A clinical utility risk-benefit analysis for HIV self-testing
Johnson et al.
Type of approach
Type of assistance
Fingerstick/whole blood, Oral-fluid
UNAIDS HIV prevalence (2017)
To assess the clinical utility of HIVST and weigh performance-related risks and benefits: sensitivity (65-99.8%), specificity (65-100%), HIV prevalence (0.01-15%), linkage to care (50-85%) and linkage to prevention (0-35%) were considered. Different scenarios characterized by varying levels of the factors listed above were simulated. A net benefit score was derived from Total Benefit (true reactive linked to care; true non-reactive linked to prevention) minus Total Risk (false reactive; false non-reactive). False non-reactives and true reactives linked to care were weighted based on expert consultation. The proportion of scenarios with positive net benefit was calculated. Sub-analysis of high (5-10%) and low (0.1-1%) prevalence scenarios, high (20-30%) and low (0-10%) linkage to prevention; and high (70-80%) and low (50-60%) linkage to care was conducted.
Summary of findings
61% of scenarios with 70% sensitivity and 90% specificity yielded greater benefit than risk. In high-prevalence scenarios, positive net benefit was observed at 80% specificity and 70% sensitivity. For low-prevalence scenarios, net benefit marginally increased when sensitivity increased from 70% to 90%. Linkage to prevention drove net benefit; when high, benefit was achieved at 80% specificity but when low 90% specificity was needed. Linkage to care had modest impact except in very high-prevalence settings, e.g. net benefit in all scenarios was not observed among female sex workers in Johannesburg until linkage to care was 50%.
Willingness to pay
Willingness to pay details
Linkage to prevention, care and treatment