Economic evaluation of safety-engineered devices and training in reducing needlestick injuries among healthcare workers in South Africa

Abstract

Background. Healthcare workers (HCWs) are at increased risk of contracting various communicable diseases. Needlestick injuries (NSIs) are a common mechanism of exposure. Training in basic universal precautions and utilisation of safety-engineered devices (SEDs) are interventions known to reduce the risk of NSI.

Objectives. To assess the cost-utility of SEDs v. a training programme in universal precautions (TP) v. a combination strategy to reduce NSIs among South African HCWs.

Methods. A Markov model comparing SEDs v. a TP v. a combination strategy against current practice was developed. A hypothetical cohort of HCWs working in the SA public sector was followed from a payer’s perspective for a period of 45 years, and discounted costs and benefits were assessed. Data were obtained from the National Department of Health, suppliers and published literature. One-way and probabilistic sensitivity analysis was conducted.

Results. Over the study time horizon, our model estimated that 2 209, 3 314 and 4 349 needlestick injuries per 1 000 HCWs could be prevented if a TP, SEDs or a combination strategy, respectively, was adopted compared with current practice. All three candidate interventions were cost-effective at a willingness to pay (WTP) of one times the gross domestic product per capita (USD6 483.90/quality-adjusted life-year (QUALY) gained). SEDs as a stand-alone intervention was dominated by a combination strategy. Compared with current practice, the incremental cost-effectiveness of training was USD32.90/QALY v. USD432.32/QALY for SEDs and USD377.08/QALY for a combination strategy. Results were sensitive to the effectiveness of the interventions. Probabilistic sensitivity analysis showed that at a WTP of USD6 483.90/QALY gained, a combination strategy would be cost-effective 95.4% of the time.

Conclusions. A combination strategy in which both SEDs and a TP are adopted is preferred.