Skip to main content

Table 2 Predictive value of HbA1c

From: Choice of HbA1c threshold for identifying individuals at high risk of type 2 diabetes and implications for diabetes prevention programmes: a cohort study

Threshold (T)

Na

Percentage of participants classified as high-risk ≥ T (n)

n ≥ T who progress to diabetes

Sensitivity

Specificity

PPV

NPV

False positives

False negatives

AUC ROC

HbA1c ≥ 39 mmol/molb (5.7%)

4227

55.8% (n = 2358)

137

95.1% (90.2, 98.0)

45.6% (44.1, 47.1)

5.8% (4.9, 6.8)

99.6% (99.2, 99.8)

54.4% (52.9,55.9)

4.9% (2.0,9.8)

70.4% (68.5,72.3)

HbA1c ≥ 42 mmol/molc (6.0%)

4227

22.0% (n = 929)

115

79.9% (72.4, 86.1)

80.1% (78.8, 81.3)

12.4% (10.3, 14.7)

99.1% (98.7, 99.4)

19.9% (18.7,21.2)

20.1% (13.9,27.6)

80.0% (76.6, 83.3)

HbA1c ≥ 44 mmol/mol (6.2%)

4227

9.0% (n = 380)

88

61.1% (52.6, 69.1)

92.8% (92.0, 93.6)

23.2% (19.0, 27.7)

98.5% (98.1, 98.9)

7.1% (6.4, 8.0)

38.9% (30.9,47.3)

77.0% (73.0, 81.0)

  1. Sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV), false positive rate, false negative rate, AUC ROC with 95% confidence intervals for progression of our cohort to diabetes over follow-up time (mean [SD] 45.0 [18.0] months) given HbA1c thresholds
  2. aN who progress to diabetes in cohort 144, bADA threshold, cIEC threshold