AIM The objective of this study is to determine factors that may contribute to hospital admissions in a cohort of medically insured South African patients in the era of HAART. METHODS This was a retrospective cohort of all HIV-positive adult and paediatric patients enrolled on a medical aid disease management programme in South Africa over a period of three years. Patient-specific demographic and clinical information were obtained from the medical aid records. Survival analysis was used to analyse time to first admission looking at admissions occurring after enrolment to the programme, during the study period of between 01 January 2006 and 31 December 2008. Only the right censored cases were included in the analyses. Descriptive analyses were conducted on the key prognostic factors. Variables that were significant in the univariate were considered in the multivariate Cox proportional hazards model. RESULTS A total of 8440 patients were included in the analysis. Half of these patients had at least one admission during the observation periods with 43.28% having had 2 or more admissions. The average admission rate was 2 admissions per patient over the 36 month observation period. Young children, adolescents and the very old (> 60 years) were significantly more likely to be admitted than the middle age groups, HR = 1.30 [95%CI 1.21 -1.40] p<0.01, 1.24 [95%CI 1.10 – 1.41] and 1.13 [95% CI 1.10 – 1.27] p<0.01 respectively. Low CD4 cell counts of < 200 cells/ µL were significantly associated with a higher likelihood of hospitalizations with hazard ratios even greater for CD4 cell counts of less than 100 cells/ µL, HR= 1.34 [95%CI 1.29 – 1.39], p<0.01. Cases were more likely to be admitted by a clinical haematologist or gynaecologist than by other specialist categories.HR =1.58 [95%CI 1.29 –1.94] and 1.17[95%CI 1.08 – 1.27] respectively with p<0.01. CONCLUSION Factors that are associated with hospital admissions in this private sector, medically insured population are a younger and older age, low CD4 cell counts and admission by a clinical haematologist and gynaecologist. These results suggest that disease management strategies should be intensified for the younger and older age groups. All HIV-positive patients should be closely monitored for CD4 deterioration so that treatment is initiated timeously. Routine haematological investigations should be recommended for all HIV-positive patients in order to pick up and treat haematological conditions before they result in a hospital admission. Evidence based guidelines, outlining the place of caesarian section deliveries in the HIV population, should be developed for use by gynaecologists specifically in the private sector. Copyright / Dissertation (MSc)--University of Pretoria, 2011. / Clinical Epidemiology / unrestricted