Magister Public Health - MPH / Background: Multidrug resistant tuberculosis (MDR TB) is a growing threat globally. The large increase in the incidence and prevalence of MDR TB in South Africa in recent years has impacted on the way in which MDR TB is managed within the health services. It became logistically difficult to manage MDR TB by treating all patients as in-patients in a specialized tuberculosis (TB) hospital. The clinics, which are run by nurses and/or general medical officers, are then required to manage this more complex form of TB, with limited resources, less experience and assumingly with less MDR TB knowledge. Of particular concern is that shifting of the patient management from specialized TB hospitals to Primary Health Care clinics which might worsen the already poor MDR TB treatment outcomes. There has been minimal assessment of the management of MDR TB at clinic level and hence the comparison of treatment outcomes for those patients initiated on treatment in clinics compared to in-patients in specialized TB hospitals is urgently needed. Aim: To compare the treatment outcomes and the effectiveness of medication regimens provided to MDR TB patients initiated on treatment in specialized TB hospitals as inpatients, to that of MDR TB patients initiated on treatment as outpatients at community clinics within the Western Cape Province, South Africa. Methodology Study Design: A retrospective cohort study was undertaken, as the length of treatment for a MDR TB patient can be for 24 months or longer and this study was based on treatment outcome data. Study Population and sample: The study population was uncomplicated MDR TB patients initiated on treatment in hospitals and clinics from January 2010 to December 2012. The sample comprised of 568 participants that were laboratory confirmed to have MDR TB and had the outcomes of their treatment recorded in an electronic database or a paper register. Data Collection: The researcher collected MDR TB information from standardized MDR TB registers as well as an electronic MDR TB database. Analysis: Data was analyzed comparing the exposed (clinic initiated) and unexposed (hospital initiated) cohorts incidence of 4 key treatment outcomes, namely: successfully treated, failed treatment, died and defaulted treatment. Bivariate analysis (relative and absolute) was done to determine the cumulative incidence ratio and cumulative incidence difference and multivariate logistic regression analysis for the adjusted odds ratio to control for confounders and effect modifiers. Ethics: Permission to conduct this research was obtained from the relevant authorities. The confidentiality of the participants as per the Department of Health policy and in adherence to general ethical guidelines was strictly maintained. The study proposal received ethical clearance and approval from the University of the Western Cape Research Committee. Results: All participants within this study received the appropriate treatment as per the MDR TB guidelines. The incidence rate for the main outcomes of this study indicated that successfully treated for the clinic initiated participants was 41% and 31% for the hospital initiated participants. ‘Defaulted’ treatment was 39% and 41%, ‘failed’ treatment 7% and 13% and ‘died’ was 14% and 16%, respectively. The clinic initiated participants appeared to have better treatment outcomes on bivariate analysis, however on multivariate analysis, there was no difference in the treatment outcomes of the clinic initiated participants compared to the hospital initiated participants, and therefore the clinic initiated treatment is seen as effective. The time to treatment initiation for clinic and hospital initiated participants is excessively long for both cohorts, with a median of 29 days, and 37 days respectively. The key findings of note in the multivariate analysis is that the Human Immunodeficiency Virus positive (HIV+) participants provided with antiretrovirals therapy (ART) were, based on adjusted cumulative incidence ratios, 6.6 times more likely to have a successfully treated outcome (95% CI 1.48-29.84), and were 0.2 times less likely to die (95% CI 0.08-0.53). Having a previous cured history of TB and no previous history of TB were 2.9 times more likely to have a successfully treated outcome (95% CI 1.48-5.56) and were 0.1 times (0.04-0.38) less likely to fail treatment. An interesting finding was that participants living in the rural districts were 2.6 times more likely to die. Conclusion: Clinic initiated treatment for uncomplicated MDR TB is as effective as hospital initiated treatment. Also, those provided with ART and those without previous TB or who had a previous bout of TB cured, had better outcomes. Main Recommendations: The Western Cape health department should continue with the decentralization of MDR TB services to the clinics and could safely consider expanding the decentralization to include uncomplicated Preextensively drug-resistant TB and Extensively drug-resistant TB patients. Offering ART to HIV+ patients should be mandatory. The delays in the time to treatment initiation of MDR TB need to be further investigated.