, 76 percent, died at home. Figure 3.2 Med ica l Ass i s tance Sought be fore Ch i ld ' s Death , and P lace of Death MEDICAL ASSISTANCE NO One Public holpital Health center Hellth bol t Private hoapitll Private doctor Pharmacy Trad. practitioner CNS/Otber PLACE OF DEATH Public boapltal Health center Private hOlpltll Home Elsewhere, milling 49 29 ~ 8 1 13 ~t 2 ~ T e ~ 8 20 40 60 80 Percent Note: Ba led oll death l among children born In the d yemrl preceding the aur~ey Bolivia ENDSA-89 17 CHAPTER 4 MORBIDITY: DIARRHEA AND ITS TREATMENT Diarrheal disease is common in children under five years of age. According to the ENDSA, 28 percent of all children under five had had diarrhea in the two weeks immediately preceding the survey. After correcting for seasonality, this corresponds to 5.8 episodes of diarrhea per child per year. For children age 6-23 months the prevalence of diarrhea was as high as 40 percent. Differentials in prevalence were moderate for most socioeconomic variables and for water source and sanitation variables. The differentials were greatest for the level of education of the mother. There were only minor differentials in the prevalence of diarrhea by mode of infant feeding. Neither the early introduction of supplements nor the use of bottle feeding appeared to be associated with a higher prevalence of diarrhea during the two weeks preceding the survey. About half of the mothers gave their children more liquids during diarrheal episodes. Examining the use of oral rehydration therapy (ORT), it was found that one in 4 mothers used a fluid prepared from a packet of oral rehydration sahs (ORS); 1 in 9 mothers prepared a homemade sugar and salt solution; and 1 in 4 children were taken for treatment to a health facility. Children were less likely to receive ORS from a private facility than from a public facility. There were marked differentials in patterns for most socioeconomic variables;for example, mothers with little or no education were less likely to treat diarrhea with ORS packets or to seek medical assistance than mothers with higher levels of education. Lack of knowledge of ORS contributed to its low rate of use. Only 70 percent of the mothers with children under five years had heard of ORS packets, while just 61 percent had ever seen a packet. Mothers with low levels of education and those living in rural areas had less knowledge of ORS packets than other women. 4.1 Methodology Mothers with children under the age of five were asked if their children had had diarrhea in the last 24 hours (defined as a current episode) and how many days previous this episode had begun. If the response was negative, the mother was asked when the last time was that the child had had diarrhea (defined as the terminated episode). The answers were recorded in days, weeks, or months. The respondent was also asked how long the terminated episode had lasted. Finally, the mother was asked whether there was blood in the stool during the episode. Diarrhea with blood is a symptom of dysentery, a gastrointestinal infection frequently caused by bacteria, which requires a different treatment from the more common viral diarrheas. All mothers whose children had been ill with diarrhea in the preceding two weeks were asked whether the child had been taken anywhere for treatment, and what type of treatment had been given. The mother was also asked whether the child had received any oral rehydration therapy (ORT), whether in the form of a fluid prepared from a packet of oral rehydration salts (ORS) or as a homemade sugar and salt solution, or both. In addition, the mother was asked whether the quantity of liquids and solid foods given to the child during the diarrheal episode was increased, decreased, or unchanged. Mothers who had not used ORS were asked whether they knew about the packets; those who had used ORS at least once were asked how they prepared the solution. Finally, mothers were asked to list the places where ORS packets could be obtained, and the cost of the packets. 4.2 Findings Prevalence The questions about current and terminated diarrheal episodes and their duration can be used to assess data quality and the effects of memory loss within a two-week reeaU period. An analysis of the 19 accuracy of diarrhea recall is presented in Appendix A (page 91). The general conclusion is that there is substantial undcrreporting when recall periods are more than three days in length. Tables 4.1 and 4.2 present the prevalence of diarrhea in the two weeks preceding the survey according to socioeconomic and demographic variables. The overall prevalence of diarrhea among children under age live was 28 percent, while 16 percent had had diarrhea in the last 24 hours. By comparison, a 1983 survey in Bolivia found the two-week prevalence of diarrhea among 66,827 children under fivc to be 24 percent (Murillo and Coloma, 1984). There was considerable variation in prevalence by the age of the child (Figure 4.1). Prevalence increased during infancy, peaking at 41 percent at age 12-23 months, and then declined for older children, to less than 15 percent at age 48-59 months. This age pattern has been observed in many other studies and is thought to be associated with the weaning period. The differentials in the prevalence of diarrhea shown in Tables 4.1 and 4.2 were small fur most other variables? The largest differences were observed by mother's age and by mother's level of education. Urban-rural differences and regional differences were minor. Piped drinking water and a flush toilet were associated with somewhat lower prevalence, but the prevalence of diarrhea was no lower among cbildrcn in households with latrines than among children in households without sanitary facilities) The proportion of children with blood in their stools during the diarrheal episode was 3.6 percent and varied from 1 to 5 percent. The prevalence was very low among children less than 6 months of age (0.9 percent) and was highest at age 12-23 months. Differentials were largest according to the mother's cducation and, to a lesser extent, the household's sanitation. Prevalence was slightly higher among rural than urban children. The mean duration of an acute diarrheal episode in longitudinal studies is usually on the order of 5 to 6 days. In Bolivia, the mean duration of an episode was 3.8 days (standard deviation 3.5 days) for children whose diarrhea had ended during the two weeks before the interview. This is shorter than expected, but it may bc that mothers recall only the days when the diarrhea is at its worst and do not count the first or last days. Considering the duration-to-date for current cases of diarrhea, consistent variation was observed by socioeconomic variables and for those having poorer sanitary environments (data not shown). A direct estimate of mean duration from the current status data is not possible? To estimate the number of episodes of diarrhea per child per year from prevalence and duration, seasonality of diarrhea needs to be known as well (WHO, 1989). Seasonality of diarrhea can be assessed by using data on the prevalence of diarrhea from health facilities in Bolivia between 1975 and 1985 (Pederscn ct al., 1987). During that period diarrhea was at its peak from October through November and reached its lowest level from June through July. After correcting for seasonal fluctuations, the average number of episodes per child per year was 5.8, based on ENDSA data collected from February through Junc. The melm duration of diarrheal episodes was assumed to be 5.5 days (WHO, 1989). This corresponds to 32 days with diarrhea per child per year. This is true particularly if sampling errors are taken into account. For example, the confidence limits (plus or minus 2 standard errors) for diarrhea prevalence among all children are 26.5-29.5 percent. z The ENDSA questionnaire did not distinguish between other facilities (other than flush toilet or latrine) and no facilities. The majority in the other/none category is assumed to be households without facilities. 3 This is because there is "length-biased sampling" in the durations of diarrhea episodes in the sample of children with current diarrhea (called a prevalence series). The mean duration of diarrhea episodes estimated from the prevalence series is 7.1 days. For details, see Freeman and Hutchison, 1980. 20 Table 4.1 Diarrhea prevalence by socioeconomic characteristics: Among children under five, the percentage who had diarrhea in the two weeks preceding the survey by socioeconomic characteristics, Bolivia ENDSA, 1989 Characteristic Diarrhea in the two weeks prec~ling the survey Diarrhea Number All with blood of diarrhea a in stools children RESIDENCE Urban 28.1 2.9 2535 Rural 28.1 4.2 2626 REGION Ahiplano 27.0 3.2 2603 Vaile 29.2 3.6 1422 Llanos 29.2 4.5 1137 MOTHER'S EDUCATION None 28.9 5.0 1061 1-5 years 30.6 4.3 2354 6-8 years 27.0 2.5 707 9+ years 22.3 1.4 1040 FATHER'S OCCUPATION White collar 27.3 3.1 1448 Blue collar 29.3 3.3 1494 Agriculture 28.7 4.6 1860 Other 23.1 1.9 360 MOTHER WORKING Yes 25.6 2.0 983 No 28.7 4.0 4178 MEDIA EXPOSURE TV daily 27.5 3.0 2242 Radio daily 26.9 4.2 1694 Neither 30.9 3.9 1226 LANGUAGE Spanish 28.6 3.2 3663 Indian 26.9 4.5 1499 DRINKING WATER Piped 26.8 3.4 3101 Non-piped 30.0 3.9 2060 SANITATION Flush toilet 24.1 2.6 1344 Latrine 33.4 4.9 892 Other, none 28.3 3.6 2926 TOTAL 28.1 3.6 5161 Note: Figures are for children born 1-59 months before the survey. a Includes diarrhea with blood in stools. 21 Table 4,2 Diarrhea prevalence by demographic characteristics: Among children under five, the percentage who had diarhea in die two weeks preceding the survey by demographic characteristics, Bolivia ENDSA, 1989 Diarrhea in the two weeks preceding the survey Characteristic Diarrhea Number All with blood of diarrhea a in stools children SEX Male 27.5 3.2 2582 Female 28.7 4.0 2580 CHILD'S AGE 0-5 months 24.7 0.9 483 6-11 months 39.4 4,1 572 12-23 months 41.4 5.3 1110 24-35 months 28,8 3.9 1017 36-47 months 20.8 3.6 993 48-59 months 14.9 2.4 986 MOTHER'S AGE 15-19 years 34.8 2.4 256 20-34 years 28.9 3,7 3616 35+ years 24.6 3.7 1289 BIRTH ORDER First 28.0 2.3 1053 2-3 25.9 2.7 1742 4-5 30.0 4.4 1188 6+ 29.5 5.3 1179 BIRTH INTERVAL <24 months 29.5 4.3 1093 24-35 months 29.5 4,8 1429 36-47 months 26.9 3.1 704 48+ months 25.1 2.5 877 TOTAL 28.1 3.6 5161 Note: Figures are for children born 1-59 months before the survey. a Includes diarrhea with blood in stools Many studies have found that the prevalence of diarrhea is lower among children who are completely breastfed than among children who are partially breastfed or not breastfed at all (Feachem and Koblinsky, 1984). The lower prevalence of diarrhea found among children younger than 6 months, as compared with children 6-23 months old, also suggests an effect of breastfeeding. In the ENDSA, the mother was asked if she was still breasffeeding the child and, if so, if she had given water, herbal tea, juice or sugar water, powdered milk, goat's milk or cow's milk, other fluids, mushy food, or solid food in the last 24 hours. In addition, the mother was asked whether any of the liquids had been given in a bottle with a nipple. This information was only collected for the last-born child. 22 Figure 4.1 Diarrhea Prevalence in the Two Weeks Preceding the Survey Percent 50 , 2O 10 O 15 6 - t l 12-23 24-35 36-47 Age o! Child (Months) I - -A l l Diarrhea DIsrrheawith Blood 48-59 Bolivia ENDBA-B9 The data on feeding is limited, however, because it refers only to the last 24 hours. Mothers may change feeding patterns in response to diarrhea. Keeping this in mind, Table 4.3 shows no strong association between breastfeeding and the prevalence of diarrhea. Only at age 3-5 and 6-11 months is the prevalence of diarrhea moderately lower among children exclusively breastfed compared with children who were partially breastfed or not breastfed at all. Among children younger than 3 months, the prevalence of diarrhea was lower for those receiving supplements than for those breastfed exclusively. The use of a bottle to give supplements did not have an effect on the prevalence of diarrhea in any of the three age groups in Table 4.3. Table 4.3 Diarrhea prevalence by feeding patterns: Among children under 12 months, the percentage who had diarrhea in the two weeks preceding the survey by feeding patterns and age, Bolivia ENDSA, 1989 Children 2 months and under Children 3-5 months Children 6-11 months Feeding pattern Diarrhea Number Diarrhea Number Diarrhea Number BREASTFEEDING Breaat feeding only 22.8 126 24.8 106 33.4 69 With supplements 12.3 76 33.9 153 40.7 424 With bottle 12.4 55 31.5 144 41.7 172 NO BREASTFEEDING a 3 a 19 37.8 79 Note: Figures are for children born 1-11 months before the survey. aLess than 20 cases 23 While these results do not unequivocally show the beneficial effects of breasffeeding, a number of factors were not available for consideration. As already mentioned, information about supplemental foods is only available for the 24-hour period immediately preceding the survey. In addition, there are no data on the severity of the diarrhea, and the number of young infants who received no breast milk is small. Case Management Figures 4.2 and 4.3 summarize the feeding and treatment patterns for diarrheal episodes occurring in the two weeks before the survey. 4 Overall, more than 90 percent of mothers who were breastfeeding their children continued to do so during the diarrheal episode: 88 percent in urban areas and 96 percent in the rural areas. Almost half the mothers reported increasing the amount of liquids given to the child, but 11 percent cut back on liquids. Thirty-two percent reduced the amount of solid foods given to the child. Figure 4.2 Feeding and Treatment Patterns for Children with Diarrhea in the Two Weeks Preceding the Survey Breaetfeeding ORS !oackete Home eolution Public health fac, Private health fa¢. Pharmacy Trad. practitioner Other Not taken for treatment 92 26 16 ~8 I, ro 20 40 60 Percent 80 tO0 Bolivia ENDSA-89 Figure 4.2 also shows the extent to which ORS packets and homemade sugar and salt solutions were used to prevent diarrheal dehydration. ORS packets were given to 1 in 4 children who had diarrhea in the preceding two weeks, while homemade sugar and salt solutions were used in 11 percent of the episodes. Mothers used the oral rehydration solutions for an average of three days. Almost a quarter of the children with diarrhea were taken to a medical facility for treatment. Public hospitals and health centers were the most frequently visited facilities, followed by private doctors 4 The analysis of the accuracy of diarrhea recall presented in Appendix A found that reporting errors and memory loss were important even within the two-week recall period. These problem may also affect the recall of treatment patterns. Therefore, the feeding and treatment patterns were initially analyzed for children who had diarrhea in the 24 hours preceding the survey and whose diarrheal episode had started at least two days before the interview. The resulting patterns of diarrhea case management were very similar to the results obtained when using a two-week recall period. Since the number of children with diarrhea is cut in half if only current cases are included, the two-week recall period has been used in this report. 24 Figure 4.3 Changes in the Amount of Liquids and Solids Given to Children with Diarrhea in the Two Weeks Preceding the Survey Don't know 2% L I I I 11% Same 42% More 46% Liquids Don't know 1% Limm 32% Same e1% More 7% 8olids Nots: Qulnt l t lee I re relative to the ulluel Imount | g[~ln. Bo l lv [ l END6A-89 and hospitals. Pharmacies and traditional medical practitioners were less frequently consulted in cases of childhood diarrhea. Of the children with blood in their stools, 44 percent were taken to a health facility (data not shown). Tables 4.4 and 4.5 examine socioeconomic and demographic differentials in diarrhea case management. While both ORS packets and homemade solutions appeared to be used somewhat more in the Llanos than in the Altiplano and Valle regions, there were no urban-rural differentials. There was considerable variation by age of the child, however: ORT was rarely used for children under 6 months and was most frequently used for children 12-23 months. Changes in fluid intake and the utilization of health services varied by mother's education and, to a lesser extent, by most other socioeconomic variables. In general, mothers with higher levels of education were more likely to give additional fluids to children with diarrhea, to use ORS, and to take a child with diarrhea to a health facility for treatment. The reasons for not taking children suffering from diarrhea to a health facility for treatment are shown in Figure 4.4. In urban areas, most mothers said the diarrhea was not serious; only 2 percent said the facility was too far away. In contrast, the distance to medical facilities was the main reason given in rural areas for not taking a child for treatment. 25 Table 4.4 Treatment practices for childhood diarrhea by socioeconomic characteristics: Among chilth'en under five who had diarrhea in the two weeks preceding the survey, the percentage who were treated with ORS, homemade solution, increased liquids, or were taken to a medical facility, by socioeconomic characteristics, Bolivia ENDSA, 1990 Characteristic Treatment received Number of Taken to Not taken children ORS Homemade Increased health for with packets solution liquids facility treatment diarrhea RESIDENCE Urban 26.6 11.7 52.9 28.9 64.0 715 Rural 24.6 10.0 36.7 18.6 75.5 737 REGION Altiplano 20.7 9.6 40.0 18,5 76.3 704 Valle 29.8 8.3 49.3 29,6 62.9 414 Llanos 30.6 16.6 48.9 27,2 64.8 334 MOTHER'S EDUCATION None 19.7 12.2 24.6 14.7 81.6 307 1-5 years 25.4 8.8 45.4 20.8 72.2 722 6-8 years 31.2 11.1 54.5 28.1 63.5 192 9+ years 29.4 15.2 61.0 40.8 52.0 232 FATHER'S OCCUPATION White collar 29.0 11.3 58.2 31.6 63.6 395 Blue collar 27.1 9.5 49.6 29.5 62.4 439 Agriculture 22.5 11.8 30.9 13.2 81.0 535 Other 21.2 9.4 43.5 22.5 66.7 83 MOTHER WORKING Yes 33.8 12.3 53.5 27.6 64.8 252 No 23.9 10.5 42.8 22.8 70.9 1201 MEDIA EXPOSURE TV daily 30.2 13.5 55.7 35.7 58.6 618 Radio daily 25.1 8.6 41.2 17.3 75.7 456 Neither 18.6 9.2 30.9 11.7 81.1 379 LANGUAGE Spanish 27.7 11.7 50.3 28.7 64.4 1049 Indian 20.0 8.5 30.1 10.6 83.8 403 DRINKING WATER Piped drink water 28.8 12.0 49.5 29.7 63.2 833 Non-piped 21.3 9.3 38.3 15.6 78.8 619 SANITATION Flush toilet 33.6 13.5 57.7 40.6 51.5 325 Latrine 23.9 11.3 52.7 24.5 66.8 298 Other, none 23.0 9.6 36.7 16.7 78.1 828 TOTAL 25.6 10.8 44.7 23.7 69.8 1452 Note: Figures are for children born 1-59 months before the survey. 26 Table 4.5 Treatment practices for childhood diarrhea by demographic characteristics: Among children under five who had diarrhea in the two weaks preceding the survey, the percentage who were treated with ORS, homemade solutions, increased liquids, or were taken to a medical facility, by demographic characteristics, Bolivia ENDSA, 1990 Characteristic Treatment received Number of Taken to Not taken children ORS Homemade Increased health for with packets solution liquids facility treatment diarrhea SEX Male 26.1 12,2 49.8 24.3 68.1 712 Female 25.1 9.6 39.8 23.0 71.4 740 CHILD'S AGE 0-5 months 12.1 3.4 30.7 18.7 77,0 121 6-11 months 19.2 10.5 39.1 25.0 68.8 225 12-23 months 34.6 11.7 44.8 27.8 63.2 459 24-35 months 23.5 12.1 48.4 25.1 71.5 293 36-47 months 25.0 11.6 52.2 20.9 72.0 206 48-59 months 23.1 11.3 46.4 13,6 79.7 147 MOTHER'S AGE 15-19 years 22.6 6.3 42.1 32.0 56,2 89 20-34 years 26,5 11.0 45.7 23.8 70.8 1046 35+ years 23.4 11.6 42.2 20.7 70.6 317 BIRTH ORDER First 24.1 8,2 47.5 29.5 61.4 295 2-3 23.8 10.2 47.2 24.8 70.5 451 4-5 28.4 10.6 44,5 23.2 72.5 356 6+ 26.2 14.1 39.2 17.6 73.3 350 BIRTH INTERVAL < 24 months 25.7 9.1 42.5 24,2 69.1 324 24-35 months 27.5 12.8 45.4 21.9 73.6 421 36-47 months 24,2 13.2 44.9 19.6 75.3 190 48+ months 25,2 11.3 42.2 22.2 70.0 220 TOTAL 25.6 10.8 44.7 23.7 69.8 1452 Note: Figures are for children bom 1-59 months before the survey. 27 Figure 4.4 Reasons for Not Taking Children with Diarrhea to a Health Facility Diarrhea Dl~c~rh~ ~.~ Mother had mild 21% Mother had m ~ o time 15% ~ 0 time 8% No serv icerS / NO serv ice \~. ~/Other 30% near 2% ~ near 41% Other 40% Urban Rural Table 4.6 examines the treatments given at liealth facilities as reported by the mothers. Overall, 46 percent of the children with diarrhea who visited a health facility received ORS packets. In public facilities, this proportion rose to 59 percent, but in private facilities (doctors and private hospitals), only 21 percent received ORS packets. Syrups were the single most commonly given treatment: 54 percent in public facilities and 74 percent in private facilities. Injection was more common in private health facilities. Table 4.6 Treamaent of childhood diarrhea by type of health facility visited: Among children under five who had diarrhea in the two weeks preceding the survey, the percentage who received specific treatments by type of health facility visited (public/private), Bolivia ENDSA, I989 All Public Private health health health Type of treatment facilities facility facility ORS packets 46.1 59.2 20.7 Syrup 60.9 54.0 73.8 Tablets 19.2 2I .4 15.0 Injection 17.7 15.1 23.1 Intravenous fluids 0.7 0.5 1.0 Other 8.4 5.2 14.8 Nothing 1.6 1.0 2.6 NUMBER OF CHILDREN 343 227 115 Note: Figures are for children born 1-59 months before the survey who were taken to a health facility. 28 Knowledge and Preparation of ORS Packets Some 70 percent of the 3,563 mothers with children under five knew of ORS packets (Figure 4.5), including 9 percent who had never seen a packet, but had heard of it. Forty-three percent of the mothers had used ORS packets at least once to treat a child. Knowl- edge of ORS packets was substan- tially higher in urban areas and a- mong more educated women. Dif- ferences between the three regions of Bolivia were smaller, with the Llanos having slightly higher lev- els of knowledge of ORS. Wom- en were also asked if they had a packet in the house: 11 percent said they had a packet, but only 6 percent could show it to the inter- viewer (data not shown). Figure 4.5 Knowledge and Use of ORS Packets Among Women with Children Under Five RESIDENCE Urb&n Rural REGION AltlpIano Val~e Llanos MOTHER S EDUCATION No educat ion 15 yoara E8 yeJr l 9 • yeera Tota l ~9 ; 80 • 4 80 H 168 ~l i 70 ~t T8 PD i 47 168 ~lf L81 ~r i TM 20 40 66 80 100 Percent [ - - Gave ORG to Ch,ld [~K . . . . . boutORS i Bo l iv ia ENDSA 89 Almost two-thirds of the women wlao knew of ORS packets said that it was used to prevent dehy- dration, while about one-third thought they were given to cure diarrhea (Figure 4.6). A large majority, 87 percent, used one liter of water to prepare the solution, and fully 99 percent reported boiling the water before adding the contents of the ORS packet. One in ten women used the same batch of ORS solution for more than one day, but the majority prepared a new solution each day. About 1 in 5 women paid for the ORS packets, most of them 1 Boliviano (about U.S. $0.35) or more per packet. Figure 4.6 Use of ORS Packets and Cost of Packets among Women Who Know about ORS WHY ORS USED Prevent dehydrat ion Cure d iar rhea Both Don t know VOLUME OF WATER One l i te r Half • l i te r Other Don ' t know COST OF PACKET Rece ived t ree Pa id under 1 Bol PIId 1 SoL or more s6 ~4 ~2 87 B6 ms I 81 20 40 60 Percent 80 100 Bol~vla BNDSA-E9 29 Virtually all the women who had heard of ORS packets knew of a source for them (see Figure 4.7). The most frequently mentioned sources were public health facilities (hospitals, health centers, and health posts) and pharmacies. The community health worker (RPS) was mentioned by about 7 percent of the respondents. Figure 4.7 Sources of ORS Packets among Women Who Know about ORS Public hospltel Health center Health po l t Private olJnic/holp Private doctor ORT unit Pharmacy Oomm, health worker Other 23 2O 2O 18 W3 21 m 7 6 10 18 20 26 30 Note: Percenteg8 of women who mentioned each Iouroe; multJpte Teaponlal allowed. Bolivia ENDSA-89 30 CHAPTER 5 MORBIDITY: RESPIRATORY INFECTIONS AND THEIR TREATMENT Acute respiratory infection, (cough accompanied by breathing problems) occurred among 20 percent of children under five years of age during the two weeks preceding the survey. The prevalence rate was about 25 percent among children under two years and 14 percent among children four years of age. There were large differences in the proportion of sick children for whom health care was sought according to socioeconomic and demographic characteristics. Background variables associated with lower socioeconomic status were correlated with lower rates of health care utilization and lower treatment rates. For example, half of the children whose mothers had nine or more years of schooling were taken for treatment to a health facility or provider, compared with only one in eight children whose mothers had no education. 5.1 Methodology Acute lower respiratory infection (ALRI), primarily pneumonia, is a common cause of morbidity and death among young children in Bolivia. In an effort to reduce the morbidity and mortality caused by ALRI, the govenunent of Bolivia has embarked on a program for early diagnosis and treatment of these infections, following the guidelines suggested by the World Health Organization (1990a). An important element of the ALRI control program is the early diagnosis of pneumonia in children and prompt initiation of antibiotic therapy. As part of this program, peripheral health workers are taught to identify children who are likely to have pneumonia, using a few simple clinical signs such as cough accompanied by a rapid respiratory rate and difficulty breathing. Although only some of the children with cough and rapid or difficult breathing have pneumonia, while the remainder have an upper respiratory infection (i.e., the common cold), all these children should be evaluated by a health worker, and antibiotic therapy should be begun for those who fulfill specified diagnostic criteria. It is crucial to have estimates of the prevalence of cough accompanied by difficult or rapid breathing in order to assess the demands which will be placed on the health care services as a result of the new ALRI control programs and 1o help in manpower planning. The questions about cough and other symptoms of respiratory infection in the ENDSA questionnaire aim to provide such an estimate? The respondents were asked whether their children under five years had been sick with a cough during the 24 hours preceding the survey. If there was no cough during the last 24 hours, the mother was asked how much time had elapsed since the last time the child was sick with a cough. The answer was recorded in number of days, weeks or months. For all children with a cough, the mother was asked whether the cough was accompanied by difficult breathing, and whether the child had been breathing fastcr than usual. The duration of the cough and of the difficult breathing was also ascertained and recorded in numbcr of days. The questions about difficult and rapid breathing were asked in order to provide an indication of the severity of the illness. All mothers whose children had been ill with a cough were also asked whether the child had been taken anywhere for treatment and what type of treatment the child had received when he or she was sick. 5.2 Findings Prevalence Tables 5.1 and 5.2 show the prevalence of cough according to socioeconomic and demographic characteristics. The results are shown separately for children who had a cough accompanied by difficult Other studies indicate that mothers are able to report the rapid breathing and dyspnea (difficulty breathing) which is associated with pneumonia in children (Campbell et al., 1988; Cherian et al., 1988). 31 Table 5.1 Prevalence of cough mad cough with breathing problems by socio- economic characteristics: Percentage of children under five with cough and cough with breathing problems in the two weeks preceding the survey, by socioeconomic characteristics, Bolivia ENDSA, 1989 Cough with Number Cough breathing Any of Characteristic only problems cough children RESIDENCE Urban 17.0 21.2 38.2 2535 Rural 24.4 19.6 40.9 2626 REGION Ahiplano 21.7 17.0 38.7 2603 Valle 26.4 14.5 40.9 1422 Llanos 11.5 35.6 47.1 1137 MOTHER'S EDUCATION None 24.0 17.3 41.3 1061 1-5 years 19.8 22.0 41.8 2354 6-8 years 21.5 24.5 46.0 706 9+ years 19.1 17.2 36.3 1040 FATHER'S OCCUPATION White collar 17.9 22.1 40.0 1447 Blue collar 22.1 21,5 43.6 1495 Agriculture 21.9 18.6 40.5 1860 Other 20.6 18,6 39.2 360 MOTHER WORKING Yes 21.2 17,0 38.2 983 No 20.6 21,2 41.8 4178 MEDIA EXPOSURE TV daily 19.5 20.9 40.4 2242 Radio daily 17.8 21.6 39.4 1694 Neither 27.2 18.0 45.2 1226 LANGUAGE Spanish 18.3 23.2 41.5 3663 Indian 26.9 13.7 40.6 1499 DRINKING WATER Piped 19.7 20.5 40.2 3101 Non-piped 22.4 20.3 42.7 2060 SANITATION Flush toilet 18.4 20.7 39.1 1343 Latrine 16.9 29.8 46.7 892 Other. none 23.0 17.4 40,4 2926 TOTAL 20.8 20.4 41.2 5161 Note: Figures are for children born 1-59 months before the survey, Breathing problems refers to rapid and/or difficult breathing. 32 Table 5.2 Prevalence of cough and cough with breathing problems by demographic characteristics: Percentage of children under five with cough and cough with breathing problems in the two weeks preceding the survey, by demographic characteristics, Bolivia ENDSA, 1989 Cough with Number Cough breathing Any of Characteristic only problems cough children SEX Male 20.8 21.0 41.8 2582 Female 20.7 19.8 40.5 2580 CHILD'S AGE 0-5 months 16.2 22.9 39.1 483 6-11 months 27.1 25.9 53.0 572 12-23 months 23.9 26.0 47.9 1110 24-35 months 21.1 19.2 40.3 1017 36-47 months 17.7 17.9 35.6 993 48-59 months 18.5 13.5 32.0 987 MOTHER'S AGE 15-19 years 24.5 25.3 49.8 256 20-34 years 20.3 21.0 41.3 3616 35+ years 21.3 17.7 39.0 1289 BIRTH ORDER 1 22.2 20.0 42.2 1053 2-3 21.2 19.1 40.3 1742 4-5 19.5 21.7 41.2 1188 6* 20.1 21.5 41.6 1179 BIRTH INTERVAL <24 months 22.7 22.6 45.3 1093, 24-35 months 20.4 21.7 42.1 1429 36-47 months 17.6 17.7 35.3 704 48+ months 19.8 18.2 38.0 877 TOTAL 20.8 20.4 41.2 5161 Note: Figures are for children born 1-59 months before the survey. Breathing problems refers to rapid and/or difficult breathing. and/or rapid breathing and for children who only had a cough. The prevalence rate refers to all children who were reported to have had a cough during the two weeks preceding the survey, including the children with a cough in the past 24 hours. Approximately half of aU the children with cough were also reported to have had breathing problems (i.e., difficult and/or rapid breathing). Only a few differences from this pattern emerge when the prevalence rates are examined according to socioeconomic characteristics. A higher proportion of cough episodes are accompanied by breathing problems among children under six months of age, from the Llanos region, and from households without a latrine. A lower proportion of breathing problems is found among children from the Valle region, from families with least media exposure, and from families reporting an Indian language as their primary language. 33 The following discussion of differences in prevalence rates refers to the children who were reported to have had both a cough and breathing problems. As seen in Tables 5.1 and 5.2, 20 percent of all children under five years were reported to have had a cough accompanied by breathing problems during the two weeks preceding the survey. Three-quarters of these children were ill during the 24 hours preceding the survey (data not shown). There were differences in prevalence by the age of the child, with illness reported for about 25 percent of children un- der two years of age compared with only 14 percent of children four years of age (Figure 5.1). There were also marked regional differences, with a prevalence rate of 36 percent among children from the Llanos compared with 15 percent in the Valle and 17 percent in the Altiplano. There was also a higher reported prev- alence rate among children from families with a latrine as opposed to those with either a flush toilet or no facilities. Lower preva- lence rates were reported among respondents who spoke an Indian language as their primary tongue Figure 5.1 Prevalence of Cough with Breathing Problems among Children Under Five PliD rCO n| 301 251 2ol 151 lol 5[ oL c6 6-11 12-23 24-35 36-47 48-59 Age of Chi ld (Months) Note: Inoludee children elck In the two wQekl preceding the eurvey. Bolivia ENDSA'89 than among Spanish-speakers. There were only small differences in prevalence for the remaining variables, including sex, urban-rural residence, and the mother's level of education. Treatment Children with cough and breathing problems were much more likely to have been in contact with the health system (29 percent, see Tables 5.3 and 5.4) than children reported to have only cough (16 percent, data not shown). Figure 5.2 shows the proportion of children who were taken somewhere for treatment according to whether the child had cough only, or cough with breathing problems. Pharmacies and traditional practitioners were rarely mentioned as sources of care. Overall, public hospitals were visited most frequently, followed by private physicians, health centers, and private hospitals. Treatment pattems vary markedly by socioeconomic characteristics (Table 5.3). Children of mothers with higher levels of education were much more likely to be taken for treatment to a health facility than children of mothers with little or no education. Figure 5,3 shows that among children with cough and breathing problems, there is a strong positive relationship between the level of the mother's education and the utilization rates of health services. Half the children of mothers with the highest level of education (nine or more years of schooling) were taken for health care, compared with only 12 percent of the children of mothers with no education. More urban children receive health care for cough and breathing problems (37 percent) than do children from rural areas (20 percent). An even greater differential is observed according to the respondent's primary language: among those who usually speak one of the Indian languages, only 7 percent of the sick children were taken for care, compared with 34 percent of those who usually speak Spanish. 34 Table 5.3 Place of treatment mad type of treatment for children with cough and breathing problems by socioeconomic characteristics: Among children under five with cough end breathing problems in the two weeks preceding the suwey, the percentage who were taken to a health facility and the type of treatment received, by socioeconomic characteristics, Bolivia ENDSA, 1990 Characteristic Percentage Place of treatment Type of treatment taken to Numbor a health Public Health Physi- Private of facility a hospital center cien hospital Injection Tablets Syrup children RESIDENCE Urben 37.3 11.7 10.5 11.1 4.1 16.3 6.8 32.3 537 R~ai 19.8 9.8 4.5 4.9 0.6 6.7 6.9 17.1 515 REGION Altiplano 23.8 6.7 7.8 6.9 2.5 8.1 7.4 17.9 441 Valle 28.0 16.4 4.1 5.9 1.7 12.4 9.1 24.6 207 Llanos 34.5 12.3 9.2 10.5 2.5 15.0 5.2 32.6 405 MOTHER'S EDUCATION None 12.4 8.3 3.6 0.2 0.4 6.8 3.6 10.1 184 1~5 years 25.1 11.4 5.7 6.0 2.0 10.8 5.7 22.4 517 6-8 years 33.5 11.6 10.6 8.8 2.4 12.1 11.4 27.7 173 9+ years 51.4 10.5 14.4 21.5 5.3 18.2 9.3 44.5 179 FATHER'S OCCUPATION White collar 41.8 11.1 12.2 13.5 5.2 15.5 8.8 36.3 320 Blue collar 33.5 12.8 8.4 10,2 2.2 14.9 8.1 28.5 321 Agriculture 13.9 8.1 3.5 2.0 0.3 5.8 3,9 11.7 345 Other 20.3 13.5 3.4 3,5 0.0 6.4 6.4 20.7 67 MOTHER WORKING Yes 38.0 16.8 8.3 12,1 1.1 15.1 9.2 34.7 167 No 27.0 9.6 7.5 7,3 2.6 10.9 6.4 23.0 886 MEDIA EXPOSURE TV daily 39.3 11.2 10.8 13.2 4.2 14.7 7.7 33.3 467 Radio daily 25.8 13.2 6.8 5.2 0.6 12.1 8.0 21.4 365 Neither 11.2 5.8 2.1 1.8 1.5 4.0 3.1 12.7 220 LANGUAGE Spanish 34.0 11.8 9.4 10.0 2.9 13,6 7.6 29.7 848 Indian 7.0 6.6 0.3 0.0 0.0 3.0 3,6 4.9 205 DRINKING WATER Piped 35.9 13.3 9.4 10.2 3.2 13.9 8.3 30.0 635 Non-piped 17.8 6.9 4.9 4.9 1.1 8.0 4.7 17.1 418 SANITATION Flush toilet 44.7 11.8 10.6 19.2 3,1 18.6 6.8 38.9 279 Laixine 34.5 13.1 9.0 10.1 2.3 15.6 5.8 32.0 155 Other, none 17,0 9.0 5.2 0.9 2.0 5.6 7.4 13.5 509 TOTAL 28.7 10.8 7.6 8.1 2.4 11.6 6.9 24.9 1053 Note: Figures axe for children born 1-59 months before survey. Breathing problems refers to rapid end/or difficult breathing. apublic or private hospital, health center, end/or private doctor 35 Table 5.4 Place of treatment and type of treatment for children with cough and breathing problems by dem- ographic characteristics: Among children under five with cough and breathing problems in the two weeks preceding the survey, the percentage who were taken to a health facility and the type of treatment received, by demographic characteristics, Bolivia ENDSA, 1990 Characteristic Percentage Place of treatment Type of treatment taken to Number a health Public Health Physi- Private of facility a hospital center cian hospital Injection Tablets Syrup children SEX Male 28.7 11.9 7.2 7.9 1.7 11.6 7.2 26.5 542 Female 28.8 9.5 8.0 8.2 3.1 11.6 6.5 23.1 511 CHILD'S AGE 0-5 months 32.2 6.9 8.1 12.5 4.6 10.5 8.3 25.2 111 6-11 months 35.2 13.5 10.7 8.5 2.4 13.2 6.7 26.4 148 12-23 months 26.2 10.4 6.5 7.6 1.9 13.4 4.4 24.3 288 24-35 months 29.4 9.4 11.8 5.0 3.2 9.3 10.7 27.3 195 36 47 months 29.0 12.7 4.8 10.2 1.3 9.4 8.0 28.5 178 48-59 months 22.9 11.0 3.7 6.7 1.6 12.9 3.8 15.9 133 MOTHER'S AGE 15-19 years 38.0 21.2 8.6 8.3 0.0 9.9 10.3 33.0 65 20-34 years 30.6 10.5 7.6 9.8 2.8 12.4 7.1 26.3 760 35+ years 19.8 8.7 7.3 2.2 1.6 9.2 4.9 18.0 228 BIRTH ORDER First 36.1 15.0 7.8 12.5 0.8 9.1 9.2 31.9 210 2-3 31.7 10.0 9.3 8.8 3.6 14.4 6.1 25.5 332 4-5 29.3 10.6 7.1 9.1 2.7 12.9 6.7 26.6 258 6+ 18.3 8.4 5.8 2.4 1.7 8.9 6.0 16.5 253 BIRTH INTERVAL <24 months 26.2 13.0 7.1 4.0 2.1 12.7 6.0 25.7 246 24-35 months 26.8 8.9 6.7 9.0 2.4 11.1 7.8 22.4 310 36-47 months 22.5 4.6 7.6 6.8 3.5 11.8 6.0 16.4 125 48+ months 32.0 10.1 9.9 7.9 4.0 14.0 3.9 26.2 159 TOTAL 28.7 10.8 7.6 8.1 2.4 11.6 6.9 24.9 1053 Note: Figures are for children bern 1-59 months before survey. Breathing problems refers to rapid and/or difficult breathing. apublic or private hospital, health canter, and/or private doctor. 36 Figure 5.2 Health Care Received by Children with Cough by Type of Cough Public hospital Health center Private doctor Private holpltal Pharmacy Trad. practitioner Other ~ J J J J J J J J J J J J J J J J J J J~ ~ J J J f J J J J~ ~ J J J J J J J J J J J J~ 2 4 6 8 10 12 14 Percentage Who Received Care [mCough *SteEthJng Prb. ~ Cough onJy Note: Include: children sick in the two wwek| preceding the survey. Bo]lvl8 ENDSA-ES Figure 5.3 Health Care Received by Children with Cough and Breathing Problems by Characteristics of Mother Percantloe Who Received Care eO SO 40 30 EO 10 O 34 3;' 34 20 25 indian 8psnlah Rural Urban None 1-5 8-8 E- LANGUAGE RESIDENCE MOTHER'S EDUCATION (YR8) b m Public Ho|pltal E Health Center | I [~ Doctor E Private Holpltal Note: Include! children sick in the two weekl preoeding the eurvey, Bolivia ENDSA-89 37 Differences according to demographic characteristics are less marked (Table 5.4). A smaller proportion of children of mothers age 35 years or older and of bil~'a order 6 and higher were taken for treatment. A positive correlation was found between the level of the mother's education and the percentage of children with a cough and breathing problems who received treatment (Figure 5.4). Higher treatment rates were observed for urban compared with rural children and for chil- dren from families in which Spanish was the primary lan- guage. Ideally, children with signs and symptoms of ALRI, or pneumonia, should be treated with antibiotics, since bacteria are the most common cause of fatal pneumonia. However, moth- er's recall may not be adequate to determine the proportion of chil- 50 40 30 20 10 O Figure 5.4 Treatment Received by Children with Cough and Breathing Problems by Characteristics of Mother Percentage Who Received Treatment Indian Spanish Rural Urban None 1-5 6°8 9. LANGUAGE RESIDENCE MOTHER'S EDUCATION (YRS) I ~1 Tablet, ~ ,njection ~Byrup i Note: Includel children l ick in the two weeks preceding the survey: multiple reaponsel allowed. Bolivia BNDBA-B9 dren who were given antibiotics. While injections are likely to represent treatment with antibiotics, other types of medication are often given as syrups or tablets. The type of treatment differed depending on the type of cough (Figure 5.5). Among children with both cough and breathing difficulties, 12 percent received an injection; this figure was just 5 percent for children with cough only. Tablets were also given more frequently to children with breathing problems (7 percent) than to those with cough only (3 percent). Figure 5.5 Specific Treatment Received by Children with Cough by Type of Cough Percentage Who Received Treatment Cough Only Cough wilh Breathing Problem ] mTab lets ~ln Jec t lon ~Byrup ~HomeRemady ~Othe , I Bolivia ENDBA-8g 38 There was little difference among the various health facilities in the types of treatment given to children with a cough accompanied by breathing problems (Table 5.5 and Figure 5.6). However, among children with cough only, those who were taken to a public hospital or health center were much more likely to have received an injection (41 percent) than were children taken to a private hospital or doctor (17 percent). Table 5.5 Type of treatment received for respiratory illness in children according to the place where treatment was sought: Among children under five with cough, the percentage who received specific treatment by place of treatment, end severity of the cough, Bolivia ENDSA, 1989 No Number Injec- Home treat- of chil- Place of treatment tlon Tablets Syrup remedy Other merit then COUGH ONLY Public hospital 52.4 17.7 75,0 5.5 3.5 0.0 66 Health center 18.7 31.0 62,8 0.0 8.7 0.0 35 Physician 20.7 17.1 74.7 3.0 13.3 2.4 47 Private Hospital 8.6 18.1 98.9 0.0 2.5 0.0 19 Pharmacy b b b b b b b Tradil~onal practitioner b b b b b b 2 Other b b b b b b 12 COUGH WITH BREATHING PROBLEMS a Public hospital 38.7 25.1 77.9 1.7 6.3 0.0 113 Health center 36.5 23.5 72.1 4.3 17.7 0.0 80 Physician 40.4 11.4 80.3 2.2 18.6 0.0 85 Private hospital 35.6 10.8 84.5 0.0 20.6 2.5 25 Pharmacy b b b b b b 17 Traditional practitioner b b b b b b 7 Other b b b b b b 17 aCough accompanied by rapid end/or difficult breathing bLess than 20 cases 39 Figure 5.6 Specific Treatment Received by Children with Cough by Type of Cough and Type of Facility Peraentage Who Received Tre~ttment public o Pr l¥ | te* , COUGH ONLY 81 7o publl¢o private** COUGH WITH BREATHING PROBLEM8 ~ 1~ Tlblet¢ ~ Injection [~ Syrup] • Public hospital or dOCtor • * private hospital or health center Bolivia ENDSA-89 40 CHAPTER 6 NUTRITIONAL STATUS Undernutrition is common among children in Bolivia. Thirty-eight percent of children age 3-36 months were classified as stunted (very short for their age), while 13 percent were considered underweight (low weight-for- age). Wasting (low weight-for-height), was not found to be a problem among children in any subgroup. Substantial deterioration in nutritional status occurred during the first two years of life. The prevalence of stunting increased markedly both during the first and second years of life. While less than 10 percent of the youngest children were stunted, about one-fifth of those age 6-12 months and half of the children two years of age were stunted. There were large differences in the prevalence of undernutrition according to socioeconomic and demographic characteristics, Forty-five percent of rural children were stunted, compared with 32 percent of urban children. The differences were even greater for mother's education and increased with age: only a small percentage of the youngest children (3-5 months) were stunted (regardless of mother's education); however, among children 33-36 months, 27 percent of those whose mothers had nine or more years of education were stunted and 79 percent of those whose mothers had no education. 6.1 Methodology All children age 3 through 36 months were weighed and measured by specially trained personnel who accompanied the interviewer teams during the survey. They were taught to obtain the child's weight to within 100 grams using a spring balance scale and to measure the child's height to within 5 mm using a measuring board. Although this report uses the term "height," the children were actually measured lying down on an adjustable wooden board as recommended by the United Nations (1986). Emphasis was also placed on obtaining the correct year and month of birth for the children. Since age misreporting can result in the misclassification of a child as either malnourished or well-nourished (depending on the direction of the error), the interviewers were trained to obtain the correct birth date. Children without a reported month of birth were excluded from this analysis. For purposes of comparison, the anthropometric data are expressed in terms of the reference population recommended by the World Health Organization (WHO)? Use of the reference population allows a comparison between children of different ages from different subgroups within the population and facilitates comparison with other survey data. Work by Martoren and Habicht (1986) has shown the applicability of this international reference population to infants and young children from many different populations. As recommended by WHO, the following three indices are used to describe the child's anthropometric status: height-for-age, weight-for-age, and weight-for-height. In the case of the first two indices, the child's measurement is compared with the reference population for the appropriate age and, if it is more than two standard deviations below the median for the reference population, the child is classified as undernourished. The term "stunted" is applied to children whose height-for-age falls in this range (i.e., children who are very short), while "underweight" is the term used to describe children whose weight-for-age falls below this cut-off point. For the third indicator, the child's weight is related to his or her height (weight-for-height), and if the ratio is more than two standard deviations below the median for Developed by the U.S. Centers for Disease Control based on data from the U.S. National Center for Health Statistics (Dibley et al., 1987a; Dibley et al., 1987b). 41 the reference population, the child is classified as "wasted" (i.e., very thin). In a well-nourished population such as the reference population only 2.3 percent of the infants and young children are more than two standard deviations below the median. Hence, the percentage of children who are below the cut- off point gives an indication of the degree of undemutrition in a population or in a subgroup of children. Stunting reflects chronic undemutrition, while wasting is often used as an indicator of acute undemutrition. Weight-for-age is a composite measure which captures both types of undernutrition. Undernutrition is often caused by inadequate nutritional intake and by recurrent bouts of infectious disease. 6.2 Findings Levels of Undernutrition The anthropometric status of 2,536 children age 3 through 36 months is included in this analysis; this is 83 percent of the children in this age group. The remaining 17 percent were not included for the following reasons: the child was not available when the measurer visited the house, only weight or only height was recorded, or the height or weight recorded was grossly improbable, due either to a measurement or recording error. The survey showed no evidence indicating misreporting of birth dates or age heaping (on commonly preferred months such as 6, 12, or 24). The children were evenly distributed throughout the age group with the exception of the first and the last months, which had somewhat smaller numbers of children. When the height-for-age and weight-for-age of young children in Bolivia are related to the reference population, it is clear that a large percentage of the children are more than two standard deviations below the median (Table 6.1 and Figure 6.1). Fully 38 percent of the children were stunted, and 13 percent were classified as underweight. However, they were not thinner than the children in the reference population, since less than 2 percent were below the cut-off point for weight-for-height. Figure 6.2 shows how nutritional status deteriorates with increasing age in the first two years of life. The percentage of children classified as stunted increased steadily from less than 10 percent of children 3-6 months of age to half of all children as they approached their second birthday; thereafter, it plateaued. The percentage of children classified as underweight increased from less than 2 percent for the youngest children to one-fifth of children in the beginning of their second year of life. During the remainder of that year, the proportion of underweight children remained fairly constant, and at the end of the second year a decline commenced which continued into the third year. The deterioration in nutritional status appears to halt earlier in terms of weight (weight-for-age) than in terms of linear growth (height- for-age). Wasting was uncommon for children of any age. Anthropometric data were collected in six DHS surveys in Latin American and the Caribbean during the period 1986-89. The proportion of children 3-35 months classified as stunted was higher in Bolivia than in northeast Brazil (29 percent), Colombia (25 percent), the Dominican Republic (21 percent) and Trinidad and Tobago (5 percent). However, Guatemala had the highest level of stunting overall (58 percent). 42 Table 6.1 Prevalence of undemutIition by socioeconomic characteristics: Percentage of children age 3-36 months who were classified as stunted, underweight, or wasted by socioeconomic characteristics, Bolivia ENDSA, 1989 Percentage of children classified as: Characteristic Stunted Underweight Wasted Number of children RESIDENCE Urban 31.5 10.7 1.2 1273 Rural 45.0 15.9 2.0 1264 REGION Alfiplano 44.0 13.4 1.6 1257 Valle 40.7 16.1 1.5 736 Llanos 21.4 9.5 1.7 544 MOTHER'S EDUCATION None 56.0 23.1 2.8 456 1-5 years 41.9 13.1 1.1 1197 6-8 years 29.7 11.3 1.0 373 9+ years 19,9 6.6 2.2 511 FATHER'S OCCUPATION White collar 26.2 8.8 1.0 698 Blue collar 39,5 12.6 1.6 774 Agriculture 48,7 18.1 2.0 862 Other 30.3 11.4 1.8 202 MEDIA EXPOSURE TV dally 28.1 8.3 1.1 1136 Radio daily 44.2 14.9 1.3 826 Neither 49.7 21.0 3.1 574 LANGUAGE Spanish 32.6 10.0 1.5 1874 Indian 54.1 22.8 2.0 662 DRINKING WATER Piped 34.8 10.6 1.2 1551 Non-piped 43.5 17.6 2.2 986 SANITATION Flush toilet 19.5 6.7 1.7 650 Latrine 30.8 12.1 1.4 459 Other, none 49.1 16.7 1.6 1427 TOTAL 38.2 13.3 1.6 2536 Note: Children were classified as stunted, underweight, or wasted when their z-score was 2 or more standard deviations below the median for the WHO/CDC/NCHS International Reference Population (for height-for-age, weight-for-age, or weight- for-height), 43 Percent 25 [ F igure 6.1 Nutritional Status of Children 3-36 Months 2O 'ol ,J -6.76 "5,00 -4.00 -3.00 -2.00 "tOO O.OO "1.00 *2,00 *3>00 *4.00 *5.00 *5.76 Z-ioore I - -~ Height-for-Age ~ Weight-for-Age i Weight-for-Height ~ International Ref, Note: Z- lcore I I the number of standard d lwl i t lon units from the median of the Internetlonll Reference Populstlon Bolivia ENDSA-SI Figure 6.2 Percentage of Children 3-36 Months Classified as Undernourished by Age Percent (moving average) 60 • . . . . . . . / 3 6 9 12 16 18 21 24 27 30 33 36 Age (Months) = i i ~ Stunted - - Underweight + Wasted Note: Undernourished children ere those tWO or more standard deviation= below the median of the International Ref i renoe Population. Bolivia ENDSA-Si Table 6.2 compares the results of the ENDSA with a national survey carried out in 1981. The age groups considered are largely comparable. The results of the two surveys are very similar, indicating that the nutritional status of children under three years of age has remained the same in Bolivia during the 1980s, in both urban and rural areas. The only change is a decrease in the proportion who were underweight in rural areas from 21 to 16 percent. 44 Table 6.2 Prevalence of undemuUition in 1981 and 1989 by residence: Percentage of children who were stunted, underweight, or wasted in 1981 (children age 6-36 months) and 1989 (children age 3-36 months) by residence, Bolivia Place of Under- Number of residence Year Stunted weight Wasted children URBAN 1981 29.7 11.5 1.1 1597 1989 31.5 10.7 1.2 1273 RURAL 1981 44.5 20.7 1.0 1886 1989 45.0 15.9 2.0 1989 Note: Children were classified as stunted, underweight, or wasted when their z-score was 2 or more standard deviations below the median for the WHO/CDC/NCHS International Reference Population (for height-for-age, weight-for-age, or weight- for-height). Source of data for 1981: National Institute of Food and Nutrition, Government of Bolivia; and USAID Bolivia National Nutritional Status Data Survey, 1981: Summary Report, Washington, D.C. 1982. Socioeconomic and Demographic Differentials For most of the socioeconomic and demographic variables, there were marked differences in the prevalence of undemutrition (Tables 6.1 and 6,3). In most instances, the same pattem was found for both stunted and underweight children. The two exceptions were the variables for age of child (discussed above), and the region in which the respondent lived. There was little evidence of wasting in any of the subgroups. All the demographic variables demonstrated the expected pattern. Higher maternal age was associated with increasing undemutrition: 30 percent of children whose mothers were 15 to 19 years were stunted compared with 43 percent of children whose mothers were 35 years or older. A similar pattem was seen for the percentage of children classified as underweight. As for the length of the preceding birth interval, it was found that children born 48 months or more after the previous child had the lowest rates of undemutrition, with 29 percent classified as stunted and 9 percent as underweight. Over 40 percent of children bom 24 to 47 months after the previous child were stunted, and almost 15 percent were underweight. Birth order shows similar results: nutritional status was generally worse among children of higher birth order. Thirty-one percent of firstborn children were categorized as stunted and 12 percent as underweight; among children with a birth order of six and higher, 45 percent were stunted and 18 percent were underweight. Both types of undemutrition were slightly less common among girls than among boys. There were also differences in the prevalence of undemutrition according to the place of residence. Urban children fared better then their rural counterparts, with 31 percent stunted and 11 percent underweight, compared with 45 and 16 percent, respectively, in rural areas. Regional differences were even greater, with children from the Llanos having the lowest rates of undernutrition. The other two regions had similar overall rates of undemutrition, but the components differed slightly, with the highest 45 Table 6.3 Prevalence of undernutrition by demographic characteristics: Per- centage of children age 3-36 months who were classified as stunted, underweight, or wasted, by demographic characteristics, Bolivia ENDSA, 1989 Percentage of children classified as: Characteristic Stunted Underweight Wasted Number of children SEX Male 40.2 15.0 1.7 1257 Female 36.3 11.7 1.5 1280 CHILD'S AGE 3-5 months 7.6 1.5 2.8 194 6-11 months 20.3 8.6 1.2 498 12-23 months 42.4 19.1 2.1 985 24-36 months 50.7 12.1 1.0 859 MOTHER'S AGE 15-19 years 30.5 13.7 1.6 182 20-34 years 37.5 12.1 1.7 1809 35+ years 43.1 17.2 1.2 546 BIRTH ORDER First 31.4 11.9 1.8 533 2-3 35.1 10.5 1.7 819 4-5 42.3 13.6 1.3 598 6+ 44.6 18.3 1.7 586 BIRTH INTERVAL <24 months 43.8 16.9 1.2 523 24-35 months 42.2 14.0 2.0 709 36-47 months 44.0 13.9 0.9 343 48+ months 28.7 9.3 1.7 426 TOTAL 38.2 13.3 1.6 2536 Note: Children were clmsified as stunted, underweight, or wasted when their z-score was 2 or more standard deviations below the median for the WHO/ CDC/NCHS International Reference Population (for height-for-age, weight-for- age, or weight-for-height). proportion of stunting (44 percent) found among children in the Altiplano and the highest proportion of underweight children (16 percent) in the VaUe. There was a much higher rate of undernutrition among children whose mothers spoke an Indian language as their primary tongue than among those whose mothers spoke Spanish. For the variables reflecting socioeconomic level, higher rates of undemutrition were found for children at lower socioeconomic levels. Over half (56 percent) of the children whose mothers had received no education were stunted, compared with just 20 percent of children whose mothers had attended nine or more years of school. The corresponding percentages for children classified as underweight were 23 and 7 percent, respectively. Not surprisingly, the percentage of children who were undernourished was lowest among the group who watched television daily and highest among those who were exposed to neither radio nor television. With regard to the father's occupation, the lowest 46 percentage of undemutrition was found among children of white collar workers and the highest among those listing agriculture. There were lower rates of undemutrition when the drinking water was piped and a flush toilet was available. In a further analysis, the extent of stunting was examined both by the child's age and by the mother's level of education (Figure 6.3). The percent of children classified as very short for their age was relatively low for the youngest age group at all levels of education. However, there were marked differences as the children grew older, with almost four of five children in the oldest age group classified as stunted when the mother had received no education, compared with only about a quarter of the children whose mothers had gone to school for nine or more years. Several of the demographic and socioeconomic variables discussed above are associated, and a more extensive analysis is needed to elucidate the causative relationships (Bicego and Boerma, 1990). Percent Figure 6.3 Percentage of Children 3-36 Months Classified as Stunted by Age and Mother's Education None 1-5 Year= 6-8 Years 9* Yearl Mot her's Education I BIB 3-5 Months ~ 11-13 Months ~ 33-36 Months i Note= Stunted children are those two or more standard deviations below the International Reference, Bolivia ENDSA-89 47 CHAPTER 7 INFANT FEEDING Almost all children (96 percent) are breastfed in Bolivia. Overall, the median duration of breast.feeding was 17 months although there was substantial variation between population subgroups. For example, the median duration was 12 months for infants whose mothers had 9 or more years of schooling, compared with 21 months among infants of mothers lacking education. Over half of the infants were breastfed exclusively for the first 3 months of life. By the end of the first year, about 60 percent of infants were receiving both breast milk and solid or mushy foods. At the time of the survey, 43 percent of children age 18-23 months were still breastfed. Powdered milk, a common breast milk supplement, was given to 20-25 percent of breastfed children. One-third of all breastfed children were also bottle fed. 7.1 Methodology The respondents were asked whether their child(ren) had ever been breastfed and, if so, for how long. The reasons for never breastfeeding and for stopping breastfeeding were also determined. These questions were asked with regard to each of the respondent's children born during the five years preceding the survey, regardless of whether the child was alive at the time of the interview. For last-bom children the current status regarding breasffeeding was recorded, and the following information was collected about breastfeeding practices during the 24-bour period before the interview: 1) the number of times the child was fed at the breast, 2) the types of supplemental foods and liquids given, and 3) the number of times each of these foods and liquids were given. The interviewer inquired about the following specific items: plain water, water with herbs, juice or sugar water, powdered milk, cow's or goat's milk, other liquids, and solid food. The mothers were also asked if any of these were given in a bottle with a nipple. The median duration of breastfeeding was calculated based on the current breastfeeding status of children alive at the time of the survey and was defined as the age when 50 percent of children did not receive any breast milk. 7.2 Findings Prevalence and Duration of Breastfeeding Fully 96 percent of all babies in Bolivia were breastfed for a period of time (ever breastfed). There were negligible differences in the percentage of children ever breastfed according to socioeconomic and demographic characteristics (Tables 7.1 and 7.2). Based on current status data, the median duration of breastfeeding was estimated at 17 months, but it differed considerably between various subgroups. For example, the median duration was only 12 months for children whose mothers had nine or more years of education, compared with 21 months among those whose mothers lacked education. Similar, or slightly smaller differentials were found for the other socioeconomic characteristics that were examined (Table 7.1). Longer median durations of bmastfeeding, of 18 to 20 months, were found in rural areas, in the Altiplano, among families not exposed to television and radio, and among respondents whose primary tongue was an Indian language. Shorter median durations, of about 16 months, were found among urban dwellers, families with television, white collar workers, and respondents who spoke Spanish. 49 Table 7.1 Initiation and duration of bre~ffeeding by socioeconomic characteristics: Percentage of children under five ever breastfed and median duration of breastfeeding by socioeconomic characteristics, Bolivia ENDSA, 1989 Characteristic Median Percentage Percentage duration of <4 months ever breasffeeding breast fed breastfed (months) exclusively RESIDENCE Urban 95.1 15.4 46.5 Rural 97.5 18.1 65.4 REGION Ahiplano 97.0 19.7 73.2 Valle 96.8 16.4 51.8 Llanos 94.2 13.2 30.8 MOTHER'S EDUCATION None 97.7 20.6 68.9 1-5 years 96.8 18.3 60.5 6-8 95.1 16.0 51.3 9 or more 94.4 11.7 40.7 FATHER'S OCCUPATION White collar 95.6 15.9 41,5 Blue collar 95.7 18,8 64.6 Agriculture 97.3 17.1 65.1 Other 96.7 17.0 49.6 MEDIA EXPOSURE TV daily 95.7 15.6 49.5 Radio daily 96.7 17.5 61.6 Neither 96,9 20.0 63.2 LANGUAGE Spanish 95,7 15.8 47.8 Indian 97.6 19.9 76.3 DRINKING WATER Piped 96.1 16.1 57.2 Non-piped 96.7 18.0 55.4 SANITATION Flush toilet 95.1 14.3 45.5 Latrine 94.9 15.0 40.7 Other, none 97.4 19.3 66.9 TOTAL 96.4 16.9 56.4 NUMBER OF CHILDREN 5731 5161 308 Note: Ever-breasffed status refers to all children born in five years preceding the survey (including deceased children). Median duration is based on the current breastfeeding status of living children born 1-59 months before the survey. 50 Table 7.2 Initiation mad duration of breastfeeding by demographic charac- teristics: Percentage of children under live ever breastfed and median duration of breastfeeding by demographic characteristics, Bolivia ENDSA, 1989 Characteristic Median Percentage Percentage duration of <4 months ever breasffeeding breast fed breasffed (months) exclusively SEX Male 95.6 17.2 54.9 Female 97.0 16.6 58.2 MOTHER'S AGE 15-19 years 94.6 16.9 56.4 20-34 years 96.6 16.4 57.7 35+ years 95.8 21.1 50.5 BIRTH ORDER First 95.7 14.4 54.2 2-3 96.0 17.0 57.5 4-5 97.2 19.1 55.8 6+ 96.4 18.0 57.2 BIRTH INTERVAL <24 months 94.7 17.3 72.2 24-35 months 97.3 18.6 60.3 36-47 months 98.1 17.9 49.8 48+ mouths 96.2 15.9 37.4 TOTAL NUMBER OF CHILDREN 96.4 16.9 56.4 5731 5161 308 Note: Ever-breasffed status refers to all children born in the five years preceding the survey (including deceased children). Median duration is based on the current breastfeeding status of living children born 1-59 months before the survey. Similar differentials exist for the demographic characteristics examined (Table 7.2). Children of birth order 4 and higher, children born after a birth interval of 24-47 months, and children of mothers 35 years and older were breastfed for a median duration of 18 to 21 months. In comparison, firstborn children, children bom after a birth interval of less than 24 months, and children of mothers under 35 years of age had a shorter median duration of breastfeeding (14 to 17 months). Reasons for Never Breastfeeding and for Stopping Breastfeeding Over one-third of children who were never breastfed died during the first week of life (Figure 7.1). These children were probably too iU to breastfeed. For an additional 3 percent who died at an older age, the child's death was listed as the reason for never breastfeeding. Other reasons for never breastfeeding fall into three main categories: the infant's illness or refusal to suckle (11 percen0, the mother's illness and reported inability to nurse the infant (42 percent), and other factors, including the mother working (8 percent). 51 Figure 7.1 Reasons for Never Breastfeeding ot for Stopping Breastfeeding Child died Child died within 3% ~enie.t 2% Nipple probleme 5% Note: Inoludell children born 1-59 months before the survey who were not breaatfed. Bolivia ENDSA-8B Table 7.3 and Figure 7.2 show the reasons for stopping breasffeeding among children who were still alive at the time of the interview. Most mothers stopped breastfeeding because they felt that the child had reached weaning age (60 percent). The remaining reasons for stopping can again be divided into three main categories related to: illness in the child or refusal to continue breastfeeding (7 percent), illness in the mother, reported inability to breastfeed and new pregnancy (25 percent), and other factors, including the mother's job (7 percent). Bmastfeeding is known to be an important determinant of child survival. Therefore, it is of interest to examine the reasons for stopping breastfeeding among children who died (Table 7.3). Three- quarters of these mothers said that they stopped breastfeeding because the child died, or that the child was breastfed until death. Among the remaining 25 percent, the distribution of the reasons given was similar to that for living children (Figure 7.2). The only difference was, as expected, a higher proportion of mothers who said they stopped because the child was sick and a smaller proportion who said they stopped because the child had reached weaning age. 52 Table 7.3 Reasons for stopping breastfeeding by survival status of child: Percent distribution of reasons for stopping breastfee(Fmg by survival status among children born in the five years preceding the survey, Bolivia ENDSA, I989 Reasons for stopping Living Dead breast feeding children children Child diod/breasffed until death 75.3 Child sick 2.9 6.1 Child refused 4.0 1.0 Reached weaning age 61.1 8.6 Milk insufficient 7.6 1.7 New pregnancy 13.8 3.4 Mother sick 3.2 1.3 Mother worksfmconvenient 3.4 1.1 Other 4.0 2.5 TOTAL 100.0 100.0 NUMBER OF CHILDREN 3473 444 Note: Figures are for children born 1-59 months before survey. Frequency of Breastfeeding Infants age 4-6 months were selected for the analysis of the frequency of breastfeeding, because breastfeeding should be well established and the proc- esses of weaning has not yet begun in most cases. The fre- quency of breastfeeding in the 24 hours before the interview dif- fered between subgroups. It tended to decrease with increases in the level of maternal educa- tion; over 60 percent of mothers without education nursed their infant ten times or more per day compared with 38 percent of mothers with nine or more years of schooling. However, when seven or more feedings was chos- en as the cut-off point, there was no consistent trend. Figure 7.2 Reasons for Stopping Breastfeeding by Survival Status of Children Other 7% Maternal rea ion l ° 25% Weaning age 61% Child l ick, refu|ed 7% Living Dead Other 11% Maternal reaiona° 26% Weaning age 35% Child sick, refused 29% *lncufflcJent milk, mother IJGk/pregnant Bollvla ENDSA-ag 53 The relationship between the frequency of breastfeeding and the length of the preceding birth interval is of interest because breasffeeding promotes longer periods of postpartum amenorrhea. Under the assumption that the frequency of breastfeeding observed for the index child reflects the mother's feeding practices for the preceding child, Figure 7.3 suggests a correlation between more frequent nursing and longer birth intervals. Almost 90 percent of infants born after a birth interval of at least 36 months were fed at least seven times per day, and 60 percent received ten or more feedings. The corresponding figures for infants born after a birth interval of less than 24 months were only 59 and 37 percent, respectively. Figure 7.3 Frequency of Breastfeeding over a 24-hour Period by Mother's Education and Length of Prior Birth Interval Percent None 1-5 6-8 9* 124 24-35 36-47 48* MOTHER'S EDUCATION (YEARS) PRIOR BIRTH 'NTERVAL (MONTHS) [ Number °f TimesFed i BIB lO+ t~ r-9 ~4-6 ~1-4 Note:Include8 children 4-6 months. Bolivia ENBSA-S9 Supplementation and Bottle Feeding About half the infants were breastfed exclusively for the first three months (Tables 7.1 and 7.2 and Figure 7.4). Exclusive breastfeeding was less common in urban areas and among more educated women. Particulariy striking were the differences by region: exclusive breastfeeding for three months was more than twice as common in the Altiplano as in the Llanos, with the Valle intermediate. The feeding practices by age for all children are summarized in Table 7.4 and Figure 7.5. Giving only water in addition to breast milk was relatively rare and was practiced by less than 5 percent of the mothers with children under 6 months of age. The percentage of infants who received both breast milk and solids increased from about 10 percent or less for infants under 5 months to 60 percent of infants age 9-17 months. The proportion of infants receiving both breast milk and other supplements (such as milk, sugar water or juice, herbal tea, and/or other liquids) but no solids ranged from one-third among babies under 5 months of age to 9 percent or less of children older than 9 months. At the time of the survey, 33 percent of children age 18-23 months were still breastfed. 54 Figure 7.4 Percentage of Children Still Breastfed by Age of Child and Type of Supplement Percent(moving average) O 3 6 9 12 15 18 21 24 27 30 33 Age (Mont ha) [ ~ Breastfed Only ~ J P~alrt Water Only m Any Supplement Bolivia ENDSA-89 Table 7.4 Feeding practices for children under three years: Among children born less then three years preceding the survey, the percentage in each feeding category by eun'ent age, Bolivia ENDSA, 1989 Breast Breast Breast Breast milk milk milk Number Not milk and and solid and other of Ageofchild breastfed exclusively water food supplements a children 1-2 months 1.6 61.5 3.4 0.3 33.2 206 3-4 months 6.I 42.7 4.6 11.2 35.4 190 5-6 months 8.6 23.5 5.0 42.6 20.3 191 7-8 months 12.7 7.4 3.5 59.8 16.6 176 9-11 months 16.7 12,1 0.5 61.7 9.0 294 12-17 months 35.1 0.7 0.0 60.0 4.2 573 18-23 months 66.7 1.0 0.8 30.1 1.3 536 24-35 months 94.6 0.2 0.5 4.2 0.5 1017 Note: Figures are for children born 1-35 months before the survey. Categories are exclusive. aHerbal water, sugar water, juice, and/or other liquids Among children still breastfed at the time of the survey the percentage who had received specific foods and liquids during the 24 hours before the interview is summarized in Table 7.5. Powdered milk, a common breast mi lk supplement, was given to one-fifth to one-quarter of the breastfed babies under 18 months of age. Cow's milk or goat's milk was given to less than 10 percent of infants under 7 months of age, and to 12-18 percent of children 7-18 months of age. The percentage of breastfed children who were given solids increased from 12 percent of those age 3-4 months, to 47 percent of those 5-6 months nf age, to over 70 percent of (breastfed) children over 9 months of age. By the end of the first year, 85 percent of breastfed infants had received either solid food or a liquid. 55 Figure 7.5 Feeding Pract ices among Young Chi ldren by Cur rent Age Percent / / [ / loO 80 BO +0 I I 20 3-4 7-8 18-23 Age (Months) • Milk, Juice, herbal or sugar water, other liquids [~ Not Brealt fed Brit. Milk & Solidi Brat. Milk & Other* Brlt, Milk & Water m Brit. Milk Only Bolivia BNDSA-Stl One-quarter to one-third of breastfed babies of all ages were given a bottle. This proportion did not vary substantially by the age of the child (Table 7.5). If the assumption is made that all infants under six months received the bottle if they were not breasffed, the prevalence of bottle feeding in Bolivia can be estimated to be 32 percent for this age group. 1 Table 7.6 shows the number of times each food was given during the 24 hours before the interview for breastfed children who received each of the respective items. Three-quarters of the breastfed children who were given solids, received the foods once or twice a day. Half of the children were given cow's milk or goat's milk only once a day, while powdered milk was given more frequently, with 38 percent receiving more than three feedings per day. 1 This is much lower than observed in six other DHS surveys in Latin America where bottle use----estimated using the same method--was 60 to 85 percent (8oerma et al., 1991a). 56 Table 7.5 Feeding practices for children under three years who were still breasffeeding at the time of the survey: Among children born less then three years preceding the survey who were breastfed, the percentage in each feeding category end the percentage who received various supplements, by age of the child, Bolivia ENDSA, 1989 Age of child Feeding category a Percentage of breastfed chilc~en who received supplements" Breast Breast- milk mad Juice Cow's or Number fed water or sugar Powdered goat's Other Solid of only only water milk milk liquids food Bottle children 1-2 months 62.5 3.4 1.7 18.7 1.7 16.3 0.3 27.2 202 3-4 months 45.5 4.9 8.4 24.7 4.9 25.8 11.9 31.7 179 5-6 months 25.7 5.4 20.3 20.2 8.3 34.3 46.6 25.5 174 7-8 months 8.5 4.0 27.1 24.5 11.9 43.3 68.4 40.7 153 9-11 months 14.6 0.6 23.5 22.2 10.8 54.1 74.1 35.1 244 12 17 months 1.1 0.1 26.0 25.2 18.1 57.3 92.4 34.1 372 18-23 months 3.0 2.5 29.7 17.5 12.6 54.9 90.5 21.2 178 24-36 months 3.4 9.0 30.0 17.7 29.8 64.0 78.5 20.9 55 Note: Figures are for children born 1-35 months before the survey. aCategories are exclusive hMore than one response allowed Table 7.6 Frequency of breast milk supplements for children under five: Among children born less than five years preceding the survey, who were still breastfeeding at the time of the survey, the percentage who received specific food supplements during the 24 hours before the interview by frequency given, Bolivia, ENDSA, 1989 Plain Herbal Juice or Powdered Cow's or Other Solid Frequency given water water sugar water milk goat's milk liquids foods Once 46.6 57.7 54.9 26.0 49.5 25.6 22.1 Twice 35.6 31.4 30.7 35.4 33.7 24.3 52.4 Three times 12.6 8.4 10.6 25.0 10.8 6.1 20.8 Four times or more 4.5 1.9 2,9 12.9 6.1 2.7 4.0 Don't know/missing 0.6 0.6 0,9 0.7 0.0 41.3 0.7 NUMBER OF CHILDREN 525 368 322 345 178 692 951 Note: Figures are for children born 1-59 months before the survey. 57 CHAPTER 8 CHILDHOOD IMMUNIZATIONS Only 19 percent of children age 12-23 months were fully immunized, and just 7 percent had completed the vaccination schedule during the first year of life. Coverage for measles vaccine, which requires only a single dose, was 58 percent. Extremely high dropout rates, however, limited coverage for those vaccines requiring multiple doses--the vaccines against diphtheria, whooping cough or pertussis, and tetanus (DPT) and poliomyelitis (polio). Thus, three-quarters of the children who received the first dose of DPT did not receive the third dose in the series before their first birthday. The survey results suggest that the figures for immunization coverage based on health facility reports may have overestimated coverage in Bolivia. Mothers could present a child health card for only about one child in five. This is one of the lowest card levels found in the DItS surveys carried out between 1986 and 1989. Immunization coverage levels showed an increase in 1987, which continued throughout 1988. The increase was substantial for all vaccines and doses, except for third doses of DPT and polio. Thus, dropout rates did not decrease. None of the subgroups identified had high coverage levels. In all groups, for example, less than half of the children 12-23 months of age were vaccinated with three doses of DPT. Socioeconomic and demographic variables identified some groups, however, with extremely low coverage; these included children of mothers with no education and children living in rural areas. Campaigns were the best-known source for immunizations, followed by health centers, public hospitals, and health posts. Less than 2 percent of the women with children under five years did not know of any source of immunization, although this figure was somewhat higher among women with no education, women speaking Indian languages, and women living in rural areas. 8.1 Methodology Data on immunization coverage were collected for all of the respondents' living children under the age of five years. If the mother could present a child health card, the interviewer copied the dates of all immunizations from the card. If the mother could not present a card, she was asked to recall the specific vaccines given to her child(ren), including the number of doses of DPT and polio vaccines. Respondents were not asked to recall the ages at which each vaccine was given. Although the written record provides the most reliable vaccination information, additional information based on the mothers' recall is helpful in estimating the overall vaccination coverage level. 8.2 Findings Coverage A child health card was presented for only 19 percent of all children under five. The percentage of children with cards in Bolivia is one of the lowest for any DHS country surveyed, Only Mali and the Dominican Republic have lower figures (Boerma et al., 1990). Among children age 12-23 months, only 23 percent had a child health card, so infonnation reported by mothers represents the largest part of the coverage estimates. Over 50 percent of the children had received the single-dose vaccines, BCG (against tuberculosis) and measles (Table 8.1). More than half of the children had also been given two doses of DPT and polio vaccines. Coverage rates for the 59 third doses of DPT and polio were, however, just 28 percent and 38 percent respectively. ~ The dropout rate between the first and third doses of DPT was high, 60 percent; for children with a health card, the dropout rate was lower, 40 percent. Only 19 percent of children 12-23 months old were fully vaccinated. According to the recommended schedule, all immunizations should be given during the first year of life, and children should be fully vaccinated by their first birthday. Table 8.1 estimates the extent of immunization coverage for children who were 12-23 months at the time of the survey, and the percentage vaccinated by 12 months of age in the same age group. 2 The coverage estimates drop considerably for all immunizations, especially for measles and the third doses of DPT and polio vaccines. Only one-third of children had received measles vaccine by age 12 months. While about half the children had been given the first doses of the DPT and polio vaccines, only 13 and 20 percent, respectively, had received three doses before their first birthday. Only 7 percent of the children were fully vaccinated by their first birthday. Table 8.1 Immunization coverage by source of information: Percentage of children 12-23 months who had received specific vaccines and doses by source of information, and percentage vaccinated by 12 months of age, Bolivia ENDSA, 1989 Type of vaccine Source of information Vaccinated Health Maternal by 12 Reported card recall Both months data VACCINES BCG 14.4 40.6 55.0 37.0 DPT 1 20.4 50.0 70.4 52.8 DFF 2 17.2 34.2 51.3 33.1 DPT 3 12.1 16.2 28.4 12.7 oRor-ov'r ~rl~ tur'rln~et'3) 40.7 67.6 59.7 76.0 Polio 1 21.6 57.7 79.3 57.6 Polio 2 18.2 42.6 60.8 40.1 Polio 3 13.4 24.4 37.8 19.7 Polio 4 5.4 4.1 9.5 2.8 27.0 39.0 40.0 44,0 MEASLES 16.6 40.8 57.5 32.9 FULLY VACCINATED a 7.6 11.1 18.8 6.9 aChildrcn who have received BCG, 3 doses of DPT, at least 3 doses of polio, and measles gaccines are classified as fully vaccinated. Data reported by health facilities in 1987 and 1988 (WHO, 1990b) The immunization schedule recommended by the World Health Organization includes four doses of polio vaccine: in addition to the three doses given simultaneously with DPT at 6, 10, and 14 weeks of age, one dose at birth is recommended. 2 To obtain coverage estimates for all children, it was assumed that immunizations reported by mothers were given at the same ages as immunizations copied from the health cards of other children. 60 The immunization coverage estimates from the ENDSA differ considerably from the coverage figures derived from Bolivian health facility reports (Table 8.1). The data reported by health facilities refer to vaccinations given to children under one year of age during 1987 and 1988 (WHO, 1990b). The differences are particularly large for the third doses of DPT and polio. For the third dose of DPT, for example, health facilities report three times the coverage (by age 12 months) that was found in the ENDSA national sample. This suggests that the coverage figures based on health facility reports overestimate vaccination levels in Bolivia. Figure 8.1 illustrates how much higher immunization cov- erage levels are when the mother has a written record, i.e., the child health card. Vaccination with BCG showed the smallest differ- ence, 9 percentage points, fol- lowed by measles and the first dose of polio vaccine, for which the difference was 19 percentage points. The third dose of DPT showed the greatest disparity, 31 percentage points. Although the lower coverage levels among children without a vaccination card are expected, the inability of mothers to remember which vac- cines her child(ran) had received may have contributed to the ob- served differences. Figure 8.1 Immunization Coverage for Children 12-23 Months by Source of Data Psrcsfll 86 04 eo 75 6 7e 62 5 56 72 3 eo $ 62 6 4o 5 1o 21 2 33 o BCG 1 D~T S 1 p~lo 3 [ I H n i t ' Car. ~ Mother'e ReCSil I Nots: Coverage raise i re shown | sp l r s te ly for children with health cards and those without. MaXillaS All Bollvls ENDSA'Sg Trends in Coverage Table 8.2 and Figure 8.2 show the trends in immunization coverage during the period 1985-1988. Using the information from the child health cards, which includes dates of immunization, the proportion of children vaccinated by 12 months of age was estimated for four age groups: 12-23, 24-35, 36-47 and 48-59 months) Coverage started to increase in 1987 and continued to increase during 1988. The increase was most pronounced for the first doses of DPT and polio vaccines and, to a lesser extent, measles immunization. Increases in the third dose of DPT and, to a lesser extent, the third dose of polio lagged behind. Thus, the dropout rates between the first and third doses of these vaccines did not decrease at all. The percentage of children fully vaccinated during the first year of life remained very low throughout the four-year period. Again, it was assumed that children with immunizations reported by the mother were immunized at the same ages as children with a health card. 61 Table 8.2 Immunization coverage by age 12 months for children 12-59 months of age: Percentage of children 12-59 months immunized for specific vaccines by age 12 months (health card information and maternal recall), Bolivia ENDSA, 1989 Age of child (months) Type of vaccine 12-23 24-35 36-47 48-59 PERCENTAGE WITH CARDS 23.1 23.1 16.3 15.1 VACCINES BCG 37.0 32.7 26.4 27.8 DPT 1 52.8 40.7 26.0 23.2 Dirt 2 33.1 25.4 16.5 17.4 DPT 3 12.7 11.7 7.8 10.0 Polio 1 57.6 39.8 29.0 24.3 Polio 2 40.1 27.7 20.5 21.6 Polio 3 19.7 13,6 9.2 10.8 Polio 4 2.8 0.0 0.0 0.0 Measles 32.9 20.9 10.5 13.3 FULLY VACCINATED a 6.9 2.5 4.0 7.1 DROPOUT RATE (DPTI/DPrJ) 76.0 71.3 70.0 56.8 NUMBER OF CHILDREN 1110 1017 993 986 Note: Figures are based on information from health cards and maternal recall. aChildren who received BCG, 3 doses of DPT, at least 3 doses of polio, and measles vaccines are classified as fully vaccinated. 62 Figure 8.2 Trends in Immunization Coverage among Children Immunized by 12 Months Percent 80 40 20 0 , 1985 1988 1987 1988 1989 --+-- Mel la le l l - - eGG -~" Po l io I ~ Pol io 3 I Bolivia EN DSA-89 Socioeconomic and Demographic Differentials Tables 8.3 and 8.4 show the socioeconomic and demographic differentials in immunization coverage for the first and third doses of DPT and for measles. (The pattern for polio vaccine was similar to that shown for DPT.) Regarding the first dose of DPT, no subgroup had a coverage of less than 50 percent, and differentials were generally small, except by mother's level of education, language, and media exposure. In contrast, coverage was less than 50 percent for the third dose of DPT in every subgroup, and differentials were large. Selected variables have been summarized in Figure 8.3. Measles immunization coverage follows the same pattern as DPT3, but coverage levels were higher, ranging from 41 to 73 percent. Overall, 15 percent of children age 12-23 months had not received any vaccinations. This figure dropped to less than 10 percent among children of mothers with more than five years of education, children of fathers with white collar jobs, children in the Valle region, and children of mothers who watch television daily. The proportion rose, however, to more than 25 percent among children of mothers without education, children of fathers working in agriculture, children of mothers who speak Indian language, and children of mothers who neither watch television nor listen to the radio. A number of low-coverage groups can be identified: The largest differentials exist for mother's level of education (Figure 8.3). Children of mothers who have no education have the lowest level of immunization coverage overall (see Table 8.3). The second most important factor is media exposure, which is related to various socioeconomic factors. Children of women who are not regularly exposed to radio or television have very low coverage. Children of mothers who speak an Indian language are less likely to have been immunized than children of mothers who speak Spanish. 63 Table 8.3 Differentials th immunization coverage by socioeconomic characteristics: Percentage of children 12-23 months who were immunized for specific vaccines and the per- centage who had a health card, by socioeconomic characteristics, Bolivia ENDSA, 1989 Type of immunization received Characteristic DPT1 DPT3 Measles Percentage with a Number health of None card children RESIDENCE Urban 73.7 35.9 61.4 10.4 27.6 542 Rural 67,2 21.2 53,7 19.8 18.7 568 REGION Alfiplano 67.2 22.8 52.2 19.3 15.2 556 Valle 75.0 39.0 62.9 9,5 26.5 318 Llanos 71.8 27.3 62.4 13,3 36.9 236 MOTHER'S EDUCATION None 52,2 16.8 41.6 34.0 15.1 207 1-5 years 69.9 24.0 53.7 14.0 20.9 514 6-8 years 75.9 32.4 66.8 8.2 25,7 146 9+ years 83.5 45.1 73.4 6.0 32.8 242 FATItER'S OCCUPATION White collar 75,2 34.1 64,8 7.3 28.3 310 Blue collar 73.1 34.6 59.5 10.3 25.6 328 Agriculture 64.1 20.1 49.2 25,6 17.1 379 Other 70.8 21.3 59.5 I6.2 20.8 93 MOTIIER WORKING Yes 72.6 33.8 57.5 13.2 19.9 206 No 69.9 27,1 57.4 15.6 23.8 904 MEDIA EXPOSURE TV daily 80.5 40.7 66.6 8.0 29.8 486 Radio daily 65.8 20.6 57.1 16.7 18.0 358 Neither 58.1 16.3 41.3 26.4 17.6 265 LANGUAGE Spanish 76.8 34.1 63.1 10.3 25.6 783 Indian 55.0 14,8 44.0 27.0 17.1 327 TOTAL 70.4 28.4 57.5 15.2 23.1 1110 Note: Figures are based on information from health cards and maternal recall. Coverage is lower in rural than urban areas. Likewise, children of fathers who work in the agricultural sector are less likely to receive immunizations than children of fathers who have non-agricultural occupations. There are also regional differences. Immunization coverage is lower in the Altiplano region than in either the Valle or Llanos regions. Children of mothers age 35 years and over and children of birth order 4 or higher are less likely than other children to receive immunizations (Table 8.4). The preceding birth interval and sex of the child did not, however, have any substantial effect on coverage. 64 Table 8.4 Differentials in immunization coverage by demographic characteristics: Percentage of ehil&en 12-23 months who ware immunized for specific vaccines and the tmrcentaga who had a health card, by demographic characteristics, Bolivia F,,NDSA, 1989 Percentage Type of immunization received: with a Number health of DPTI DPT3 Measles Nona card children Characteristic SEX Male 68.2 28.0 54.1 16.6 23.3 574 Female 72.7 28.8 61.0 13.7 22.8 535 MOTHER'S AGE 15-19 67.2 22.9 62.1 18.0 17.4 81 20-34 73.3 31.0 60.4 13,0 25.0 796 35+ years 61.6 21.4 45.8 21.9 18.3 232 BIRTH ORDER First 73.3 34.3 61.9 12.6 31.5 231 2-3 71.3 28.8 61.1 13.1 23.4 380 4-5 71.9 25.1 52.6 16.0 18.9 250 6+ 64.8 25.5 52.6 20.2 18.8 248 BIRTH INTERVAL <24 months 73.0 31.6 61.1 14.5 20.1 227 24-35 months 65.7 25,3 53.1 17.7 18.5 303 36-47 months 68.8 21,7 57.8 13.9 21.1 142 48+ months 72.2 27,2 54.5 16.3 24.9 206 TOTAL 70.4 28,4 57.5 15.2 23.1 1110 Nota: Figures are based on information from health cards and mammal recall. Figure 83 DPT3 Coverage Differentials for Children 12-23 Months REBtDENCE L l lno I MOTHER'a EDUCATION No education 1-5 y le r l years %: ~are LANGUAG 8panle~ Indian M aa TV ~l i l y Rad io dally Neither Feml • MOTHER'8 AGE 16"19 yr l 20-34 yr l 36" yr l BIRTH ORDER 1 2-3 4 -5 e* Note : Overall Coverage $1 26.4 percent, i m i ra m mm m mm m m -20 -15 -10 -5 m =lie m m m m m m I I$ la 15 20 Percent e l f~lrenc l from Overall CQver|ge Bolivia ENDSA-89 65 Sources of Immunization All women with children under five years were asked what sources they knew of for immuni- zations. Campaigns were the source most frequently men- tioned, 42 percent of the women; 34 percent mentioned health cen- ters, 22 percent mentioned public hospitals, and 17 percent men- tioned heaith posts (Figure 8.4). There were distinct regional d i f ferences in knowledge: campaigns were mentioned by 56 percent of the women in the Valle, 43 percent in the Altiplano, and just 20 percent in the Llanos. Private sources were not common, 10 percent of the women with at least nine years of education mentioned private doctors; 10 percent of this group also mentioned private hospitals. Figure 8.4 Knowledge of Immunization Sources among Women with Children Under Five Campaign Public hospital ~ Health center Health post ~ School ~ 9 Comm, Health Worker ~ 4 Pvt hoap/doctor, CNS ~ 6 Other* ~2 4 SotJrc(~ not known 0 10 Note: Percentage of women who mentioned each souroe; multiple reiponses allowed. • Includea ORT unit and pharmacy 22 I lr 20 30 40 50 Bollvls ENDSA-89 On/y 2 percent of the women said they did not know of any source for immunizations. This figure was somewhat higher in the rural areas (3 percent), among women without education (6 percent), and among women speaking an Indian language (6 percent). 66 CHAPTER 9 PRENATAL CARE AND DELIVERY ASSISTANCE More than half of all pregnant women did not receive any prenatal care during the period 1984-88, and only 1 in 5 women received an injection to prevent tetanus in the newborn. This low level of tetanus toxoid (I7) vaccination is due in part to the general lack of prenatal care, but also in part to missed opportunities. Many pregnant women who had a prenatal visit still did not receive a IT injection. There were substantial socioeconomic differentials in prenatal care, with rural women and less educated women the least likely to have a prenatal check- up. Sixty percent of the babies born in the last five years were born at home. Home deliveries were much mare common in the rural areas and among women with specific socioeconomic characteristics. Almost half of all babies were delivered at home and without prenatal care. Half of these mothers did not know of any modern method of family planning. One in 13 babies was delivered by cesarean section, but this rate varied markedly according to socioeconomic variables. In urban areas, 1 in 8 babies was born by cesarean section, and among women with at least nine years of education, 1 in 5. 9.1 Methodology Women who had given birth in the five years preceding the survey were asked a series of questions concerning the type of health care they received during pregnancy and the subsequent delivery. Respondents were asked whether or not they had seen anyone for a check-up during the pregnancy and whether they had received a tetanus toxoid injection. Women were also asked where each baby was delivered and who assisted during the delivery. If more than one type of care-giver provided prenatal care or attended the delivery, interviewers were instructed to record the most qualified person. In addition, women were asked whether their last baby was born by cesarean section, Women who were currently pregnant also reported on whether they had received prenatal care and/or a tetanus toxoid vaccination. 9.2 Findings Prenatal Care and Tetanus Toxoid Coverage Figure 9.1 shows the percent distribution of births in the five years preceding the survey according to (1) the type of prenatal care received by the mother and (2) whether the mother received at least one tetanus toxoid injection during pregnancy. The overall level of prenatal care was low: over half (53 percent) of the women did not receive any prenatal care. Doctors were by far the most common providers of prenatal care; one percent of the women were visited by traditional midwives. Only 20 percent of the women said they received an injection during pregnancy to prevent tetanus in the newbom. Generally, a tetanus toxoid ('IT) injection is given during a prenatal visit. However, for 6 percent of "all pregnancies, women responded that they had received a TT injection but had not seen a medical person during the pregnancy (data not shown). These women may have received the "IT injections during immunization campaigns. In addition, 30 percent of the women received prenatal care, but still did not get a "IT injection. This may have been due to shortages of the vaccine, but it might also reflect missed opportunities for vaccination. Missed opportunities appeared to be more common in urban areas, where 41 percent of the women did not receive a "IT injection although they visited a doctor, nume, or trained midwife during their pregnancy (data not shown). 67 Figure 9.1 Type of Prenatal Care Received by Women PRENATAL CARE NO prenatal care Doctor Nurae/mldwife Trsdition&l midwife Other MisiJng . ~ 63 43 ~2 11 It 0 TETANUS TOXOID Received inject}on ~ EO Prenatal care/ ~ 30 but no TT Injection 0 10 20 30 40 SO EO Percent Note: Includes births 1-59 months preceding the Ilurvey, eotiva ENDSA-S9 Based on information from women pregnant at the time of the interview, the first prenatal care visit was when they were a median of 5.0 months pregnant. Among women who were in their last two months of pregnancy at the time of the interview, 43 percent had a prenatal visit, 15 percent had received one dose of TT, and 15 percent had received at least 2 doses of' IT (Table 9.1). Health facilities were the main source for TT immunization for these women, providing the injection for 49 percent of those who received TT; 38 percent received the TT injections during campaigns, 7 percent in schools, and 7 percent from other sources. Table 9.1 Prenatal care and tetanus toxoid immunization among currently pregnant women: Percentage of currently pregnant women who had at least one prenatal visit and the percentage who were immunized with tetanus toxoid by duration of the pregnancy, Bolivia ENDSA, 1989 Tetanus toxoid vaccine Number Duration of Prenatal One Two or of pregnancy (months) visit dose more doses women <5 months 23.4 14.9 6.5 193 5-7 months 27.3 14.2 8,0 285 8+ months 42.9 15.1 15.1 123 There are substantial socioeconomic differentials in the utilization of prenatal care (Table 9.2). The proportion of pregnant women who did not receive prenatal care was twice as high in the rural areas as in the urban areas. Women riving in the Llanos region were more likely than women living elsewhere to receive prenatal care. Prenatal care was least frequently received in the Altiplano. As might be expected, the most educated women were far more likely to receive prenatal care: only 12 percent of their pregnancies went without any care compared with 81 percent among women with no education. The 68 other socioeconomic variables showed similar patterns. Women who did not speak Spanish, women who did not watch television or listen to the radio, and women whose husbands worked in agriculture were all less likely to receive prenatal care. Table 9.2 Prenatal care and tetanus toxoid immtmization by socioeconomic characteristics: Percent distribution of births in the five years preceding the survey by prenatal care, and the percentage of births for which women received at least one tetanus toxoid O"T) injection, by socioeconomic characteristics, Bolivia ENDSA, 1989 Characteristic None Prenatal care Received Number T'I" of Medical a Other Total injection births RESIDENCE Urban 35.7 62.3 2.0 100.0 25.6 2767 Rural 69.0 28.8 2.2 100.0 15.0 2965 REGION Ahiplano 59.5 38.7 1.8 100.0 15.4 2865 Valle 50.8 46.7 2.5 100.0 19.4 1601 Llanos 40.7 57.1 2.2 100.0 31.7 1265 MOTHER'S EDUCATION None 81.4 15.5 3.1 100.0 9.5 1204 1-5 years 61.7 35.7 2.6 100.0 18.5 2667 6-8 years 35.5 63.6 1.1 100.0 29.5 767 9+ years 12.4 86.8 0.8 100.0 29.3 1093 FATHER'S OCCUPATION White collar 30.8 67.8 1.4 100.0 26.4 1564 Blue collar 48.7 48.9 2.4 100.0 21.4 1673 Agriculture 73.2 24.3 2.5 100.0 15.0 2101 Other 50.6 47.0 2.4 100,0 17.4 394 MEDIA EXPOSURE TV daily 30.3 67.1 2.6 100.0 26.6 2443 Radio daily 63.7 34.5 1.8 100.0 18.1 1903 Neither 78.0 20.2 1.8 t00.0 11.5 1385 LANGUAGE Spanish 41.9 56.2 1.9 100.0 24.6 4029 Indian 78.9 18.5 2.6 100.0 9.5 1702 TOTAL 52.9 44.9 2.2 100.0 20.1 5731 Note: Figures are for births 1-59 months before the survey. aprenatal care provided by doctor, nurse or trained midwife. Variation by demographic variables is smaller (Table 9.3). The most important differences were observed for mother's age and birth order. Women age 35 years and over were less likely to have received prenatal care, as were high parity women (6 or more live births). 69 Table 9.3 Prenatal care and tetanus toxoid immunization by demographic characteristics: Percent distribution of births in the five years preceding the survey by prenatal care, and the percentage of births for which women received at least one tetanus toxoid O'T) injection, by demographic characteristics, Bolivia ENDSA, 1989 Characteristic Prenatal care Received Number T r of None Medical a Other Total injection births SEX Male 52.2 46.1 1.7 100.0 20.1 2879 Female 53.6 43.9 2.5 100.0 20.1 2853 MOTHER'S AGE 15-19 years 50.1 49.6 0.3 100.0 22.2 292 20-34 years 50.1 48.0 1,9 100.0 21.6 4015 35+ years 61.5 35.5 3.0 100.0 15.6 1424 BIRTH ORDER First 40,1 59.0 0.9 100.0 21.5 1149 2-3 48.4 49.5 2.1 100.0 20.7 1944 4-5 57.2 41.1 1.7 100.0 20.3 1307 6+ 66.3 30.1 3.6 100.0 17,9 1332 BIRTH INTERVAL <24 months 57.4 38.9 3.7 100.0 19.5 1315 24-35 months 60.8 37,1 2.1 100.0 20.7 1576 36-47 months 57.6 40.1 2.3 100.0 18.4 765 48+ months 45,3 53.4 1.3 100.0 20.0 921 TOTAL 52.9 44.9 2.2 100.0 20.1 5731 Note: Figures are for births 1-59 months before the survey. aprenatal care provided by doctor, nurse or trained midwife Assistance at Delivery Over a third of all deliveries took place in health facilities, including 10 percent which occurred in private hospitals (Table 9.4 and Figure 9.2). Sixty percent of the babies born from 1984 to 1988 were born at home. In a number of cases, home deliveries were assisted by medically qualified personnel. Two percent of all births were home deliveries assisted by a doctor, and three percent were home deliveries assisted by a nurse or auxiliary nurse. Not all mothers understood the ranking of health personnel or the status of trained vs. untrained traditional midwives and birth attendants. Thus, the information collected on who assisted at the delivery may be inaccurate in a number of cases. However, Figure 9.2 shows that relatives were the most common attendants at delivery (39 percent), followed by doctors (38 percent). Traditional birth attendants (mostly untrained or with minimal training) assisted at 12 percent of all deliveries, while 5 percent of the women said they delivered without an attendant. The place where delivery occurred varied considerably, depending on the women's socio- economic characteristics (Table 9.4). In urban areas, for example, 58 percent of babies were delivered at health facilities compared with 18 percent in rural areas. In the Llanos, 53 percent were delivered at health facilities, and in the Altiplano, 30 percent. Substantial differentials also existed for mother's level 70 of education, language, media exposure, and father's occupation. Deliveries at private health facilities were more common among women with at least nine years of education, women who watch television regularly, and women with husbands employed in white collar jobs. Five percent of all deliveries took place at home but were attended by health personnel. This proportion varies little according to socioeconomic variables; however, when home deliveries attended by trained health personnel are examined as a percentage of total home deliveries, the expected socioeconomic differences emerge. For example, 28 percent of home deliveries among women with at least nine years of schooling were attended by health personnel, compared with only 4 percent among women without education. Table 9.4 Place of delivery for births by socioeconomic characteristics: Percentage of births in the five years preceding the survey that were delivered in a health facility and the percentage delivered at home (with and without a health attendant), by socioeconomic characteristics, Bolivia ENDSA, 1989 Delivery in health facility Delivery at home Chirr acteristic Public Private Total Health No No Number atten- health infor- of dant a attendant Total Other mation births RESIDENCE Urban 42.3 15.9 58.2 4.6 34.9 39.5 1.9 0.4 2767 Rur,,d 13.6 4.6 18.3 5.3 74.6 79.9 0.8 1.0 2965 REGION Altiplano 21.4 8.7 30.0 5.0 63.1 68.1 1.2 0.7 2865 Valle 28.9 10.0 39.0 4.7 54.1 58.9 1.3 0.9 1601 Llanos 39.5 13.3 52.8 5.2 39.6 44.8 1.6 0.8 1265 MOTHER'S EDUCATION None 7.2 0.4 7.6 3.6 87.0 90.6 0.8 1.0 1204 1~5 years 21.7 4.7 26.5 5.8 65.5 71.3 1.2 1.0 2667 6-8 years 46.7 11.8 58.5 5.5 34.2 39.7 1.7 0.1 767 9 or more 50.4 32.5 82.9 4.1 10.7 14.8 1.8 0.5 1093 FATHER'S OCCUPATION White collar 42.5 20.6 63.0 4.9 29.8 34.7 1.8 0.5 1564 Blue collar 33.1 10.7 43.7 6.2 48.2 54.4 1.5 0.4 1673 Agriculture 10.4 2.3 12.7 4.3 81.2 85.5 0.6 1.2 2101 Other 35.4 7.3 42.7 3.1 50.1 53.2 2.7 1,4 394 LANGUAGE Spanish 36.3 13.8 50.1 5.4 42.0 47.4 1.6 0,9 4029 Indimt 6.6 1.3 7.9 3.9 87.1 91.0 0.5 0,6 1702 MEDIA EXPOSURE TV daily 42.8 19.4 62.2 5.5 30.0 35.5 1.4 0.9 2443 Radio daily 19.7 4.1 23.9 5.8 67.8 73.6 1.8 0.7 1903 Neither 11.1 1.8 12.9 2.9 83.1 86.0 0.4 0.7 1385 TOTAL 27.5 10.1 37.6 5.0 55.4 60.4 1,3 0.7 5731 Note: Figures are for births 1-59 months before the survey. aDoctor, nurse or trained midwife 71 Figure 9.2 Place of Delivery end Person Attending the Delivery PLACE Public hospital Health center Pr4v|ts ho|pitml Home Other Missing DELIVERY ABBIBTAN T Doctor (Auxl l l l ry) Nurse Birth i t tsndant Rs l s t l~ Other No One Missing B 24 I: o Do le 8e 20 4O E0 Percent,ge of All DalE,erie| I ~ With Medbcal Anbst. 80 Nots: In©l .~s births 1-69 monthl preceding the sur ly . Bollvls ENOSA-89 Older women (35 or over) were less likely to give birth in health facilities than younger mothers (Table 9.5). Differences by birth order were more pronounced: firstborn babies were more than twice as likely to be delivered at a health facility as babies of birth order 6 or higher. Figure 9.3 shows the practices of health facilities concerning the separation of mother and child during the first day after delivery. In public hospitals, 71 percent of the mothers kept their babies in their own room or bed, 16 percent were separated for a few hours, 6 percent were only together with the baby during feeding, and 7 percent did not see the baby at all during the first day. Mothers and babies were more frequently separated in private hospitals: 56 percent of the mothers kept their babies with them throughout the day, 19 percent were separated for only a few hours, 12 percent were with the baby only during feedings, and 12 percent did not see the baby at all during the first day after delivery. 72 Table 9.5 Place of delivery for births by demographic characteristics: Percentage of births in the five years preceding the survey that were delivered in a health facility end the percentage delivered at home (with and without a health attendant), by demographic characteristics, Bolivia ENDSA, 1989 Delivery in health facility Delivery at home Characteristic Public Private No Health health No Number atten- atten- in for- of Total dant a dent Total Other mation births SEX Male 28.4 10.1 38.4 5.5 53.5 59.0 1.5 1.1 2879 Female 26.5 10.1 36.7 4.5 57.3 61.8 1.1 0.4 2853 MOTHER'S AGE 15-19 years 36.4 7.0 43.4 5.8 48.7 54.5 2.1 0.0 292 20-34 years 28.0 11.5 39.5 5.3 53.2 58.5 1.4 0.6 4015 35+ years 24.0 6.8 30.8 3.8 62.9 66.7 1.0 1.5 1424 BIRTH ORDER First 37.3 15.0 52.4 5.0 40.8 45.8 1.7 0.2 1149 2-3 28.2 13.6 41.8 5.1 50.7 55.8 1.6 0.8 1944 4-5 25.5 7.9 33.3 5.1 60.3 65.4 0.8 0.5 1307 6+ 19.9 2.8 22.8 4.5 70.0 74.5 1.1 1.6 1332 BIRTH INTERVAL <24 months 22.9 8.4 31.4 5.5 59.9 65.4 1.3 1.9 1315 24-35 months 22.5 6.9 29.5 4.2 64.2 68.4 1.6 0.5 1576 36-47 months 23.7 7.3 31.0 5.6 61.4 67.0 1.0 1.0 765 48+ months 33.3 13.8 47.1 5.0 47.1 52.1 0.6 0.3 921 TOTAL 27.5 10.1 37.6 5.0 55.4 60.4 1.3 0.7 5731 Note: Figures are for births 1-59 months before the survey. aDoctor, nurse or trained midwife 73 Figure 9.3 Mother/Child Contact the First Day After Delivery by Type of Health Facility Other 2% aep i t l ted "~ ell d ly 7% 8•13. • few hr•. 15% Together 71% Publio Private Othe¢ 3~ ~ S P rated St i r i ted •xegpt to ~led 12% Sep. • ~w hr| 19% Together 56% Bolivia ENDSA-8O Maternity Care and Family Planning Many of the maternal and child health variables and family planning indicators are closely related. For example, women who received prenatal care were more likely to deliver in health facilities, and women who used prenatal care and delivery services were more likely to use family planning. Table 9.6 examines the use of a modem method of contraception with respect to maternity care and several key socioeconomic variables. Maternity care is defined as prenatal and/or delivery care provided in a health facility or by trained health personnel for at least one birth in the five years preceding the survey. The family planning indicator chosen was ever use of a modem method of contraception. Overall, 16 percent of the women had received maternity care and also had ever used a modem method of contraception. Another 44 percent had received maternity care, but had never used a modem family planning method. Two percent reported ever use of a modem method of contraception but no maternity care, while 39 percent had neither received matemity care nor ever used a modem contraceptive method. There were marked differentials according to mother's education, place of residence, and region (Table 9.6 and Figure 9.4). Over one-third of women with at least nine years of schooling reported receiving maternity care and using a modem method of contraception; 56 percent had used matemity services but had never used modem contraception; and only 6 percent had done neither. In contrast, only 3 percent of women without any education had both received maternity care and used modem methods of family planning; another 21 percent received maternity care only; and 75 percent reported neither. The very low rate of contraceptive use for some subgroups may reflect the lack of contact with the health care system. 74 Table 9.6 Maternity ca,re and use of convaception by socioeconomic characteristics: Percent distribution of women 15-49 who had at least one llve birth in the five years preceding the survey by whether or not they received maternity cars and if they had ever used a modem method of contraception, according to selected socioeconomic characteristics, Bolivia ENDSA, 1990 Characteristic Received Did not receive maternity care maternity care Number Ever Never Ever Never of used used used used Total women RESIDENCE Urban 25.7 52.3 1.6 20,4 100.0 1882 Rural 5.6 34.4 1.5 58,5 100.0 1810 REGION Ahiplano 7.8 43.4 1.6 47,3 100.0 1879 Vallc 14.1 45.1 0.9 39,9 I00.0 1008 Llanos 36.9 41.9 2.3 18.9 100.0 805 MOTHER'S EDUCATION None 2.8 21.2 1.2 74.8 100.0 764 1-5 years 9.2 43.2 2.1 45.6 100.0 1601 6-8 years 23.0 57.6 1.5 17.9 100.0 510 9+ years 36.6 56.4 0.8 6.2 100.0 818 TOTAL 15.9 43.5 1.5 39.1 100.0 3692 Note: Maternity care is defined as prenatal care provided by health personnel and/or delivery in a health facility, or delivery at home attended by health personnel (doctor, nurse or midwife). Percent Figure 9.4 Maternity Care and Ever Use of a Modern Family Planning Method Rural Urban None 1-15 6-8 9* RESIDENCE MOTHER'S EDUCATION (YRS) I "Matern i ty C . . . . ~ Modern Contt.caption i Note: Includll womln who had a birth in the five y l l r l prlcldlng the lurvly, • Prenatlll and/or delivery care provided by heslth personnel Bolivls ENDSA-89 75 About 1 in 13 last-born babies was delivered by cesarean section (Table 9.7). In the urban areas, this ratio rose to 1 in 8 births. The practice was twice as common in the Llanos as in the other two regions. Among women with at least nine years of education, 1 in 5 babies was delivered by cesarean seclion. There were substantial socioeconomic differentials. Most of the women who had had a cesarean section had seen a doctor during pregnancy: only 2 percent of those women who had not seen anyone for a prenatal checkup went on to have cesarean sections. 76 Table 9.7 Prevalence of caesarean section by socioeconomic characteristics: Percentage of births in the five years preceding the survey that were delivered by caesarean section, by socioeconomic characteristics, Bolivia ENDSA, 1989 Births Number delivered by of Characteristic caesarean section births RESIDENCE Urban 12.3 1882 Rural 3.4 1811 REGION Altiplano 5.7 1879 Valle 6.9 1009 Llanos 14.4 805 MOTHER'S EDUCATION None 2.5 764 1-5 years 5.0 1601 6-8 years 6.9 510 9+ years 19.3 818 FATHER'S OCCUPATION White collar 13.2 1057 Blue collar 9.3 1056 Agriculture 2.2 1250 Other 8.1 329 MOTHER WORKING Yes 11.3 784 No 7.0 2908 MEDIA EXPOSURE TV daily 13.9 1649 Radio daily 4.4 1147 Neither 1.3 866 LANGUAGE Spanish 10.5 2622 Indian 1.5 1070 PRENATAL CARE No one 2.0 1804 Doctor 14.7 1734 Trained nurse 1.3 58 Trained midwife 0.0 34 Birth auendant 0.5 39 Other 0.0 18 TOTAL 7.9 3692 77 CHAPTER 10 MATERNAL MORTALITY Maternal mortality is difficult to measure since it is a relatively rare phenomenon. Two methods of calculation based on data on the survival status of the respondent's female siblings were used to estimate maternal mortality. The direct method yielded an estimate of 332 maternal deaths per 100,000 live births for 1975-88, while the indirect method resulted in an estimate of 373 deaths per 100,000 live births in 1977. These maternal mortality estimates do not include single women, who may account for as much as 30 percent of maternal deaths, according to another Bolivian study. Therefore, they may underestimate the true extent of maternal mortality. Women giving birth in their forties, irrespective of the number of children they already had, were at greater risk of dying. There are two important obstacles to reducing maternal mortality in Bolivia: the limited extent of prenatal care (less than half of the pregnant women) and, to a lesser extent, the low proportion of deliveries taking place in health facilities (about 40 percent). 10.1 Other Sources of Data on Maternal Mortality The only national data on maternal mortality that are available for Bolivia refer to the period from 1973 to 1977. At that time, maternal mortality was estimated to be 480 per 100,000 live births) In 1988 a survey was carried out in the province of Avarao, Oruro Department, to estimate the level of maternal mortality. The matemal mortality estimate, using the sisterhood method, was extremely high: 1,379 matemal deaths per 100,000 live births. At this rate, women had a 1 in 10 lifetime risk of dying of maternal causes (Simons et al., 1989). Another study, which included eight of the twelve health units in Bolivia, examined 118 matemal deaths that took place in hospitals or were recorded in civil registers. It found that 30 percent of the maternal deaths were among single women (De la Galvez Murillo and Castillo, 1982). Of the deaths, 27 percent were due to abortion (induced or spontaneous), while hemorrhage before and after delivery was the Icading cause of death (38 percent). Eleven percent of the maternal deaths were associated with infections, and 13 percent with toxemia. In a study of 70 obstetric deaths which took place in medical institutions in Cochabamba between 1979 and 1986, 20 percent were due to induced abortion, 21 percent to hemorrhage, 26 percent to infection, and 10 percent to toxemia (Salinas, 1987). This study reported a maternal mortality ratio of 250 per 10l),00l) live births. 10.2 Methodology Maternal mortality is defined as deaths among women while pregnant or within 42 days of termination of pregnancy from any cause related to or aggravated by the pregnancy or its management (WHO 1979). Maternal death is much less common than childhood death and therefore more difficult to measure. These data were reported by WHO (1986) and were drawn from WHO (1980). They are based on health facility statistics. 79 Data on the survivorship of sisters of ENDSA respondents was used to estimate maternal mortality. The level of maternal mortality can be estimated indirectly, using the so-called sisterhood method (Graham et at., 1989). In addition, a new approach, which entailed collecting more information about the sisters of the respondents than is required by the sisterhood method, was used for direct estimation of the matemal mortality. The main difference between these two methods is that the indirect method provides only overall estimates of maternal mortality, while the direct method allows the calculation of rates for different time periods. 2 The survey respondents were asked to list all their siblings and to give the sex, survival status, age and, if applicable, the age at death and the time period when death took place. In addition, marital status and parity were ascertained for female siblings who died after the age of 12 years. Deaths among married female siblings were categorized according to whether they occurred during pregnancy or delivery, or within two months following delivery or termination of a pregnancy. This definition of a death as being due to matemal causes differs from the def'mition of maternal mortality cited above in two respects: 1) all deaths within a specific time period are defined as maternal deaths even if they were not due to pregnancy-related causes, and 2) the time period has been extended from 42 days to two months. This simplified definition was chosen in order to minimize underreporting. Extending the time period is supported by other data which suggest that a significant number of maternal deaths take place in the period 42 to 90 days following the termination of a pregnancy (Rochat, 1985). 10.3 F ind ings Indirect Estimates The sisterhood method uses the proportion of the respondent's adult sisters who have died during pregnancy, childbirth or within two months after childbirth, as reported by the ENDSA respondents, to estimate the level of maternal mortality. Table 10.1 presents the results of the indirect estimation procedure. A total of 7,454 female respondents age 15-49 years reported 11,934 ever-married sisters and 116 maternal deaths. Seventy-nine percent of the maternal deaths occurred during pregnancy or childbirth, and 21 percent occurred in the two months after delivery. The estimated lifetime risk of maternal death was 0.023 (i.e., 116 maternal deaths taken as a proportion of 5103 sister units of risk) (Graham et al., 1989). In other words, the lifetime risk of dying of pregnancy-related causes was 1 in 44. From these data and knowledge of the total fertility rate (TFR) it is possible to estimate the maternal mortality ratio, that is, maternal deaths per 100,000 live births 3. The resulting estimate of the maternal mortality ratio is 373 deaths per 100,000 live births and refers to a point in time about 12 years before the survey, 1977. The last cohurm of Table 10.1 shows the proportion of dead sisters who died from maternal causes. Almost a quarter of all deaths to sisters of reproductive age were ascribed to maternal causes, which is relatively high. There was no decline in the proportion due to maternal causes with increasing age of the respondent. 2 For details see Rutenberg et al., 1990. 3 The maternal mortality ratio (MMR) can be approximated as follows: MMR = l-((Probability of survival)La~a). Since the MMR estimate refers to 12 years before the survey, the TFR for the period 10-14 years before the survey was used. That rate was 6.177. 80 Table 10.1 Indirect estimates of maternal mortality: Estimates of maternal mortality, Bolivia ENDSA, 1989 Number Number Proportion of of Number Sister Life-time of sisters Number sisters ever- of units risk of who died of of 15 years married maternal Adjustment of risk maternal maternal Age group respondents and older sisters a deaths factor exposure death causes 15-19 1524 1774 2440 b 2 0.107 261 0.009 0.08 20-24 1282 2074 2053 b 6 0.206 423 0.014 0.19 25-29 1289 2346 2064 b 16 0.343 708 0.023 0.21 30-34 1039 2046 1584 19 0.503 797 0.024 0.26 35-39 101l 1973 1642 20 0.664 1090 0.019 0.22 40-44 680 1295 1140 25 0.802 914 0.028 0.30 45-49 628 1128 101I 27 0.900 910 0.029 0.22 TOTAL 7454 12635 11934 116 5103 0.023 0.23 Maternal mortality ratio (MMR) c 373 aThe number of ever-married sisters was calculated by applying the proportion ever-married by five-year age groups in the ENDSA to the age distribution of living sisters and summing to obtain the number of ever-married living sisters. The gumber of sisters who reached age 15 and married but subsequendy died were then added to the sum Derived by multiplying the number of repondents by the average number of ever-married sisters per respondent for respondents age 30-49, that is 1.60 CMMR = 1 - (1 - Life-time risk) vrFR assumes TFR 10-14 years before survey equal to 6.177 Direct Estimates In addition to the questions required for the sisterhood method, four other questions were asked to obtain a direct estimate of maternal mortality. These questions concerned the age of all surviving siblings, the age at death of those sisters who died, the period in which the death occurred, and the parity of the sisters who died. The sex of the sibling was missing for only two cases, and survival status was unknown for only 106 cases (0.33 percent). These cases were excluded from the analysis. There was no age given for 3.3 percent of all living female siblings. The year of death was missing for 18 percent, and the age at death for 10 percent. The missing data were imputed according to standard DHS procedures (Rutenberg et al., 1990). From data on the ages of the female siblings and the period in which death occurred, maternal mortality rates can be estimated. The rates are presented in Table 10.2. Since there was considerable heaping on 5 and 10 years before the survey, the periods 0-6 years and 7-13 years before the survey were used as reference periods. It appears that the maternal mortality rate increased from 0.60 per 1,000 women age 15-49 in the period 1975-81 to 0.65 per 1,000 in the period 1982-88. The general fertility rate, which was based on data from all women in the ENDSA, declined from 216 live births per 1,000 women 15-49 years for the period 1975-81 to 173 for 1982-88. The maternal mortality ratio is equal to the maternal mortality rate divided by the general fertility rate; thus it increased from 278 to 373 deaths per 100,000 live births. A possible explanation for the increase in maternal mortality rates is that deaths, including maternal deaths, are more likely to be omitted the longer ago that they occurred. The maternal mortality ratio for the ten-year period preceding the survey is 332 deaths per 100,(300 live births. Studies have shown that maternal mortality risks are high for very young women and for older women, as well as for first births and births of higher orders. In this study, the maternal mortality rates cannot be computed directly by age and parity since parity was not determined for surviving sisters. However, it is 81 Table 10.2 Direct estimates of maternal mortality: Maternal mortality rates based on the survivorship of sisters of survey respondents, by time period, Bolivia ENDSA, 1989 1975-1981 1982-1988 1975-1988 Maternal Years of Maternal Years of Maternal Years of Age group deaths exposure Rate a deaths exposure Rate a deaths exposure Rate a 15-19 6.3 15342 .42 3.8 14959 .25 10.1 30301 .33 20-24 5,7 13008 ,44 5.4 15480 .35 11,1 28488 .39 25-29 5.1 11072 .46 11.3 13789 .82 16.4 24861 .66 30-34 4.8 7530 .63 12.5 11691 1.07 17.3 19221 .90 35-39 4.5 453l .99 3.6 8888 .40 8.1 13419 .60 40-44 3.7 2131 1.73 6.5 5453 1.20 10.2 7584 1.35 45 49 2.5 1047 2.48 4.1 2869 1.43 6.6 3914 1.71 TOTAL 15-49 32.8 54664 .60 47.2 73130 .65 80.0 127794 .63 Maternakmortality ratio (MMR)" 278 373 332 aRate per 1000 women for age-specific mortality and overall mortality rate for women 15-49. MMR is per 100,000 births. bCalculated as the maternal mortality rate for women 15-49 divided by the general fertility rate (GFR). The GFR, per I,O00 women, was 216, 173 and 190 for the time periods 1975-81, 1982-88 and 1975-88, respectively. possible to indirectly assess differential mortality risks by parity and age, as shown in Table 10.3. The distribution of maternal deaths by age and parity is known for the sisters of the ENDSA respondents, while ttle distribution of births by parity and age for Bolivia as a whole can be obtained from the respondents. From these distributions, the relative risk of maternal death by age and parity can be estimated. Table 10.3 shows that there is only minor variation by parity. Differentials by age of the women are large, however, and increase dramatically for pregnancies at age 40 and over. In addition, women having their first birth in their thirties appear to be at higher risk of maternal mortality, but the numbers are small. 82 Table 10.3 Relative risk of maternal death by age and parity: Maternal deaths and births by age and parity and the relative risk of maternal death by age and parity, Bolivia ENDSA, 1989 Age of woman Parity <20 20-29 30-39 40+ Total MATERNAL DEATHS P~ity 1 5.9 7.9 2.8 0.0 16.5 Parity 2-5 4.4 17.6 19.8 4.2 46.0 Parity 6* 0.0 2.3 2.6 12.8 17.7 Total 10.3 27.7 25.2 17.0 80.3 BIRTHS Parity 1 985 1233 107 6 2330 Parity 2-5 535 4503 1576 74 6688 Parity 6+ 453 1722 376 2551 Total 1520 6189 3405 455 11569 RELATIVE RISK OF MATERNAL DEATH Pa~-it y 1 0.86 0.92 3.78 0.00 1.03 Parity 2-5 1.19 0.56 1.81 8.21 0.99 Parity 6+ 0.73 0.22 4.91 1.00 TOTAL 0.98 0.65 1.07 5.38 1.00 Note: Maternal deaths are based on the sisterhood method; births are the number of births to survey respondents; relative risk of maternal death (per live birth) is calculated with total risk (l.00) as the reference category. Discussion The indirect estimate of the maternal mortality ratio was 373 deaths per 100,000 live births for a point in time about 12 years before the survey. The direct estimate for the 10-year period preceding the survey was very close: 332 per 100,000 live births. The direct method of estimation requires that a few extra questions are asked of the respondents and results in some added complexity in data processing and analysis compared with the indirect method. A major advantage of the direct method, however, is that the data on years since death allow the calculation of maternal mortality rates for more than one retrospective time period. Future direct estimates can be compared with estimates for earlier time periods for the purposes of program evaluation. Both the indirect and direct methods excluded women who had never been married. Another study in Bolivia found, however, that 30 percent of all maternal deaths involved single women and that abortion was a leading cause of maternal death. Since the ENDSA excluded unmarried women, the estimates presented here must be considered an underestimate of the true rate of maternal mortality. Maternal mortality in Bolivia appears to be among the highest in Latin America. The utilization of prenatal care in Bolivia is low: 53 percent of women do not receive any prenatal care. In addition, only about 40 percent of women delivered in health facilities. Increasing the utilization of prenatal care is the first step necessary to improve the detection and referral of women at high risk and thus reduce maternal mortality. 83 REFERENCES Alam, N., F.J. Henry, and M.M. Rahaman. 1989. Reporting Errors in One-week Diarrhoea Recall Surveys: Experience from a Prospective Study in Rural Bangladesh. International Journal of Epidemiology 18(3):697-700. Bicego, G.T. and J.T. Boerma. 1990. Maternal Education, Use of Health Services and Child Survival: An Analysis of Data from the Bolivia DHS Survey. DHS Working Papers, No. 1. Columbia, Maryland: Institute for Resource Development. Boerma, J,T., S.O. Rutstein, A.E. Sommerfelt, and G. Bicego. 1991. Bottle Use for Infant Feeding in Developing Countries: Data from the Demographic and Health Surveys: Has the Bottle Battle been Lost? Journal of Tropical Pediatrics 37(1):. (Text reference is Boerma et al., 1991a). Boerma, J.T., R.E. Black, A.E. Sommerfelt, S.O. Rutstein, and G.T. Bicegn. [1991]. Accuracy and Completeness of Mother's Recall of Diarrhea Occurrence in Preschool Children in the Demographic and Health Surveys. International Journal of Epidemiology Vol 20. Forthcoming. (Text reference is Boerma ct al., 1991b). Boerma LT., A.E. Sommeffelt, S.O. Rutstein, and G. Rojas. 1990. Immunization: Levels, Trends and Differentials. DHS Comparative Studies, No. 1. Columbia, Maryland: Institute for Resource Development. Campbell, H., P. Byass, and B.M. Greenwood. 1988. Simple Clinical Signs for Diagnosis of Acute Lower Respiratory Infections. Lancet 2(8613):742-743. Cherian, T., T.J. John, M.C. Steinhoff, and M. John. 1988. Evaluation of Simple Clinical Signs for the Diagnosis of Acute Respiratory Tract Infection. Lancet 2(8603):125-128. Coloma, R. Belmonte and B.P. de Ormachea. 1985. Encuesta de Prevalencia de Medicamentos, Bolivia 1983: Resultados Generales. La Paz, Bolivia: Consultora Boliviana de Reproducci6n Humana. Consejo Nacional de Poblaci6n. 1988. Posici6n de Consejo Nacional de Poblaci6n (CONAPO) sobre el Control de la Natalidad y la Planificaci6n Familiar (PF). Boletfn Informativo 3(September). De la Galvez Murillo, A. and J. CastiUo. 1982. Mortalidad Materna en Hospitales y Registros Civiles Ano 1980. La Paz, Bolivia: Ministerio de Previsi6n Social y Salud Pt~blica. Dibley, M.J., J.B. Goldsby, N.W. Staehling, and F.L. Trowbridge. 1987. Development of Normalized Curves for the International Growth Reference: Historical and Technical Considerations. American Journal of Clinical Nutrition 46(5):736-748. (Text reference is Dibley et al., 1987a). Dibley, M.J., N.W. Staehling, P. Nieburg, and F.L. Trowbridge. 1987. Interpretation of Z-score Anthropometric Indicators Derived from the International Growth Reference. American Journal of Clinical Nutrition 46(5):749-762. (Text reference is Dibley et al., 1987b). Feacbem, R.G. and M.A. Koblinsky. 1984. Interventions for the Control of Diarrhoeal Diseases among Young Children: Promotion of Breastfeeding. Bulletin of the Worm Health Organization 62(2) :271-291. Freeman, J. and G.B. Hutchison. 1980. Prevalence, Incidence and Duration. American Journal of Epidemiology 112(5):707-723. 85 Freij, L. 1977. Exploring Child Health and its Ecology: The Kirkos Study in Addis Ababa, An Evaluation of Procedures in the Measurement of Acute Morbidity and a Search for Causal Structure. Acta Paediatrica Scandinavica, Supplement 267:1-180. Garenne, M. and O. Fontaine. 1986. Assessing Probable Causes of Death Using a Standardized Questionnaire: A Study in Rural Senegal. Paper presented at the IUSSP Seminar on Comparative Studies of Mortality and Morbidity, Siena, Italy, July 7-12. Goldman, N., L. Moreno, and C.F. Westoff. 1989. Peru Experimental Study: An Evaluation of Fertility and Child Health Information. Columbia, Maryland: Office of Population Research (Princeton University) and Institute for Resource Development/Macro Systems, Inc. Graham, W., W. Brass, and R.W. Snow. 1989. Estimating Maternal Mortality: The Sisterhood Method. Studies In Family Planning 20(3):125-135. lnstituto Nacional de Estadfstica (INE). 1989. Bolivia: Encuesta Nacional de Poblaci6n y Vivienda 1988: Resultados Finales. La, Paz, Bolivia. lnstituto Nacional de Estadfstica (INE) and Institute for Resource Development/Macro Systems, Inc. (IRD). 1990. Encuesta Nacional de Demografla y Salud 1989. Columbia, Maryland: INE and IRD. Kalter, H., R.H. Gray, R.E. Black, and S.A. Gultiano. 1990. Validation of Postmortem Interviews to Ascertain Selected Causes of Death in Children. International Journal of Epidemiology 19(2):380-386. Koenig, M.A., M.A. Khan, B. Wojtyniak, J.D. Clemens, J. Chakraborty, V. Fauveau, J.F. Phillips, J. Akbar, and U.S. Barua. 1990. Impact of Measles Vaccination on Childhood Mortality in Rural Bangladesh. Bulletin of the World Health Organization 68(4):441-447. Martorell, R. and J.P. Habicht. 1986. Growth in Early Childhood in Developing Countries. In Human Growth: A Comprehensive Treatise Vol. 3, ed. F. Falkner and J.M. Tanner, 241-262. New York: Plenum Press. Martorell, R., J.P. Habicht, C. Yarbrough, and R.E. Klein. 1976. Underreporting in Fortnightly Recall Morbidity Surveys. Journal of Tropical Pediatrics and Environmental Child Health 22(3): 129-134. Ministerio de Previsi6n Social y Salud Ptiblica. 1983. Lista B~ica de Causas y Estructura de Edad para Tabulacidn de Morbilidad y Mortalidad. La Paz, Bolivia. Ministerio de Previsi6n Social y Salud Pdblica. 1989. Plan Nacional de Supervivencia - Desarrollo lnfantil y Salud Materna. La Paz, Bolivia. Moreno, L. and N. Goldman. 1990. An Assessment of Survey Data on Birth Weight. Social Science and Medicine 31(4):491-500. Murilio, A. and R. Belmonte Coloma. 1984. [Diarrheal Disease in U~an Areas of Bolivia: Report of a Study of Children Aged 0-5 Undertaken with the Aid of PAHOj. La Paz, Bolivia: Ministerio de Previsi6n Social y Salud Ptiblica, Divisi6n Nacional Matemo Infantil. National Institute of Food and Nutrition, Government of Bolivia and U.S. Agency for Intemalional Development (USAID). 1982. Bolivia National Nutritional Status Survey, 1981: Summary Report. Washington, D.C. 86 Pedersen, D., C. Betts, J. Mariscal, and J. Tones Goitia. 1987. Supervivencia Infantil en Bolivia: Situaci6n Actual y Prioridades para la Acci6n. Resources for Child Health (REACH) Project Report. La Paz, Bolivia: REACH. Rance, S., O. Wolowyna, and G. Pinto Aguirre. 1989. Salud y Poblaci6n: Supervivencia Infantil. La Paz, Bolivia: Consejo Nacionai de Poblaci6n and Research Triangle Institute. Rochat, R. 1985. The Magnitude of Maternal Mortality: Definitions and Methods of Measurement. Paper presented at the WHO Intenegional Meeting on Prevention of Maternal Mortality, Geneva, Switzerland, November 11-15. Rutenberg, N., T. Boerma, J. Sullivan, and T. Croft. 1990. Direct and Indirect Estimates of Matemal Mortality with Data on the Survivorship of Sisters: Results from the Bolivia DHS. Paper presented at the 1990 annual meeting of the Population Association of America, Toronto, Ontario, May 3-5. Rutstein, S.O. 1984. Infant and Child Mortality: Levels, Trends, and Demographic Differentials. WFS Comparative Studies, No. 43. Voorburg, Netherlands: International Statistical Institute. Salinas, W. 1987. Mortalidad Matema-lnstitutionai: Cochabamba 1979-86. Cochabamba, Bolivia. Simons, H., L. Wong, W. Graham, and S. Schkolnik. 1989. The Sisterhood Method to Estimate Maternal Mortality: Applications with Special Reference to Latin America. Paper presented at the General Conference of the International Union for the Scientific Study of Population (IUSSP), New Delhi, India, September 20-27. Sullivan, J.M. and S.E. Wilson. 1982. A Methodology for Detecting Bias in Indirect Mortality Estimates: The Case of Indonesia. Carolina Population Center, Paper No. 25. Chapel Hill, North Carolina: Carolina Population Center (Univ. of North Carolina). United Nations. 1983. Manual X: Indirect Techniques for Demographic Estimation. Population Studies, No. 81. New York: Department of International Economic and Social Affairs. United Nations. Young Children operation for Programme). 1986. How to Weigh and Measure Children: Assessing the Nutritional Status of in Household Surveys. New York: United Nations, Department of Technical Co- Development and Statistical Office (National Household Survey Capability World Health Organization (WHO). 1979. International Classification of Diseases, Injuries and Death, 9th ed. (ICD9). Geneva. World Health Organization (WHO). 1980. The Sixth Report on the World Health Situation 1973- 77, Part 2: Review by Country and Area. Geneva. World Health Organization (WHO). 1986. Maternal Mortality Rates: A Tabulation of Available Information, 2nd ed., Geneva. (Document FIlE 86.3). World Health Organization (WHO). 1989. Household Survey Manual: Diarrhea Case Management, Morbidity and Mortality. Geneva. World Health Organization (WHO). 1990a. Acute Respiratory Infections in Children: Case Management in Small Hospitals in Developing Countries. Geneva. (WHO/ARI/90.5) World Health Organization (WHO). 1990b. EPl--Health Information System. Geneva: EPI. 87 APPENDIX A ACCURACY OF DIARRHEA RECALL APPENDIX A ACCURACY OF D IARRHEA RECALL A.1 Heaping of Responses Mothers were allowed to respond in days, weeks, or months to the question on when their child's last episode of diarrhea had occurred. The results of the responses reported in days and weeks (if less than 3 weeks), are shown in Table A.1 (unweighted data). More than half of the children with diarrhea in the two weeks preceding the survey had diarrhea at the time of the survey. Table A-I Length of time since the last episode of diarrhea: Distribution of mother's responses in days and weeks to a question about child's last episode of diarrhea, Bolivia ENDSA, 1989 Time elapsed since last episode Percent Number DAYS 0 53.9 839 1 0.5 8 2 2.3 36 3 3.7 57 4 1.9 30 5 1.7 27 6 0.8 12 7 2.4 38 8 0.5 8 9 0.2 3 10 0.6 9 11 0.0 0 12 0.0 0 13 0.0 0 14 0.2 3 15 1.1 17 16 0.0 0 17 0.0 0 WEEKS 1 15.5 241 2 14.8 230 Current 53.9 839 Days 15.9 248 Weeks 30.2 471 TOTAL 100.0 1558 No~:lncludeso~yresponsesofless thmal8 daysor ~ss ~an3 weeks. 91 For children who did not have diarrhea in the 24 hours prior to the interview, but had diarrhea in the two weeks preceding the survey (or slightly longer ago), about one-third of the women responded to the question in days. There was heaping on 2-4 days, and on 7, and 15 days. The latter appears to be common in Latin America, where respondents consider a two-week period as 15 rather than 14 days (Goldman et al., 1989). More important is the large proportion of cases that were recalled as one or two weeks ago. In Bolivia 15 per cent of the answers were "two weeks ago." These will presumably be translated into 14-15 days ago, unless the interviewer makes a special effort to find out exactly which was the last day of the diarrheal episode. What respondents mean by two weeks ago may vary between cultures, however. Assuming that the responses of "two weeks ago" really refer to the period 11-17 days before the survey, if all the "two weeks" responses are included in the recall period, diarrhea prevalence will be overestimated by about 6 percent. Studies witl~ a one-week recall period would have similar problems. A.2 Underreporting of Diarrhea The data from Bolivia can be used to study diarrhea reporting in recent and more distant periods preceding the survey. An estimate of the daily prevalence and incidence of diarrhea can be made, since questions were asked both about the length of time which had elapsed since the last diarrheal episode and about its duration. The daily incidence and prevalence of diarrhea were calculated from the responses to these questions and are presented in Table A.2, and Figures A.1 and A.2. The mother's responses were assumed to refer to the end of the last episode. Responses of "one week ago" were distributed over the period 5-10 days before the survey. Responses of "two weeks ago" were distributed over the period 11- 17 days before the survey. Daily prevalence in Bolivia drops from over 150 cases per 10

View the publication