Performance of ML algorithm compared with risk-based screening
| Approach | Proportion of the overall patient population to be screened by each approach | Proportion of correctly identified undiagnosed HCV cases (precision) |
| To find 20.9% of undiagnosed HCV cases | ||
| R-PWID cohort | 3.70% | 0.12% |
| ML algorithm | 1.80% | 0.40% |
| To find 48.4% of undiagnosed HCV cases | ||
| Birth cohort | 34.50% | 0.03% |
| ML algorithm | 12.30% | 0.12% |
HCV, hepatitis C virus; ML, machine learning; R-PWID, risk of being a person who injects drugs.