Strategies to Improve Kaposi Sarcoma (KS) Outcomes in Zimbabwe (SIKO)

The "Strategies to Improve Kaposi Sarcoma (KS) Outcomes in Zimbabwe" (SIKO Study) will consist of both an interventional and an observational component. Eight (8) primary care sites in Zimbabwe will be randomized to receive structured training on Kaposi sarcoma (KS) diagnosis and treatment. This intervention will be evaluated through the utilization of three components; the KS Standardized Evaluation (KS-SE), integration of palliative care and an algorithm-based KS management strategy at the sites. The impact of the overall intervention will be evaluated using a step-wedge randomized cluster trial design in which the 8 primary care sites will be randomized to receive the intervention at different time points such that the intervention will be eventually be implemented at all sites during the 2 year (102 weeks) course of the study. The 2 year evaluation period will consist of a monitoring period, followed by an intervention period.

The observational component of the SIKO study will be the enrollment of all patients who are found to have Kaposi sarcoma (KS) during either the monitoring or intervention periods at any of eight study sites. After informed consent is obtained, information of HIV status, KS diagnosis and staging, demographic data and a quality of life questionnaire will be collected, along with A glycoprotein predominantly found on the surface of helper T cells. (CD4+) counts (if available), antiretroviral (ARV) medication and tuberculosis and KS treatment history. Patients will continue to be followed over the course of the study.

For power estimations we assumed each site would enroll between 0·5 to 0·7 subjects per week, resulting in final sample sizes of 158 to 221 in the Pre-Intervention Period and 318 to 445 in the Intervention Period. To evaluate the effect of the Intervention Package on T0 (T0/(T0 + T1)) we assumed T0 proportions of 0·12 to 0·15 during the Pre-Intervention Period, within site correlations of 0·10, 0·15 and 0·20, and a maximum enrollment of 100 per site. We estimated 80% power to detect a minimum increase of the T0 proportion from 0·12 to 0·37 or from 0·15 to 0·40. Power estimation for evaluation of time to loss-to-care assumed a Kaplan-Meier approach and provided 80% power to detect a hazard ratio (HR) ≥ 1·89 with a total sample of up to 512 KS cases. Sample sizes were calculated using PASS, (NCSS software, Kayesville UT), assuming a 0·05 two-sided significance level.