Employment preferences of healthcare workers in South Africa: Findings from a discrete choice experiment

There is a maldistribution of human resources for health globally, with many Lower- and Middle-Income Countries experiencing significant shortages. We examined healthcare workers' job preferences in South Africa to identify factors which potentially influence employment decisions. A discrete choice experiment was conducted among 855 South African healthcare workers critical to its national HIV testing and treatment programs. Job characteristics included workload, workplace culture, availability of equipment, training opportunities, sector and facility type, location, salary and benefits. Main effects analysis was conducted using fixed effects logistic regression. Interaction effects identified divergence in preferences. Heavy workload (OR = 0.78; 95% C.I. 0.74-0.83), poor workplace culture (odds ratio 0.66; 95% C.I. 0.62-0.69), insufficient availability of equipment (OR = 0.67; 95% C.I. 0.63-0.70) and infrequent training opportunities (OR = 0.75; 95% C.I. 0.71-0.80) had large, significant effects on worker preferences. An increase in salary of 20% (OR = 1.29; 95% C.I. 1.16-1.44) had a positive effect on preferences, while a salary decrease of 20% (OR = 0.55; 95% C.I. 0.49-0.60) had a strong negative effect. Benefits packages had large positive effects on preferences: respondents were twice as likely to choose a job that included medical aid, pension and housing contributions worth 40% of salary (OR = 2.06; 95% C.I. 1.87-2.26), holding all else constant. Although salary was important across all cadres, benefits packages had larger effects on job preferences than equivalent salary increases. Improving working conditions is critical to attracting and retaining appropriate health cadres responsible for the country's HIV services, especially in the public sector and underserved, often rural, communities. Crucially, our evidence suggests that factors amenable to improvement such as workplace conditions and remuneration packages have a greater influence on healthcare workers employment decisions than employment sector or location.