Skip to main content

Table 2 Apparent and test performance of the PONV models described in Table 1

From: Prediction models for clustered data: comparison of a random intercept and standard regression model

  Apparent performance Test performance
  Standard model Marginal risk calculation Conditional risk calculation Standard model Marginal risk calculation
Harrell’s C-index † 0.66 (0.014) 0.66 (0.014) 0.69 (0.013) 0.68 (0.014) 0.67 (0.014)
C-index within clusters ‡ 0.63 (0.089) 0.62 (0.097) 0.62 (0.097) 0.72 (0.129) 0.70 (0.127)
Calibration intercept* 0.00 0.01 0.00 0.13 0.14
Calibration intercept* within clusters ‡ 0.01 (0.329) -0.00 (0.385) 0.00 (0.000) 0.10 (0.339) 0.11 (0.380)
Calibration slope 1 0.95 1.08 1.08 0.99
Calibration slope within clusters ‡ 0.97 (0.189) 0.94 (0.251) 1.08 (0.000) 1.06 (0.060) 1.00 (0.027)
  1. † overall performance (standard error).
  2. ‡ within anesthesiologist performance (standard deviation).
  3. * With calibration slope equal to 1 (i.e. calibration in the large).