Includes bibliographical references. / Background: Food allergy (FA) is a significant and often life-threatening health problem affecting about 4-6% of children and their families globally. In some developed countries FA prevalence has reached 10% and it is believed that developing economies may follow a similar trend since there is a reported rise in the global burden of other allergic diseases like asthma, allergic rhinitis and eczema. However, there is a dearth of population studies at global level documenting challenge-proven Ig-E mediated food allergy (FA) prevalence. As such, we studied an unselected population of children attending crèches in Cape Town, South Africa. Methodology All children aged 12-36 months attending the selected crèches between February 2013 and October 2013 were eligible for the study. Participants were assessed with an allergy questionnaire, had skin prick tests (SPTs) done and if they qualified, were invited for an oral food challenge (OFC) at the Red Cross Hospital Paediatric Allergy Clinic (RCHPAC). The SPT wheal size results were categorised into ≥1mm, ≥3mm and ≥7mm. We gave a general description of the study sample with respect to the demographic characteristics and compared participants and non-participants. We reported sensitisation pattern towards foods in the panel i.e. egg white extract, peanut, cow’s milk, wheat(flour), soy, hazelnut and fish (cod) according to the SPT categories. The effects of age, ethnicity, sex and concomitant allergy on sensitisation patterns were assessed. Associations between the potential predictor variables and sensitisation were assessed by Z-test for proportions and Chi-square/Fisher’s exact. PART I presents the study protocol with a brief motivation for the relevance of the study and the methodology used. PART II presents a structured literature review on FA and FS in large populations of selected and unselected cohorts. It provides an overview of empirical evidence on prevalence estimates from both the developed and developing world, and the potential risk factors causing Fav. PART III summarises the methodology, results and interpretation of the analysis conducted in a journal-ready manuscript according to Current Allergy and Clinical Immunology Journal requirements. Results The sample consisted of; 39% black African, 20% Caucasian and 41% mixed race participants, with a median age 26 months (IQR: 22-31). Amongst 121 participants (66% response rate, 92% participation rate and 94% completion rate), the prevalence of SPT≥1mm to any food was 16%, SPT≥3mm 12% and SPT≥7mm 4%. The prevalence of challenge-proven Ig-E mediated raw egg allergy was 1.7% and peanut allergy 0.8%. Black African participants had higher sensitisation rates (23%) of SPT≥1mm to any food, when compared to Caucasian (13%) and mixed race (10%) participants despite the difference not reaching statistical significance (p=0.17). Conclusions: This study was acceptable and feasible in this population that has a low prevalence of Ig-E mediated FA that is comparable to other studies from developed countries using objective measures in unselected cohorts. The prevalence of FS is appreciably high in this sample and there are ethnic differences that require further investigation. The findings seem to suggest an existing burden of Ig-E mediated FAs in the South African context that is un-diagnosed and therefore not managed

Background Sentinel antenatal clinic (ANC) surveys remain a key source of data on HIV prevalence trends. Recently though, with an increase in the uptake of prevention of mother-to-child transmission (PMTCT) programmes, there have been debates on the prospects of using PMTCT data for reporting antenatal HIV prevalence and trends. Aim To describe the HIV prevalence trends for the Western Cape Province and to compare prevalence from ANC surveys to PMTCT programmes. Methods HIV prevalence and 95% confidence intervals were estimated from ANC surveys from 2001-2012 for the province as well as the 6 health districts and the 8 City of Cape Town Metropolitan sub-districts in the province. HIV prevalence from expanded provincial ANC survey sampling was compared to the nationally reported provincial and district estimates, before and after re-weighting to account for differences between the realized sample and updated sampling frame. A regression line was fitted with calendar year included as both a linear and quadratic term to create smoothed trend lines of the change in HIV prevalence over time by province, district, sub-district and age group. A multivariable logistic regression model was fitted to the multi-year ANC survey data to explore associations with HIV prevalence. ANC survey HIV prevalence estimates were compared to those from routinely reported HIV testing data from the PMTCT program for 2009-2012.

Introduction: Historically in South Africa, civil society has played a key role within the health system, including advocating for equitable and quality health care services. The purpose of this research is to explore the implementation of a pilot health systems strengthening intervention in primary health facilities in Khayelitsha, South Africa. The study is built on Treatment Action Campaign, a civil society organization, which has recently implemented a health system monitoring tool within health care facilities in Khayelitsha. Specifically, this study considers the functioning and potential impact of the monitoring tool introduced as a community accountability mechanism at the local level. The development and implementation of the monitoring tool can also be seen as part of a policy implementation process. Methods: Using an action research approach, the researcher engaged with implementing actors in the development and implementation of the monitoring tool. Qualitative methods were used to explore: the understandings of various stakeholders about the tool, their interests or concerns, potential positions, power and influence on its implementation. Quantitative data allowed for the ability to track potential improvements in clinic performance in terms of operational research. The challenges during tool development and implementation and how these were overcome were also explored. Results: Analysis of the stakeholders demonstrated how actors exerted their power in various ways to influence the development and implementation of the tool. Results suggest it can be an empowering process for members of civil society and there is a role for civil society in improving health system performance. Findings have highlighted the need for civil society organization monitoring tools to be not only methodologically sound but, more importantly, accepted by the activist. If carefully considered and driven by civil society itself, rather than imposed, there does seem to be some tentative examples of service delivery improvement and scope for their engagement. Conclusions: The findings offer relevant and useful insights for understanding how this tool acts as an accountability mechanism at a local level within Khayelitsha sub-district. Such findings may have implications for further adaptations to the tool, potential scale-up by Treatment Action Campaign and for other low and middle income contexts.

Child stunting remains one of the biggest public health concerns in Zambia and other low and middle-income countries (LMICs). A formidable challenge faced in improving child health outcomes in LMICs includes persistent socioeconomic and residential disparities. Despite achieving an overall decline in the prevalence of child stunting over the past decades, children residing in rural areas and less-privileged households continue to fall behind their peers from urban areas and wealthier households in Zambia and other LMICs. Notably, studies have shown that children residing in rural areas and less privileged households have a higher risk and burden of stunted growth in sub-Saharan Africa (SSA). However, basic rural-urban differentiation in child stunting can potentially conceal wealth differentials that exist within rural and urban areas. Specifically, cross country analyses have revealed that wealth differentials were higher in urban areas compared to rural areas; and higher than the overall urban-rural odds of stunting among children under five years of age. Using data from the 2013/14 Zambia Demographic Health Survey (ZDHS), differences in the relationship between socioeconomic status and child stunting in urban and rural areas of Zambia were assessed in this study. Furthermore, the study examines the effect of socioeconomic status and residence type in predicting child stunting prevalence in Zambia. To achieve these, the thesis used chi-square tests and logistic regression analysis. To the best of my knowledge, this is the first single-country analysis primarily focused on Zambia that has disaggregated the effect of predictors of child stunting by residence type. It is anticipated that the results of this dissertation will broaden the knowledge-base on wealth and residential differentials in child nutritional outcomes in Africa and thereby provide useful information to policymakers and technocrats in Zambia. Overall, the findings indicate that children under five years who reside in urban areas and poorer households have a higher likelihood of becoming stunted compared to their peers in rural and wealthier households. However, the relationship between child stunting and household wealth (SES) differs slightly after segregating by residence type. In both rural and urban areas, there is a consistent inverse relationship between the odds of stunted growth among under-fives and SES. Furthermore, these findings indicate that socioeconomic differentials are wider in rural areas compared to urban areas and much wider than the overall rural-urban odds ratios in Zambia. These findings could possibly be because of socioeconomic inequalities in child stunting that are higher in rural areas than urban areas. However, there is a need for further research to examine the causes of differentials in child stunting that may exist in rural and urban locations of Zambia.

Includes bibliographical references. / This qualitative study was conducted in Cape Town, South Africa in 2010. It examines the effectiveness of promotional educational pamphlets on the awareness, understandings and practice of the right to health among eight civil society organizations and their constituents.

Includes bibliographical references. / To develop an understanding of the factors that promotes and inhibits child sexual abuse (CSA) disclosure and its impact on the child and caregiver. Young girls and adolescents between the ages of 8-17 years who experienced penetrative sexual abuse were recruited from two sexual assault centres in the Western Cape, South Africa. On arrival at the centre, caregivers were approached and informed about the study and informed consent was obtained from those who agreed to participate in the study. This study has shown that factors promoting CSA disclosure include circumstances around the incident, caregiver concerns around the child's behavior, which made caregivers suspicious and allowed them to create an enabling environment to facilitate disclosure for the child as well as coercion by caregivers to talk.

Background: Tuberculosis is a major cause of morbidity and mortality as an estimated 8.6 million people developed TB and 1.3 million died from the disease in 2012. The number of deaths is high given that TB can be prevented. Health care workers are an at - risk group, since they are frequently in contact with infectious patients and/or work with infectious products. The World Health Organisation has declared the importance of finding innovative tools and strategies to prevent TB and implementing them successfully, especially for those with a high risk of TB transmission. Methods: This systematic review aims to undertake a quantitative review of tuberculosis interventions for health care workers in health care settings, so as to assess whether these interventions are effective in reducing the transmission of tuberculosis infection and disease. We will preferably include experimental studies, such as, randomised - controlled trials, but observational studies, such as controlled before and after studies and cohort studies will also be included in the absence of randomised - controlled studies. We will search databases, such as Medline, Scopus, Trip, LILACS and various trial registries. A hand search of reference lists of identified articles, abstracts, conference proceedings and campaign materials will be performed. Grey literature sites will also be used for the search. Data will be extracted using a single form. The quality of each study will be assessed in terms of selection bias, performance bias, attrition bias and detection bias. Thereafter a meta - analysis will be produced and subgroups will be analysed according to the three intervention types. Clinical and statistical significance will be determined for the included studies, and descriptive narratives of heterogeneous studies will be written. Discussion: Our results will be useful to policy - makers and public health officials for the prioritisation of those interventions identified as effective and critical .

Background: Task-shifting of care and first-line treatment for people living with HIV (PLHIV) from doctors to nurses is feasible, effective and acceptable in sub-Saharan Africa. Since first-line resistance to antiretroviral therapy (ART) is increasing, comprehensive HIV care should be informed by viral load (VL) monitoring to ensure timely detection of treatment failure and switch to second-line ART if appropriate. Patients identified with a first elevated VL (FEVL) enter a cycle with extra counselling sessions, VL re-testing, and potential regimen switch, putting additional strain on scarce human resources. However, identification of treatment failure and its management remains largely doctor-led and there are limited results from programmes with nurse-led management of treatment failure. Health service-delivery in the Democratic Republic of Congo (DRC) is challenging in a weak health system, pressured by epidemics and political unrest. Despite low HIV prevalence in the capital Kinshasa (1.6%), many PLHIV present with advanced disease and early mortality is high. HIV care and treatment, VL testing and second-line switch in decentralised facilities in Kinshasa is exclusively managed by nurses, but results have so far not been documented. Methods Patient outcome data in three primary care facilities in Kinshasa, were routinely collected since ART introduction in 2002 and entered in the national Tier.net database. A protocol (Section A: Study protocol) was developed with detailed methods and procedures for a retrospective analysis of patient outcomes after having had a FEVL (≥1000 copies/ml). The protocol includes analysis of predictors for favourable outcomes (i.e. retained and with VL re-suppressed at 12 months after first high VL or administratively censored or transferred with suppressed VL), and analysis of compliance to existing protocols. The protocol received approval by ethics boards of the Ministry of Health of the DRC and the University of Cape Town. A structured literature review was conducted (Section B: Literature review), critically appraising peer-reviewed publications describing outcomes of PLHIV on ART after first-line failure in sub-Saharan Africa under programmatic conditions. The review included studies where treatment failure and switch were managed by medical doctors, due to paucity of data of nurse-led management of treatment failure. Predictors for outcomes after failure and switch to secondline were also extracted from available literature. A journal-ready manuscript (Section C: Manuscript) presents the findings of the analysis conducted on patients with a FEVL in three decentralised nurse-led facilities in Kinshasa, and predictors for favourable outcomes. Results Of 294 adults with FEVL who did not switch to second-line before confirmatory VL, 82% had a second VL (VL2) done within 24 months of FEVL at a median (interquartile range [IQR]) of 4.0 (3.1-5.6) months) after FEVL. Among patients with VL2 done, 69% had VL2 ≥1000copies/ml, of whom 75% switched to second-line a median of 1.1 (IQR, 0.7-2.0) months after VL2. Among the 85% of patients who were not deceased, LTFU or transferred out by 6 months after second-line switch, 82% had VL6 versus ≤3 months after FEVL and switching 1-3 versus ≤1 month after VL2 ≥ 1000copies/ml were independently associated with lower odds of a favourable outcome. Conclusion Exclusively nurse-managed detection of virologic failure and switch to second-line ART in decentralised health facilities yielded acceptable outcomes in our cohort in urban Kinshasa. Early detection and fast switch can help improve retention and viral suppression following virologic failure. Task-shifting along the viral load cascade is a feasible and safe strategy in settings with limited human resources and growing viral resistance.

Includes bibliographical references (leaves 135-147). / This thesis presents the findings of a research on the economic burden of malaria morbidity to rural households in Mtoko district of North-East Zimbabwe. The main objective of this study was to ascertain the household level impacts of direct costs (medical costs, consultation costs, transport costs and other related costs) and indirect costs (lost productive time by malaria sufferers whilst sick, lost time by caretakers whilst caring for the sick) due to malaria sickness. A cross sectional study with both descriptive and analytical features was carried out and the main finding from the research was that the economic costs of seeking malaria care were regressive. In other words the poor were using a higher percentage of their income whilst seeking malaria care. In addition, access to care was very limited for the poor as they either could not afford to access the care because of prohibitive costs or they were geographically too far away from sources of care to easily access it. Furthermore, indirect costs were far higher than direct costs as they constituted a greater percentage of total malaria costs. It was concluded that measures meant to exempt the poor from paying for malaria treatment and care were needed to limit the economic burden of malaria morbidity on poor households. The need to ensure that cheap affordable malaria drugs were available to the affected rural people is imperative.

Bibliography: leaves 105-112. / Many developing countries have or are in the process of implementing decentralisation of financial management systems. The reform can take various forms: self-financing or cost recovery through user charges and co-financing or expansion of local revenue through taxation. This study examines the impact of decentralisation of fiscal financial management of cost sharing revenue on hospital efficiency. Potentially, decentralisation of financial management of cost sharing is expected to improve coverage and accessibility of health services, quality of services and efficiency in the delivery of health care. The main aim of this study was to review the impact of fiscal decentralisation or the cost sharing reform on the efficiency of Kenya's health care delivery and to identify factors that need to be addressed in order to enhance the success of the reform policy.