Serbia Multiple Indicator Cluster Survey 2005: Monitoring the Situation of Children and Women
Publication date: 2005
SERBIASERBIA MULTIPLE INDICATORMULTIPLE INDICATOR CLUSTER SURVEYCLUSTER SURVEY 20052005 MONITORING THE SITUATION OF CHILDREN AND WOMEN Publisher UNICEF Belgrade For publisher Ann-Lis Svensson UNICEF Area Representative for Serbia, Croatia and Montenegro Editor Oliver Petrovic UNICEF Belgrade Programme Specialist Proofreading Rowland Palairet Design Rastko Toholj Cover photo Zoran Jovanovic Maccak Printed by Stojkov, Novi Sad Print run 600 Published in May, 2007 Contributors to the report MICS Global team Oliver Petrovic Ivana Bjelic Dragisa Bjeloglav Dragana Djokovic–Papic The Multiple Indicator Cluster Survey (MICS) in Serbia was carried out by the Statistical Office of the Republic of Serbia (SORS) and Strategic Marketing Research Agency (SMMRI). Financial and technical support was provided by the United Nations Children’s Fund (UNICEF). The survey has been conducted as part of the third round of MICS (MICS3), carried out around the world in more than 50 countries, in 2005–2006, following the first two rounds of MICS surveys that were conducted in 1995 and 2000. Survey tools are based on the models and standards developed by the global MICS project, designed to collect information on the situation of children and women in countries around the world. Additional information on the global MICS project may be obtained from www.childinfo.org. Suggested citation Statistical Office of the Republic of Serbia and Strategic Marketing Research Agency. 2006. Republic of Serbia Multiple Indicator Cluster Survey 2005, Final Report. Belgrade, Republic of Serbia: Statistical Office of the Republic of Serbia and Strategic Marketing Research Agency. SERBIASERBIA MULTIPLE INDICATOR MULTIPLE INDICATOR CLUSTER SURVEY 2005CLUSTER SURVEY 2005 AcknowledgementsAcknowledgements The completion of the survey and this report would not be possible without the commitment and professionalism of teams of people, at home and abroad. UNICEF’s Global MICS team conceptualized and designed the MICS survey. Tessa Wardlaw, Edilberto Loiaiza, Trevor Croft, Marco Segone assisted in the survey design and provided excellent technical support throughout the whole process. Emma Holmberg and Ngagne Diakhate gave excellent on-line technical support on data processing and analysis. Turgay Unalan made some very valuable contributions to the draft report. Particularly valuable was the assistance provided by Attila Hancioglu and George Sakvarelidze. Their leadership, timely advice and team spirit were of critical importance in overcoming numerous constraints during both the complex and quality research. The Area representative for Serbia, Croatia and Montenegro, Ann-Lis Svensson, made some very thoughtful inputs in adapting the survey to national priorities, particularly with regards to the inclusion of the most excluded population group in the survey design. The Director of the Statistical Office of the Republic of Serbia, Dragan Vukmirovic, accepted the challenge of implementing this very comprehensive survey for the first time. Their team was very competently led by Dragana Djokovic–Papic. The other implementing partner, the Strategic Marketing Research Agency was led by Dragisa Bjeloglav, whose experience and knowledge was especially helpful in overcoming constraints. Ivana Bjelic of the Strategic Marketing Team, took on the challenge of operational coordination and communication with the global team. Both partners showed great commitment and flexibility in adopting all the changes emerging throughout the process, as well as professionalism in completing all the tasks to the highest standard and in a timely manner. Oliver Petrovic from UNICEF’s Belgrade Office successfully dealt with the challenge of overall coordination and reporting. Support of the line ministries, the Ministries of Health, Education and Social Welfare, was very important not only during the survey design phase, but also for the use of data for evidence-based policy changes. Particularly valuable were contributions from the Poverty Reduction Strategy implementation team, who supported us in maintaining focus on the most marginalized families in the country and who have already used the MICS3 findings for the second national PRS report. The technical contributions of the members of the Council for Children’s Rights in Serbia were very important, and their role in data use and dissemination is crucial. The National Roma Council and local Roma NGOs played an important role in conducting the survey with Roma families in the field. The implementation of the survey was made possible thanks to the Canadian International Developmental Agency, the Swedish International Developmental Agency, and UNICEF Set Aside Funds, who contributed with the funds required. ContentsContents LIST OF ABBREVIATIONS . . . . . . . . . . . . . . 7 SUMMARY TABLE OF FINDINGS . . . . . . . . . . 9 EXECUTIVE SUMMARY. . . . . . . . . . . . . . . . 13 INTRODUCTION . . . . . . . . . . . . . . . . . . . 19 Background . . . . . . . . . . . . . . . . . . . . . . 19 Survey Objectives . . . . . . . . . . . . . . . . . . 19 SAMPLE AND SURVEY METHODOLOGY. . . . . . . 21 Sample Design . . . . . . . . . . . . . . . . . . . . 21 Questionnaires . . . . . . . . . . . . . . . . . . . . 22 Training and Fieldwork . . . . . . . . . . . . . . . . 23 Data Processing . . . . . . . . . . . . . . . . . . . . 23 SAMPLE COVERAGE AND THE CHARACTERISTICS OF HOUSEHOLDS AND RESPONDENTS . . . . . . . . . . . . . . . . . 24 Sample Coverage . . . . . . . . . . . . . . . . . . . 24 Characteristics of Households . . . . . . . . . . . . 24 Characteristics of Respondents . . . . . . . . . . . 26 CHILD MORTALITY . . . . . . . . . . . . . . . . . . 28 NUTRITION . . . . . . . . . . . . . . . . . . . . . . 30 Nutritional Status . . . . . . . . . . . . . . . . . . . 30 Breastfeeding . . . . . . . . . . . . . . . . . . . . . 33 Low Birth Weight . . . . . . . . . . . . . . . . . . . 37 CHILD HEALTH . . . . . . . . . . . . . . . . . . . . 38 Immunization . . . . . . . . . . . . . . . . . . . . . 38 Oral Rehydration Treatment . . . . . . . . . . . . . 40 Care Seeking and Antibiotic Treatment of Pneumonia . . . . . . . . . . . . . . . 41 Solid Fuel Use . . . . . . . . . . . . . . . . . . . . . 42 ENVIRONMENT . . . . . . . . . . . . . . . . . . . . 44 Water and Sanitation . . . . . . . . . . . . . . . . . 44 Security of Tenure and Durability of Housing . . . . . . . . . . . . . . . 47 REPRODUCTIVE HEALTH . . . . . . . . . . . . . . . 48 Contraception . . . . . . . . . . . . . . . . . . . . . 48 Unmet Need . . . . . . . . . . . . . . . . . . . . . . 49 Antenatal Care . . . . . . . . . . . . . . . . . . . . 50 Assistance at Delivery . . . . . . . . . . . . . . . . 51 CHILD DEVELOPMENT . . . . . . . . . . . . . . . . 52 EDUCATION . . . . . . . . . . . . . . . . . . . . . . 54 Pre-School Attendance and School Readiness . . . . . . . . . . . . . . . . . 54 Primary and Secondary School Participation . . . . . . . . . . . . . . . . . . . . . . 55 Adult Literacy . . . . . . . . . . . . . . . . . . . . . 56 CHILD PROTECTION . . . . . . . . . . . . . . . . . 57 Birth Registration . . . . . . . . . . . . . . . . . . . 57 Child Labour . . . . . . . . . . . . . . . . . . . . . . 57 Child Discipline . . . . . . . . . . . . . . . . . . . . 58 Early Marriage . . . . . . . . . . . . . . . . . . . . . 59 Domestic Violence. . . . . . . . . . . . . . . . . . . 60 Child Disability . . . . . . . . . . . . . . . . . . . . . 60 HIV/AIDS AND SEXUAL BEHAVIOUR . . . . . . . . 61 Knowledge of HIV Transmission and Condom Use . . . . . . . . . . . . . . . . . . . 61 Sexual Behaviour Related to HIV Transmission . . . . . . . . . . . . . . . . . 64 List of References . . . . . . . . . . . . . . . . . . . 65 APPENDICES Appendix A SAMPLE DESIGN . . . . . . . . . . . . . . . . . . . 69 Appendix B LIST OF PERSONNEL INVOLVED IN THE SURVEY . . . . . . . . . . . . . 74 Appendix C ESTIMATES OF SAMPLING ERRORS . . . . . . . . . . . . . . . . . . . . . . . . 75 Appendix D DATA QUALITY TABLES . . . . . . . . . . . . . . . 86 Appendix E MICS INDICATORS: NUMERATORS AND DENOMINATORS . . . . . . . . 97 Appendix F QUESTIONNAIRES . . . . . . . . . . . . . . . . . . 105 TABLE HH.1 Results of household and individual interviews . . . . . . . . . 143 TABLE HH.2 Household age distribution by sex. . . . . 144 TABLE HH.3 Household composition . . . . . . . . . . 145 TABLE HH.4 Women’s background characteristics . . . 146 TABLE HH.5 Children’s background characteristics . . . 147 TABLE CM.1 Child mortality . . . . . . . . . . . . . . 148 TABLE CM.2 Children ever born and proportion dead. . . . . . . . . . . . 149 TABLE NU.1 Child malnourishment . . . . . . . . . . . 150 TABLE NU.2 Initial breastfeeding . . . . . . . . . . . . 151 TABLE NU.3 Breastfeeding . . . . . . . . . . . . . . . 152 TABLE NU.4 Adequately fed infants . . . . . . . . . . 153 TABLE NU.5 Low birth weight infants . . . . . . . . . 154 TABLE CH.1 Vaccinations in first year of life . . . . . . 155 TABLE CH.1a Vaccinations in first year of life (BCG vaccine calculated to mother’s report only) . . . . . . . . . . 156 TABLE CH.2 Vaccinations by background characteristics . . . . . . . . . . . . . . 157 TABLE CH.3 Oral rehydration treatment . . . . . . . . 158 TABLE CH.4 Home management of diarrhoea . . . . . 160 TABLE CH.5 Care seeking for suspected pneumonia . . . . . . . . . . . . . . . . 162 TABLE CH.6 Antibiotic treatment of pneumonia . . . . 163 TABLE CH.6a Knowledge of the two danger signs of pneumonia . . . . . . . . . . . . 164 TABLE CH.7 Solid fuel use . . . . . . . . . . . . . . . 165 TABLE CH.8 Solid fuel use by type of stove or fire . . . 166 TABLE EN.1 Use of improved water sources . . . . . . 167 TABLE EN.2 Household water treatment . . . . . . . . 168 TABLE EN.3 Time to source of water . . . . . . . . . 169 TABLE EN.4 Person collecting water . . . . . . . . . . 170 TABLE EN.5 Use of sanitary means of excreta disposal . . . . . . . . . . . . 171 TABLE EN.6 Disposal of child’s faeces . . . . . . . . . 172 TABLE EN.7 Use of improved water sources and improved sanitation. . . . . . . . . . 173 TABLE EN.8 Security of tenure . . . . . . . . . . . . . 174 TABLE EN.9 Durability of housing . . . . . . . . . . . 175 TABLE EN.10 Slum housing . . . . . . . . . . . . . . . 176 TABLE RH.1 Use of contraception . . . . . . . . . . . 177 TABLE RH.2 Unmet need for contraception. . . . . . . 178 TABLE RH.3 Antenatal care provider . . . . . . . . . . 179 TABLE RH.4 Antenatal care . . . . . . . . . . . . . . 180 TABLE RH.5 Assistance during delivery . . . . . . . . 181 TABLE CD.1 Family support for learning . . . . . . . . 182 TABLE CD.2 Learning materials. . . . . . . . . . . . . 183 TABLE CD.3 Children left alone or with other children . . . . . . . . . . . . . . . 184 TABLE ED.1 Early childhood education . . . . . . . . . 185 TABLE ED.2 Primary school entry . . . . . . . . . . . 186 TABLE ED.3 Primary school net attendance ratio. . . . . . . . . . . . . . 187 TABLE ED.4 Secondary school net attendance ratio. . . . . . . . . . . . . . 188 TABLE ED.4w Secondary school age children attending primary school . . . . . . . . . 189 TABLE ED.5 Children reaching grade 5 . . . . . . . . . 190 TABLE ED.6 Primary school completion and transition to secondary education. . . 191 TABLE ED.7 Education gender parity . . . . . . . . . . 192 TABLE ED.8 Adult literacy . . . . . . . . . . . . . . . 193 TABLE CP.1 Birth registration . . . . . . . . . . . . . 194 TABLE CP.2 Child labour . . . . . . . . . . . . . . . . 195 TABLE CP.3 Labourer students and student labourers . . . . . . . . . . . 196 TABLE CP.4 Child discipline . . . . . . . . . . . . . . 197 TABLE CP.5 Early marriage . . . . . . . . . . . . . . . 198 TABLE CP.6 Spousal age difference . . . . . . . . . . 199 TABLE CP.7 Attitudes toward domestic violence . . . . 200 TABLE CP.8 Child disability . . . . . . . . . . . . . . 201 TABLE HA.1 Knowledge of preventing HIV transmission . . . . . . . . . . . . . 202 TABLE HA.2 Identifying misconceptions about HIV/AIDS . . . . . . . . . . . . . . 203 TABLE HA.3 Comprehensive knowledge of HIV/AIDS transmission . . . . . . . . . 204 TABLE HA.4 Knowledge of mother-to-child HIV transmission . . . . . . . . . . . . . 205 TABLE HA.5 Attitudes toward people living with HIV/AIDS . . . . . . . . . . . . . . 206 TABLE HA.6 Knowledge of a facility for HIV testing . . . . . . . . . . . . . . 207 TABLE HA.7 HIV testing and counselling coverage during antenatal care . . . . . . 208 TABLE HA.8 Sexual behaviour that increases risk of HIV infection . . . . . . . . . . . . 209 TABLE HA.9 Condom use at last high-risk sexual encounter . . . . . . . . . . . . . 210 List of TablesList of Tables List of FiguresList of Figures FIGURE HH.1 Age and sex distribution of household population, Serbia, 2005 . . . . . . . . . . . . . . . . 25 FIGURE NU.1 Percentage of children under-5 who are undernourished, Serbia, 2005 . . . . . . . . . . . . . . . . 31 FIGURE NU.1a Distribution of weight-for-age among children under five, Serbia, 2005 . . . . . . 32 FIGURE NU.1b Distribution of height-for-age among children under five, Serbia, 2005 . . . . . . 32 FIGURE NU.1c Distribution of weight-for-height among children under five, Serbia, 2005 . . . . . . 32 FIGURE NU.1d Prevalence of malnutrition in children under five, Serbia, 1996–2005 . . . . . . . 33 FIGURE NU.2 Percentage of mothers who started breastfeeding within one hour and within one day of birth, Serbia, 2005 . . . . . . . 34 FIGURE NU.3 Infant feeding patterns by age: Percentage distribution of children under the age of 3 by feeding pattern by age group, Serbia, 2005 . . . . . . . . 35 FIGURE NU.3a Progress in major breastfeeding indicators, Serbia, 1996–2005 . . . . . . . 36 FIGURE CH.1 Percentage of children aged 18–29 months who received the recommended vaccination by the age of 12 months (by 18 months in case of MMR). . . . . . . 39 FIGURE EN.1 Percentage distribution of household members by source of drinking water, Serbia, 2005 . . . . . . . . . . . . . . . . 45 FIGURE RH.1 Use of contraception, modern and traditional contraceptive methods, Serbia, 2005 . . . . . . . . . . . . . . . . 49 FIGURE ED.1 Percentage of children aged 36–59 months who are attending some form of organised early childhood education program, Serbia, 2005 . . . . . . . . . . . 54 FIGURE HA.1 Percentage of women who have comprehensive knowledge of HIV/AIDS transmission, Serbia, 2005 . . . . . . . . . 62 FIGURE HA.2 Sexual behaviour that increases risk of HIV infection, Serbia, 2005 . . . . . . . 64 7MONITORING THE SITUATION OF CHILDREN AND WOMEN List of AbbreviationsList of Abbreviations AIDS Acquired Immune Deficiency Syndrome BCG Bacillis-Cereus-Geuerin (Tuberculosis) CSPro Census and Survey Processing System DPT Diphteria Pertussis Tetanus EPI Expanded Programme on Immunization GPI Gender Parity Index HIV Human Immunodeficiency Virus IPV Inactivated Polio Vaccine IUD Intrauterine Device LAM Lactational Amenorrhea Method MDG Millennium Development Goals MICS Multiple Indicator Cluster Survey MMR Measles, Mumps, and Rubella MoH Ministry of Health NAR Net Attendance Rate NGO Non Governmental Organization ORS Oral Rehydration Solution ORT Oral Rehydration Therapy SPSS Statistical Package for Social Sciences SRSWoR Simple Random Sampling Without Replacement STI Sexually Transmitted Infection UNAIDS United Nations Programme on HIV/AIDS UNDP United Nations Development Programme UNFPA United Nations Population Fund UNGASS United Nations General Assembly Special Session on HIV/AIDS UNICEF United Nations Children’s Fund WFFC World Fit For Children WHO World Health Organization 9MONITORING THE SITUATION OF CHILDREN AND WOMEN SummarySummary Table of FindingsTable of Findings Multiple Indicator Cluster Surveys (MICS) and Millennium Development Goals (MDG) Indicators, Serbia, 2005 TO P I C M I C S INDICATOR NUMBER M D G INDICATOR NUMBER I N D I C AT O R VA LU E CHILD MORTALITY Child mortality 1 1 13 Under-five mortality rate – Roma in Roma settlements 28 per thousand 2 14 Infant mortality rate – Roma in Roma settlements 25 per thousand NUTRITION Nutritional status 6 4 Underweight prevalence 1.6 percent 7 Stunting prevalence 5.9 percent 8 Wasting prevalence 3.3 percent Breastfeeding 45 Timely initiation of breastfeeding 17.4 percent 15 Exclusive breastfeeding rate 15.1 percent 16 Continued breastfeeding rate at 12–15 months 22.4 percent at 20–23 months 8.2 percent 17 Timely complementary feeding rate 39.0 percent 18 Frequency of complementary feeding 33.5 percent 19 Adequately fed infants 24.6 percent Low birth weight 9 Low birth weight infants 5.0 percent 10 Infants weighed at birth 97.9 percent CHILD HEALTH Immunization 25 Tuberculosis immunization coverage 74.1 percent 26 Polio immunization coverage 88.2 percent 27 DPT immunization coverage 89.7 percent 28 15 Measles (MMR) immunization coverage 84.1 percent 31 Fully immunized children 43.6 percent 1 The module on child mortality was used only for Roma living in Roma settlements 10 MICS3 FULL TECHNICAL REPORT TO P I C M I C S INDICATOR NUMBER M D G INDICATOR NUMBER I N D I C AT O R VA LU E Care of illness 33 Use of oral rehydration therapy (ORT) 94.0 percent 34 Home management of diarrhoea 26.2 percent 35 Received ORT or increased fluids, and continued feeding 71.0 percent 23 Care seeking for suspected pneumonia 92.5 percent 22 Antibiotic treatment of suspected pneumonia 56.8 percent Solid fuel use 24 29 Solid fuels 33.5 percent ENVIRONMENT Water and Sanitation 11 30 Use of improved drinking water sources 98.9 percent 13 Water treatment 4.5 percent 12 31 Use of improved sanitation facilities 99.8 percent 14 Disposal of child’s faeces 36.7 percent Security of tenure and durability of housing 93 Security of tenure 16.5 percent 94 Durability of housing 0.4 percent 95 32 Slum household 18.9 percent REPRODUCTIVE HEALTH Contraception and unmet need 21 19c Contraceptive prevalence 41.2 percent 98 Unmet need for family planning 28.5 percent 99 Demand satisfied for family planning 59.1 percent Maternal and newborn health 20 Antenatal care 98.2 percent 44 Content of antenatal care 99.0 percent Blood test taken 95.6 percent Blood pressure measured 95.5 percent Urine specimen taken 95.4 percent Weight measured 93.5 percent 4 17 Skilled attendant at delivery 99.0 percent 5 Institutional deliveries 98.8 percent CHILD DEVELOPMENT Child development 46 Support for learning 84.4 percent 47 Father’s support for learning 70.0 percent 48 Support for learning: children’s books 79.3 percent 49 Support for learning: non-children’s books 75.7 percent 50 Support for learning: materials for play 20.9 percent 51 Non-adult care 8.8 percent 11MONITORING THE SITUATION OF CHILDREN AND WOMEN EDUCATION Education 52 Pre-school attendance 32.5 percent 53 School readiness 89.0 percent 54 Net intake rate in primary education 93.6 percent 55 6 Net primary school attendance rate 98.4 percent 56 Net secondary school attendance rate 83.8 percent 57 7 Children reaching grade ﬁ ve 99.8 percent 58 Transition rate to secondary school 97.1 percent 59 7b Primary completion rate 90.7 percent 61 9 Gender parity index primary school 1.00 ratio secondary school 1.08 ratio Literacy 60 8 Adult literacy rate 95.6 percent CHILD PROTECTION Birth registration 62 Birth registration 98.9 percent Child labour 71 Child labour 4.4 percent 72 Labourer students 93.4 percent 73 Student labourers 4.5 percent Child discipline 74 Child discipline Any psychological/physical punishment 72.7 percent Early marriage 67 Marriage before age 15 0.8 percent Marriage before age 18 8.4 percent 68 Young women aged 15–19 currently married/in union 5.8 percent 69 Spousal age diﬀ erence Women aged 15–19 26.3 percent Women aged 20–24 13.7 percent Domestic violence 100 Attitudes towards domestic violence 6.2 percent Disability 101 Child disability 11.3 percent TO P I C M I C S INDICATOR NUMBER M D G INDICATOR NUMBER I N D I C AT O R VA LU E 12 MICS3 FULL TECHNICAL REPORT HIV/AIDS AND SEXUAL BEHAVIOUR HIV/AIDS knowledge and attitudes 82 19b Comprehensive knowledge about HIVprevention among young people 42.3 percent 89 Knowledge of mother-to-child transmission of HIV 57.0 percent 86 Attitude towards people with HIV/AIDS 36.5 percent 87 Women who know where to be tested for HIV 69.4 percent 88 Women who have been tested for HIV 7.0 percent 90 Counselling coverage for the preventionof mother-to-child transmission of HIV 14.3 percent 91 Testing coverage for the preventionof mother-to-child transmission of HIV 8.9 percent Sexual behaviour 84 Age at ﬁ rst sex among young people 1.1 percent 92 Age-mixing among sexual partners 7.4 percent 83 19a Condom use with non-regular partners 74.4 percent 85 Higher risk sex in the last year 60.8 percent TO P I C M I C S INDICATOR NUMBER M D G INDICATOR NUMBER I N D I C AT O R VA LU E 13MONITORING THE SITUATION OF CHILDREN AND WOMEN The Serbia Multiple Indicator Survey 2005 is a nationally representative sample survey of households, women and children. The results pertain to October 2005 – January 2006, when the fieldwork was conducted. Child mortality � The infant mortality rate among the Roma children living in Roma settlements is esti- mated at 25 per thousand live births, while the probability of dying under the age of 5 is around 28 per thousand live births (almost three times higher than national average). Nutritional status � 15 percent of children under the age of five are overweight. � The prevalence of child malnourishment (moderate and severe) in Serbia is relatively low: the prevalence of underweight is nearly 2 percent, 6 percent of children are stunted and 3 percent are wasted. � The opposite nutritional status is found among Roma children living in Roma settlements – while prevalence of malnourishment is several times higher than the national average (nearly 8 percent of children are underweight and 20 percent stunted), the prevalence of obesity is over two times lower (nearly 7 percent of Roma children are moderately or severely obese). Breastfeeding � Approximately 23 percent of children aged under four months were exclusively breastfed, whereas all infants should still be breastfed exclusively at this age. The percentage of ex- clusively breastfed children aged under six months is significantly lower, at 15 percent. � Only every sixth child was breastfed for the first time within one hour of birth. The practice is much less present in Belgrade (9 percent) than in West and South-East Serbia (23 and 25 percent, respectively). 67 percent of children in Serbia started breastfeeding within one day of birth. � Between the ages of 6–9 months, 39 percent of children are receiving breast milk and solid or semi-solid foods. By the ages of 12–15 months and 20–23 months respectively, 22 and 8 percent of children are still being breastfeed. Roma children from Roma settlements are more likely to continue to be breastfed than the entire population, since 60 percent of Roma children by age 12–15 months and 34 percent by age 20–23 months are still breastfed. Low birth weight � Out of 98 percent of weighed live births, 5 percent were below 2500 grams. In the Roma population, only 90 percent were weighed, and 9 percent of those live births were underweight. Immunization � Roughly two thirds of children had their own personal health card recording their vaccines. � Only 74 percent of children aged 18–29 months received a BCG vaccination in the first year EXECUTIVE SUMMARYEXECUTIVE SUMMARY 14 MICS3 FULL TECHNICAL REPORT of life. This result is quite unexpected and is most probably due to the lack of information on the BCG vaccination from their vaccination cards, since according to the mother’s report, the BCG vaccine was received by nearly all children aged 18–29 months. � All three doses of DPT and oral polio were given to 90 percent and 88 percent, respective- ly. 84 percent of children aged 18–29 months received a measles vaccine in the form of the measles-mumps-rubella (MMR) vaccine by the age of 18 months. � Only 44 percent of children had all eight recommended vaccinations according to the national immunization schedule. The low coverage is mostly due to the low immuniza- tion coverage against tuberculosis. The figure is not realistic and mainly shows that per- sonal immunization records are not a reliable source of information. � There are significant territorial differences in immunization coverage – the lowest percentage of fully immunized children is in South-East Serbia, at 31 percent, and the highest in Vojvodina and East Serbia, at 81 and 63 percent, respectively. Children liv- ing in urban settlement s are more likely to be fully immunized; 62 percent of children living in urban, compared to 52 percent of children from rural areas have received all the recommended vaccinations. � Ethnicity is strongly related to immuniza- tion coverage. While 57 percent of Serbian children aged 18–29 have been fully im- munized, only 27 percent of Roma children living in Roma settlements have received all vaccinations. Oral rehydration treatment � Overall, 5 percent of children under the age of five had diarrhoea in the two weeks preceding the survey. High prevalence of diarrhoea is noticed among children living in Roma settlements and in the poorest households, at 13 and 7 percent respectively. � The use of oral rehydration therapy (ORT), which is the application of ORS fluid or other recommended home made fluids is satisfacto- rily high, at 94 percent. � However, home management of diarrhoea is only 26 percent in Serbia, with significant gen- der (29 percent of boys compared to 23 percent of girls) and urban/rural differences (29 percent and 22 percent respectively) among children re- ceiving home management of diarrhoea. Also, home management of diarrhoea is much lower (18 percent) among children from the poorest households. � 71 percent of children who had diarrhoea in two weeks prior to the survey received ORT (or increased fluids) and continued feeding, with the same pattern as for home manage- ment of diarrhoea. Care seeking and antibiotic treatment of pneumonia � 3 percent of children aged 0–59 months had suspected pneumonia during the two weeks preceding the survey. 92 percent of children with suspected pneumonia were taken to an appropriate health provider (94 percent of boys and 90 percent of girls). � 57 percent of children with suspected pneumo- nia received antibiotics in the last two weeks. Antibiotics were given more to boys (61 percent) than girls (52 percent) and more in urban (59 percent) than in rural (54 percent) areas. � Every third mother/caretaker recognizes two danger signs of pneumonia (fast and difficult breathing). Knowledge is strongly correlated to the region, ethnicity and wealth index. While 57 percent of mothers/caregivers in Belgrade recognize the two danger signs of pneumonia, only 15 percent of mothers/caregivers in West Serbia have similar knowledge. Poorer knowledge is noticed among Muslim/Bosnians and Roma mothers – only 8 and 12 percent respectively, re- cognize both signs of pneumonia. 28 percent of mothers from the poorest households, compared to 40 percent from the richest know those signs. 15MONITORING THE SITUATION OF CHILDREN AND WOMEN � Interestingly, fever is considered as the most dangerous sign of child illness in Serbia – 82 percent of mothers think that their child should be taken to a health facility as soon as it develops a fever. Solid fuel use � One third of households in Serbia are using solid fuels for cooking. A much higher propor- tion of solid fuels usage is noticed among the poorest socio-economic classes (86 percent), in Roma households (85 percent) and in rural areas (61 percent). Water sanitation � 99 percent of the population have access to an im- proved drinking water source, if one uses a broad definition of access which, by improved drinking water source, includes piped water, a public tap/ standpipe, a tubewell/borehole, a protected well and spring. 77 percent of the population uses wa- ter piped into their dwelling or yard from a public or local water supply as the main source of drink- ing water. Such access is much higher in urban areas (91 percent) than in rural areas (60 percent). In rural areas, 15 percent of the population have a tubewell/borehole with a pump, and 18 percent have a protected well or spring. � Virtually the entire population uses sanitary means of excreta disposal. 89 percent have a flush toilet connected either to a sewage system or septic tank. Septic tanks are much more common in rural areas; 64 percent of the rural population uses a septic tank, compared to 13 percent of the urban population. Usage of traditional pit latrines is quite common among the Roma population (48 percent of them use one). 5 percent of the Roma popu- lation do not even have sanitation facilities. Security of tenure and durability of housing � 13 percent of households do not have formal residential documentation. Regional differ- ences are significant, with the best situation in Belgrade, where 7 percent of households do not have formal documentation, and worst in South-East Serbia, where 18 percent of house- holds are missing formal documentation. The least security of tenure is found among Roma living in Roma settlements where 38 percent of households do not have formal residen- tial documentation, and among the poorest households, with 26 percent of the house- holds without formal documentation. � Almost the entire population lives in durable housing – less than 1 percent of households and household members are living in dwell- ings which are considered non-durable. The situation in Roma settlements is quite differ- ent, with as much as 12 percent of the Roma population living in non-durable households. Contraception � Current use of any contraception was reported by 41 percent of married women/in union women. Among women in Serbia, traditional methods are more popular than modern ones, 23 compared to 19 percent. The most popular traditional methods are withdrawal and peri- odical abstinence (14 and 8 percent respective- ly), while the most popular modern method is the condom (8 percent). � Contraceptive prevalence ranges from 27 per- cent in Central Serbia to 54 percent in East Serbia. Contraceptive prevalence increases with age, up to the age of 40, and then there is a pattern of decreased contraception use. Women’s education level, ethnicity and wealth index are significantly linked to con- traceptive prevalence. � Contraceptive usage is highest among Hungarian women, where 57 percent of women are using mainly modern methods of contraception. Only one in four married Roma women are using any contraceptive method, usually withdrawal (every fifth). Traditional methods of contraception are also present among Muslim/Bosnian women whose main choice is periodical abstinence. 16 MICS3 FULL TECHNICAL REPORT Unmet need � Almost one third of married or in union women in Serbia have an unmet need for con- traception. The need for contraception is less satisfied among the “highest risk” population: Roma women living in Roma settlements, and the poorest and most uneducated women. Antenatal care � 99 percent of pregnant women received antenatal care one or more times during pregnancy. Lower antenatal care coverage is noticed among Roma, the youngest and less educated women, and women from the poorest households. � 98 percent of women aged 15–49 who gave birth in the two years preceding the survey received antenatal care from skilled personnel. Antenatal care provided by medical person- nel, especially a doctor, is significantly lower among Roma and less educated women. Assistance at delivery � With 99 percent of women, skilled personnel assisted at the delivery. Approximately the same percentage of women was delivered in a health facility. Roma women from Roma settlements and Muslim/Bosnian women are behind the national average – 93 and 94 percent of those women respectively were delivered by skilled personnel. � Medical doctors assisted in 87 percent of the deliveries and an auxiliary midwife assisted in 10 percent of cases. Child development � For about 84 percent of children under the age of 5, an adult engaged in four or more activities that promote learning and school readiness during the 3 days preceding the survey. The average number of activities was 5. The involvement of fathers in such activities was significantly lower, 70 percent, with the average number of activities at 2.3. � Children from the poorest and Roma families, as well as children whose parents are less well educated are less likely to be involved in activi- ties that promote learning than average. � About 9 percent of children under the age of 5 in Serbia were left with inadequate care in the week preceding the survey. Female, older chil- dren and children from rural areas are more often left with inadequate care. This prac- tice is particularly present among Roma and Muslim/Bosnian families (18 and 16 percent of children left without proper care, respectively). Pre-school attendance and school readiness � 32 percent of children aged 36–59 months were attending some form of organised early childhood education programme. Attendance was three times higher in urban than rural ar- eas. Higher education of a child’s mother and higher household wealth status imply a higher proportion of children attending pre-school. � Roma children attend pre-school eight times less than the rest of the population. Primary and secondary school participation � 94 percent of children of primary school entry age in Serbia are currently attending first grade. Roma children start primary education on time more rarely than other children (66 percent). � 98 percent of children of primary school age attend primary school. Only 74 percent of Roma children of this age attend school. � Almost all of the children who entered the first grade of primary school eventually reach grade five. � About 84 percent of children in Serbia, be- tween the ages of 15 and 18, are attending secondary or higher school. Only 10 percent of Roma children from Roma settlements are attending school at that age. Children of secondary school age from wealthier house- holds are more likely to attend secondary school or higher. 17MONITORING THE SITUATION OF CHILDREN AND WOMEN � The Gender Parity Index (GPI) for primary school in Serbia is 1.0, indicating no difference in primary school attendance of girls and boys. However, the indicator goes up to 1.1 for sec- ondary education. � The disparity of girls is only pronounced in Roma settlements, where the GPI for primary and sec- ondary schools are 0.94 and 0.42, respectively. Adult literacy � The literacy rate of females, aged between 15 and 24 in Serbia is 96 percent. The literacy level is lower among women with no or primary edu- cation. Younger women aged 15–19 are less lit- erate than women aged 20–24 (93 compared to 98 percent). Roma women living in Roma settle- ments showed a lower level of literacy; only 52 percent of young Roma women are considered literate. The second and middle wealth index quintile classes are the most literate. Birth Registration � The births of 99 percent of children under five years of age in Serbia have been registered. There are no significant variations in birth reg- istration across sex, age, or education categories. Only Roma from Roma settlements are some- what less likely to have their births registered than other children (95 percent registered). Child Labour � 4 percent of children aged between 5 and 14, in Serbia are involved in child labour. � Children living in rural areas are twice as involved in child labour activities (6 percent), than urban children (3 percent). Even 8 percent of children from the poorest households and 7 percent of Roma children are involved in child labour. � Out of the 92 percent of children aged between 5 and 14 who are attending school, 5 percent are involved in child labour activities. On the other hand, out of the 4 percent of children classified as child labourers, the majority of them are also attending school (93 percent). Child Discipline � In Serbia, 73 percent of children aged be- tween 2 and 14 were subjected to at least one form of psychological or physical pun- ishment by their mothers/caretakers or other household members. � 62 percent of children were exposed to psy- chological punishment, 51 percent to minor physical punishment and 7 percent were sub- jected to severe physical punishment. Severe physical punishment is a frequently used dis- ciplinary method among Roma (21 percent) and Muslims/Bosnians (15 percent). � 6 percent of mothers/caretakers in Serbia believe that in order to raise their children properly, they need to physically punish them. � Every fifth child in Serbia has been disciplined through non-violent methods. � Male children were subjected more to both minor and severe physical discipline (53 and 8 percent) than female children (49 and 5 percent). Violent disciplining is most prac- tised in South-East Serbia. Less educated mothers and mothers from the poorest households are using violent methods more in child disciplining. Early marriage � The proportion of women aged between 15 and 49 who got married before the age of 15 is very low (1 percent). 6 percent of women aged 15 to 19 are currently married or in union. This practice is more present in East Serbia, rural areas, among the less educated, the poorest and particularly among Roma, where almost half of Roma women from Roma settlements were married before the age of 18. � 26 percent of young married women aged 15 to 19 are married to a partner 10 or more years older. The percentage is much lower (14 percent), among married women aged 20 to 24 years. This phenomenon is mainly correlated to poverty and lower education. 18 MICS3 FULL TECHNICAL REPORT Domestic violence � 6 percent of women in Serbia feel that their husband/partner has a right to hit or beat them for at least one of a variety of reasons. Women who approve their partner’s violence, in most cases agree and justify violence in in- stances when they neglect the children (6 per- cent), or if they demonstrate their autonomy, e.g. go out without telling their husbands or argue with them (2 percent). � Domestic violence is more accepted in South and West Serbia (approximately 12 percent) than in other parts of the country. Acceptance is more present among the poorest and less educated, and also currently married women. It is also strongly correlated to ethnicity – ev- ery third Muslim/Bosnian and Roma woman believes it is justified for a husband to beat his wife/partner. Child disability � 11 percent of children aged between 2 and 9 have at least one reported disability. Child disabilities are more frequent among Roma children from Roma settlements (23 percent), children from the poorest households (17 percent) and those whose mothers are less educated (15 percent). Knowledge of HIV transmission � In Serbia, 98 percent of the women inter- viewed have heard of AIDS. Yet, the number of women who knew all three main ways of preventing HIV transmission (having only one faithful uninfected partner, using a condom during each act of sexual intercourse, and abstaining from sex) is only 36 percent. � Only half of the women correctly identify the two most common misconceptions about HIV transmission (that HIV cannot be transmit- ted by sharing food and that HIV cannot be transmitted by mosquito bites) and know that a healthy looking person can be infected. This percentage is higher among more educated and wealthier women. � Only 37 percent of women in Serbia have compre- hensive knowledge about HIV transmission (iden- tify 2 prevention methods and 3 misconceptions). 42 percent of women aged 15–24 have compre- hensive knowledge about HIV transmission. � 57 percent of women of reproductive age know the three ways in which HIV can be transmit- ted from mother to child. � 69 percent of women know a place to get test- ed for AIDS. Seven percent have been tested, of whom 84 percent have been given the result. � 64 percent of women expressed a discrimina- tory attitude towards people with HIV/AIDS. Sexual Behaviour � Only 1 percent of women aged 15 to 19 had sex before the age of 15, while 19 percent of women aged 20 to 24 reported having sex be- fore the age of 18. A different pattern is found among Roma women from Roma settlements (16 percent had sex before the age of 15 and 51 percent before the age of 18) and girls who terminate their education at primary school (3 percent had sex before the age of 15 and 42 percent before the age of 18). � Having sex with a partner 10 or more years older is reported by one in twelve women. It appears that women with primary or less education and women from the poorest house- holds are more inclined to these kinds of rela- tionship (20 and 12 percent respectively). � Over 60 percent of women aged 15–24 report having sex with a non-regular partner in the 12 months prior to the MICS. Three-quarters of those women reported condom usage when they had sex with the high risk partner. 19MONITORING THE SITUATION OF CHILDREN AND WOMEN Background This report is based on the Serbia Multiple Indicator Cluster Survey, conducted in 2005 by UNICEF, the Statistical Office of the Republic of Serbia and the Strategic Marketing Research Agency. The survey provides valuable information on the situation of children and women in Serbia. It was based, to a large extent, on the need to monitor progress towards goals and targets deriving from recent international agreements: the Millennium Declaration, adopt- ed by all 191 United Nations Member States in September 2000, and the World Fit for Children Plan of Action, adopted by 189 Member States at the United Nations Special Session on Children in May 2002. Both of these commitments build upon promises made by the international com- munity at the 1990 World Summit for Children. In signing these international agreements, gov- ernments committed themselves to improving conditions for their children and to monitoring progress towards that end. UNICEF was assigned a supporting role in this task. The Government of the Republic of Serbia ad- opted the Poverty Reduction Strategy Plan in 2003, the National Plan of Action for Children (NPA) in 2004, and in 2005 the United Nations Millennium Declaration. By undertaking these international obligations, the Republic of Serbia committed itself to monitor and assess progress towards interna- tionally defined goals and targets. MICS3 is the third round of a nation-wide household Multiple Indicator Cluster Survey. The survey provides the largest single source of data for reporting on progress towards the aforementioned goals, which will give a rich foundation of comparative data for comprehensive progress reporting. This final report presents the results of the indicators and topics covered in the survey. Survey Objectives The 2005 Serbia Multiple Indicator Cluster Survey has the fallowing primary objectives: � To provide up-to-date information for assessing the situation of children and women in Serbia; � To learn about existing disparities in the status of children and women in Serbia; � To provide data needed for monitoring progress towards goals established by the Millennium Development Goals, the goals of A World Fit For Children (WFFC), and other internationally agreed-upon goals, as a basis for future action; � To contribute to the improvement of data and monitoring systems in Serbia and to strengthen technical expertise in the design, implementation, and analysis of such systems. INTRODUCTIONINTRODUCTION 20 MICS3 FULL TECHNICAL REPORT A Commitment to Action: National and International Reporting Responsibilities The governments that signed the Millennium Declaration and the World Fit for Children Declaration and Plan of Action also committed themselves to monitoring progress towards the goals and objectives they contained: “We will monitor regularly at the national level and, where appropriate, at the regional level and assess progress towards the goals and targets of the present Plan of Action at the national, regional and global levels. Accordingly, we will strengthen our national statistical capacity to collect, analyse and disaggregate data, including by sex, age and other relevant factors that may lead to disparities, and support a wide range of child-focused research. We will enhance international cooperation to support statistical capacity-building efforts and build community capacity for monitoring, assessment and planning.” (A World Fit for Children, paragraph 60) “…We will conduct periodic reviews at national and subnational levels of progress in order to address obstacles more effectively and accelerate procedures…” (A World Fit for Children, paragraph 61) The Plan of Action (paragraph 61) also calls for the specific involvement of UNICEF in the preparation of periodic progress reports: “…As the world’s lead agency for children, the United Nations Children’s Fund is requested to continue to prepare and disseminate, in close collaboration with Governments, relevant funds, programmes and the specialized agencies of the United Nations system, and all other relevant actors, as appropriate, information on the progress made in the implementation of the Declaration and the Plan of Action.” Similarly, the Millennium Declaration (paragraph 31) calls for periodic reporting on progress: “…We request the General Assembly to review on a regular basis the progress made in implementing the provisions of this Declaration, and ask the Secretary-General to issue periodic reports for consideration by the General Assembly and as a basis for further action.” 21MONITORING THE SITUATION OF CHILDREN AND WOMEN Sample Design The sample for the Serbia Multiple Indicator Cluster Survey (MICS) was designed to provide estimates on a large number of indicators on the situation of children and women at the national level, for urban and rural areas, and for six regions: Vojvodina, Belgrade, West, Central, East and South-East Serbia. Belgrade has a large population (almost one-quarter of the total) and its predomi- nantly urban characteristics make it necessary to separate it from the rest of Central Serbia, to which it administratively belongs. In order to look more deeply into existing ethnic disparities and to provide national estimates, a separate sample was designed for Roma living in Roma settlements. Important factors which influenced the sample design of both Serb and Roma samples are the fertility rate and number of household members. For example, one generation of Serb children makes up less than 1 percent of the population, and the average number of Serb household mem- bers is around three. But the situation in Roma settlements is quite different; the average house- hold size is around five and there is a significant- ly higher proportion of children under the age of 5 in the total Roma population. Because of these differences, the sample plan had to be modified. For both samples, regions were identified as the main sampling domains and the samples were selected in two stages. SAMPLE AND SURVEY SAMPLE AND SURVEY METHODOLOGYMETHODOLOGY In the case of Serbia without the Roma settle- ments sample, 400 census enumeration areas within each region with probability proportional to size were selected during the first stage. Since the sample frame (Census 2002) was not up to date, household lists in all selected enumera- tion areas were updated prior to the selection of households. Owing to the low fertility rate and small household size, households were stratified into two categories. One category of households consists of households with under 5 children, while the other category consists of households without children under 5. The allocation of the sample in the category of households with chil- dren was significantly greater than the allocation of the sample in the category of households with- out children. Based on the updated information, selected units were divided into clusters of 18 households on average, plus 3 backup households. Backup households were interviewed only if some of the first 18 households were not found. In the event that a household refused to be interviewed, a backup household was not contacted. In each cluster, the number of households with children was selected with probability proportional to size. In the case of the Roma population, the uni- verse could be defined only for Roma who live in separate settlements. During the first stage, 106 census enumeration areas were selected. The updating of household lists was done prior 22 MICS3 FULL TECHNICAL REPORT to household selection, but there was no need for sample stratification of households with and without children under 5. The average number of households selected in each cluster was 18 on average, plus 3 backup households. Secondly, after the household listing was car- ried out within the selected enumeration areas, a systematic sample of 7.974 households in Serbia without Roma from Roma settlements and 1.979 Roma households from Roma settlements was drawn up, which makes a total of 9.953 sampled households. Each selected enumeration area was visited during the field work period. The Serbia Multiple Indicator Cluster Survey sample is not self-weighted. For reporting of national level re- sults, sample weights were used. A more detailed description of the sample design can be found in Appendix A. Questionnaires Three sets of questionnaires were used in the survey: 1) a household questionnaire which was used to collect information on all de jure house- hold members, the household, and the dwelling; 2) a women’s questionnaire; and 3) an under-5’s questionnaire. The Household Questionnaire included the following modules: � Household listing � Education � Water and Sanitation � Household characteristics � Child Labour � Child Discipline � Child Disability � Roma in Roma settlements The Questionnaire for Individual Women was completed by all women aged 15–49 living in the households, and included the following modules: � Women’s information panel � Child Mortality (this module was used only in Roma households living in Roma settlements) � Maternal and Newborn Health � Security of tenure � Marriage/Union � Contraception � Attitudes toward domestic violence � Sexual behaviour � HIV/AIDS The Questionnaire for Children Under Five2 was completed by mothers of under-5 children; in cases when the mother was not listed in the household roster, a primary caretaker for the child was iden- tified and interviewed. The questionnaire included the following modules: � Under-Five Child’s Information Panel � Birth Registration and Early Learning � Child Development � Breastfeeding � Care of Illness � Immunization � Anthropometry The questionnaires are based on the MICS3 model questionnaire3. Certain changes were made ac- cording to country specific situations. Some of the original modules (tetanus, polygamy, Vitamin A modules, etc.) were not covered by this sur- vey since they were not applicable. On the other hand, some of the modules were expanded with additional questions in order to obtain data that 2 The terms “children under 5”, “children age 0–4 years”, and “children aged 0–59 months” are used interchangeably in this report. 3 The model MICS3 questionnaire can be found at www.childinfo.org. 23MONITORING THE SITUATION OF CHILDREN AND WOMEN is missing, but important for finding out more about existing practices and for future activity planning. For example, the Education module was expanded with additional questions concern- ing commuting to school – distance, cost, means and duration of the journey. Questionnaires used among the Roma population differ slightly from those used in non-Roma settlements. They were expanded with a new module which included ques- tions about the language spoken in the household, the type of Roma settlement, the main source of household income, etc. Questionnaires from the MICS3 model English version were translated into Serbian. After adaptation, they were translated into Albanian and Hungarian, and were pre-tested in several places in Serbia: Belgrade, Novi Sad, Subotica, Kraljevo, Kragujevac, Valjevo and Nis, at the end of September 2005. Based on the results of the pre-test, modifications were made to the word- ing of the questionnaires. A copy of the Serbia MICS questionnaires is provided in Appendix F. In addition to the administration of question- naires, fieldwork teams measured the weight and height of children under 5 years of age. Details and findings of these measurements are provided in the respective sections of this report. Training and Fieldwork The Statistical Office of the Republic of Serbia and the Strategic Marketing Research Agency were responsible for data collection. The whole territory of Serbia was divided into 18 districts according to the regional network of responsible institutions. In each district a team of people was selected – one supervisor for the district and the interviewers (whose number depended on the number of clusters in the region). The interviewers and supervisors were recruited according to their qualifications, communication skills, experience in fieldwork and knowledge of the region where the research was to be conducted. All field staff were selected before the updating of household listing started. Training of supervisors was conducted in September 2005, before the pre-test. Towards the end of the supervisor training period, supervisors spent five days to practise interviewing and check- ing questionnaires and methodology in several places: Belgrade, Novi Sad, Subotica, Kraljevo, Kragujevac, Valjevo and Nis. The field interview- ers, all highly skilled professionals with previous experience in similar surveys, were trained for three days in October 2005. Training included lectures on interview techniques and the contents of the questionnaires, and the interactive approach of practising interviews between trainees to gain practice in asking questions. A separate process was applied for the interviewers that were to conduct interviews in Roma settlements. Roma women included in the fieldwork received addi- tional special training. The data was collected by 47 teams; each com- prised of three to four interviewers. Each team in charge of data collection in Roma settlements included one or two members from the Roma women’s network, and one professional inter- viewer. One supervisor was in charge of two or three teams. Fieldwork began in October 2005 and concluded in January 2006. Data Processing Data was entered using the CSPro software into twenty computers by twenty data entry operators and four data entry supervisors. In order to ensure quality control, all questionnaires were entered twice and internal consistency checks were performed. Procedures and standard programmes developed under the global MICS3 project and adapted to the Serbia questionnaire were used throughout. Data processing began simultaneously with data collection in October 2005 and was completed in March 2006. Data was analysed using the Statistical Package for Social Sciences (SPSS) software programme, Version 14, and the model syntax and tabulation plans devel- oped by UNICEF for this survey. 24 MICS3 FULL TECHNICAL REPORT Sample Coverage O f the 9.953 households selected for the sample, 9.372 were found to be occupied. Of these, 8.730 were successfully inter- viewed for a household response rate of 93 percent. In the interviewed households, 7.895 women (aged 15–49) were identified. Of these, 7.516 were successfully interviewed, yielding a response rate of 95 percent. In addition, 3.838 children under the age of five were listed in the household questionnaire. Questionnaires were completed for 3.777 of these children, which corresponds to a response rate of 98 percent. Overall response rates of 89 and 92 percent are calculated for the women’s and under-5s’ inter- views respectively (Table HH.1). Response rates were similar across regions and areas. Despite the busy lifestyle of Belgrade resi- dents, surprisingly high response rates were re- corded in the capital. This can be attributed to a successful media campaign and interviewers’ dili- gent efforts. Although the inclusion of the Roma women’s network in fieldwork helped interviewers to gain the trust of Roma communities, the re- sponse rate in Roma settlements was still slightly lower than that of the total population. Thanks to previously updated lists of households in select- ed enumeration areas, there were no significant differences between the number of sampled and occupied households. Characteristics of Households The age and sex distribution of the surveyed population is provided in Table HH.2. The dis- tribution is also used to produce the population pyramid in Figure HH.1. In the 8.730 households successfully interviewed in the survey, 28.163 household members were listed. Of these, 13.614 were male, and 14.549 were female. These figures also indicate that the survey estimated the average household size at 3.2 persons. The age and sex distribution of the surveyed pop- ulation is in accordance with the 2002 Census and indicates negative population growth during the past decade. Serbia is characterised by a low proportion of under-five children and a high pro- portion of elderly. The majority of the population SAMPLE COVERAGE AND SAMPLE COVERAGE AND THE CHARACTERISTICSTHE CHARACTERISTICS OF HOUSEHOLDS OF HOUSEHOLDS AND RESPONDENTSAND RESPONDENTS 25MONITORING THE SITUATION OF CHILDREN AND WOMEN is in the 50–55 age group, i.e. the post Second World War baby boom generation. Children (up to 18 years of age) constitute 22 percent of the population. The single year age distribution (Table DQ.1 and Figure DQ.1a in Appendix D) shows a decrease in population after the fifties (expected, natural), and a constant decline in the young population. The male-female ratio shows small variations in the first 50 years of life after which the number of women increases and exceeds that of men. The Roma population single year distribution is completely different (Figure DQ.1b in Appendix D), and is comparable with other underdeveloped populations. The proportion of children under 5 in the Roma population is highest (about 12 percent), and then in each subsequent age band the propor- tion of the population progressively decreases. As a basic check on the quality of age reporting, the percentage of missing data is shown in the Table DQ.6 in Appendix D. The age of almost all the surveyed population was collected. 1 percent of all eligible women interviewed did not know their complete date of birth (ie. day, month and year). Yet, the year of birth and age were gathered for these women. For all children under 5, the complete date of birth (month and year) was col- lected, because field supervisors were instructed to repeat interviews in case information was missing. Table HH.3 provides basic background informa- tion on the households. Within households, the area (Serbia without Roma in Roma settlements or Roma in Roma settlements), the sex of the head of the household, the region, urban/rural status, the number of household members, and the ethnicity4 of the household head are shown in the table. These background characteristics are also used in subsequent tables in this report; the figures in the table are also intended to show the numbers of observations by major categories of analysis in the report. In addition to being useful in interpretation of survey results, background characteristics serve as a basic check on sample implementation. The totals of weighted and unweighted number of households are equal, since sample weights were normalised (See Appendix A). Table HH.3 also shows the proportions of households where at least one child under 18, at least one child under 5, and at least one eligible woman aged 15–49 were found. Looking at the number of households in each back- ground category, we can notice a significant differ- ence between weighted and unweighted numbers of Roma households from Roma settlements (the weighted number is 85 and the unweighted 1716). According to the 2002 Census data, the proportion of Roma in the general population is approximately 1 percent. In order to gain qualitative results for the Roma population it was necessary to do over-sam- pling, and later by using sample weights, the model was adjusted to Census data. About 60 percent of households are urban, while the rest are rural. The regional distribution is in accordance with Census data. The Vojvodina re- gion comprises the largest number of households with nearly one third of the total, while the smallest 4 This was determined by asking the respondents what ethnic group the head of household belonged to. MALES FEMALES 8 6 4 2 0 2 4 6 8 Figure HH.1 Age and sex distribution of household population, Serbia, 2005 0–4 5–9 10–14 15–19 20–24 25–29 30–34 35–39 40–44 45–49 50–54 55–59 60–64 65–69 70–74 75–79 80+ 26 MICS3 FULL TECHNICAL REPORT number of households is from East Serbia at 9 percent. In two thirds of interviewed households the head of the household is male, which corre- lates strongly with Serbian tradition in which the male figure represents the guardian and bread- winner of the family. Household distribution by the ethnicity of the head of the household shows that the biggest ethnic group is Serbian (87 per- cent), while the proportion of other ethnic groups is less than 5 percent each. The majority of households have between two and five members. In 56 percent of interviewed house- holds, lives at least one woman aged 15 to 49, and in 14 percent of interviewed households at least one child under 5 years of age. The fact that in 38 percent of households lives only one child under 18, shows that the Serbian population is “old”. Characteristics of Respondents Tables HH.4 and HH.5 provide information on the background characteristics of female respon- dents 15–49 years of age and of children under the age of 5. In both tables, the total numbers of weighted and unweighted observations are equal, since sample weights have been normalised (standardised). In addition to providing useful information on the background characteristics of women and children, the tables are also intended to show the number of observations in each back- ground category. These categories are used in the subsequent tabulations of this report. Like in the description of households’ background information, the weighted numbers of Roma women and children under 5 are significantly lower than unweighted numbers, due to over-sampling. By using sample weights, the model was adjusted to Census data. Table HH.4 provides background characteristics of female respondents 15–49 years of age. The table includes information on the distribution of women according to area, region, urban-rural set- tlements, age, marital status, motherhood status, education5, ethnicity and wealth index quintiles.6 Approximately, one third of interviewed eli- gible women live in Vojvodina, and one fifth in Belgrade. Distribution among the other four re- gions is approximately equal, with the exception of East Serbia where the proportion of eligible women in the interviewed population is 9 per- cent. This pattern is expected and follows the Census data. The majority of the sample are women 30–34 years of age, around 17 percent. The proportion of young women is significantly lower: there are 13 percent of both 15–19 and 20–24 age groups which could be the consequence of systematical negative natality growth in past years. Around two thirds of all women in this sample are married, while 30 percent have never been married. Distribution by motherhood status is similar: 65 percent of women have given birth, compared to 35 percent that have never given birth. The majority of interviewed women have secondary education (59 percent), while the pro- portion of women with none or primary educa- tion, and those with higher education is approxi- mately the same (20 percent). Additional analysis which is not shown in the table HH.4 indicates a 5 Unless otherwise stated, “education” refers to the educational level attended by the respondent throughout this report when it is used as a background variable. 6 Principal components analysis was performed by using information on the ownership of household goods and amenities (assets) to assign weights to each household asset, and obtain wealth scores for each household in the sample. The assets used in these calculations were as follows: number of rooms for sleeping per member; floor, roof and walls material of dwelling; type of water and sanitations; the type of fuel used for cooking; radio, mobile, phone, fridge, washing machine, dishwasher, computer, air conditioner, central heating and car. Each household was then weighted by the number of household mem- bers, and the household population was divided into five groups of equal size, from the poorest quintile to the richest quintile, based on the wealth scores of households they were living in. The wealth index is assumed to include the underlying long-term wealth through information on the household assets, and is intended to produce a ranking of households bv wealth, from poorest to richest. The wealth index does not provide information on absolute poverty, current income or expenditure levels, and the wealth scores calculated are applicable for only the particular data set they are based on. Further information on the construction of the wealth index can be found in Rutstein and Johnson, 2004, and Filmer and Pritchett, 2001. 27MONITORING THE SITUATION OF CHILDREN AND WOMEN strong correlation between woman’s educational level and the type of settlement she lives in. While only one in five women in urban settlements have either none or primary education, in rural areas there are as many as 30 percent of such women. Conversely, while approximately one third of ur- ban women have high education, this applies to only one in ten women living in rural settlements. It is important to notice that education levels are categorised, according to the country’s educational system, into three groups: none or primary, sec- ondary and higher/high. In accordance with household ethnic distribu- tion, the majority of women live in households where the ethnicity of the head of the house- hold is Serbian. The proportion of other ethnic groups is under 5 percent. As far as wealth index quintiles are concerned, less women live in the poorest households – about 15 percent – while the largest group are those living in the richest households, 23 percent. Some background characteristics of children un- der 5 are presented in Table HH.5. These include distribution of children according to several attributes: area of residence, sex and region; ur- ban/rural settlements; age in months; mother’s or caretaker’s education, ethnicity and wealth. The proportion of male and female children in the under-5 sample is approximately the same. Age distribution of children 0–59 months is well bal- anced. The proportion of mothers with secondary education is approximately the same as in the women’s sample. This is not the case with moth- ers with higher education; it seems that women with higher education tend to have less children, since the proportion is lower than in the women’s sample and is around 17 percent. For children whose mother did not live in the household the educational level of the caretaker was taken into consideration. The majority of children live in households where the ethnicity of the head of the household is Serbian (82 percent). The number of children living in Roma and Muslim – ie. Bosnian fami- lies is higher than average. As a result, the percentage of these children is higher than the percentage of household or women’s distribution of these ethnic groups. 28 MICS3 FULL TECHNICAL REPORT One of the overarching goals of the Millennium Development Goals (MDGs) and the World Fit for Children (WFFC) is to reduce infant and under-five mortality. Specifically, the MDGs call for the reduction in under-five mortality by two-thirds between 1990 and 2015. Monitoring progress towards this goal is an important but difficult objective. Measuring childhood mortality may seem easy, but attempts using direct questions, such as “Has anyone in this household died in the last year?” give inaccurate results. Alternatively, indirect methods developed to measure child mortality produce robust estimates that are comparable with the ones obtained from other sources. Indirect methods minimise the pitfalls of memory lapses, inexact or misinterpreted definitions, and poor interviewing technique. The infant mortality rate is the probability of dying before the first birthday. The under-five mortality rate is the probability of dying be- fore the fifth birthday. In MICS surveys, infant and under-five mortality rates are calculated, based on an indirect estimation technique known as the Brass method (United Nations, 1983; 1990a; 1990b). The data used in the es- timation is: the mean number of children ever born to five year age groups of women from aged 15 to 49, and the proportion of these children who are dead, similarly for five-year age groups of women. The technique converts this data into probabilities of dying by tak- ing into account both the mortality risks to which children are exposed and their length of exposure to the risk of dying, assuming a par- ticular model age pattern of mortality. Based on previous information on mortality in Roma settlements in Serbia, the East model life table was selected as the most appropriate. All esti- mates have been calculated by averaging mor- tality estimates obtained from women aged 25–29 and 30–34, and refer to mid-2003. The module on child mortality was used just for Roma living in Roma settlements for several reasons: there was no data on Roma child mor- tality in the regular statistics, and the estima- tion was that it was higher than the national average. Secondly, it was estimated that regular statistics provide accurate data for the general population, and also that mortality is low, so with this research technique we could not compile reliable data. Table CM.1 provides estimates of child mortal- ity by various background characteristics, while Table CM.2 provides the basic data used in the calculation of mortality rates for Roma in Roma settlements. The infant mortality rate is estimated at 25 per thousand, while the probability of dying under 5 years of age is around 28 per thousand live births. There is a significant difference be- tween the likelihood of dying among males and CHILDCHILD MORTALITYMORTALITY 29MONITORING THE SITUATION OF CHILDREN AND WOMEN females: the infant mortality rate among boys is 32, and among girls 20, and the under-five mortality rate among boys is 36, compared to 23 per thousand among girls. Infant and under- 5 mortality rates are lowest in Vojvodina, while the figures for Central Serbia7 are almost twice as high as that of Vojvodina. There are also significant differences in mortality in terms of educational levels and wealth. However, all those differences should be regarded with cau- tion, owing to the small number of children in the sample. 7 Since all results are estimated using the Brass method, in order to obtain qualitative findings, grouping of some background variables was necessary. As the number of Roma women who gave birth was not big enough to provide estimates on the level of six regions, the results are shown on three regional levels: Vojvodina, Belgrade and Central Serbia (which includes Central, West, East and South East Serbia). 30 MICS3 FULL TECHNICAL REPORT Nutritional Status Children’s nutritional status is a reflection of their overall health. When children have access to an adequate food supply, are not exposed to repeated illness, and are well cared for, they reach their growth potential and are considered well-nourished. Malnutrition is associated with more than half of all children’s deaths worldwide. Undernourished chil- dren are more likely to die from common childhood ailments, and those who survive have recurring sicknesses and faltering growth. Three-quarters of children who die from causes related to malnutri- tion were only mildly or moderately malnourished, showing no outward sign of their vulnerability. The Millennium Development target is to reduce by half the proportion of people who suffer from hunger between 1990 and 2015. The World Fit for Children goal is to reduce the prevalence of malnutrition among children under five years of age by at least one-third (between 2000 and 2010), with special attention to children under 2 years of age. A re- duction in the prevalence of malnutrition will as- sist in the goal to reduce child mortality. In a well-nourished population, there is a refer- ence distribution of height and weight for children under the age of five. Under-nourishment in a population can be gauged by comparing children to a reference population. The reference popula- tion used in this report is the WHO/CDC/NCHS reference, which was recommended for use by UNICEF and the World Health Organization at the time the survey was implemented. Each of the three nutritional status indicators can be expressed in standard deviation units (z-scores) from the median of the reference population. Weight-for-age is a measure of both acute and chronic malnutrition. Children whose weight-for- age is more than two standard deviations below the median of the reference population are con- sidered moderately or severely underweight while those whose weight-for-age is more than three standard deviations below the median are classi- fied as severely underweight. Height-for-age is a measure of linear growth. Children whose height-for-age is more than two standard deviations below the median of the ref- erence population are considered short for their age and are classified as moderately or severely stunted. Those whose height-for-age is more than three standard deviations below the median are classified as severely stunted. Stunting is a reflec- tion of chronic malnutrition as a result of failure to receive adequate nutrition over a long period and recurrent or chronic illness. Finally, children whose weight-for-height is more than two standard deviations below the median of the reference population are classified as mod- erately or severely wasted, while those who fall more than three standard deviations below the median are severely wasted. Wasting is usually the result of a recent nutritional deficiency. The indicator may exhibit significant seasonal shifts associated with changes in the availability of food or disease prevalence. On the other hand, NUTRITIONNUTRITION 31MONITORING THE SITUATION OF CHILDREN AND WOMEN children whose weight-for-height is two or more standard deviations above the median of the ref- erence population are considered as moderately or severely obese. Obesity is mostly a result of bad nutritional practices (low intake of proteins, fruit and vegetables, high intake of saturated fats and sugar…) and is a risk factor for some of the chronic diseases in future life, like cardiovascu- lar diseases and diabetes. In assessing child nutritional status the following determinants were used: height (in centimetres), weight (in kilos), age (in months) and sex. Table NU.1 shows the percentages of children classified into each of the described malnutrition categories, and also the percentage of children who are overweight. Children who were not weighed and measured (approximately 12 percent of children) and those whose measurements are outside a plausible range are excluded. Almost 2 percent of children under the age of five in Serbia are moderately underweight and almost no child is classified as severely underweight (Table NU.1). 6 percent of children are stunted or too short for their age and 3 percent are wast- ed or too thin for their height. By each measure – underweight, wasting and stunting – figures for children from the poor- est and Roma households significantly exceed the average. Stunting prevalence among Roma from Roma settlements is 20 percent, which is three times higher than average. Underweight and stunting is two times higher among chil- dren from poor households and children whose mothers are less educated. Regional distribution shows that children in Belgrade are more likely to be undernourished than other children, which is quite an unexpected result. In contrast, the percentage of all indicators is lowest in the West region. The age pattern shows that the child becomes vulnerable to wasting during the first six months of life when complementary food is introduced and breastfeeding is discontinued. That is also the age when stunting begins, which is very difficult to recover from (Figure NU.1). 15 percent of children are overweight. There is a significant difference between urban and rural prevalence: 17 percent of children from rural ar- eas compared to 14 percent of urban children are overweight. Looking at age distribution we can conclude that obesity starts right after birth, and constantly continues to grow, reaching its maxi- mum at a child’s second year, when every fourth child becomes overweight. The distribution of indicators, shown in Figures NU.1a, NU.1b and NU.1c confirms that the main nutrition problem in Serbia is not malnutrition, which is even below the values of the reference population, but obesity. Analysing the trends in malnutrition, we found that malnutrition remains at a very low level, with small insignificant variations. The decrease in prevalence of underweight and wasting since the year 2000 can be explained by the differ- ent seasons the survey was conducted in, which entail insensitivity to infections, consumption of different types of food etc (differences shown 0 1 2 3 4 5 6 7 8 9 6 12 18 24 30 36 42 48 54 60 Figure NU.1 Percentage of children under-5 who are undernourished, Serbia, 2005 Age (in months) UNDERWEIGHT STUNTED WASTED [%] 32 MICS3 FULL TECHNICAL REPORT on the Figure NU.1d are not statistically signifi- cant). There is a continuation of the rising trend of stunting prevalence during the past ten years. Overweight prevalence has also shown some growth, but the differences are not statistically significant. Figure NU.1a Distribution of weight-for-age among children under five, Serbia, 2005 Figure NU.1b Distribution of height-for-age among children under five, Serbia, 2005 Figure NU.1c Distribution of weight-for-height among children under five, Serbia, 2005 [%] 0 10 20 30 40 –5.5 –4.5 –3.5 –2.5 –1.5 –0.5 0.5 1.5 2.5 3.5 4.5 5.5 Standard deviations SERBIA REFERENCE POPULATION +2SD –2SD [%] 0 10 20 30 40 –5.5 –4.5 –3.5 –2.5 –1.5 –0.5 0.5 1.5 2.5 3.5 4.5 5.5 Standard deviations SERBIA REFERENCE POPULATION –2SD +2SD [%] 0 10 20 30 40 –3.5 –2.5 –1.5 –0.5 0.5 1.5 2.5 3.5 4.5 5.5 Standard deviations SERBIA REFERENCE POPULATION –2SD +2SD 33MONITORING THE SITUATION OF CHILDREN AND WOMEN Breastfeeding Breastfeeding for the first few years of life protects children from infection, provides an ideal source of nutrients, and is economical and safe. However, many mothers stop breastfeeding too soon and there are often pressures to switch to infant for- mula, which can contribute to growth difficulties and micronutrient malnutrition and is unsafe if clean water is not readily available. WHO/UNICEF have the following feeding recommendations: � Exclusive breastfeeding for the first six months � Continued breastfeeding for two years or more � Safe, appropriate and adequate complemen- tary foods beginning at 6 months � Frequency of complementary feeding: 2 times per day for 6–8 month olds; 3 times per day for 9–11 month olds It is also recommended that breastfeeding be initiated within one hour of birth. Figure NU.1d Prevalence of malnutrition in children under five, Serbia, 1996–2005 2.74 2.10 2.79 4.86 3.72 6.84 1.06 1.10 1.59 2.54 2.59 2.30 4.96 1.90 1.60 2.19 3.70 3.30 3.01 6.00 3.40 4.70 5.90 0.70 0.50 0.60 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 Underweight 1996 Underweight 2000 Underweight 2005 Wasting 1996 Wasting 2000 Wasting 2005 Stunting 1996 Stunting 2000 Stunting 2005 % o f c hi ld re n w ith m od er at e or se ve re m al nu tr iti on 4.01 34 MICS3 FULL TECHNICAL REPORT The indicators of recommended child feeding practices are as follows: � Exclusive breastfeeding rate � Timely complementary feeding rate � Continued breastfeeding rate � Timely initiation of breastfeeding � Frequency of complementary feeding � Adequately fed infants Also, additional indicators are calculated (pre- dominant breastfeeding and total breastfeeding rate) in order to compare the results with the MICS 2000. The goal of this module was to assess the exist- ing feeding practices of young children in Serbia. The MICS remain the only national source of data for this purpose. Table NU.2 shows the proportion of women who started breastfeeding their infants within one hour of birth, and women who started breastfeeding within one day of birth (in- cluding those who started within one hour). Although a very important step in manage- ment of lactation and establishment of a physical and emotional relationship between the baby and the mother, only every sixth baby is breastfed for the first time within one hour of birth, while 67 percent of newborns in Serbia start breastfeeding within one day of birth. The practice of breastfeeding within one hour of birth is much less present in Belgrade (9 percent) than in West and South-East Serbia (23 and 25 percent, respectively). The percent- age falls, the higher the mothers’ education and socioeconomic status of the household. Roma children from Roma settlements are twice as likely to be breastfed within one hour of birth. 67 73 65 56 76 67 64 74 63 71 67 17 33 1717 9 23 18 11 25 16 19 Figure NU.2 Percentage of mothers who started breastfeeding within one hour and within one day of birth, Serbia, 2005 Se rb ia w ith ou t R om a in R om a se tt le m en ts Ro m a in R om a se tt le m en ts Vo jv od in a Be lg ra de W es t Ce nt ra l Ea st So ut hE as t Ur ba n Ru ra l Se rb ia [%] WITHIN ONE HOURWITHIN ONE DAY 35MONITORING THE SITUATION OF CHILDREN AND WOMEN In Table NU.3, breastfeeding status is based on the reports of mothers/caretakers concern- ing their children’s consumption of food and fluids in the 24 hours prior to the interview. Exclusively breastfed refers to infants who re- ceived only breast milk (and vitamins, mineral supplements, or medicine). The table shows ex- clusive breastfeeding of infants during the first six months of life (separately for 0–3 months and 0–5 months), as well as complementary feeding of children 6–9 months and continued breastfeeding of children at 12–15 and 20–23 months of age. Approximately 15 percent of children aged less than six months are exclusively breastfed, a level significantly lower than recommended. Exclusive breastfeeding is more frequent in urban settlements and among girls. However, all the conclusions should be regarded with caution, owing to the small sample size. At age 6–9 months, 39 percent of children are receiving breast milk and solid or semi-solid foods. By the ages of 12–15 and 20–23 months respectively, 22 and 8 percent of children are still being breastfed. Roma children from Roma settlements are more likely to continue to be breastfed than the entire population, since 60 percent of Roma children by age 12–15 months and 34 percent by age 20–23 months are still breastfed. Figure NU.3 shows the detailed pattern of breastfeeding status by the child’s age in months. Even at the earliest ages, the majority of children are receiving liquids or foods other than breast milk. 32 percent of infants aged 0 – 1 months were exclusively breastfed, and this proportion drops off rapidly until it falls below 4 percent by the fifth month. Looking at the trends over the past decade, an increasing trend of key breastfeeding indicators is noted (Figure NU.3a). Exclusive breastfeeding 0–1 2–3 4–5 6–7 8–9 10–11 12–13 14–15 16–17 18–19 20–21 22–23 24–25 26–27 28–29 30–31 32–33 34–35 Figure NU.3 Infant feeding patterns by age: Percentage distribution of children under the age of 3 by feeding pattern by age group, Serbia, 2005 [%] Age (in months) WEANED (NOT BREASTFED) BREASTFED AND COMPLEMENTARY FOODS BREASTFED AND OTHER MILK/ FORMULA BREASTFED AND NON-MILK LIQUIDS BREASTFED AND PLAIN WATER ONLY EXCLUSIVELY BREASTFED 36 MICS3 FULL TECHNICAL REPORT during the first four months rose from 3 percent in 1996, up to 23 percent in 2005, which is almost 8 times higher. The continued and predominant breastfeeding rates have also improved, while the percentage of the total number of breastfed chil- dren is slightly decreasing (although the change is not statistically significant). The adequacy of infant feeding in children under 12 months is provided in Table NU.4. Different criteria for adequate feeding are used depend- ing on the age of the child. For infants aged 0–5 months, exclusive breastfeeding is considered to be adequate feeding. Infants aged 6–8 months are considered to be adequately fed if they are receiv- ing breastmilk and complementary food at least twice a day, while infants aged 9–11 months are considered to be adequately fed if they are receiv- ing breastmilk and eating complementary food at least three times a day. When we add all the above recommended practices together, we will come to the sum indicator – the percentage of infants aged 0–11 months who are adequately fed. Every third child aged 6–8 months has re- ceived breastmilk and complementary food according to the recommended schedule. This practice is more used for girls than boys, more in rural areas, and much more in Vojvodina than in the rest of the country. The recom- mendation is more practised by more edu- cated mothers, and by the middle class. While every tenth Muslim/Bosnian child is receiv- ing breastmilk and complementary food at this age, much more Roma (42 percent) and Hungarian (66 percent) are benefiting from this practice. The percentage is growing, but not significantly between the ages of 9 and 11 months, showing similar patterns, as already explained. In total, when we look at the feeding pattern for all infants (aged 0–11), only each fourth infant is adequately fed, following the recom- mendations. Girls are better fed than boys; infants in Vojvodina and Central Serbia are better fed than infants from the West and East. There are no urban/rural differences. Interestingly, Roma children from Roma settle- ments have a better chance of being adequately fed, mainly owing to the higher percentage of breastfed children, but also owing to the fact that they follow more closely the other feeding recommendations. The worst feeding practices are noted among the poorest and the richest children, but the reasons are completely dif- ferent. While only 17 percent of the poorest children aged 0–11 months are adequately fed, mainly owing to the fact that they are not even receiving complementary food the minimum recommended times a day, children from the richest households are not appropriately fed mainly because very few of them are exclusive- ly breastfed. A mother’s education and appro- priate child nutrition are strongly correlated. The more educated the mother is, the better the child’s chance of being adequately fed. The percentage of children under 12 months who are adequately fed rises from 20 percent among those children whose mother has primary or no education to 24 percent among children whose mothers have secondary education, and up to 33 percent among mothers with higher education. Figure NU.3a Progress in major breastfeeding indicators, Serbia, 1996–2005 2000 20051996 Exclusive breastfeeding Continued breastfeeding rate 12–15 months Predominant breastfeeding rate Ever breastfed [%] 37MONITORING THE SITUATION OF CHILDREN AND WOMEN Low Birth Weight Weight at birth is a good indicator not only of a mother’s health and nutritional status but also the newborn’s chances for survival, growth, long-term health and psychosocial development. Low birth weight (less than 2500 grams) carries a range of grave health risks for children. Babies who were undernourished in the womb face a greatly in- creased risk of dying during their early months and years. Those who survive have an impaired immune function and increased risk of disease; they are likely to remain undernourished, with reduced muscle strength throughout their lives, and suffer a higher incidence of diabetes and heart disease in later life. Children born under- weight also tend to have a lower IQ and cogni- tive disabilities, affecting their performance in school and their job opportunities as adults. In the developing world, low birth weight stems primarily from the mother’s poor health and nutrition. Three factors have most impact: the mother’s poor nutritional status before concep- tion, short stature (due mostly to under nutri- tion and infections during her childhood), and poor nutrition during the pregnancy. Inadequate weight gain during pregnancy is particularly important since it accounts for a large propor- tion of foetal growth retardation. Moreover, diseases such as diarrhoea and malaria, which are common in many developing countries, can significantly impair foetal growth if the mother becomes infected while pregnant. In the industrialised world, cigarette smoking during pregnancy is the leading cause of low birth weight. In developed and developing countries alike, teenagers who give birth when their own bodies have yet to finish growing run the risk of bearing underweight babies. The percentage of births weighing below 2500 grams is estimated from two items in the ques- tionnaire: the mother’s assessment of the child’s size at birth (i.e., very small, smaller than average, average, larger than average, very large) and the mother’s recall of the child’s weight or the weight as recorded on a health card if the child was weighed at birth8. Overall, 98 percent of births were weighed at birth and approximately 5 percent of infants are estimated to weigh less than 2500 grams at birth (Table NU.5). While there are no significant dif- ferences in the percentage of weighed infants at birth (with the exception of Roma babies, where 10 percent are not weighed at birth) there are certainly differences in the low birth weight measured. Roma children (9 percent) and the poorest children (9 percent) are twice as likely to weigh less than 2500 grams at birth than average children. The indicator improves, the higher the mother’s education. The percentage of low birth weight does not vary much by region or urban and rural areas. 8 For a detailed description of the methodology, see Boerma, Weinstein, Rutstein and Sommerfelt, 1996. 38 MICS3 FULL TECHNICAL REPORT Immunization The Millennium Development Goal (MDG) 4 is to reduce child mortality by two thirds between 1990 and 2015. Immunization plays a key part in this goal. Immunizations have saved the lives of millions of children in the three decades since the launch of the Expanded Programme on Immunization (EPI) in 1974. Worldwide there are still 27 million children overlooked by routine immunization and as a result, vaccine-preventable diseases cause more than 2 million deaths every year. A World Fit for Children goal is to ensure full im- munization of children under one year of age at 90 percent nationally, with at least 80 percent coverage in every district or equivalent administrative unit. According to UNICEF and WHO guidelines, a child should receive a BCG vaccination to pro- tect against tuberculosis, three doses of DPT to protect against diphtheria, pertussis, and tetanus, three doses of the polio vaccine, and a measles vaccination by the age of 12 months. The national immunization calendar in Serbia differs slightly, with the measles vaccine being ad- ministered between the ages of 12 and 18 months in the MMR form (Measles, Mumps, and Rubella). Mothers were asked to provide vaccination cards for children under the age of five. Interviewers copied vaccination information from the cards onto the MICS3 questionnaire. Overall, 71 percent of children aged 18 to 29 months have vaccination cards. If the child did not have a card or the card was not shown, the mother was asked to recall whether or not the child had received each of the vaccinations and, for DPT and Polio, how many times. The percent- age of children aged 18 to 29 months who had re- ceived each of the vaccinations is shown in Table CH.1. The denominator for the table is comprised of children aged 18–29 months, so that only chil- dren who are old enough to be fully vaccinated are counted. In the top panel, the numerator includes all children who had been vaccinated at any time before the survey, according to the vaccination card or the mother’s report. In the bottom panel, for BCG, DPT and polio vaccines, only those who had been vaccinated before their first birthday are included. For MMR vaccine, in the bottom panel, the numerator includes only those children who were vaccinated before 18 months of age. For children without vaccination cards, the proportion of vaccinations given before the first birthday (18 months in the case of MMR) is assumed to be the same as for children with vaccination cards. According to survey results, only 74 percent of children aged 18–29 months received a BCG vaccine by the age of 12 months. This result is quite unexpected and probably not accurate. It is much more an indicator of weaknesses in the BCG registration than of accurate BCG vac- cination coverage. Namely, only 44 percent of children aged 18 to 29 months who had the vac- cination card have the BCG vaccine registered, while according to the mother’s report all chil- dren received a BCG. Most likely the problem CHILD HEALTHCHILD HEALTH 39MONITORING THE SITUATION OF CHILDREN AND WOMEN arises from lack of communication between different health facilities where the vaccine is given and where the vaccination card is pro- vided for a child. Usually, a child receives the BCG vaccine in a maternity ward, and gets the vaccination card later, in the health post. The first dose of DPT was given to 97 percent of children. The percentage declines for sub- sequent doses of DPT to 96 percent for the second dose, and 90 percent for the third dose (Figure CH.1). Similarly, 95 percent of children received polio 1 by age 12 months and this declines to 88 percent by the third dose. Polio and DPT are normally administered simulta- neously. The slight difference in coverage rates can be explained by the introduction of the new practice whereby children are receiving IPV instead of the oral polio vaccine. The cov- erage for the MMR vaccine for children was somewhat lower: 84 percent of children aged 18–29 months received the measles vaccine before 18 months of age. The percentage of children who had all eight recommended vaccinations by the specified age is 44 percent, far below the international and na- tional target. When we change the denominator and include all children, we see that 57 percent of children aged 18 to 29 months have received all the recommended vaccinations. That indicates delayed vaccination, where some children, prob- ably due to false temporary contra-indications, are receiving vaccines after the defined time. Since there is a suspicion in validation of BCG vaccination coverage results, additional calcu- lation of fully immunized children was done. The figures for fully immunized children are shown in the table CH.1a, and they indicate children receiving vaccinations according to all the aforementioned criteria and, in the case of all vaccines except the BCG, are based on infor- mation from both cards and mothers’ reports. According to the aforementioned analysis of BCG vaccine registration, in this calculation, information on the BCG is only based on the mothers’ report, while the probability that the child received it at the recommended time was based on the vaccination card. Taking this “mix” of information into account was necessary for estimation of the BCG, and consequently esti- mation of full immunization coverage of chil- dren in Serbia. Based on this calculation, the proportion of fully immunized children accord- ing to the recommended schedule in Serbia is 60 percent, while 80 percent of children aged 18 to 29 months of age received vaccinations at some point preceding the survey. Tables CH.2 show vaccination coverage rates among children 18–29 months by background characteristics. The figures indicate children receiving the vaccinations at any time up to the date of the survey, and are based on informa- tion from both the vaccination cards and moth- ers’/caretakers’ reports. Regional distribution indicates that the lowest percentage of fully immunized children is in South-East Serbia, at 31 percent, while the highest proportion of children who received all recom- mended vaccinations is in Vojvodina and East Serbia, at 81 and 63 percent, respectively. Figure CH.1 Percentage of children aged 18–29 months who received the recommended vaccinations by 12 months (by 18 months in case of MMR), Serbia, 2005 BCG DP T1 DP T2 DP T3 Po lio 1 Po lio 2 Po lio 3 MMR All 74 97 96 90 95 93 88 84 44 [%] 40 MICS3 FULL TECHNICAL REPORT Children living in urban settlements are more likely to be fully immunized; 62 percent of children living in urban, compared to 52 per- cent of children from rural areas have received all recommended vaccinations. Whether the child will receive all the recommended vac- cinations, depends on the mother’s education. The percentage of children who received all vaccinations rises from 54 percent for chil- dren whose mothers are less educated to 65 percent for children whose mothers have at- tended university. As far as household wealth is concerned, the majority of children from the middle and fourth quintile have been fully immunized, while nearly every second child from other socio-economic classes has re- ceived all the recommended vaccinations. A specially jeopardized group of Roma chil- dren are those living in Roma settlements, with only 27 percent of them having received all vaccinations. No significant differences between boys’ and girls’ immunization coverage have been found. The findings presented are much lower than routine statistical data shows. Oral Rehydration Treatment Diarrhoea is the second leading cause of death among children under five worldwide. Most di- arrhoea-related deaths in children are due to de- hydration from loss of large quantities of water and electrolytes from the body in liquid stools. Management of diarrhoea – either through oral rehydration salts (ORS) or a recommended home fluid (RHF) – can prevent many of these deaths. Preventing dehydration and malnutri- tion by increasing fluid intake and continuing to feed the child are also important strategies for managing diarrhoea. The goals are to: 1) reduce by half deaths due to diarrhoea among children under five by 2010 compared to 2000 (A World Fit for Children); and 2) reduce by two thirds the mortality rate among children under five by 2015 compared to 1990 (Millennium Development Goals). In addition, the World Fit for Children calls for a reduction in the incidence of diarrhoea by 25 percent. The indicators are: � Prevalence of diarrhoea � Oral rehydration therapy (ORT) � Home management of diarrhoea � (ORT or increased fluids) AND continued feeding Official figures show that in the past few years acute diarrhoea diminishes as one of the most fre- quent causes of child mortality. In Serbia, the last case of death caused by diarrhoea of a child under 5 was registered in 2004. In the MICS questionnaire, mothers (or caretak- ers) were asked to report whether their child had had diarrhoea in the two weeks prior to the sur- vey. If so, the mother was asked a series of ques- tions about what the child had had to drink and eat during the episode of diarrhoea. Overall, 5 percent of under-five children had had diarrhoea in the two weeks preceding the survey (Table CH.3), which would correspond to 1.4 episodes per child annually (assuming no seasonal variations and that the child only had had one diarrhoeal episode during the previ- ous two weeks). High prevalence of diarrhoea is noticed among Roma and the poorest chil- dren, at 13 and 7 percent respectively. Regional distribution shows that the smallest number of children with diarrhoea was in Belgrade and East Serbia, while in South East Serbia almost one out of twelve children under five had had diarrhoea. A strong correlation between the mother’s education and the child’s health is noticeable; the percentage of children with di- arrhoea falls from 7 percent for children whose mothers have primary or no education, to 5 and 4 percent for those whose mothers have 41MONITORING THE SITUATION OF CHILDREN AND WOMEN secondary, ie. higher education. With regards to a child’s age, the peak of diarrhoea preva- lence occurs in the weaning period, among children aged 6–23 months. Table CH.3 also shows the percentage of chil- dren receiving various types of recommended liquids during the episode of diarrhoea. For this, so-called Oral Rehydration Therapy varie- ties of liquids are recommended for use during the episode of diarrhoea in Serbia: ORS fluid, mother’s milk, porridge, soup, yoghurt, tea, sugar and salt solution and unsweetened juice. Since mothers were able to name more than one type of liquid, the percentages do not nec- essarily add up to 100. Using this wide defini- tion of ORT, we found that a high percentage of under-five children (94 percent) are receiving some of the recommended liquids during diar- rhoea episodes. Mainly children are receiving porridge or soup (63 percent) or yoghurt, tea, sugar and salt solution (68 percent). Use of ORS is relatively low, only 17 percent of children with diarrhoea received ORS fluids, being the highest in Belgrade – 26 percent. There are some variations in ORT use according to the type of settlement and mothers’ education, but owing to the small number of children with diarrhoea in each sample group, those variations are not statistically significant. The only significance is the difference with Roma children, where one out of ten Roma children with diarrhoea is left without treatment. More than one third (36 percent) of under- five children with diarrhoea drank more than usual while 64 percent drank the same or less (Table CH.4). Three-quarters of children ate somewhat less, the same or more (continued feeding), but one-quarter ate much less or ate almost nothing. In all, only 26 percent of chil- dren received home management of diarrhoea. Regional differences are significant, with half of the children in South East Serbia receiving home management of diarrhoea and only 15 percent of children in Vojvodina. Urban/ru- ral differences are also strongly marked, with 29 percent of urban and 22 percent of rural children being managed by the recommended model. Roma children from Roma settlements are poorly treated; only every seventh Roma child received home management of diarrhoea. A child’s treatment is strongly dependent on the mother’s educational level. The more edu- cated the mother is, the more likely the child will be appropriately cared for. Overall 71 percent of children with diarrhoea received ORT or increased fluids and continued feeding, with significant urban/rural differ- ences (76 and 64 percent respectively). Care Seeking and Antibiotic Treatment of Pneumonia Pneumonia is the leading cause of death in children and the use of antibiotics in under-5s with suspect- ed pneumonia is a key intervention. A World Fit for Children goal is to reduce by one-third the deaths due to acute respiratory infections. Children with suspected pneumonia are those who had an illness with a cough accompanied by rapid or difficult breathing and whose symptoms were NOT due to a problem in the chest and a blocked nose. The indicators are: � Prevalence of suspected pneumonia � Care seeking for suspected pneumonia � Antibiotic treatment for suspected pneumonia � Knowledge of the danger signs of pneumonia In the MICS questionnaire, children with acute respiratory infection are defined as those who had an illness with a cough accompanied by rapid or difficult breathing, and whose symptoms were due to a problem in the chest, or both a problem in the chest and a blocked nose, or whose mother did not know the source of the problem. Table CH.5 presents the prevalence of suspect- ed pneumonia and, if care was sought outside 42 MICS3 FULL TECHNICAL REPORT the home, the location of that care. 3 percent of children aged 0–59 months were reported to have had symptoms of pneumonia during the two weeks preceding the survey. The propor- tion of children with suspected pneumonia is significantly higher among Roma children, at 10 percent. Of all children with suspected pneumonia, 93 percent were taken to an appropriate provider. Boys and younger children were taken to an appropriate health provider more often. Table CH.6 presents the use of antibiotics for the treatment of suspected pneumonia in under-5s by sex, age, region, residence and socio-economic factors. In Serbia, 57 percent of under-5 children with suspected pneumo- nia had received an antibiotic during the two weeks prior to the survey. Antibiotic treat- ment of suspected pneumonia is very low among Roma households, at only 45 percent. Mothers’ knowledge of the danger signs of pneumonia are presented in Table CH.6a. Overall, every third woman knows the two danger signs of pneumonia – fast and diffi- cult breathing. 39 percent of mothers identi- fied fast breathing and 48 percent of moth- ers identified difficult breathing as symp- toms for taking children immediately to a health care provider. The most commonly identified symptom for taking a child to a health facility is when the child develops a fever (82 percent). The percentage of mothers who know the two danger signs of pneumonia is extremely low among Roma and Muslim – ie. Bosnian mothers – at only 12 and 7 percent, respec- tively. Regional distribution shows that wom- en living in Belgrade are most familiar with these two signs (57 percent), while in West Serbia this figure is only 15 percent. A moth- er’s educational level is an important factor in recognising symptoms, since a higher number of women with secondary or higher education named both signs. Solid Fuel Use More than 3 billion people around the world rely on solid fuels (biomass and coal) for their basic energy needs, including cooking and heat- ing. Cooking and heating with solid fuels leads to high levels of indoor smoke, a complex mix of health-damaging pollutants. The main prob- lem with the use of solid fuels is products of incomplete combustion, including CO, polyaro- matic hydrocarbons, SO2, and other toxic ele- ments. Use of solid fuels increases the risks of acute respiratory illness, pneumonia, chronic obstructive lung disease, cancer, and possibly tuberculosis, low birth weight, cataracts, and asthma. The primary indicator is the propor- tion of the population using solid fuels as the primary source of domestic energy for cooking. Overall, one third of all households in Serbia use solid fuels for cooking. Use of solid fuels is characteristic of Roma settlements, where 85 percent use solid fuels, almost three times above the national average. Use of solid fuels is lower in urban areas (14 percent), than in ru- ral households (61 percent), as shown in Table CH.7. Differentials with respect to household wealth and the educational level of the head of the household are also significant. The propor- tion of households that use solid fuels for cook- ing rapidly decreases from 86 percent among poorest households to 21 percent among the middle class, while hardly any rich household uses solid fuel for cooking. Cooking with electricity is highest in Belgrade, where 84 percent of interviewed households use electricity, and lowest in West Serbia at 27 percent. In Vojvodina the usage of electricity and gas is approximately the same, with 43 per- cent of households in Vojvodina cooking with electricity and 38 percent using gas. With re- gards to household wealth, as expected, cook- ing with electricity is inversely proportional to solid fuel usage: 89 percent of the richest, as opposed to 8 percent of the poorest households cook with electricity. 43MONITORING THE SITUATION OF CHILDREN AND WOMEN Solid fuel use alone is a poor proxy for indoor air pollution, since the concentration of the pollutants is different when the same fuel is burnt in differ- ent stoves or fires. Use of closed stoves with chim- neys minimises indoor pollution, while an open stove or fire with no chimney or hood means that there is no protection from the harmful effects of solid fuels. The type of stove used with a solid fuel is depicted in Table CH.8. Approximately 96 percent of households that use solid fuels for cooking have a closed stove with a chimney. The proportion of other types of stoves are 1 or below 1 percent, with the exception of Belgrade, where 6 percent of households have an open stove or fire with a chimney. Still, barely a single household cooks on a type of stove, without protection from the effects of solid fuels. 44 MICS3 FULL TECHNICAL REPORT Water and Sanitation Safe drinking water is a basic necessity for good health. Unsafe drinking water can be a signifi-cant carrier of diseases such as trachoma, chol- era, typhoid, and schistosomiasis. Drinking water can also be tainted with chemical, physical and radiological contaminants with harmful effects on human health. In addition to its association with disease, access to drinking water may be particu- larly important for women and children, especially in rural areas, who bear the primary responsibility for carrying water, often over long distances. The MDG goal is to reduce by half, between 1990 and 2015, the proportion of people with- out sustainable access to safe drinking water and basic sanitation. The World Fit for Children goal calls for a reduction in the proportion of house- holds without access to hygienic sanitation facili- ties and affordable and safe drinking water by at least one-third. The list of indicators used in MICS are as follows: Water � Use of improved drinking water sources � Use of adequate water treatment methods � Time to source of drinking water � Person collecting drinking water Sanitation � Use of improved sanitation facilities � Sanitary disposal of child faeces The distribution of the population according to the source of drinking water is shown in Table EN.1 and Figure EN.1. The population using improved sources of drinking water are those using any of the following types of supply: piped water (into their dwelling, yard or plot, from a public or local piped system), a public tap/standpipe, a tubewell/bore- hole, a protected well and spring. Bottled water is considered an improved water source only if the household is using an improved water source for other purposes, such as hand washing and cooking. Overall, 99 percent of the population uses an improved source of drinking water – 99 percent in urban areas and 98 percent in rural areas. The source of drinking water for the population slightly varies according to region (Table EN.1). In West Serbia, 66 percent of the population uses drinking water from a public or local water sup- ply, and there is the highest percentage of tube- well/borehole and protected well or spring water use at 32 percent. In Vojvodina, 14 percent of the household population that has access to improved water sources uses bottled water, while in the West region use of bottled water is nearly zero. The more interesting information for the country is the proportion of the population with a water source that is piped into the dwelling or yard from a water supply system. 77 percent of the popula- tion uses water from a public or local water sup- ply as a main source of drinking water. There is a significant difference in terms of water supply between urban and rural settlements. 91 percent of people in towns in Serbia use water from public ENVIRONMENTENVIRONMENT 45MONITORING THE SITUATION OF CHILDREN AND WOMEN or local water-supply installations in their dwell- ings or yards as a main source of drinking water. In rural settlements only 60 percent of people get their drinking water in the dwelling/yard from a public or local water-supply system. There are regional differences in Serbia: whereas 87 percent of the population in the Belgrade uses water piped into the dwelling or yard, the proportion in West Serbia is lowest, at 66 percent. If we look at this restrictive definition, we can notice that the poorest households are a more jeopardized group. Only 49 percent of the poorest compared to 89 percent of the richest population has access to water from public or local water supply systems. Use of in-house water treatment is presented in Table EN.2. Households were asked about ways they may be treating water at home to make it safer to drink – boiling, adding bleach or chlorine, using a water filter, and using solar disinfection were considered to be proper drinking water treatments. The table shows the percentage of household members using appropriate water treatment methods, separately for all households, and for households using improved and unim- proved drinking water sources. In Serbia, only 4 percent of the population uses appropriate water treatment. Use of appropriate water treatment greatly depends on the region the household lives in. While in Belgrade, 9 percent of the population treats water to make it safer, only 1 percent of the East Serbian popu- lation does the same. More importantly, only 8 percent of the population that has unimproved water sources apply appropriate water treat- ment methods. The socioeconomic status and educational level of the head of the household have a great influence on practising water treat- ment. Wealthier households and those with highly educated heads of the household appro- priately treat their drinking water more often than poorer households and those with less educated heads of the family. The Roma popu- lation living in Roma settlements represents the most unaware group of users of correct water treatment methods – only 1 percent of them use any water treatment. The amount of time it takes to obtain water is presented in Table EN.3, and the person who usually collected the water in Table EN.4. Note that these results refer to one round trip from home to the drinking water source. Information on the number of trips made in one day was not collected. In most of the households (96 percent) the drinking water source is on the premises. For 3 percent of all households, it takes less than 30 minutes to get to the water source and bring water, while about 1 percent of households spend 30 minutes or more for this purpose. When we exclude the households with water on the prem- ises, the average time to the source of drinking water is 22 minutes. Unexpectedly, more time for collecting water is spent in urban settlements (33 minutes) than in rural ones (19 minutes). It is also interesting that the average time to collect water in Belgrade as a typical urban city is quite high, 31 minutes. Although the percentage of the Roma population that has water on the premises is under the national average – 86 percent – they spent less time collecting water (19 minutes), since they usually have a source of water in the Roma settlement they live in. In the majority of households (53 percent), an adult male is usually the person collecting the water, when the source of drinking water is not on the premises. Adult women collect water in 41 percent of cases, while in about 1 percent of households, a female or male child under the age of 15 collects water. In Roma settlements, Figure EN.1 Percentage distribution of household members by source of drinking water Serbia, 2005 BOTTLED WATER TUBEWELL/BOREHOLE PUBLIC TAP/STANDPIPE LOCAL WATER SUPPLY PROTECTED WELL OR SPRING UNPROTECTED WELL OR SPRING OTHER UNIMPROVED PUBLIC WATER SUPPLY 62.8 13.8 1.4 7.1 5.5 8.5 0.4 0.5 46 MICS3 FULL TECHNICAL REPORT the situation is a little different. In nearly two thirds of Roma households, an adult woman is the person in charge of collecting water. Inadequate disposal of human excreta and per- sonal hygiene is associated with a range of dis- eases. Improved sanitation facilities for excreta disposal include: flush or pour flush to a piped sewer system, a septic tank, no flush with a wa- ter-proof septic tank and traditional pit latrine. Nearly 100 percent of the population of Serbia is liv- ing in households using improved sanitation facili- ties (Table EN.5), with no differences between the population living in urban and rural settlements. A more appropriate analysis for the country is when breaking down the sanitary means according to type: toilets linked to sewage systems; toilets linked to septic tanks, and latrines. The safest way of disposing of human excreta and liquid waste is by means of a sewage system; 53 percent of the population live in a household with such means. 37 percent use a toilet that flushes into a septic tank. Altogether, 89 percent of the population has flush toilets either linked to a sewage system or a septic tank. Geographical distribution shows all the varia- tions within the country: in Vojvodina households mostly used a flush to a septic tank – 55 percent – while in Belgrade (76 percent), Central (56 per- cent) and South-East Serbia (56 percent) a flush to piped sewer system is predominant. In West and East Serbia both systems are equally used. A high- er percentage of traditional pit latrines are used in the poorest regions (in the South East, 19 percent; and in the West, 15 percent). There are important urban/rural differences when it comes to sanitation facilities. The urban population mostly (84 percent) uses a flush toilet linked to a sewage system, while in rural areas the population mainly (64 percent) uses a flush to a septic tank, and then traditional pit latrine (19 percent). Only 15 percent of rural households have a toilet in a piped sewer system. Use of a flush to the sewer system, as the most appropriate type of toilet facility is much higher among the richest households – 97 percent, com- pared to 7 percent of the poorest who are con- nected to the piped sewer system. Half of the Roma living in Roma settlements use the traditional pit latrine. A third of their house- holds are connected to the piped sewer system – mainly those in peri-urban areas. 5 percent live without toilet facilities. Safe disposal of child faeces is the child’s last stool which was disposed of, using a toilet or rinsed into a toilet or latrine. Disposal of the faeces of children between 0 and 2 years of age is presented in Table EN.6. In one out of three households with children aged 0 to 2 years the child faeces are safely disposed of. Namely, in 11 percent of households, children use a toilet, and in 26 percent of the cases their faeces were put/inserted into the toilet of a latrine. Findings show that the most frequently used meth- od of disposing of child faeces in Serbia is throwing it into the rubbish (58 percent of the households). Regional distribution shows that the lowest number of safe disposals of child faeces is in East Serbia, at 21 percent. This practice is more used in Vojvodina, West and Central Serbia, at about 40–42 percent. Only 20 percent of Roma households with chil- dren aged 0 to 2 years are disposing of their children’s faeces safely. 62 percent of them throw child faeces into the rubbish. An overview of the percentage of households with improved sources of drinking water and sanitary means of excreta disposal is presented in Table EN.7. Overall, 99 percent of the population in Serbia has access to improved water sources and sanitation. Exceptions are Roma and the poorest households, where slightly lower percentages have access to both improved water sources and sanita- tion facilities. 47MONITORING THE SITUATION OF CHILDREN AND WOMEN Security of Tenure and Durability of Housing Target 11 of MDG is the achievement of significant improvements in the lives of at least 100 million slum dwellers, and the related indicator is the pro- portion of urban household members living in slum housing. In MICS, three indicators were introduced to measure issues related to slum housing: security of tenure, durability of housing, and the proportion living in slum households. An urban household is considered a slum in MICS if it fulfils one of the following conditions: improved drinking water sources are not used; improved sanitation facilities are not used; the living area is not sufficient; hous- ing is not durable, or security of tenure is lacking. Lack of security of tenure is defined as the lack of formal documentation for the residence or the perceived risk of eviction. Table EN.8 is on the security of tenure. In urban areas covered in Serbia MICS, 13 percent of households do not have formal documentation for their residence, and 7 percent of respondents to the household questionnaire indicated that there is a risk of evic- tion. Combining these figures, it is observed that 17 percent of households do not have security of tenure. As additional information, the table also shows that 13 percent of household members have indeed been evicted from a dwelling they were residing in during the last 15 years. Lack of security of tenure is highest among the Roma population living in Roma settlements. As many as 38 percent of Roma households live in dwellings without formal documentation, and every fifth Roma household feels that there is a risk of eviction. Security of tenure is highly linked to household wealth. The wealthier the household is, the more secure the tenure. The proportion of households that do not have security of tenure decreases from 31 to 11 percent from the poorest to the richest house- holds. These results are not only a consequence of lack of formal documentation; they are highly cor- related with the perceived risk of eviction. Finding that every eighth household was forced to move in the last 15 years is not so surprising con- sidering the political situation and the civil wars that happened during this period. The fact that 22 percent of these households are from the poorest class, indicates that they were not in a position to improve their socio-economic status. Structures that households are inhabiting con- sidered non-durable in MICS are those where the floor material is natural and two or more bad conditions were identified, like cracks or openings in the wall; no windows or windows with broken glass; visible holes in the walls; an incomplete roof or an insecure door; or where conditions of vulnerability to accidents in terms of the dwelling’s surroundings exist, or if the structure is located in or near a hazardous area (eg. a landslide area, a flood-phone area, a river bank, a steep hill, a rubbish tip, an industrial pollution area, a railway line, power plant or flyover). Table EN.9 provides information on the findings of the survey. The proportion of households and household members that live in dwellings which are considered non-durable is very low, under 1 percent, with the exception of Roma households. As many as 12 percent of Roma live in non-durable dwellings. 14 percent of dwellings inhabited by Roma have a natural floor, and 4 percent of them are located in haz- ardous areas. Table EN.10 brings together all 5 components of slum housing (see above). Overall, 19 percent of households are considered to be slum housing. 21 percent of the population is living in those households. The Roma population represents the most jeopardized group regarding security of tenure and poor dwelling conditions. Around 60 percent of Roma households and as many as 63 percent of Roma household members live in slum housing. 48 MICS3 FULL TECHNICAL REPORT Contraception Appropriate family planning is important to the health of women and children by: 1) preventing pregnancies that are too early or too late; 2) ex- tending the period between births; and 3) limiting the number of children. A World Fit for Children goal is access for all couples to information and ser- vices to prevent pregnancies that are too early, too closely spaced, too late or too numerous. Current use of contraception was reported by less than half (41 percent) of women currently married or in union (Table RH.1). Women in Serbia are more frequently using traditional contraceptive methods – 23 percent. The most popular contraceptive methods are withdrawal and periodic abstinence, at 14 and 8 percent, re- spectively. The next most popular contraceptive method is the condom at 8 percent. 6 percent use intrauterine devices, 3 percent use the pill and less than 1 percent of married women reported the use of some other method for preventing pregnancy. Contraceptive prevalence ranges from 27 percent in Central Serbia to 54 percent in East Serbia. Each area has its own characteristics in terms of practices used. For example, in East and South- east Serbia, which are considered underdeveloped parts of the country, women predominantly use traditional contraceptive methods, while in Belgrade and Vojvodina, the most developed areas, modern methods of contraception are prevalent. Contraceptive prevalence of both traditional and modern methods seemed to increase with age, up to the age of 40, and then there is a pattern of decreased contraception use. Women’s education level, ethnicity and wealth index are significantly linked to contraceptive prevalence (Figure RH.1). The percentage of women using any method of contraception rose from 33 percent among those with primary or no education to 41 percent among women with secondary education, and up to 52 percent among women with higher education. Also, as the educational level increases, the propor- tion using modern methods – particularly the con- dom – increases, and the proportion using with- drawal as a method decreases. A similar pattern is noticed with regards to wealth: women from the richest households are using modern contraceptive methods more than average. The use of any contra- ceptive method decreases as poverty increases. Contraceptive usage is highest among Hungarian women, where 57 percent of women are using mainly modern methods of contraception. Only one in four married Roma women are using any contraceptive method, usually withdrawal (every fifth). Only 4 percent of Roma women use modern methods. Traditional methods of contraception are also present among Muslims – i.e. Bosnians – but their main choice is periodic abstinence. In both of these ethnic groups, condom usage is ex- tremely low. REPRODUCTIVEREPRODUCTIVE HEALTHHEALTH 49MONITORING THE SITUATION OF CHILDREN AND WOMEN Unmet Need The unmet need9 for contraception refers to fe- cund women who are not using any method of contraception, but who wish to postpone the next birth or who wish to stop childbearing altogether. Unmet need is identified in MICS by using a set of questions eliciting current behaviour and prefer- ences pertaining to contraceptive use, fecundity, and fertility preferences. Women with an unmet need for spacing include women who are currently married (or in union), fe- cund (are currently pregnant or think that they are physically able to become pregnant), are currently not using contraception, and want to space their births. Pregnant women are considered to want to space their births when they did not want the child at the time they got pregnant. Women who are not pregnant are classified in this category if they want to have a(nother) child, but want to have the child at least two years later, or after marriage. Women with an unmet need for limiting are those women who are currently married (or in union), fecund (are currently pregnant or think that they are physically able to become pregnant), are cur- rently not using contraception, and want to limit their births. The latter group includes women who are currently pregnant but had not wanted the pregnancy at all, and women who are not currently pregnant but do not want to have a(nother) child. The total unmet need for contraception is simply the sum of the unmet need for spacing and the unmet need for limiting. Using information on contraception and unmet need, the percentage of demand for contracep- tion satisfied is also estimated from the MICS data. The percentage of demand for contraception satisfied is defined as the proportion of women currently married or in union who are currently using contraception, out of the total demand for contraception. The total demand for contra- ception includes women who currently have an unmet need (for spacing or limiting), plus those who are currently using contraception. Table RH.2 shows the results of the survey on contraception, unmet need, and the satisfied demand for contraception. 28 percent of married women or women in union in Serbia have an unmet need for contraception. Since there is a close link to contraception use, the findings according to background character- istics are very similar to those of contraceptive prevalence. We notice that as women’s education and wealth decrease, so too does their satisfac- tion of their demand for contraception. Above all, needs for contraception are not satisfied among the “most at risk” population: Roma women liv- ing in Roma settlements, uneducated women, and women from the poorest households. Regional distribution shows that an unmet need for contraception is lowest in East Serbia. Figure RH.1 Use of contraception, modern and traditional contraceptive methods, Serbia, 2005 SERBIAN HUNGARIAN MUSLIM \ BOSNIAN ROMA OTHER PRIMARY OR NONE SECONDARY UNIVERSITY POOREST SECOND MIDDLE FOURTH RICHEST SERBIA 40.6 56.6 38.2 27.1 46.9 33.4 41.4 51.7 33.1 38.3 42.2 42.6 48.6 41.2 9 Unmet need measurement in MICS is somewhat different to that used in other household surveys, such as the Demographic and Health Surveys (DHS). In DHS, more detailed information is collected on additional variables, such as postpartum amenhorrea, and sexual activity. Results from the two types of surveys are strictly not comparable. 50 MICS3 FULL TECHNICAL REPORT Conversely, women from West Serbia, to the larg- est degree, have unmet needs for contraception. Unmet need for contraception mainly manifests as unmet need for limiting, with the exception of younger women, 15 to 24 years old, whose needs are mainly manifested as a need for spacing. Antenatal Care The antenatal period presents important op- portunities for reaching pregnant women with a number of interventions that may be vital to their health and well-being, and that of their infants. Better understanding of foetal growth and development and its relationship to the mother’s health has resulted in increased at- tention to the potential of antenatal care as an intervention to improve both maternal and newborn health. For example, if the antenatal period is used to inform women and families about the danger signs and symptoms and the risks of labour and delivery, it may provide the route for ensuring that pregnant women do, in practice, deliver with the assistance of a skilled health care provider. The antenatal period also provides an opportunity to supply information on birth spacing, which is recognized as an important factor in improving infant survival. The prevention and management of anaemia during pregnancy and treatment of STIs can significantly improve foetal outcomes and im- prove maternal health. Adverse outcomes such as low birth weight can be reduced through a combination of in- terventions to improve women’s nutritional status and prevent infections (e.g., STIs) during pregnancy. More recently, the potential of the antenatal period as an entry point for HIV pre- vention and care, in particular for the preven- tion of HIV transmission from mother to child, has led to renewed interest in access to and use of antenatal services. The WHO recommends a minimum of four antenatal visits based on a review of the effective- ness of different models of antenatal care. WHO guidelines are specific on the content of antenatal care visits, which include: � Blood pressure measurement � Urine testing for bacteriuria and proteinuria � Blood testing to detect syphilis and severe anaemia � Weight/height measurement (optional) Coverage of antenatal care (by a doctor, nurse, or midwife) is almost universal in Serbia, with 99 percent of women receiving antenatal care at least once during pregnancy. Lower antenatal care cov- erage is noticed among Roma, the youngest and less educated women, and women from the poor- est households. Looking at the content of antenatal care received, almost all women in Serbia have their blood pres- sure measured, a blood sample and urine speci- men taken. 6 percent of women are not weighed. A Papanicolau test is performed on half of the women in Serbia during the antenatal care pro- vided, and is performed mostly in Belgrade. Those interventions are performed more frequently on the more educated and richest women. Coverage of Roma women with almost all reported inter- ventions is 20 percent lower than average. The type of personnel providing antenatal care to women aged 15–49 years who gave birth in the two years preceding is presented in Table RH.3. 98 percent of women have received antenatal care from a professional health worker (doctor, nurse or midwife). In the majority of cases, care is provided by a medical doctor (98 percent), while other medical personnel represent 1 percent or less. Contrary to these general results, antenatal care provided by medical personnel, especially a doctor is significantly lower among Roma and less educated women. 51MONITORING THE SITUATION OF CHILDREN AND WOMEN Assistance at Delivery Three quarters of all maternal deaths occur dur- ing delivery and the immediate post-partum pe- riod. The single most critical intervention for safe motherhood is to ensure that a competent health worker with midwifery skills is present at every birth, and that transport is available to a referral facility for obstetric care in case of emergency. A World Fit for Children goal is to ensure that women have ready and affordable access to skilled attendance at delivery. The indicators are the pro- portion of births with a skilled attendant and the proportion of institutional deliveries. The skilled attendant at delivery indicator is also used to track progress towards the Millennium Development target of reducing the maternal mortality rate by three quarters between 1990 and 2015. The MICS included a number of questions to as- sess the proportion of births attended by a skilled attendant. A skilled attendant includes a doctor, nurse, midwife or auxiliary midwife. About 99 percent of births occurring in the two years prior to the MICS survey were delivered by skilled personnel (Table RH.5). There are no statistically significant differences in any of the categories, except ethnicity. Roma women from Roma settlements and Muslim/Bosnian women are slightly behind the national average – 93 and 94 percent of those women respectively were delivered by skilled personnel. In most of the cases (87 percent) doctors assisted at the delivery. Doctors assisted at the delivery in a higher percentage to the youngest, more edu- cated and richest women. On average, an auxiliary midwife assisted in 10 percent of births in two years prior to the MICS survey. The other type of assistance during delivery is practically absent in Serbia (the exception are Roma again where 6 per- cent were delivered without assistance or with the assistance of a relative or friend). Roma women from Roma settlements more frequently delivered with the help of an auxiliary midwife. In Serbia, 99 percent of births which occurred in the past two years were delivered in health facilities. Similar to all other indicators, women from the “most at risk population”, Roma in Roma settlements happened to deliver in health facilities less than the average female population, 93 per- cent. Although there are some disparities regard- ing women’s education and living standard, they are not of statistical significance. 52 MICS3 FULL TECHNICAL REPORT It is well recognized that a period of rapid brain development occurs in the first 3–4 years of life, and the quality of home care is the major determinant of a child’s development during this period. In this context, adult activities with children, the presence of books in the home, for the child, and the conditions of care are impor- tant indicators of quality of home care. A World Fit for Children goal is that “children should be physically healthy, mentally alert, emotionally secure, socially competent and ready to learn.” Information on a number of activities that sup- port early learning was collected in the survey. These included the involvement of adults with children in the following activities: reading books or looking at picture books, telling sto- ries, singing songs, taking children outside the home, compound or yard, playing with chil- dren, and spending time with children naming, counting, or drawing things. For more than four in five (84 percent) un- der-five children, an adult engaged in more than four activities that promote learning and school readiness during the 3 days pre- ceding the survey (Table CD.1). The average number of activities that adults engaged in with children was 5. The father’s involvement in such activities was somewhat limited, with about 70 percent of fathers involved in, on average, 2 activities with the child. Around 5 percent of under-5 children in Serbia live in a household without their fathers. There are no gender or urban/rural differences found in family support for learning. The most significant differences concerning the parents’ involvement in development activities were in the terms of the child’s age. Children aged 2 to 5 years are more likely to be supported in different activi- ties by their parents, than younger children (94 as opposed to 69 percent). The higher the parents’ education is, the more they are involved in sup- porting child development. The same statement goes for wealth index – the richer families are, the more they are involved in child development (particularly fathers). While more or less all ethnic groups are showing similar behaviour patterns, Roma parents are showing a different habit – every second Roma child received development support from family members, with also a lower (on aver- age 3) number of activities. Paternal involvement in child activities among the Roma population is much lower – only one third of fathers are involved in child development, with less than one activity. Exposure to books in early years not only provides the child with greater understanding of the nature of print, but may also give the child opportunities to see others reading, such as older siblings doing school work. The presence of books is important for later school performance and IQ scores. In Serbia, in four out of five households, three or more books (children’s and non-children’s) are present (Table CD.2). The median number of books is 10. There are no gender differences, but it is found that a child’s exposure to books CHILDCHILD DEVELOPMENTDEVELOPMENT 53MONITORING THE SITUATION OF CHILDREN AND WOMEN is strongly influenced by household wealth and the mother’s education. While only 54 percent of children whose mother has primary or less education live in a household where more than 3 non-children’s books are present, that percentage is as much as 90 among children whose moth- ers are highly educated. Socio-economic status shows a similar pattern regarding book owner- ship. Only 50–54 percent of the poorest house- holds own 3 or more books from both categories, while this figure is 91–94 percent in rich families. The number of children’s books present in the household varies according to the child’s age. 87 percent of children aged between 2 and 5, as op- posed to 67 percent of younger children live in a household where 3 or more children’s books are present. Regional distributions show the highest percentage of books present in Belgrade (about 90 percent), and the lowest in Central Serbia (about 69 percent). Books are more present in urban than rural households. There are significant dif- ferences regarding exposure to books between Roma children and the rest of the population. One in four Roma children is living in a house- hold with books. The median number of books present in Roma households is zero. Table CD.2 also shows that 21 percent of children aged 0–59 months had 3 or more playthings to play with in their homes, while 5 percent had no playthings. The playthings in MICS included household objects, homemade toys, toys that came from a store, and objects and materi- als found outside the home. Most children, 88 percent, play with toys that come from a store; one third of them play with household objects or objects and materials found outside the house. Just 20 percent of children are playing with homemade toys. Gender differences are not noticed. Interestingly, playthings are less present in Belgrade than in other regions. As expected in rural areas, there are more objects and materials found outside the home than in urban areas (34 vs 27 percent). As many as 13 percent of Roma and of Muslim children do not have any playthings. Roma children play with objects and homemade toys above average, and with toys that came from a store below average. The percentage of children from the poorest households that play with toys that came from a store is significantly below aver- age, but they more often play with homemade toys and objects and materials found outside the home. Leaving children alone or in the presence of other young children is known to increase the risk of accidents. In MICS, two questions were asked to find out whether children aged 0–59 months were left alone during the week preceding the inter- view, and whether children were left in the care of other children under 10 years of age. Table CD.3 shows that 9 percent of children were left with inadequate care during the week preced- ing the survey – 7 percent of children aged 0–59 months were left in the care of other children, while 5 percent were left alone during the week preceding the interview. Female and children from rural areas are more often left with inadequate care. This practice is more present in Belgrade, West and Central Serbia (10–11 percent) than in Vojvodina (5 percent). Older children (aged 24–59 months) were left with inadequate care more than younger ones. Mothers with primary or no educa- tion and the poorest tended to leave the child with inadequate care more than others. This practice is also more present in Roma and Muslim/Bosnian families (18 and 16 percent of children left without proper care, respectively). 54 MICS3 FULL TECHNICAL REPORT Pre-School Attendance and School Readiness Attending pre-school education in an organ-ised learning or child education programme is important for children’s readiness for school. One of the World Fit for Children goals is the promotion of early childhood education. In Serbia, every third child aged between 3 and 5 years is attending some form of organised early education programme (Table ED.1). While there are no gender differences, all other differences are very prominent (see Figure ED.1). The small national coverage is even smaller among Roma (4 percent coverage), the poorest, mothers with primary or no education (7 per- cent) and in rural areas (14 percent). Regionally it is more prevalent in Belgrade (57 percent), and lowest in the East and South East region (18 and 21 percent respectively). Those facts are strong evidence that the system favours the better-off, the richest, and working families in the main urban centres. The table ED.1 also shows the proportion of children in the first grade of primary school who attended pre-school the previous year (Table ED.1), an important indicator of school readiness. Overall, 89 percent of children who are currently aged 7 and attending the first grade of primary school were attending pre- school the previous year. This proportion was somewhat expected, mainly owing to the fact that the pre-school programme before first grade is obligatory in Serbia. Again, the biggest difference is coverage of Roma and children from the poorest households, where just 62 percent of Roma and 77 percent of the poorest children attended pre-school programmes one year prior to the first grade. No significant dif- ferences in terms of child gender, region and type of settlement were found. EDUCATIONEDUCATION Figure ED.1 Percentage of children aged 36–59 months who are attending some form of organised early childhood education program, Serbia, 2005 Se rb ia w ith ou t R om a fr om Ro m a se tt le m en ts Ro m a in R om a se tt lm en ts Ur ba n Ru ra l Pr im ar y o r n on e Se co nd ar y Un iv er sit y Po or es t Se co nd M id dl e Fo ur th Ri ch es t Se rb ia 33.4 3.9 45.2 14.4 7.4 33.1 57.6 7.4 11.9 25.6 45.9 64.1 32.5 55MONITORING THE SITUATION OF CHILDREN AND WOMEN Primary and Secondary School Participation Universal access to basic education and the achievement of primary education by the world’s children is one of the most important goals of the Millennium Development Goals and A World Fit for Children. Education is a vital prerequisite for combating poverty, empowering women, protect- ing children from hazardous and exploitative labour and sexual exploitation, promoting human rights and democracy, protecting the environ- ment, and influencing population growth. The indicators for primary and secondary school attendance include: � Net intake rate in primary education � Net primary school attendance rate � Net secondary school attendance rate � Net primary school attendance rate of children of secondary school age � Female to male education ratio (GPI) The indicators of school progression include: � Survival rate to grade five � Transition rate to secondary school � Net primary completion rate Of children who are of primary school entry age (children that are to turn 7 in the observed cal- endar year) in Serbia, 94 percent are attending the first grade of primary school (ED.2). The only significant difference concerning intake rate in primary education is detected in the case of Roma children. Only 66 percent of them enter primary school on time. Table ED.3 shows the percentage of children of primary school age attending primary or second- ary school. The majority of children of primary school age are attending school (98 percent of children that are to turn 7 to 14 in the observed calendar year). Similar to the intake rate findings, the only category that is significantly lower is the category of Roma children. The rate of pri- mary school attendance among this population is significantly lower, at 74 percent. On average, there are no significant differences between boys and girls in regard to elementary education at- tendance. Again, exceptions are Roma children, where 76 percent of boys compared to 71 percent of girls are attending primary school. Among Roma children, the child’s age has a strong influ- ence on the school attendance rate. The atten- dance rate is lowest in the first grade. The maxi- mum is reached between the ages of 8 and 10, and after that, a serious drop is visible at the age of 11 with girls, and at the age of 12 with boys. The difference between the rate of primary education and the primary school attendance rate indicates that there are certain children who start their education later than expected. This occurrence is more frequent among Roma children, since the difference between rates is higher in their case than the average (8 to 5 percent). 84 percent of children of secondary school age are attending secondary school (Table ED.4). Of the remaining 16 percent, 1.5 percent are attend- ing primary school (Table ED.4w), while the rest are out of school. The proportion of children attending secondary school among the Roma population is drastically below average, at 10 percent (Table ED.4). Again, among this ethnic group, gender differences are significant: the proportion of boys in second- ary school is twice as high as the proportion of girls (14 and 6 percent), which is the complete opposite of the rest of the population, where the percentage of girls in secondary school is slightly higher. Secondary school attendance is influenced by household wealth. About 64 percent of chil- dren from the poorest households are attending secondary school, while that is the case with 94 percent of children from the richest households. Children from urban areas are more likely to continue their education after primary school. 56 MICS3 FULL TECHNICAL REPORT The percentage of children entering first grade who eventually reach grade 5 is presented in Table ED.5. Nearly all children starting grade one will eventually reach grade five. The only exception is Roma children from Roma settlements. Data shows that 97 percent of Roma children entering first grade will eventually reach grade five. No other significant differences were observed. Notice that this number includes children that repeat grades and that eventually move up to reach grade five. The net primary school completion rate and tran- sition rate to secondary education is presented in Table ED.6. At the time of the survey, 91 percent of children of primary completion age (14 years) were attending the last grade of primary educa- tion. A significantly lower percentage of Roma children (28 percent) attended the last grade of primary education. This value should be distin- guished from the gross primary completion ratio which includes children of any age attending the last grade of primary. The majority of children who successfully completed the last grade of pri- mary school (97 percent) were found to be attend- ing the first grade of secondary school. No signifi- cant gender, regional or socioeconomic influences on education continuance were detected. The ratio of girls to boys attending primary and secondary education is provided in Table ED.7. These ratios are better known as the Gender Parity Index (GPI). Notice that the ratios in- cluded here are obtained from net attendance ratios rather than gross attendance ratios. The latter ratios provide an erroneous description of the GPI mainly because in most of the cases the majority of over-aged children attending primary education tend to be boys. The table shows that gender parity for primary school is close to 1.0, indicating no difference in the at- tendance of girls and boys to primary school. However, the indicator goes up to 1.1 for sec- ondary education. The disparity of girls is only pronounced in the Roma in Roma settlements, where the GPI for primary and secondary school are 0.94 and 0.42, respectively. Adult Literacy One of the World Fit for Children goals is to assure adult literacy. Adult literacy is also an MDG indicator, relating to both men and women. In MICS, since only a women’s ques- tionnaire was administered, the results are based only on females aged 15–24. Literacy was assessed on the ability of women to read a short simple statement, or on school attend- ance. The literacy percentage is presented in Table ED.8. The literacy rate of females, aged 15–24 years in Serbia is 96 percent. As expect- ed, the literacy level is lower among women with no or primary education. Younger women aged 15–19 are less literate than women aged 20–24 (93 compared to 98 percent). 52 percent of Roma women aged 15 to 24 years are literate – much lower than the national average. The women from the second and middle wealth index quintile classes are the most literate. 57MONITORING THE SITUATION OF CHILDREN AND WOMEN Birth Registration The Convention on the Rights of the Child states that every child has the right to a name and a nationality and the right to pro- tection from being deprived of his or her iden- tity. Birth registration is a fundamental means of securing these rights for children. The World Fit for Children states the goal of developing systems to ensure the registration of every child at or shortly after birth, and fulfil his or her right to acquire a name and a nationality, in accordance with national laws and relevant in- ternational instruments. The indicator is the percentage of children under 5 years of age whose birth is registered. The births of 99 percent of children under five years in Serbia have been registered. There are no significant variations in birth registration across sex, age, type of settlement, mother’s education or wealth index. Only Roma from Roma settlements are somewhat less likely to have their births regis- tered than other children (95 percent registered). Child Labour Article 32 of the Convention on the Rights of the Child states: “Parties recognize the right of the child to be protected from economic ex- ploitation and from performing any work that is likely to be hazardous or to interfere with the child’s education, or to be harmful to the child’s health or physical, mental, spiritual, moral or social development…” The World Fit for Children mentions nine strategies to combat child labour and the MDGs call for the protection of children against exploitation. In the MICS questionnaire, a number of ques- tions addressed the issue of child labour, that is, children 5–14 years of age involved in labour activities. A child is considered to be involved in child labour activities at the moment of the survey if during the week preceding the survey: � Ages 5–11: at least one hour of economic work or 28 hours of domestic work per week. � Ages 12–14: at least 14 hours of economic work or 28 hours of domestic work per week. This definition allows one to differentiate child labour from child work and to identify the type of work that should be eliminated. As such, the esti- mate provided here is a minimum of the prevalence of child labour since some children may be involved in hazardous labour activities for a number of hours that could be less than the numbers specified in the criteria explained before. Table CP.2 presents the results of child labour by the type of work. 4 percent of children aged 5 to 14 years in Serbia are involved in child labour, mainly unpaid and working for the family business. There is a strong correlation between child labour and the type of settlement a child is liv- ing in. Children living in rural areas are twice as involved in child labour activities (6 percent), than CHILD PROTECTIONCHILD PROTECTION 58 MICS3 FULL TECHNICAL REPORT urban children (3 percent). The poorest children and Roma children from Roma settlements are the most exploited group when it comes to child labour. As many as 8 percent of children from the poorest households and 7 percent of Roma children are involved in child labour. The pattern is the same – it is mainly unpaid, family business- type work. The exception are Roma children, who work outside home, doing an equal amount of paid and unpaid work. Children whose mothers have secondary or higher education are less likely to be involved in labour (4 percent), than children whose mothers have primary or no education (6 percent). Table CP.3 shows the percentage of children clas- sified as student labourers or as labourer students. Student labourers are children attending school that were involved in child labour activities at the time of the survey. More specifically, of the 92 percent of children aged 5–14 attending school, 4 percent are also involved in child labour activities. The proportion of student labourers is almost two times higher among the poorest and Roma chil- dren, at 8 percent and 6 percent respectively. On the other hand, out of the 4 percent of chil- dren classified as child labourers, the majority of them are also attending school (93 percent). This percentage is slightly lower among the poorest children; 89 percent of those who work are also attending school, while among Roma children this percentage is significantly lower (58 percent). Statistically, it is important to emphasise that only 90 percent of children from the poorest house- holds and 67 percent of Roma children aged 5 to 14 are attending school. Child Discipline As stated in A World Fit for Children, “children must be protected against any acts of violence…” and the Millennium Declaration calls for the protection of children against abuse, exploitation and violence. In the Serbia MICS survey, moth- ers/caretakers of children aged between 2 and 14 were asked a series of questions on the methods parents tended to use to discipline their children when they misbehaved. Note that for the child discipline module, one child aged 2–14 per house- hold was selected randomly during fieldwork. Out of these questions, the following indicators used to describe aspects of child discipline are: 1) the number of children aged 2–14 that experi- ence psychological aggression as punishment or minor physical punishment or severe physical punishment; and 2) the number of parents/care- takers of children 2–14 years of age that believe that in order to raise their children properly, they need to physically punish them. In Serbia, 73 percent of children aged 2–14 years were subjected to at least one form of psychologi- cal or physical punishment by their mothers/care- takers or other household members. Two thirds of those children were exposed to psychological punishment, half of them to minor physical pun- ishment and 7 percent were subjected to severe physical punishment. The latter figure correlates with the percentage of mothers/caretakers who believed that children should be physically pun- ished, which unfortunately implies that they are applying this belief in practice. Every fifth child in Serbia has been disciplined through non-violent methods, and 6 percent of children have neither been punished nor disciplined which leaves an open question: how are they disciplined? Male children were subjected more to both minor and severe physical discipline (53 and 8 percent) than female children (49 and 5 percent). Violent disciplining is most practised in South-East Serbia, where two-thirds of children experienced minor, and one in seven children severe physical punishment. A seed of a good practice can be seen in Belgrade, where 27 percent of caregivers are using non-violent disciplinary methods. Physical punishment decreases with the age of the child. The mother’s education is strongly correlated with the child disciplining methods used – less educated mothers use non-violent methods less, and violent methods more in child disciplining. The same pattern is seen looking at the wealth index – use of violent methods decreases and use of non-violent methods increases according to the degree of wealth. 14 percent of the poorest car- egivers use severe physical punishment for disci- 59MONITORING THE SITUATION OF CHILDREN AND WOMEN plining the child – twice the national average. The ethnicity of the head of the household is strongly correlated with the disciplinary method used. While Serbs and Hungarians show similar prac- tices (the difference is that Hungarians use more psychosocial and less physical punishment), Roma and Muslims/Bosnians practise completely dif- ferent methods – 89 and 86 percent respectively use some form of punishment in child disciplin- ing. Severe physical punishment is a frequently used disciplinary method (22 among Roma and 15 among Muslims/Bosnians). Only 6 percent of Roma and 9 percent of Muslim/Bosnian children are disciplined using non violent methods. Early Marriage Marriage before the age of 18 is a reality for many young girls. According to UNICEF’s worldwide estimates, over 60 million women aged 20–24 were married/in union before the age of 18. Factors that influence child marriage rates include: the state of the country’s civil registration system, which pro- vides proof of age for children; the existence of an adequate legislative framework with an accompa- nying enforcement mechanism to address cases of child marriage; and the existence of customary or religious laws that condone the practice. In many parts of the world parents encourage the marriage of their daughters while they are still children in the hope that the marriage will benefit them both financially and socially, while also re- lieving financial burdens on the family. In actual fact, child marriage is a violation of human rights, compromising the development of girls and often resulting in early pregnancy and social isolation, with little education and poor vocational training reinforcing the gendered nature of poverty. The right to ’free and full’ consent to a marriage is rec- ognized in the Universal Declaration of Human Rights – with the recognition that consent cannot be ’free and full’ when one of the parties involved is not sufficiently mature to make an informed de- cision about a life partner. The Convention on the Elimination of all Forms of Discrimination against Women mentions the right to protection from child marriage in article 16, which states: “The betrothal and the marriage of a child shall have no legal effect, and all necessary action, including legislation, shall be taken to specify a minimum age for marriage…” While marriage is not consid- ered directly in the Convention on the Rights of the Child, child marriage is linked to other rights – such as the right to express their views freely, the right to protection from all forms of abuse, and the right to be protected from harmful traditional practices – and is frequently addressed by the Committee on the Rights of the Child. Young married girls are often required to perform large amounts of domestic work, are under pres- sure to demonstrate fertility, and are responsible for raising children while still children them- selves. Women who married at younger ages were more likely to believe that it is sometimes accept- able for a husband to beat his wife and were more likely to experience domestic violence themselves. The age gap between partners is thought to con- tribute to these abusive power dynamics and to increase the risk of untimely widowhood. Closely related to the issue of child marriage is the age at which girls become sexually active. Women who are married before the age of 18 tend to have more children than those who marry later in life. Pregnancy related deaths are known to be a leading cause of mortality for both married and unmarried girls between the ages of 15 and 19, particularly among the youngest of this group. Two indicators are to estimate the percentage of women married before 15 years of age and the percentage married before 18 years of age. The percentage of women married at various ages is provided in Table CP.5. In Serbia, according to the Family Law, marriage is not allowed before the age of 19. However, under special circumstances marriage is allowed after the age of 16. Nearly 1 percent of women aged 15 to 49 had married before the age of 15. The tendency towards early marriage is significantly higher among Roma women from Roma settlements, 60 MICS3 FULL TECHNICAL REPORT where 12 percent of interviewed women married before the age of 15. 8 percent of all women aged 20 to 49 married be- fore the age of 18. This practice is more present in East Serbia, rural areas, among the less educated, the poorest and particularly among Roma. For example, almost half of Roma women from Roma settlements were married before the age of 18, ev- ery third woman with primary or no education and every fifth woman from the poorest households. Another component is the spousal age difference with an indicator being the percentage of married/ in union women with a difference of 10 or more years of age compared to their current spouse. Table CP.6 shows the results of the age difference between husbands and wives. While 26 percent of young married women aged 15 to 19 are married to a partner 10 or more years older, among married women aged 20 to 24 years this percentage is much lower – 14 percent. Interestingly this phenomenon is mainly correlated with poverty and lower educa- tion – every third young women is married to a husband who is more than 10 years older, for ex- ample. On the other hand, Roma women, although entering marriage very early, are mainly married to slightly older partners, and very seldom to partners 10 years older. Domestic Violence A number of questions were asked of women aged 15–49 to assess their attitudes towards whether husbands are justified to hit or beat their wives/partners for a variety of scenarios. These questions were asked to have an indication of cultural beliefs that tend to be associated with the prevalence of violence against women by their husbands/partners. The main assumption here is that women that agree with the statements indicating that husbands/partners are justified to beat their wives/partners under the circumstances described in reality tend to be abused by their own husbands/partners. The responses to these questions can be found in Table CP.7. Overall, 6 percent of women in Serbia feel that their husband/partner has a right to hit or beat them, mostly in cases when they neglect the chil- dren (6 percent), or if they demonstrate their au- tonomy, e.g. go out without telling their husbands or argue with him (2 percent). Regionally, domestic violence is more accepted in South and West Serbia (approximately 12 percent) than in other parts of the country, owing to the same reasons as stated above. Acceptance is more present among the poorest and less educated, and also currently married women. However, it is mainly correlated with ethnicity – every third Muslim/Bosnian and Roma woman believes it is justified for a husband to beat his wife/partner, mainly when she neglects the children or goes out or argues with him, and in the case of Roma women also if she refuses to have sex with him. Child Disability One of the World Fit for Children goals is to pro- tect children against abuse, exploitation, and vio- lence, including the elimination of discrimination against children with disabilities. For children aged between 2 and 9 years, a series of questions were asked to assess the number of disabilities/impair- ments, such as sight impairment, deafness, and dif- ficulties with speech. This approach is based on the concept of functional disability developed by the WHO and aims to identify the implications of any impairment or disability for the development of the child (e.g. health, nutrition, education, etc.). Table CP.8 shows the results of these questions. According to the mothers’ report, every tenth child aged between 2 and 9 in Serbia displays some kind of disability. This appears to be more frequent among Roma children from Roma set- tlements (23 percent), children from the poorest households (17 percent) and those whose moth- ers are less educated (15 percent). Also, child dis- abilities are more frequent among children from rural (13 percent), than children from urban settlements (9 percent). 61MONITORING THE SITUATION OF CHILDREN AND WOMEN Knowledge of HIV Transmission and Condom Use One of the most important prerequisites for reducing the rate of HIV infection is accurate knowledge of how HIV is transmitted and strategies for preventing transmission. Correct in- formation is the first step towards raising aware- ness and giving young people the tools to protect themselves from infection. Misconceptions about HIV are common and can confuse young people and hinder prevention efforts. The UN General Assembly Special Session on HIV/AIDS (UNGASS) called on governments to improve the knowledge and skills of young people to protect themselves from HIV. The indicators to measure this goal as well as the MDG of reducing HIV infections by half include improving the level of knowledge of HIV and its prevention, and chang- ing behaviour to prevent further spread of the disease. The HIV module was administered to women 15–49 years of age. One indicator which is both an MDG and UNGASS indicator is the percentage of young women who have comprehensive and correct knowledge of HIV prevention and transmission. Women were asked whether they knew of the three main methods of HIV transmission – hav- ing only one faithful uninfected partner, using a condom every time, and abstaining from sex. HIV/AIDS AND SEXUALHIV/AIDS AND SEXUAL BEHAVIOURBEHAVIOUR The results are presented in Table HA.1. In Serbia, 98 percent of the interviewed women had heard of AIDS. Although this percentage is significantly high, the percentage of women who know all three main ways of preventing HIV transmission is only 36 percent. Comprehensive and correct knowledge of HIV prevention is more frequent among young women, women who are wealthier and have secondary or higher education. On the other hand, women who have primary educa- tion, or belong to the poorest class are less aware of AIDS existence, and their correct knowledge about HIV transmission is below average. Roma women from Roma settlements are fairly unfa- miliar with AIDS. Only 72 percent of them had ever heard of AIDS, and only 14 percent of them are aware of all three ways of transmission. An alarming fact is that one in two Roma women are unaware of any way HIV transmission can be prevented. 86 percent of women know about using a condom every time, 67 percent about having one faith- ful uninfected sex partner, and 50 percent know about abstaining from sex as the main ways of preventing HIV transmission. While 91 percent of women know at least one way, almost one in ten do not know any of the three ways. 62 MICS3 FULL TECHNICAL REPORT The results show a high correlation between women’s place of residence and knowledge of HIV. Women living in urban settlements are likely to be better informed about HIV and its preven- tion, as well as those living in Belgrade (a highly urban city) and in Vojvodina. On the other hand, 16 and 10 percent of women respectively, living in Central and East Serbia, do not know any way of preventing transmission. Table HA.2 shows the percentage of women who can correctly identify misconceptions concerning HIV. The indicator is based on the two most com- mon and relevant misconceptions in Serbia: that HIV can be transmitted by sharing food with an infected person and mosquito bites. The table also provides information on whether women know that HIV cannot be transmitted by supernatural means, and that HIV can be transmitted by shar- ing needles. Only half of the interviewed women reject the two most common misconceptions and know that a healthy-looking person can be infect- ed. The table shows that the most common mis- conception is that mosquitoes can transmit HIV, since almost one third of women in Serbia believe in this kind of transmission. 79 percent of women know that HIV can not be transmitted by sharing food with an infected person and 77 percent of women know that a healthy-looking person can be infected. One in ten women believes that HIV can be transmitted by supernatural means. This misconception is particularly common among Roma and the poorest women; every second Roma woman and every fourth woman from a poor household believes in this misconception. In general, misconceptions are more common in rural areas, and among poorer and less educated women. Looking at territory distribution, we notice that the most informed are women from Belgrade and Vojvodina. Table HA.3 summarises information from Tables HA.1 and HA.2 and shows the percentage of women who know two ways of preventing HIV transmission and reject three common miscon- ceptions. Although a large proportion of women have heard of AIDS, only 37 percent of women in Serbia have comprehensive knowledge regarding this topic. This figure indicates that comprehensive knowl- edge of HIV prevention methods and transmis- sion in Serbia is still fairly low, although there are differences according to age groups, areas of residence, education, and women’s wealth. There is a significant lack of knowledge among Roma women from Roma settlements. Only 5 percent of Roma women know two ways of preventing transmission and reject the three most common misconceptions. The highest level of knowledge is registered among adolescents and younger women, but still, it is below 50 percent. Women from urban areas have a slightly higher knowledge of HIV transmission. As expected, the percentage of women with comprehensive knowledge in- creases with the woman’s education level (Figure HA.1). While in Vojvodina, Belgrade and West Serbia almost every second women has above average knowledge (from 44 to 46 percent), in Central, East and South-East Serbia, comprehen- sive knowledge is registered only in every fourth woman (from 24 to 28 percent). Knowledge of mother-to-child transmission of HIV is also an important first step for women to seek HIV testing when they are pregnant to Figure HA.1 Percentage of women who have comprehensive knowledge of HIV/AIDS transmission, Serbia, 2005 Primary or none Secondary University Serbia KNOWS 2 WAYS TO PREVENT HIV IDENTIFY 3 MISCONCEPTIONS COMPREHENSIVE KNOWLEDGE 45 28 19 67 52 38 70 72 53 63 51 37 63MONITORING THE SITUATION OF CHILDREN AND WOMEN avoid infection of the baby. Women should know that HIV can be transmitted during pregnancy, delivery, and through breastfeeding. The level of knowledge among women aged 15–49 years concerning mother-to-child transmission is presented in Table HA.4. Overall, 87 percent of women know that HIV can be transmitted from mother to child. The percentage of women who know all three ways of mother-to-child transmis- sion is 57 percent, while 11 percent of women did not know of any specific way. The knowledge of HIV transmission during pregnancy is most com- mon among women. When two other ways are in question, the values decrease slightly. 74 percent of women know that HIV can be transmitted from mother to child at delivery, and 62 percent are aware of HIV transmission through breast- feeding. As in the case of other indicators, lack of mother-to-child HIV transmission is more promi- nent among Roma women, women from the poor- est households and women who are less educated. Only one in five of Roma women are aware of the ways of mother-to-child transmission of HIV. The indicators on attitudes towards people living with HIV measure stigma and discrimination in the community. Stigma and discrimination are low if respondents report an accepting attitude to the following four questions: 1) would care for a family member suffering from AIDS; 2) would buy fresh vegetables from a vendor who was HIV positive; 3) thinks that a female teacher who is HIV positive should be allowed to teach in school; and 4) would not want to keep the HIV status of a family member a secret. Table HA.5 shows the attitudes of women towards people living with HIV/AIDS. In Serbia, 64 percent of women who have heard of AIDS agree with at least one discriminatory statement. According to the survey results, the most common discriminative attitude is rejec- tion of buying food from a person with HIV/ AIDS. Half the women who have heard of AIDS would not buy food from an infected person. Less educated women and those from the poor- est households have more severe prejudices than the ones with a better wealth status and higher education. Ethnicity is strongly correlated to discriminatory attitudes. Data shows that about 80 percent of Muslims/Bosnians and Roma agree with at least one of the discriminatory state- ments. A very low percentage of women who have heard about AIDS would not take care of a family member infected with HIV (2 percent). But Roma women from Roma settlements have a more discriminative attitude towards this statement: as many as 7 percent of Roma women wouldn’t take care of a family member who was suffering from AIDS, although the percentage of them who would keep it a secret is lower than the national average. All this data confirms the facts that lack of knowledge is a serious source of fear and not an acceptance of differences. Another important indicator is the knowledge of where to be tested for HIV and use of such servic- es. Questions related to knowledge among women of a facility for HIV testing and whether they have ever been tested is presented in Table HA.6. 69 percent of women know where to be tested, while 7 percent have actually been tested. This informa- tion is known only to every fifth Roma woman from a Roma settlement and about every second woman with a low level of education and from the poorest households. A slightly larger percent- age applies to women from rural settlements (59 percent) and from South-East Serbia (52 percent). Although the knowledge of a place to get tested is not so low, only 7 percent of all women have been tested. This percentage is significantly higher among women living in Belgrade and in East Serbia, as well as among more educated women and those from wealthier households. Among women who had given birth within the two years preceding the survey, the percentage who had received counselling and HIV testing during antenatal care is shown in Table HA.7. Although the coverage of antenatal care in Serbia is quite high (98 percent), information about HIV was provided to only every seventh women (14 percent), and only every tenth women has been tested and received her results within the ante- natal care program. East Serbia is the only region where HIV counselling and testing during ante- natal care is well provided: 35 percent of women from East Serbia have been tested and received 64 MICS3 FULL TECHNICAL REPORT results during antenatal care. On the other hand, the population from Roma settlements barely receive any antenatal care at all. Only 2 percent of Roma women received their test results within the antenatal care program. Sexual Behaviour Related to HIV Transmission Promoting safer sexual behaviour is critical for reducing HIV prevalence. The use of condoms during sex, especially with non-regular partners, is especially important for reducing the spread of HIV. In most countries over half of new HIV infections are among young people between the ages of 15 and 24, thus a change in behaviour among this age group will be especially important to reduce new infections. A module of questions was administered to women aged 15–24 to assess their risk of HIV infection. Risk factors for HIV include sex at an early age, sex with older men, sex with a non-marital non-cohabitating partner, and failure to use a condom. The frequency of sexual behaviour that increases the risk of HIV infection among women is pre- sented in Table HA.8 and Figure HA.2. With the exception of Roma women from Roma settlements, the number of young women who had sex before the age of 15 in Serbia is very low. Only 1 percent of women aged 15 to 19 had sex before the age of 15. On the other hand Roma women from Roma settlements start their sexual activities as very young girls, 16 percent of them had sex before the age of 15. The situation is simi- lar regarding sex at the age of 18. Namely, while in Serbia 19 percent of women aged 20 to 24 report- ed having sex before the age of 18, among Roma women this percentage is 51, which is over two times higher. Girls who terminate their education at primary school level tend to have sex at an early age; 3 percent before the age of 15 and 42 percent before the age of 18. Having sex with a partner 10 or more years older is reported by one in twelve women. It appears that women with primary or less education are more inclined to these kinds of relationships. One in every five women from the poorest house- holds had had sex with a partner 10 or more years older in the 12 months prior to the MICS. Condom use during sex with men other than husbands or live-in partners (non-marital, non-cohabiting) was assessed in women aged 15–24 who had had sex with such a partner in the previous year (Table HA.9). Over 60 percent of women aged 15–24 report having sex with a non-regular partner in the 12 months prior to the MICS. Three-quarters of those women reported condom usage when they had sex with the high risk partner. Figure HA.2 Sexual behaviour that increases risk of HIV infection, Serbia, 2005 Urban Rural Serbia WOMEN 15–19 WHO HAD SEX BEFORE AGE 15 WOMEN 20–24 WHO HAD SEX BEFORE AGE 18 WOMEN 20–24 WHO HAD SEX IN LAST 12 MONTHS WITH A MAN 10 YEARS OR MORE OLDER 18 6 20 9 111 19 7 65MONITORING THE SITUATION OF CHILDREN AND WOMEN List of ReferencesList of References Boerma, J. T., Weinstein, K. I., Rutstein, S. O., and Sommerfelt, A. E., 1996. Data on Birth Weight in Developing Countries: Can Surveys Help? Bulletin of the World Health Organisation, 74(2), 209–16. Filmer, D. and Pritchett, L., 2001. Estimating wealth effects without expenditure data – or tears: An application to educational enrolments in states of India. Demography 38(1): 115–132. Rutstein, S.O. and Johnson, K., 2004. The DHS Wealth Index. DHS Comparative Reports No. 6. Calverton, Maryland: ORC Macro. UNICEF, 2006. Monitoring the Situation of Children and Women. Multiple Indicator Cluster Survey Manual, New York. United Nations, 1983. Manual X: Indirect Techniques for Demographic Estimation (United Nations publication, Sales No. E.83.XIII.2). United Nations, 1990a. QFIVE, United Nations Program for Child Mortality Estimation. New York, UN Pop Division United Nations, 1990b. Step-by-step Guide to the Estimation of Child Mortality. New York, UN www.childinfo.org. 67MONITORING THE SITUATION OF CHILDREN AND WOMEN AppendicesAppendices APPENDIX A SAMPLE DESIGN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 APPENDIX B LIST OF PERSONNEL INVOLVED IN THE SURVEY . . . . . . . . . 74 APPENDIX C ESTIMATES OF SAMPLING ERRORS . . . . . . . . . . . . . . . . . 75 APPENDIX D DATA QUALITY TABLES . . . . . . . . . . . . . . . . . . . . . . . . . 86 APPENDIX E MICS INDICATORS: NUMERATORS AND DENOMINATORS. . . . . . . . . . . . . . . . . 97 APPENDIX F QUESTIONNAIRES . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 68 MICS3 FULL TECHNICAL REPORT 69MONITORING THE SITUATION OF CHILDREN AND WOMEN The major features of sample design are de-scribed in this appendix. Sample design features include target sample size, sample allocation, sample frame and listing, choice of domains, sampling stages, stratification, and the calculation of sample weights. The primary objective of the sample design for the Serbia Multiple Indicator Cluster Survey was to produce statistically reliable estimates of most indicators, at national level, for urban and rural areas, and for the six regions of the coun- try: Vojvodina, Belgrade, West, Central, East and South-East Serbia. In order to look more deeply into ethnic disparities and to provide national estimates, a separate sample was designed for Roma living in Roma settlements. A stratified, two-stage random sampling approach was used for the selection of the survey sample. Sample Size and Sample Allocation The Serbia MICS3 sample was created by merg- ing two samples: Serbia without Roma from Roma settlements and Roma living in Roma settlements. The average household size and the percentage of children under five in the total population were the factors which caused slightly different sample design for each of the samples. Serbia is characterised by a very low fertility rate and a small number of household members. For example, one generation of children born makes up less than 1 percent of the population, and the average number of household members is around 3. Owing to these facts, the modification of the recommended sample plan had to be made, and that was the stratification of households in se- lected census block units into two categories: households with children and households without children under 5. The allocation of the sample in the category of households with children was sig- nificantly bigger than the allocation of the sample in the category of households without children. In the case of the Roma population, the universe could be defined only for Roma who live in sepa- rate settlements. The birth rate and household size among Roma living in Roma settlements is higher than in other population groups, so key determinants were different to those used in calculations for Serbia. The target sample size for the Serbia MICS was calculated as 7200 households in Serbia exclud- ing Roma settlements and 1900 Roma households from Roma settlements. For the calculation of the sample size, for both Serbia without Roma and Roma samples, the key indicator used was the percentage of children aged 0–4 years who had had Acute Respiratory infections. The following formula was used to estimate the required sample size for these indicators: n = [4 (r) (1–r) (f) (nr)] (1)[(me)2 (r)2 (p) (nh)] where � n is the required sample size, expressed as the number of households � 4 is a factor to achieve the 95 per cent level of confidence � r is the predicted or anticipated prevalence (coverage rate) of the indicator Appendix A Sample DesignSample Design 70 MICS3 FULL TECHNICAL REPORT � nr is the factor necessary to raise the sample size by 100(nr – 1) percent for non-response � f is the shortened symbol for deff (design effect) � me * r is the margin of error to be tolerated at the 95 percent level of confidence, defined as me percent of r (relative sampling error of r) � p is the proportion of the total population upon which the indicator, r, is based � nh is the average household size. As far as the sample size for Serbia excluding Roma settlements is concerned, the following levels of parameters were included: r (percentage of children aged 0–4 years who had had Acute Respiratory infections) was assumed to be 12 percent. The expected non-response rate nr, was determined at 15 percent. The value of deff (de- sign effect) was taken as 1.5 based on estimates from previous surveys. The maximum relative error allowed (me) was 12 percent, p (percentage of children aged 0–4 years in the total popula- tion) was taken as 4.5 percent and nh (average household size) was taken as 3. For the Roma sample: r (percentage of children aged 0–4 years who had had Acute Respiratory infections) and the expected non-response rate (nr) were assumed to be 12 and 15 percent, respec- tively, like the Serbia sample. Also the predicted value for design effect was the same, 1.5. The rela- tive margin of error wanted (me) was 20 percent. The percentage of children 0–4 years in the total population, p was taken as 9.5, and nh (average household size) as 4.7. The resulting number of households from these exercises was as follows: In the case of Serbia, excluding Roma settlements, the calculated sample size was 26000 households. Only a sample of that size would provide a signifi- cant number of children under 5 for drawing reli- able conclusions. Therefore, in order to cut down the number of households in the sample, but not to lose estimation reliability, the stratification of the sample into categories with and without chil- dren aged 0–4 years was needed. For calculation of the necessary number of households in each category, the following formula was used: n = (ns) (nc) (ps) (2) where � n is the required sample size, expressed as the number of households � ns is the expected number of households with, or the number of households without children under 5 in a cluster, depending on what cat- egory the calculation is used � nc is the number of clusters in the sample, and � ps is the probability of selection of the house- hold in each category. Taking into account that the proportion of children under 5 in the total population, p was 4.5 percent, and if the average household size is 3, the estimated number of households with children was 13.5 per 100 households (the average number of households in each cluster). So the ns was assumed to be 13.5 for the category with children, and 86.5 for the category without children. The probability of selec- tion of a household (ps) with at least one child out of all households with children was assumed to be 0.67, and the probability of selecting a household without children from all households with children in each cluster was 0.1. Supposing that 400 clus- ters were about to be selected, the total number of households was calculated at 3600 households with, and 3400 of households without children un- der 5, which makes a total of 7000 households. For the Roma sample, the stratification of primary units was not needed. Using formula (1), the cal- culated sample size was 1800 households. The average cluster size in the Serbia MICS was determined as 18 households, plus 3 backup households for both Serbia and Roma samples. Back-up households were to be interviewed only if some of the first 18 households were not found. In cases where a household refused to be inter- viewed, the substitution with a back-up household was not possible. The calculation was based on a 71MONITORING THE SITUATION OF CHILDREN AND WOMEN number of considerations, including the budget available, and the time that would be needed per team to complete one cluster. Dividing the total number of households by the number of house- holds per cluster, it was calculated that the selec- tion of a total number of 400 clusters in Serbia without Roma from Roma settlements and 106 clusters in Roma settlements would be needed in all regions. Table SD.1 Allocation of Sample Clusters (Primary Sampling Units) to Sampling Domains Region Population (Census 2002) Roma Population (Census 2002) Number of Clusters Serbia without Roma Number of Clusters Roma in Roma settlements Total Urban Rural Total Urban Rural Urban Rural Total Urban Rural Total Vojvodina 2031992 1152295 879697 29057 12593 16464 62 45 107 13 13 26 Belgrade 1576124 1281801 294323 19191 16040 3151 67 17 84 17 5 22 West Serbia 835225 322919 512306 6294 1345 4949 18 26 44 1 3 4 Central Serbia 1301656 636412 665244 7320 5446 1874 36 37 73 4 2 6 East Serbia 694905 326326 368579 8452 6924 1528 17 20 37 7 1 8 South-East Serbia 1058099 506143 551956 37879 24415 13464 29 26 55 27 13 40 Total 7498001 4225896 3272105 108193 66763 41430 229 171 400 69 37 106 Sampling Frame and Selection of Clusters The 2002 Serbian Population Census framework was used for the selection of clusters. Census enu- meration areas (app. 100 households) were de- fined as primary sampling units (PSUs), and were selected from each of the sampling domains by using systematic pps (probability proportional to size) sampling procedures, based on the estimated sizes of the enumeration areas from the 2002 Population Census. The first stage of sampling was thus completed by selecting the required number of enumeration areas from each of the 6 regions by urban and rural areas separately. Listing Activities Since the sample frame (the 2002 Population Census) was not up to date, household lists in all selected enumeration areas were updated prior to the selection of households. For this purpose, listing teams were formed, who visited each enu- meration area, and listed the occupied households. The Statistical Office of the Republic of Serbia and The Strategic Marketing Research Agency were responsible for updating household lists. The list- ing exercise was performed by teams which were the direct implementers of the field work during the course of data collection that came later. The whole territory of Serbia was divided into 18 dis- tricts according to the regional network of institu- tions responsible for listing and fieldwork. In each district a team of people was selected – one super- Allocation of the total sample size to the six re- gions was targeted with probability proportional to the regions’ size. Therefore, 400 Serbia exclud- ing Roma, plus 106 Roma sample clusters were allocated across the regions, with the final sample size calculated at 9108 households [(400+106) clusters x 18 households per cluster]. In each re- gion, the clusters (primary sampling units) were distributed to urban and rural domains, propor- tional to the size of the urban and rural popula- tions in that region. The table below shows the allocation of clusters to the sampling domains. 72 MICS3 FULL TECHNICAL REPORT visor for the district and the interviewers (whose number depended on the number of clusters in the region). Criteria for the selection of the inter- viewers and supervisors were their qualifications, communication skills, experience in fieldwork and knowledge of the region where research was to be conducted. A total of 47 teams were formed. For each team, the list of all households in the selected cluster from the last census was provided. The interviewers’ task was to go to the addresses listed and to mark any change that had happened, e.g. the dwelling didn’t exist any more, the household had moved away from the dwelling and another household was living there, and to note the num- ber of children under five living in the house- hold. The listing process was performed during September 2005. Besides providing updated infor- mation on households, updating household lists made interviewers more acquainted with the field. Selection of Households Lists of households were prepared by the listing teams in the field for each enumeration area. The households were then sequentially numbered from 1 to n (the total number of households in each enumeration area) at the Strategic Marketing Research Agency. Selection of 18 plus 3 back-up households with equal probability in each enu- meration area was carried out using the method of random start and equal random walk (simula- tion of the SRSWoR scheme). In the case of the Serbia without Roma from Roma settlements sample, before the selection of households, up- dated census block units were put into two cat- egories: households with children and households without children under 5. Calculation of Sample Weights The Serbia Multiple Indicator Cluster Survey sample is not self-weighted. In order to report the results at the national level sample weights were used. The use of sample weights was needed for the following reasons: � one sample was created by merging two samples (Serbia without Roma in Roma settlements and Roma in Roma settlements) � stratification of each of the separate samples by region and type of settlement. For the Serbia without Roma in Roma settlements sample, two more strata were included – households with under-5 children and households without un- der-5 children. Calculated sample weights were used in the subsequent analyses of the survey data. The major component of weight for both samples is the reciprocal value of the sampling fraction employed in selecting the number of sample households in that particular sampling domain: Wh = 1 / fh (3) The term fh, the sampling fraction at the h–th stratum, is the product of probabilities of selec- tion at every stage in each sampling domain: fh = P1h * P2h (4) where Pih is the probability of selection of the sampling unit in the i–th stage for the h–th sampling domain. Since the estimated numbers of households per enumeration area prior to the first stage selec- tion (selection of primary sampling units) and the updated number of households per enumera- tion area were different, individual sampling fractions for households in each enumeration area (cluster) were calculated. The sampling fractions for households in each enumeration area (cluster) therefore included the probability of selection of the enumeration area in that par- ticular sampling domain, and the probability of selection of a household in the sample enumera- tion area (cluster). A second component which has to be taken into account in the calculation of sample weights is the level of non-response for household and indi- vidual interviews. The adjustment for household non-response is equal to the inverse value of: 73MONITORING THE SITUATION OF CHILDREN AND WOMEN RR = Number of interviewed households / Number of occupied households listed (5) After completion of the fieldwork, response rates were calculated for each sampling domain. These were used to adjust the sample weights calculated for each cluster. Response rates in the Serbia Multiple Indicator Cluster Survey are shown in Table HH.1 in this report. Similarly, the adjustment for non-response at the individual level (women and under-5 children) is equal to the inverse value of: RR = Completed women’s (or under-5’s) questionnaires / Eligible women (or under-5’s) (6) The numbers of eligible women and under-5 children were obtained from the household listing in the Household Questionnaire in households where interviews were completed. The unadjusted weights for the households were calculated by multiplying the above factors for each enumeration area. These weights were then standardised (or normalised), one purpose of which is to make the sum of the interviewed sam- ple units equal to the total sample size at national level. Normalisation is performed by multiplying the aforementioned unadjusted weights by the ratio of the number of completed households to the total unadjusted weighted number of house- holds. A similar standardisation procedure was followed in obtaining standardised weights for the women’s and under-5s questionnaires. Adjusted (normalised) weights varied between 0.36 and 1.5 in the 400 enumeration areas (clusters) in Serbia without Roma from Roma settlements, and be- tween 0.39 and 3.47 in the 106 clusters among Roma settlements in Serbia. For merged Serbia and Roma samples, additional customisation of calculated weights was performed. Since according to the 2002 Census, the proportion of Roma households in the total Serbia household population is 1 percent, the proportion of women aged 15 to 49 is 1.3 percent and the proportion of children under 5 is 3.4 percent; the final weights are products of normalised weights and the ratio of pro- portion of both the Roma and non-Roma population. Sample weights were appended to all data sets and analyses were performed by weighting each house- hold, woman and under-5 with t
Looking for other reproductive health publications?
The Supplies Information Database (SID) is an online reference library with more than 2000 records on the status of reproductive health supplies. The library includes studies, assessments and other publications dating back to 1986, many of which are no longer available even in their country of origin. Explore the database here.