Background
Methods
Results
Male | Female | All | *p value | |
---|---|---|---|---|
n
| ||||
Age (years) | 45.5 (15.8) | 45.0 (14.5) | 45.2 (15.1) | 0.363 |
FPG (mmol/l) | 4.9 (1.0) | 4.8 (1.2) | 4.9 (1.1) | 0.375 |
2-h PG (mmol/l) | 5.7 (2.9) | 6.4 (2.8) | 6.1 (2.8) | < 0.001 |
BMI (kg/m2) | 20.9 (3.7) | 22.0 (4.5) | 21.6 (4.2) | < 0.001 |
Waist circumference (cm) | 77.4 (10.8) | 76.1 (12.1) | 76.6 (11.6) | < 0.001 |
Male sex (%) | – | – | 40.0 | – |
Physical inactivity (%) | 14.5 | 7.9 | 10.5 | < 0.001 |
Family history of diabetes (%) | 21.3 | 24.5 | 23.2 | 0.016 |
Hypertension (%) | 24.3 | 24.5 | 24.4 | 0.955 |
Gestational diabetes (%) | 1.2 | – | – | |
Osmotic symptoms (%) | 14.1 | 13.6 | 13.8 | 0.661 |
Balanitis or vulvitis (%) | 1.1 | 5.4 | 3.7 | < 0.001 |
Development | Validation | All | *p value | |
---|---|---|---|---|
n
| ||||
Age (years) | 45.3 (15.1) | 45.1 (14.9) | 45.2 (15.1) | 0.743 |
FPG (mmol/l) | 4.9 (1.1) | 4.9 (1.2) | 4.9 (1.1) | 0.796 |
2-h PG (mmol/l) | 6.1 (2.9) | 6.2 (2.8) | 6.1 (2.8) | 0.796 |
BMI (kg/m2) | 21.6 (4.3) | 21.6 (4.2) | 21.6 (4.2) | 0.815 |
Waist circumference (cm) | 76.6 (11.8) | 76.6 (11.3) | 76.6 (11.6) | 0.994 |
Male sex (%) | 39.6 | 40.5 | 40.0 | 0.560 |
Physical inactivity (%) | 10.7 | 10.2 | 10.5 | 0.971 |
Family history of diabetes (%) | 23.3 | 23.2 | 23.2 | 0.971 |
Hypertension (%) | 24.6 | 24.1 | 24.4 | 0.706 |
Gestational diabetes (%) | 0.7 | 0.8 | 0.7 | 0.979 |
Osmotic symptoms (%) | 13.0 | 15.4 | 13.8 | 0.027 |
Balanitis or vulvitis (%) | 3.6 | 3.9 | 3.7 | 0.677 |
Model development
Variables (p value for diabetes) | Category | Odds ratio (95 % CI) | β | Contribution to Risk Score (10 × β) |
---|---|---|---|---|
Age (years) (p = 0.000) | <30 | Reference | 0 | |
30–39 | 2.6 (0.9–7.8) | 0.95 | 10 | |
40–49 | 4.8 (1.7–13.7) | 1.57 | 16 | |
≥50 | 5.0 (1.8–14.2) | 1.61 | 16 | |
BMI (kg/m2) (p = 0.000) | <18.5 | Reference | 0 | |
18.5–22.9 | 1.7 (0.7–3.9) | 0.52 | 5 | |
≥23.0 | 2.1 (0.8–5.2) | 0.72 | 7 | |
Waist circumference (cm) (p = 0.000) | F < 70 / M <75 | Reference | 0 | |
F (70–79)/M (75–84) | 1.9 (0.9–4.1) | 0.65 | 7 | |
F ≥ 80/M ≥ 85 | 2.9 (1.3–6.6) | 1.00 | 10 | |
Hypertension (p = 0.000) | Absent | Reference | 0 | |
Present | 2.3 (1.6–3.4) | 0.84 | 8 | |
Family history of diabetes (p = 0.000) | Present | Reference | 0 | |
Absent | 1.7 (1.1–2.5) | 0.52 | 5 | |
Physical activity level (p = 0.008) | Sufficiently active | Reference | 0 | |
Moderately active | 1.2 (0.8–1.8) | 0.17 | 2 | |
Inactive | 1.4 (0.8–2.4) | 0.32 | 3 | |
Gestational diabetes (p = 0.270) | Absent | Reference | 0 | |
Present | 1.4 (0.3–6.8) | 0.36 | 4 | |
Balanitis or vulvitis (p = 0.034) | Absent | Reference | 0 | |
Present | 2.0 (0.9–4.4) | 0.71 | 7 | |
Osmotic symptoms (p = 0.024) | Absent | Reference | 0 | |
Present | 1.2 (0.7–1.9) | 0.14 | 1 |
Model validation
Comparison of SLDRISK with Cambridge, Indian, Chineses and Thai Diabetes Risk Scores
SLDRISK | IDRS | CRS | |
---|---|---|---|
AUC | 0.78 | 0.72 | 0.66 |
Sensitivity (%) | 77.9 | 66.2 | 54.4 |
Specificity (%) | 65.6 | 66.1 | 59.3 |
Positive predictive value (%) | 9.4 | 8.8 | 6.2 |
Negative predictive value (%) | 98.3 | 97.5 | 96.4 |
SLDRISK | RS 1 [27] | RS 2 [28] | RS 3 [29] | |
---|---|---|---|---|
AUC | 0.78 | 0.70 | 0.77 | 0.76 |
Sensitivity (%) | 77.9 | 65.5 | 71.0 | 62.2 |
Specificity (%) | 65.6 | 64.9 | 73.0 | 69.5 |
Positive predictive value (%) | 9.4 | 7.7 | 11.1 | 8.7 |
Negative predictive value (%) | 98.3 | 97.6 | 98.1 | 97.4 |