A research report submitted to the Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, in partial fulfilment of the requirements for the degree of Master of Science in the Field of Population-Based Field Epidemiology / Socio-economic status (SES) disparities do not only exist between racial groups in South Africa but also exists within the vulnerable black population with the devastating impacts of the HIV epidemic. Households are important determinants of human welfare. However, little is known about the effect of household socio-economic status on the establishment and break-up of households within a low-resource setting and a severe HIV epidemic. It is in the midst of these challenges in rural South Africa that this study examined the effect of household SES on household formation and dissolution among the black population in rural northern KwaZulu-Natal. METHODS Using longitudinal data from the period 2003-2012 from the Africa Centre for Health and Population Studies, the study used a cross-sectional study design approach to examine the effect of household SES on household formation. It also examined the effect of household SES change (i.e. either positive, negative change or stable SES) between the start and end of observation of a household within the study period. Household formation was defined as when an individual or individuals come from different households to form a new social unit with a new household head. Dissolution occurred when all individuals in a household end their membership to a household due to death, out-migration or by joining other households. Separate regression models for the two outcomes, household formation and dissolution were explored with household SES covariates while adjusting for other household variables. RESULTS Household formation and dissolution trends both decreased over the study period. Out of a total of 18,249 households, newly formed households had a relatively higher percentage of tertiary educated household heads (10.7% versus 2.5%), unemployed household members (41.6% versus 28.5%), grant recipient household members (37.1% versus 8.5 %) and households within the average to richest wealth quintiles (44.1% versus 36.4 %) than pre-existing households. Multivariate analysis showed that tertiary educated household heads (aOR=2.96, 95% (CI) 2.26-3.89) and households within the average to richest wealth quintiles most especially the 4th quintile (aOR=3.29, 95% (CI) 2.69-4.04) were associated with a higher odds of households being newly formed. However, the lesser the employed members (aOR=0.31, 95% (CI) 0.21-0.45) and grant recipients per household size in a household (aOR=0.15, 95% (CI) 0.12-0.18) the lower the odds of formation. Furthermore, small size households (aOR=0.68, 95% (CI) 0.56-0.80) and unmarried household heads (aOR =0.47, 95% (CI) 0.40-0.55) were associated with lower odds of being newly formed. Whereas female headed households (aOR=2.23, 95% (CI) 1.93-2.57) were associated with a higher odds of household formation. With regards to household dissolution, close to a quarter of households had an increase in SES over the study period compared to households with a decreased SES (24.6% versus 8.6 %). Similar to household formation, male headed households dominated the study population with the highest proportion in dissolved households (63.8% and 61.5% at start and end of household observation respectively). Also unmarried household heads were the majority in dissolved households (62.7% and 64.1% at start and end of household observation respectively). Approximately 65.6% of households that never dissolved had an extended family type of composition compared to 36.6% of dissolved households. The area was predominantly rural with about 47.2% households in rural segment of the study area. The study has shown that households had lower odds of dissolving if there is a positive change (i.e. an increase) in household SES compared with households with an unchanged SES over the period. In exact terms, an increment in the number of employed household members over the study period was associated 49% lower odds of a household being a dissolved (aOR=0.51 95% (CI) 0.42-0.61). Also, an increment in the number of household grant recipients over the period of observation was associated with a 69% lower odds to result in the dissolution (aOR=0.31 95% (CI) 0.25-0.39). Households with an improved wealth index over the period of study were associated with 55% lower odds of dissolution (aOR =0.45, 95% (CI) 0.38-0.54). However, households with both male and female death (multiple sex) were more likely to dissolve. Similarly, peri-urban (aOR=0.71; 95% (CI) 0.58-0.86) households were more likely to dissolve compared to urban households. Surprisingly divorced, widowed and separated couples were not significantly associated with household dissolution. CONCLUSION SES is an important determinant of household existence and stability. This study has shown a complex relationship between household SES and household formation. Although education and improved household wealth index were more likely to result in household formation, an increase in the number of employed household members and household grant recipients did not necessary have an effect on household formation. Government cash transfers, education, employment of household members are valuable cushioning mechanisms necessary for household stability. There is need for government and non-governmental organisations to set up interventions to improve the socio-economic conditions of poor households prioritising rural and female headed households. This is especially critical in a high HIV prevalence area where these interventions will also mitigate against the burden of the HIV epidemic on the population. / MT2017

Includes bibliographical references / This thesis uses the National Income Dynamics Study (NIDS) data to estimate the extent of, and the factors correlated with, socio economic related health inequalities in South Africa. We extend our analysis by investigating whether income has a causal effect on health outcomes. The thesis is divided into four separate, but related chapters. In chapter two, we describe the data and the variables used in the study. We then check the quality of health related data in the NIDS by analyzing attrition trends and establishing whether attrition affects the representativeness of the data in subsequent waves. We use three health outcomes, self-assessed health, body mass index and depression, to test for the potential effects of attrition bias on parameter estimates. We test using the attrition probit and Becketti, Gould, Lillard and Welch (BGLW) tests, which are two well-known tests for attrition bias in panel data. We find that although the attrition rates of individuals from the sample are high in wave 2 and 3 (21% and 20% respectively), their attrition is random with respect to the health outcomes we use. In chapter three, we establish the socioeconomic factors correlated with health outcomes in South Africa. We use bivariate and panel data approaches. We find significant correlations between health outcomes and socioeconomic factors (income, educational attainment, and demographic factors). Income is positively correlated with self-assessed health and body mass index, and it is negatively correlated with depressive symptoms. In chapter four, we build on the findings discussed in chapter three to estimate the extent of Income Related Health Inequality (IRHI). We estimate the index of inequality using a health concentration index. We then decompose the concentration index to establish the extent to which the correlates of health outcome drive the IRHI. The panel nature of the data allows us to investigate whether IRHI is narrowing or widening. We find a positive health concentration index. This implies that better health is concentrated among the rich. The decomposition of the index reveals that these differences are explained by disparities in income and educational attainment. We also find that the IRHI has narrowed from 2008 to 2012. Most of the narrowing is unexplained but about 21% and 20% of the decrease is correlated with the changes in the distribution and response to covariates respectively. One of the socioeconomic determinants identified from the previous chapters to be correlated with health is income. In the last part of this thesis, we extend the analysis to investigate whether this relationship is causal. To do so, we use the Old Age Pension (OAP) programme as a natural experiment. The OAP is based on age eligibility. Therefore, we use this age eligibility as an exogenous income shock to isolate the effect of income on health. We apply a Regression Discontinuity Design on the NIDS data to identify this effect. We do not find any contemporaneous effect of income on three health outcomes considered, namely; self assessed health (SAH), body mass index (BMI), and depression.

No description available.

No description available.

Background: This study examines the socioeconomic factors associated with infant and postneonatal mortality in Kenya and tries to quantify these associations in order to put those factors in ranked order so as to prioritize them in health policy plans aiming to decrease infant and postneonatal mortality. The study has used wealth index, mother’s highest educational level, mother’s occupation and place of residence as exposures of interest. Methods: The study uses analytical cross-sectional design through secondary data analysis of the 2003 Kenyan Demographic and Health Survey (KDHS) dataset for children. Series of logistic regression models were fitted to select the significant factors both in urban and rural areas and for infant and postneonatal mortality, separately, through the use of backward stepwise technique. Then the magnitude of the significance for each variable was tested using the Wald’s test, and hence the factors were ranked ordered according to their overall P-value. Results: After excluding non-singleton births and children born less than one year before the survey, a sample size of 4 495 live births was analyzed with 458 infants died before the first year of life giving IMR of 79.6 deaths per 1000 live births. After adjusting for all biodemographic and other health outcome determining factors, the analyses show no significant association between socioeconomic factors and infant mortality in both urban and rural Kenya. The exclusion of deaths that occurred in the first month of ages shows that risk of postneonatal (OR 3.09; CI: 1.29 – 7.42) mortality, in urban Kenya, were significantly higher for women working in agricultural sector than nonworking women. While in rural Kenya, the risk of postneonatal (OR 0.42; CI: 0.20 – 0.90) mortality were significantly lower for mothers with secondary school level of education than mothers with no education. Conclusions: There is lack of socioeconomic differentials in infant mortality in both urban and rural Kenya. However, breastfeeding, ethnicity and gender of the child in urban areas on one hand and breastfeeding, ethnicity and fertility factors on the other hand are the main predictors of mortality in this age group. Furthermore, results for postneonatal mortality show that level of maternal education is the single most important socioeconomic determinant of postneonatal mortality in urban Kenya while mother’s occupation is the single most important socioeconomic determinant of postneonatal mortality in rural areas. Other determinants of postneonatal mortality are ethnicity and gender of the child in urban areas, while in rural areas; the other main predictors are ethnicity, breast feeding and fertility factors.

The research report is submitted to the School of Public Health, Faculty of Health Sciences, University of Witwatersrand, in partial fulfilment of the requirements for the Master of Public Health degree. Johannesburg March, 2014 / Introduction: This is the descriptive study of the socio demographic profile and other characteristics such as the burn injury details and socio economic characteristics of adult burn injury patients treated at Johannesburg Tertiary Hospitals (JTH) during the study period. Relevant stakeholders can use this information in the efforts to reduce preventable burn injuries. Method: Prospective study where all adult burn patients in the burns unit, trauma/surgical wards during the study period were eligible to participate in the study. The information was extracted from the medical files (such as hospital classification, date of birth (DOB), type of burn, type of management done to date etc), this was followed by an interview done by principal investigator only, using a questionnaire to gather the information on patients’ demographic details, socio economic information, income details and burn injury details. Descriptive statistics were used to define the profile of burn patients and other characteristics. Results: The results revealed the description and the profile of adult burn patients: a male (71%), African (94%), unmarried (70%), mean age of 35.6 years. He was most likely to have a secondary school qualification (62%), full time employed possible (51%) in the industrial sector, stays with 2-5 people in his household. He was likely to be originally from outside the Gauteng Province (58%). He sustained burns injuries of 10-29% TBSA, while at home (94%), from flames (68%). He remembered (92%), his activity prior to the burn incident and thought that the burn could have been prevented (82%). Conclusion: Burns injuries were reported to be preventable. The burn injury-prevention program should be targeted to males, in the working age groups, residing in one roomed dwelling or informal settlements. Patients with poor judgement, predisposing medical conditions such as epileptics, those that have modified their electricity supply and heat sources should be prioritised for burn injury-prevention programs.

Research report submitted to the Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, in partial fulfilment of the requirements for the degree of Master of Science (MSc) in Epidemiology in the field of Epidemiology and Biostatistics / Obesity is an emerging problem in South Africa, particularly in women for whom prevalence rates well above 40% have been reported. Parallel to this health problem, South Africa continues to experience relatively high poverty levels of 10.5% to 48.0%. The aim of this study was to estimate the prevalence of obesity and low social economic status (SES) levels at Agincourt Health and Socio-Demographic Surveillance System site (AHDSS). The study also sought to investigate the association between low SES and obesity at AHDSS. Materials and methods This was a secondary data analysis of the original Na Nakekela HIV/Non communicable disease (NCD) study conducted at AHDSS from August 2010 to May 2011. Included in the study presented in this report were residents of AHDSS aged 15 years or older during this time period. Data from 4 502 individuals (2 683 females and 1 819 males) were analysed. Age-specific prevalences of obesity (body mass index ≥ 30kg/m2), and central obesity (waist hip ratio ≥1.0 and ≥0.85 in men and women, respectively), stratified by sex and SES, were calculated. SES was assessed by ascertaining the household assets of AHDSS residents and assigning a weighted score to the household assets, using multiple correspondence analysis (MCA). The household score was then computed and used to classify the population into SES categories. The relative ranks of households, using this score, were then used as a measure of SES. The association between SES and obesity (BMI ≥ 30) was assessed by means of chi-square tests and logistic regression. Results The overall prevalence of obesity at the AHDSS in the study period was 20.4%. Overall, sex -specific prevalences of obesity were 29.3% and 7.4% in females and males, respectively. Females aged 50-59 years and males aged 45-49 years had the highest age-specific prevalence of obesity, at 40.1% and 18.3%, respectively. The overall prevalence of central obesity was 31.1%. Sex-specific prevalence of central obesity in females was 51.1%, while in males it was 4.9%. The highest age-specific prevalence of central obesity in both sexes was for those 70 years and older: 74.3% in females and 11.1% in males. Around 50% of individuals at the AHDSS were classified as belonging to lower SES categories, with females constituting 56.6% of these individuals. The highest prevalence of individuals in the high SES category was females aged 60-69 (14.5%) and males aged 70 (16.4%) years and older. After adjusting for other variables, being in a lower SES category was inversely associated with obesity as measured by BMI, as was being male and being HIV positive. The only positive predictor of high BMI was older age. No association between central obesity and lower SES was found after adjusting for confounders and other explanatory variables. However, older age was a predictor of central obesity. Being male, HIV positive and the male head of the household were factors that were inversely associated with central obesity. Discussion The high prevalence of individuals in the lower SES group (50.5%) reported in this study is similar to the Mpumalanga provincial poverty estimate of 51%.The ratio of obese females to males was at least 2.2 in every age group. The prevalence of central obesity in females of 51.1% in the AHDSS was higher than the national estimate of 47.1% for females, while the male estimate of 4.9% was lower than the 6.8% national estimate for males. In contrast to other studies, no associations between lower SES and obesity as measured by central obesity were observed. Conclusion and Recommendations Specific interventions to reduce obesity in females should be undertaken, including the provision of educational talks. This would empower them to make better informed decisions about food and lifestyle choices. These recommendations should be integrated into already existing HIV prevention programmes because HIV prevention is currently the main focus of policy makers in South Africa. Measures to reduce the number of individuals in the lower SES group, which this study reported to be very high (especially among women), e.g. through job creation, should be considered.

Thesis (Ph. D.)--University of Washington, 2003. / Vita. Includes bibliographical references (leaves 69-103).

No description available.

Little is currently known about the spatial-temporal dynamics of dengue epidemics in arid areas. This study assesses dengue outbreaks that occurred in two arid cities of Mexico, Hermosillo and Navojoa, located in northern state of Sonora. Laboratory confirmed dengue cases from Hermosillo (N=2730) and Navojoa (N=493) were geocoded by residence and assigned neighborhood-level characteristics from the 2010 Mexican census. Kernel density and Space-time cluster analysis was performed to detect high density areas and space-time clusters of dengue. Ordinary Least Square regression was used to assess the changing socioeconomic characteristics of cases over the course of the outbreaks. Both cities exhibited contiguous patterns of space-time clustering. Initial areas of dissemination were characterized in both cities by high population density, high percentage of occupied houses, and lack of healthcare. Future research and control efforts in these regions should consider these space-time and socioeconomic patterns.