Abstract:
BACKGROUND : Occupational tuberculosis (TB) continues to plague the healthcare workforce in South Africa.
A 2-year cluster randomized controlled trial was therefore launched in 27 public hospitals in Free State
province, to better understand how a combined workforce and workplace program can improve health of the
healthcare workforce.
OBJECTIVE : This mid-term evaluation aimed to analyze how well the intervention was being implemented, seek
evidence of impact or harm, and draw lessons.
METHODS : Both intervention and comparison sites had been instructed to conduct bi-annual and issue-based
infection control assessments (when healthcare workers [HCW] are diagnosed with TB) and offer HCWs
confidential TB and HIV counseling and testing, TB treatment and prophylaxis for HIV-positive HCWs.
Intervention sites were additionally instructed to conduct quarterly workplace assessments, and also offer
HCWs HIV treatment at their occupational health units (OHUs). Trends in HCW mortality, sick-time, and
turnover rates (2005 2014) were analyzed from the personnel salary database (‘PERSAL’). Data submitted
by the OHUs were also analyzed. Open-ended questionnaires were then distributed to OHU HCWs and indepth
interviews conducted at 17 of the sites to investigate challenges encountered.
RESULTS : OHUs reported identifying and treating 23 new HCW cases of TB amongst the 1,372 workers
who used the OHU for HIV and/or TB services; 39 new cases of HIV were also identified and 108 known-
HIV-positive HCWs serviced. Although intervention-site workforces used these services significantly more
than comparison-site healthcare staff (pB0.001), the data recorded were incomplete for both the intervention
and comparison OHUs. An overall significant decline in mortality and turnover rates was documented
over this period, but no significant differences between intervention and comparison sites; sick-time data
proved unreliable. Severe OHU workload as well as residual confidentiality concerns prevented the proper
implementation of protocols, especially workplace assessments and data recording. Particularly, the failure to
implement computerized data collection required OHU staff to duplicate their operational data collection
duties by also entering research paper forms. The study was therefore halted pending the implementation of a
computerized system.
CONCLUSIONS : The significant differences in OHU use documented cannot be attributable to the intervention
due to incomplete data reporting; unreliable sick-time data further precluded ascertaining the benefit potentially attributable to the intervention. Computerized data collection is essential to facilitate operational
monitoring while conducting real-world intervention research. The digital divide still requires the attention of
researchers along with overall infrastructural constraints.