Skip to main content

Table 3 Difference in overall diabetes risk (observed – true) and percent that achieved calibration (H-L χ2 <20) in 500 replications under systematic reporting error (bias) in height and weight for males (N=9,177) and females (10,618)

From: The influence of measurement error on calibration, discrimination, and overall estimation of a risk prediction model

 

Males (N=9,177)

Females (N=10,618)

Reporting bias

Difference in diabetes risk (observed - true)

Number of diabetes cases

Percent that achieved calibration*

Difference in diabetes risk (observed - true)

Number of diabetes cases

Percent that achieved calibration

Underestimate of weight by

      

   0.5 kg

−0.23%

24

95.4%

−0.22%

23

93.0%

   1.0 kg

−0.46%

49

87.2%

−0.44%

47

80.2%

   1.5 kg

−0.69%

73

73.0%

−0.67%

71

58.6%

   2.0 kg

−0.94%

100

49.0%

−0.90%

95

32.2%

   2.5 kg

−1.18%

125

23.2%

−1.13%

120

11.8%

   3.0 kg

−1.43%

152

7.4%

−1.37%

145

2.0%

Overestimate of height by

      

   0.5 cm

−0.22%

23

95.4%

−0.19%

20

93.4%

   1.0 cm

−0.45%

48

87.4%

−0.38%

40

85.6%

   1.5 cm

−0.69%

73

75.0%

−0.57%

61

69.4%

   2.0 cm

−0.92%

98

52.4%

−0.77%

82

48.2%

   2.5 cm

−1.16%

123

25.8%

−0.98%

104

25.8%

   3.0 cm

−1.41%

150

9.4%

−1.19%

126

9.2%

  1. *(H-L χ2 <20).