Skip to main content

Table 2 Median chance-corrected cause-specific mortality fraction accuracy for InSilicoVA and Tariff 2.0

From: Performance of InSilicoVA for assigning causes of death to verbal autopsies: multisite validation study using clinical diagnostic gold standards

   InSilicoVA(default probbase) InSilicoVA(InterVA training) InSilicoVA(Tariff 2.0 training) Tariff 2.0
   Median 95% UI Median 95% UI Median 95% UI Median 95% UI
Adult No HCE −59.4 (−61.7, −57.7) 4.3 (2.6, 5.4) 2.1 (0.5, 3.9) 23.1 (21.6, 24.3)
  HCE −40.1 (−41.3, −38.7) 12.1 (10.5, 13.4) 13.9 (12.6, 15.5) 37.6 (36.5, 38.9)
Child No HCE −46.2 (−48.4, −43.6) −10.4 (−13.5, −7.3) 22.3 (20.7, 23.9) 30.5 (28.4, 32.4)
  HCE −42.7 (− 47.9, −37.4) −11.5 (−13.1, −8.2) 22.4 (20.6, 23.8) 41.1 (39.2, 42.0)
Neonate No HCE −39.9 (−43.8, −32.1) 10.9 (4.9, 15.1) 37.6 (33.7, 40.8) 49.2 (47.4, 52.2)
  HCE −38.5 (−43.9, −34.0) 17.5 (12.8, 22.9) 34.4 (30.9, 37.3) 53.1 (50.9, 55.1)
  1. Table 2 shows the population-level performance as the median value and uncertainty interval (UI) across 500 test-train splits using different probbase matrices for prediction, by age group, with and without health care experience (HCE) questions included. InSilicoVA was run without training using the default probbase, with an empirical probbase derived from training data mapped to the InterVA format, and with an empirical probbase derived from training data mapped to the Tariff 2.0 format. Previously published Tariff 2.0 results are shown for comparison