A RESEARCH THESIS SUBMITTED TO THE FACULTY OF HUMANITIES, UNIVERSITY OF THE WITWATERSRAND, JOHANNESBURG, IN FULFILMENT OF THE REQUIREMENTS FOR THE AWARD OF THE DEGREE OF DOCTOR OF PHILOSOPHY IN DEMOGRAPHY AND POPULATION STUDIES. October 2013. / The burden of disease and mortality in South Africa is a challenge. In the area of HIV/AIDS, an estimated 316,900 new infections to persons aged 15 and older were reported in 2011. For persons below the age of 15 years, 63,600 new infections were detected in this year (Stats SA, 2011). South Africa is also doing poorly with regard to other development indicators. The infant mortality rate for this year was estimated at 37.9 deaths per 100,000 live births and under- five mortality is 54.3 deaths per 100,000. With adolescents in South Africa constituting approximately 20% of the total population, a substantial percentage of adolescents could then be orphans. Adolescents in contemporary South Africa are affected by HIV/AIDS, teen pregnancy, domestic violence and other reproductive health issues (Pettifor et al., 2005; Kaufman et al., 2001; King et al., 2004). They are also faced with poverty and unemployment (Ngoma, 2005). Given the overall optimistic forecast for South Africa’s economic and political growth, based on the end of Apartheid, these grave challenges that youth face persist, making South Africa a very interesting study area to examine the determinants of other-health related and mortality risks for adolescents. The main objective of this study is to identify the levels, causes and determinants of adolescent mortality in South Africa. More specifically the study examines a host of natural, unnatural and broad-underlying causes of death. In addition, individual, household and community, demographic and socioeconomic characteristics are tested as determinants of adolescent mortality. This study is a cross - sectional study using nationally representative data from the South African Census 2001; the Community Survey 2007 and Death Notification Forms as collated by Statistics SA. The Community Survey is nationally representative and collects the same information as the Census. Thus demographic, socioeconomic and mortality questions are similar in both sources and are suitable for a nationally representative study on mortality trends. Death Notification Forms articulate immediate and broad- underlying causes of death, this is useful in identifying specific risks. Adolescents aged 10 to 19 years old are covered in this study. Both adolescent males and females are studied. A total of 41,261 adolescent deaths were recorded in 2001 and 54,046 adolescent deaths were noted for 2007. Demographic and socioeconomic indicators are included for all adolescents represented in the surveys. This study is a quantitative study of adolescent mortality in South Africa. STATA 11(SE) was for the management and analysis of data in this study. The methods of analysis for this study follow the order of the study objectives. To determine levels of adolescent mortality, frequency distributions, age and sex specific mortality rates and proportional mortality ratios were used. In identifying and evaluating prevalent causes of death age- standardized death rates by cause and cause- specific mortality rates are initially used. Then multiple and associated decrement life tables ( in particular, probability of dying, survival function, life expectancy and cause- deleted life expectancy) and Years of Potential Life Lost (YPLL) are calculated. To identify determinants of adolescent mortality logistic regression and multilevel logistic regression is applied. Among other results, the main findings of this study are first the levels of adolescent mortality in South Africa increased from 2001 to 2007 by approximately 1.3%. Second among adolescents, the levels of male mortality (21,686) exceed that of their female (19,575) counterparts in 2001, however the reverse is true of 2007, where female adolescent mortality (28,517) exceeds that of males (25,529). This particular finding is unusual since male mortality during youth (15- 24 years) is generally higher among males. This is due to the increased burden of disease on females in the country, which is now affecting adolescent females too. This is shown in results that the see plight of Tuberculosis- related deaths is more concentrated among females than males. In addition, the probability of adolescent females dying from Tuberculosis increased for 1.45 in 2001 to 1.75 in 2007 in South Africa. In addition, this study shows that for specific causes of adolescent mortality the elimination of causes of death such as Tuberculosis and pneumonia could produce substantial gains in life expectancy. Females have consistently lower odds of adolescent mortality from unnatural causes of death compared to males. Finally, having a few household assets, 6 or more people living in a residence, and high ethnic diversity within the community is associated with increased odds of adolescent mortality in South Africa in 2001. A main conclusion of this study is that with adolescent mortality in the country increasing future economic growth is compromised. Adolescents will soon enter the labour force and contribute to the country’s growth. With fewer adolescent’s surviving to adulthood this means the country’s economic growth with be slow. In addition, increasing adolescent mortality compromises the economic support available to dependent populations. With fewer economically active adults, the country’s dependency ratio will be high, thus increasing the burden placed on government to provide for the elderly and children.