PubMed Central, : RMMJ Rambam Maimonides Medical Journal Rambam Health Care Campus 2026; 17(3): e0019. ISSN: 2076-9172Published online 2026 July 31. doi: 10.5041/RMMJ.10579

Table 4.

Multivariable Logistic Regression Analyses of Factors Associated with Moderate-to-High Type 2 Diabetes Risk According to FINDRISC, QDScore, and TRAQ-D Scales.

Variable	FINDRISC OR (95% CI)	QDScore OR (95% CI)	TRAQ-D OR (95% CI)	P Value
Men	1 (ref)	1 (ref)	1 (ref)
Women	2.36 (2.19–2.54)	2.15 (2.01–2.30)	1.94 (1.80–2.09)	<0.001
Age 18–29 years	1 (ref)	1 (ref)	1 (ref)
Age 30–39 years	1.29 (1.25–1.34)	1.34 (1.28–1.40)	1.42 (1.35–1.50)	<0.001
Age 40–49 years	1.98 (1.87–2.10)	1.85 (1.70–2.01)	2.12 (1.90–2.35)	<0.001
Age 50–59 years	2.58 (2.40–2.77)	2.64 (2.40–2.89)	2.93 (2.64–3.23)	<0.001
Age 60–69 years	3.60 (3.35–3.86)	3.12 (2.81–3.42)	3.70 (3.31–4.10)	<0.001
Social class I	1 (ref)	1 (ref)	1 (ref)
Social class II	1.35 (1.30–1.41)	1.41 (1.33–1.49)	1.41 (1.32–1.51)	<0.001
Social class III	1.96 (1.85–2.07)	2.10 (1.95–2.35)	1.92 (1.79–2.06)	<0.001
Non-smokers	1 (ref)	1 (ref)	1 (ref)
Smokers	1.16 (1.13–1.20)	1.29 (1.22–1.36)	1.40 (1.31–1.51)	<0.001
Yes, Mediterranean diet	1 (ref)	1 (ref)	1 (ref)
No, Mediterranean diet	1.89 (1.79–1.99)	2.89 (2.60–3.20)	2.91 (2.60–3.21)	<0.001
Yes, physical activity	1 (ref)	1 (ref)	1 (ref)
No, physical activity	3.23 (2.99–3.47)	4.12 (3.68–4.57)	5.49 (5.02–5.96)	<0.001
No insomnia	1 (ref)	1 (ref)	1 (ref)
Subthreshold insomnia	1.29 (1.24–1.35)	1.21 (1.17–1.25)	1.33 (1.25–1.41)	<0.001
Moderate insomnia	1.87 (1.76–1.98)	1.68 (1.52–1.85)	1.68 (1.50–1.87)	<0.001
Severe insomnia	2.62 (2.45–2.79)	2.15 (1.80–2.50)	2.30 (2.01–2.61)	<0.001

All variables were included simultaneously in the multivariable logistic regression models. Odds ratios (ORs) are adjusted for sex, age, social class, smoking, adherence to the Mediterranean diet, physical activity, and insomnia severity. P-values correspond to comparisons with the reference category. Significant dose–response trends were additionally observed across age group, social class, and insomnia severity (all P for trend <0.001).

Model diagnostics: variance inflation factors (VIF) were <2.5 for all covariates, indicating no multicollinearity. Model calibration was good (Hosmer–Lemeshow P>0.20), and discriminatory ability was acceptable (AUC 0.78 for FINDRISC, 0.80 for QDScore, and 0.77 for TRAQ-D).

AUC, area under the receiver operating characteristic curve; FINDRISC, Finnish Diabetes Risk Score; QDScore, QResearch-derived Diabetes Risk Score; ref, reference; TRAQ-D, Trinidad Risk Assessment Questionnaire for Type 2 Diabetes Mellitus.