Table 3.
Performance of various screening strategies
| Sensitivity TP/(TP + FN) 95% CI |
Specificity TN/(TN + FP) 95% CI |
Negative Predictive Value TN/(TN + FN) |
Negative likelihood ratio Point estimate 95% CI |
Positive Predictive Value TP/(TP + FP) |
Positive likelihood ratio Point estimate 95% CI |
|
|---|---|---|---|---|---|---|
| Venous 75 g OGTTa | 1 18/(18 + 0) (0.82–1) |
1 598/(598 + 0) (0.99–1) |
1 598/(598 + 0) |
n/a | 1 (18/18 + 0) | n/a |
| POC RBG ≥ 7 mmol/L (126.0 mg/dL) | 0.17 3/(3 + 15) (0.06–0.39) |
0.97 581/(581 + 17) (0. 95–98) |
0.97 581/(581 + 15) |
0.86 (0.70–1.1) |
0.15 3/(3 + 17) |
5.86 (1.89–18.23) |
| POC 1 h 50 g ≥ 7.8 mmol/L (140.4 mg/dL) | 0.44 8/(8 + 10) (0.25–0.66) |
0.79 470/(470 + 128) (0.75–0.82) |
0.98 470/(470 + 10) |
0.71 (0.47–1.07) |
0.06 8/(8 + 128) |
2.08 (1.21–3.56) |
| POC 1 h 50 g ≥ 7.2 mmol/L (129.6 mg/dL) | 0.56 10/(10 + 8) (0.34–0.75) |
0.64 382/(382 + 216) (0.60–0.67) |
0.98 382/(382 + 8) |
0.70 (0.41–1.17) |
0.04 10/(10 + 216) |
1.54 (1.00–2.36) |
| Venous FBG ≥ 5.1 (91.8 mg/dL) | 0.78 14/(14 + 4) (0.55–0.91) |
1 598/(598 + 0) (.99–1) |
0.99 598/(598 + 4) |
n/a | 1 14/(14 + 0) |
0.22 (0.09–0.53) |
| POC FBG ≥ 5.1 mmol/L (91.8 mg/dL) | 0.33 6/(6 + 12) (0.16–0.56) |
0.95 567/(567 + 31) (0.93–0.96) |
0.98 567/(567 + 12) |
0.70 (0.51–0.98) |
0.16 6/(6 + 31) |
6.43 (3.08–13.45) |
| POC 75 g OGTT | 0.56 10/(10 + 8) (0.34–0.75) |
0.91 542/(542 + 56) (0.88–0.93) |
0.99 542/(542 + 8) |
0.49 (0.29–0.82) |
0.15 10/(10 + 56) |
5.93 (3.66–9.61) |
| POC HbA1c ≥ 6.5% | 0.22 4/(4 + 14) (0.09–0.45) |
0.99 592/(592 + 6) (0.98–0.99) |
0.98 592/(592 + 14) |
0.79 (0.61–1.01) | 0.4 4/(4 + 6) |
22.15 (6.84–71.72) |
| Venous HbA1c ≥ 6.5% | 0.11 2/(2 + 16) (0.03–0.33) |
0.99 591/(591 + 7) (0.98–0.99) |
0.97 591/(591 + 16) |
0.90 (0.76–1.06) |
0.22 2/(2 + 7) |
9.49 (2.12–42.54) |
TP true positives, FN false negatives, TN true negatives, FP false positives, HbAIc glycated hemoglobin, OGTT oral glucose tolerance test, POC point of care, RBG random blood glucose, FBG fasting blood glucose
aAll results were compared to the venous 75 g oral glucose tolerance test to demonstrate diagnostic accuracy