Introduction: Sub-Saharan Africa remains the region worst affected by the HIV/AIDS pandemic in the world. South Africa (SA) is the country with the highest population of people living with HIV/AIDS in the world and Mpumalanga province is the province with the second highest prevalence of HIV/AIDS in SA.
The district of Gert Sibande has the highest prevalence in the province of Mpumalanga (38.9%) in 2006. Since many patients living with HIV/AIDS usually present to district hospitals as the first point of contact it is important to understand the implications of HIV/AIDS in a resource limited health system.
Study setting: The setting for this study was Amajuba Memorial Hospital (AMH) a district hospital in the Gert Sibande district of the Mpumalanga province.
Objective: To analyse direct costs of providing inpatient care to adult patients with HIV/AIDS-related illnesses at AMH from the perspective of the provider (hospital)
Study methods: The population of study comprised adult patients with HIV/AIDS-related illnesses admitted to the medical wards during the period of October 2009 and March 2010 at AMH. A detailed retrospective record review of patients admitted to the adult wards at AMH with HIV/AIDS-related illnesses over a 6-month period was conducted.
After the record review the costs were estimated using standard costs and utilisation. Demographic and clinical patient profiles were determined then descriptive statistics were calculated with total costs as an outcome variable. Subsequently univariate and multivariate regression analysis were performed.
Results: The demographic and clinical profiles revealed that most patients admitted with HIV/AIDS-related illnesses were: between the ages of 39 & 49 years (35.3%), male (54.9%), urban residents (82.0%), unemployed (87.2%), single (80.5%), were not on HAART (70.7%), had CD4 counts between 0 & 50 x 106 /L (38.3%), had pulmonary tuberculosis (PTB) (38.4%), were admitted for the first time (60.9%) and of the total admitted to hospital 79.0% survived the index admission during the study period.
Descriptive statistics of the continuous data variables were determined. Minimums, maximums, inter-quartile ratios, means and modes were determined and tabulated.
Consultation costs followed by investigation costs were the two major contributors to total admission costs (77.7% of the median total admission cost). Univariate analysis revealed these significant associations with total admission costs: admission diagnosis, discharge diagnosis, first admission, outcome, pre-admission consults and preceding admissions.
In multivariate regression, admission diagnosis and pre-admission consults were analysed. Significant associations were found between the following categories: retroviral disease versus other diseases (p=0.001), retroviral disease versus anaemia (p=0.035), no pre-admission consults versus 1 pre-admission consult (p=0.007), no pre-admission consult versus 4 pre-admission consults (p=0.039) and no pre-admission consult versus 5 or more pre-admission consults (p=0.006).
Conclusion: In our study we successfully determined demographic and clinical profiles of patients admitted with HIV-related illnesses at AMH. Emerging from the results of our study were patterns of burden of HIV disease, health seeking behaviour and risky sexual behaviour that all had implications for admission costs in the hospital. Major cost drivers were consultation and investigation costs, which were increased significantly by disease categories; other diseases, anaemia and PTB. Pre-admission consults emerged as a cost reducing parameter in our study.