Effectiveness of return-of-service schemes for human resources for health retention: a retrospective cohort study of four Southern African countries.

Background: Governments use return-of-service (RoS) schemes to train, employ and retain health professionals in the public sector. We determined the effectiveness of RoS schemes in four Southern African countries.

Methods: This retrospective cohort study used databases of RoS beneficiaries from South Africa, Botswana, Eswatini and Lesotho. We ascertained the period of funding for beneficiaries between 2000 and 2010, study programme, selection criteria, study country and if they completed their studies. Records were sought to track beneficiaries' service and fulfilment of their RoS obligations. Data were sought at the provincial level in South Africa and nationally for the other three countries. Binomial logistics regression and Kaplan-Meier survival estimates were used to determine risk factors and predictors of defaulting.

Results: Most beneficiary enrolment (eg, decision on why they were funded, socioeconomic status, disability status, high school results) and service data (eg, health facilities where they worked, how long they worked at each health facility, movement between health facilities) were not available. A total of 5616 beneficiaries were drawn from the four countries' databases. Of those with full data available, 21.7% (229/1056) were retained/served beyond their obligatory period and 20.2% (213/1056) were still serving. A total of 24.3% (95% CI: 21.7% to 26.9%; n=257/1056) of beneficiaries in the final subanalysis of two South African provinces fulfilled their contractual obligations. Only 32.2% (277/861) of beneficiaries undertook internship within their funding provinces. Governments needed to fund six beneficiaries to have one beneficiary complete their contractual obligation if they undertook internship outside their province.

Conclusion: Record keeping in all countries was poor, hampering the effectiveness of RoS schemes. Of the units with full data available, the retention rate was below 25%, and internship being undertaken outside the funding province was associated with higher defaulter rates, calling for a policy overhaul.