Table 5.
Performance of the Kunzmann, Wang and CanPredict risk algorithms in women in the QResearch validation cohort overall and by ethnic group.9,10
| Statistic | 5-year risk estimate Kunzmann (OAC) (age 50–84 years) |
5-year risk estimate CanPredict (OAC) (age 50–84 years) |
10-year risk estimate Wang (OSCC) (40–84 year) |
10-year risk estimate CanPredict (OSCC) (40–84 year) |
|---|---|---|---|---|
| Mean (95% CI) | Mean (95% CI) | Mean (95% CI) | Mean (95% CI) | |
| Harrell's C | ||||
| Overall | 0.649 (0.634–0.664) | 0.725 (0.707–0.742) | 0.679 (0.666–0.691) | 0.79 (778–801) |
| White | 0.643 (0.636–0.65) | 0.722 (0.714–0.73) | 0.663 (0.658–0.669) | 0.785 (0.779–0.79) |
| Indian | 0.746 (0.71–0.781) | 0.769 (0.712–0.826) | 0.793 (0.764–0.823) | 0.873 (0.842–0.903) |
| Pakistani | 0.603 (0.504–0.701) | 0.684 (0.495–0.874) | 0.711 (0.645–0.778) | 0.748 (0.602–0.895) |
| Bangladeshi | 0.627 (0.52–0.733) | 0.816 (0.73–0.903) | 0.773 (0.701–0.845) | 0.799 (0.728–0.87) |
| Other Asian | 0.607 (0.521–0.694) | 0.522 (0.461–0.583) | 0.753 (0.706–0.801) | 0.75 (0.694–0.806) |
| Caribbean | 0.626 (0.577–0.674) | 0.82 (0.78–0.859) | 0.78 (0.755–0.804) | 0.817 (0.785–0.849) |
| Black African | 0.779 (0.719–0.839) | 0.81 (0.743–0.878) | 0.731 (0.666–0.796) | 0.815 (0.755–0.875) |
| Chinese | 0.788 (0.714–0.862) | 0.662 (0.527–0.797) | 0.703 (0.607–0.799) | 0.832 (0.74–0.924) |
| Other | 0.786 (0.72700.845) | 0.866 (0.815–0.917) | 0.818 (0.778–0.858) | 0.871 (0.823–0.92) |
| Royston'sD statistic | ||||
| Overall | 0.828 (0.739–0.916) | 1.45 (1.33–1.58) | 1.01 (0.928–1.1) | 1.76 (1.66–1.86) |
| White | 0.798 (0.706–0.89) | 1.45 (1.31–1.58) | 0.942 (0.856–1.03) | 1.71 (1.61–1.81) |
| Indian | 1.35 (.604–2.09) | 2.09 (1.06–3.11) | 1.64 (0.899–2.38) | 2.53 (1.65–3.41) |
| Pakistani | a | a | a | a |
| Bangladeshi | a | a | a | 1.75 (0.00222–3.49) |
| Other Asian | a | a | 1.27 (0.331–2.21) | 1.7 (0.701–2.7) |
| Caribbean | a | 1.79 (0.83–2.76) | 1.41 (0.682–2.14) | 1.8 (0.989–2.61) |
| Black African | a | 2.04 (0.321–3.75) | a | a |
| Chinese | a | a | a | a |
| Other | a | 2.89 (1.13–4.64) | 2.05 (.227–3.87) | 2.97 (1.22–4.71) |
| R2 (%) | ||||
| Overall | 14.1 (11.5–16.6) | 33.6 (29.7–37.4) | 19.7 (17–22.3) | 42.5 (39.8–45.2) |
| White | 13.2 (10.6–15.8) | 33.3 (29.3–37.4) | 17.5 (14.8–20.1) | 41.2 (38.3–44.1) |
| Indian | 30.2 (6.98–53.5) | 50.9 (26.2–75.5) | 39 (17.6–60.5) | 60.3 (43.6–77.1) |
| Pakistani | a | a | a | a |
| Bangladeshi | a | a | a | a |
| Other Asian | a | a | a | 40.9 (12.4–69.4) |
| Caribbean | a | 43.3 (16.9–69.8) | 32.2 (9.79–54.6) | 43.5 (21.4–65.7) |
| Black African | a | 49.5 (5.76–93.2) | a | a |
| Chinese | a | a | a | a |
| Other | a | 65.9 (39.4–92.4) | 48.7 (6.96–90.4) | 67 (43.3–90.7) |
a
Cells with too few observations to be estimated reliably.