Includes bibliographical references. / Background and purpose: Groote Schuur Hospital (GSH) admits about 570 stroke patients annually. These patients about occupy 10% of acute medical beds. Recent investigations have demonstrated that stroke services are poorly organised Although "Organised stroke care" has been shown to improve outcomes, this has not been demonstrated locally. This study was undertaken to determine whether stroke unit care within a general medical ward improves outcomes. Study design and sampling: The study was a prospective non-randomised controlled trial, with systematic allocation by admission day. of all acute stroke admissions to the Department of Medicine from 1 January to 15 May 1996. Intervention: There were three comparison groups: in the Stroke Intervention Ward, the intervention was implemented by the author; in the Guidelines Ward, the manner in which the intervention can be set up and implemented was provided in the form of a Guidelines Booklet and in the Control Ward, stroke patients received routine care. The intervention consisted of (i) geographic/spatial unity and allocation of a designated nurse; (ii) use of a Stroke Clerking Pro forma and (iii) a multidisciplinary Team Care Plan and Post Intake Stroke Ward Round. Results: 58 patients were admitted to the Stroke Intervention Ward, 40 patients were admitted to the Guidelines Ward and 91 were admitted to the 2 Control Wards. The groups had similar baseline characteristics, except for the percentage of patients continent on admission. There were no significant differences in the Modified Barthel Index prior to admission, at discharge or at follow-up. There were no significant differences in the principal outcome measures (death, dependency, death or dependency, institutionalisation and death or institutionalisation) between the comparison groups. The median [inter-quartile range] length of hospital stay in days was significantly reduced in the Stroke Intervention Ward (6.5[5-9]) compared to the Guidelines Ward (9[7-14]) and the Control Wards (8[5-12]). The referral rates to Professionals Allied to Medicine were significantly increased in the Stroke Intervention Ward, except for referrals to the Community Liaison Sister. The referral rates to rehabilitation resources on discharge were significantly increased in the Stroke Intervention Ward, except for referrals to physiotherapy. Conclusions: Organised Stroke Care is feasible in our setting and results in improved utilisation of resources without increasing length of stay. However, principle outcome measures were not significantly altered in this study.

Includes bibliographical references. / Treatment for drug-resistant tuberculosis (DR-TB) is longer and associated with more significant side-effects than drug susceptible TB. Second line injectable therapy using kanamycin, amikacin or capreomycin is associated with irreversible hearing loss. There is a scarcity of literature regarding the frequency of hearing loss as well as associated risk factors, particularly with long term use. This study aimed to determine the incidence and risk factors for hearing loss among patients receiving second line injectable drugs.

Background: Recent international guidelines call for universal use of triple-drug antiretroviral therapy (ART) in HIV-infected women during pregnancy and postpartum. There are however concerns regarding potentially high levels of non-adherence and/or loss to follow-up (LTF) that may attenuate the benefits of ART for HIV transmission and maternal health. We investigated missed visits and LTF among women initiating ART during pregnancy in Cape Town, South Africa. Methodology: A retrospective cohort study was conducted of women starting ART between January 2011 and September 2012, at a large primary care antenatal clinic. Eligible women were identified in prevention of mother-to-child transmission (PMTCT) services based on CD4 ≤350 cells/μl, and women initiated a regimen of tenofovir, lamivudine and efavirenz. Women eligible for ART were either referred to general adult ART services nearby (January-December 2011) or received ART integrated into ANC services (January-September 2012). Outcomes were measured up to six months postpartum: (i) LTF (no attendance within 56 days of a scheduled visit) and (ii) missed visit (returning to care 14-56 days late for a scheduled visit).

Background and literature review: Introduction: Acute upper gastrointestinal bleeding is a common cause of emergency hospital admission. It is also not uncommon in patients already in hospital. The bulk of severe morbidity and mortality occurs in patients with recurrent bleeding or significant comorbid illness. The use non-steroidal anti-inflammatory drugs (NSAID), which is more common in the elderly, more than doubles the mortality associated with peptic ulcer complications. Endoscopy and endotherapy have improved outcome in patients with continued or recurrent bleeding. Clinical Predictors of Outcome: Despite improved technology in the management of upper gastrointestinal bleeding (UGIB), mortality has remained high. This has been attributed to the increase in the population of elderly people who tend to have other underlying diseases leading to the high mortality rate. From international literature, mortality varies from 4-10%. Most of the reports reflect mortality around 8%. A local retrospective study by Van Stiegmann et. al. (1983) on patients with bleeding peptic ulcer had an overall mortality of 5.4 %, and mortality of 9.8% in those treated surgically. Ideally to avoid waste of resources and time, emergency endoscopy should primarily be considered in relatively few patients at risk of further haemorrhage. In this class of patients, endoscopy is used not only for diagnosis but also for endoscopic treatment to control massive ulcer bleeding. This approach is worthy of consideration in a "resource poor" environment. Most previous studies have employed endoscopic criteria for identifying patients at high risk of poor outcome. These data are not available at the time of admission. This is therefore not helpful in the triage of patients prior to endoscopy. Groote Schuur Hospital (location for the current study) has experienced a significant staff cutback over the past 5 years. Few of our regional hospitals and none of our primary health care facilities have endoscopy facilities and endoscopic expertise is similarly limited. Given the pressures on the limited endoscopy services, there is a need for guidelines for the referral of patients to centres with these facilities. There is a possibility that the number of potential referrals of patients with upper gastrointestinal bleeding (UGIB) will increase as the primary health care services improve, a rational review of criteria for referral appeared timely. There would be need to triage patients into: i). those who must have endoscopy urgently and ii). those who could be safely referred for elective endoscopy. Most previous studies have looked at predictors of adverse outcome. To our knowledge no study has been done in South Africa to determine clinical predictors of good outcome prior to endoscopy, that would help determine patients who would recover uneventfully without urgent endoscopy. Study Rationale: Given the scarcity of both endoscopy resources and of information for the triage of patients who would recover uneventfully following UGIB without urgent endoscopy in our setting, this study set out to answer the question: Is it possible to identify clinical criteria that will predict patients with UGIB in whom endoscopy could be safely deferred? OBJECTIVES: 1. To identify clinical criteria that predict uneventful recovery. 2. To determine the accuracy of individual symptoms and signs or combinations. of symptoms or signs at the time of presentation in predicting uneventful recovery. Study population and Methods: Study design and study setting: The design was a prospective, descriptive, cross sectional study with an analytical component. The study was conducted at Groote Schuur Hospital (GSH), which is a tertiary referral teaching hospital with 14 70 beds. It is affiliated to the University of Cape Town.

Includes bibliographical references (leaves 79-85). / Objectives: to determine the prevalence of atopic eczema in a Southern African setting using both a modified version of the UK working party diagnostic criteria for atopic eczema and a clinical assessment by a dermatologist, and to measure the validity of the UK criteria.

Includes bibliographical references. / Sub-Saharan Africa holds 90% of the HIV-infected children worldwide and most of them are infected through vertical transmission. The elimination of mother-to-child transmission of HIV in this region can be achieved through complementing prophylaxis regimens with effective service delivery. The latter should involve reaching all those at risk and optimizing adherence through adequate and routine follow-up. A study set out to assess the effectiveness of preventing mother-to-child transmission (PMTCT) programs in four African countries including South Africa using the simple nevirapine-based PMTCT regimen, found that prophylaxis coverage for the PMTCT of HIV was on average poor, at only 50% among HIV-exposed infants during 2007-2008. Incomplete HIV testing in pregnancy accounted for 46% of missed opportunities for PMTCT intervention. In addition, discrepancies were found between data from cord-blood samples, which are the gold standard measure of ingestion of the prophylaxis by mothers and routinely collected data on the provision of prophylaxis at antenatal clinics. Clinic records overestimated adherence to prophylaxis which could mislead decisions about service delivery. Adherence to the simple nevirapine regimen, data and service quality should be investigated in order to identify needs for strengthening the effectiveness of WHO Option B guidelines which are being rolled out in resource-poor settings. This project set out to assess the extent of clinic-level PMTCT prophylaxis coverage in the resource-limited setting of the Free State province. Adherence to treatment as well as accuracy of clinic records so as to inform better service implementation were measured. A total of 1572 mother-infant pairs were included in a cross-sectional survey carried out in rural antenatal and delivery services from two health districts between 2007 and 2008. HIV testing and nevirapine prophylaxis data were collected by nurses and compared to anonymously linked cord-blood tests which confirmed true HIV status and ingestion of nevirapine. Logistic regression was used to assess variables from the clinic surveillance data: age, gravidity, mode of delivery, timing of HIV testing and number of antenatal visits during pregnancy, as potential predictors of adherence to nevirapine. The Kappa statistic revealed a disagreement of 10% in HIV test results and 20% in nevirapine intake between clinic records and cord-blood. The clinic records under-estimated maternal HIV prevalence by 9% (22% records versus 31% cord-blood) in mothers aged between 12 and 43 years. Also, cord-blood surveillance revealed that 19.4% of the HIV positive women recorded to have been offered nevirapine during labor, did not actually ingest it. The more frequently a woman sought antenatal care during pregnancy, the more likely she was to ingest nevirapine. Women who had at least 4 antenatal visits were 4.5 times more likely to adhere than women who attended services only once. Even though this is the simplest regimen for preventing vertical HIV transmission, compliance to the entire antenatal cascade during pregnancy is important in improving adherence to therapy and preventing missed opportunities for intervention. Inaccurate collection of routine clinic data negatively impacts on routinely reported data outcomes. Major improvements need to be undertaken both at the service provider and user levels in order to ensure that the effectiveness of the new Option-B guidelines are optimal in remote settings.

Summary: There is strong evidence suggesting that diabetes mellitus (DM) triples the risk of tuberculosis (TB) disease and worsens TB outcomes. South Africa carries a heavy burden of TB which is primarily driven by the human deficiency virus (HIV). The burden of non-communicable disease is also growing rapidly in South Africa. There is however lack of up to date data on the burden of DM and the associated risk factors among TB patients. This dissertation is based on a cross-sectional study which sought to assess the prevalence of DM and impaired glucose tolerance (IGT) and determine the risk factors associated with DM among TB patients. Methods: This cross sectional study forms part of a case control study that aimed to assess the association between DM and TB and the population attributable risk of TB due to DM in Khayelitsha, a high HIV and TB setting. The TB patients recruited in the case control study formed the population of this current cross-sectional study. Based on oral glucose tolerance test, fasting blood glucose, glycated haemoglobin and self-report the prevalence of DM was determined. Bivariate and multivariate logistic regression analyses were performed to assess risk factors associated DM among TB patients. Due to significant differences between male and females with respect to various characteristics, we also stratified the data by sex during analysis.

Includes bibliographical references (leaves 73-79). / With a growing burden of disease and disability stemming from mental disorders in South Africa, additional research into this area may be justified, with specific reference to depression. The effects of parental influences have been explored in the literature as a relevant and valid risk factor in the development and presence of depression. The literature has suggested two main underlying themes of parental styles that included parental overprotection and care. The concept of overprotection has been associated with depression. This thesis explored these constructs in their association with depression whilst highlighting the influence of control. Furthermore the influences of parental depression were also included.

Background: Antiretroviral therapy (ART) services have expanded over the past decade, providing treatment to over 15 million people globally. It is imperative that this scale-up of ART provision be accompanied by optimal treatment response monitoring strategies to timely and accurately detect treatment failure. Routine viral load (VL) monitoring is the preferred ART response monitoring tool and its use has been increasing across Africa; however, there are few insights into VL monitoring practices during pregnancy. This thesis describes public sector VL testing practices in a cohort of HIV-infected pregnant women who initiated ART before pregnancy in Cape Town, South Africa. Methods: This study was conceived in 2015 as a sub-analysis of the first phase of an on-going prospective trial: the Strategies to optimize antiretroviral therapy services for maternal and child health (MCH-ART) Study, being conducted in Gugulethu, Cape Town, South Africa. Consecutive HIV-infected pregnant women on ART before pregnancy and making their first visit to a primary care antenatal clinic between March 2013 and June 2014 were enrolled into the study. Pre-existing demographic, obstetric and ART history data collected during enrolment into the MCH-ART study were used. In addition, HIV VL results were obtained from the National Health Laboratory Service (NHLS) system from 15 months prior to the estimated date of conception to delivery. VL testing and VL results were described for the two periods: (i) before conception (from estimated date of conception to 15 months prior) and (ii) during pregnancy (from estimated date of conception to delivery). Results: Among 520 women the median age was 31 years [Interquartile range (IQR), 28-35 years] and the median duration of ART use was 2.7 years [IQR, 1.5-5.1 years]. Before conception, 66% (n=311) of women had at least one VL test done in routine adult ART services, and 9% of these results (n=29) were >1000 copies/mL. During the pregnancy, 80% (n=415) of women had at least one VL test done and 12% (n=49) of these results were >1000 copies/mL. Pregnant women with elevated VL >1000 copies/mL were more likely to have been on ART for longer (p=0.049), report at least 2 missed ART doses in the preceding 30 days (p=0.043) and be on a protease inhibitorbased regimen (p=0.016). Among women with VL >1000 copies/mL during pregnancy, 59% (n=29) had a repeat VL done at a median of 3.5 months after the initial test (IQR, 2.1-4.4 months) with 52% (n=15) of these women having a VL >1000 copies/mL on this second test. Conclusion: While coverage of VL monitoring appears high in this setting, a substantial fraction of women with elevated VL in pregnancy were never retested. With increasing numbers of HIV positive women using ART, greater attention is needed to design and implement effective strategies for VL monitoring in pregnancy.

Includes bibliographical references. / Background: The Western Cape Province of South Africa has the highest reported rates of Foetal Alcohol Syndrome (FAS) in the world. Reported statistics on FAS in this province show that 40.5 to 46.4 per 1000 children aged 5 to 9 years have FAS compared to developed nations that reported 0.5 to 2 cases per 1000 births. The loss in human potential is immeasurable and various studies have shown that the financial cost is formidable. Each child affected by FAS may require an estimated $1 million to $2 million over the course of their lifetime to support remedial medical, educational and social costs. Primary prevention programmes targeted to women at risk of alcohol-exposed pregnancies could lead to measurable reductions in the incidence of FAS. An alcohol-exposed pregnancy (AEP) is a pregnancy that results when a sexually active woman is not on effective contraception and is involved in risky drinking. Aim: To evaluate the effectiveness of a training intervention to improve screening, identification and management of women at risk of alcohol exposed pregnancies.