Table 5.
AUC (bCI) | Sens (%) | Screen (%) | Prev (%) | NNSnew | ||
Hindustani Surinamese | 0.58 (0.49–0.70) | |||||
Total population | 6.0 | |||||
≥ 5 points | 94.4 | 97.3 | 6.6 | 16 | ||
≥ 8 points | 94.4 | 80.7 | 8.0 | 13 | ||
≥ 11 points | 38.9 | 42.0 | 6.3 | 16 | ||
African Surinamese | 0.79 (0.70–0.89) | |||||
Total population | 3.1 | |||||
≥ 5 points | 93.8 | 86.8 | 3.3 | 31 | ||
≥ 8 points | 87.5 | 58.9 | 4.5 | 22 | ||
≥ 11 points | 81.3 | 25.8 | 9.6 | 11 | ||
Ethnic Dutch | 0.77 (0.68–0.85) | |||||
Total population | 3.1 | |||||
≥ 5 points | 100 | 56.2 | 5.5 | 19 | ||
≥ 8 points | 64.3 | 24.9 | 8.0 | 13 | ||
≥ 11 points | 14.3 | 6.4 | 6.9 | 15 |
a Persons with known DM were excluded from the analysis. Simplified score is presented in additional file 1: Risk score for DM SUNSET.pdf. Cut off values (number of points) chosen based on the 20–50–75 percentiles in the total population.
DM = diabetes mellitus (fasting plasma glucose ≥ 7.0 mmol/l and/or self-reported), sens = sensitivity of the criteria at the given cut-off, screen = percentage of the total population that is screened at the given cut-off, prev = prevalence in the screened population (the predictive value positive), NNSnew = number needed to identify a case of newly detected DM, i.e. a previously undiagnosed case of DM in this study, AUC = area under the curve, a measure of diagnostic accuracy, bCI = 'confidence interval' estimated via a bootstrapping procedure [20], the 2.5 and 97.5 percentiles are listed between brackets.