Improving Short‐ and Long‐Term Type 2 Diabetes Risk Prediction for Middle‐Aged British Adults
Authors' conclusion
We developed and externally validated a tool to estimate 5-, 10- and 15-year diabetes risk in British middle-aged adults, which may support primary care decision-making.
Study context
- Population
- Middle-aged British adults; development and internal validation in Whitehall II (Training n = 4502; Test n = 1929), external validation in English Longitudinal Study of Ageing
- Comparison
- LASSO-selected vs random survival forests (RSF)-selected vs best subset Cox models
- Outcome
- Incident diabetes
Methodological detailsinternal validation, external validation, LASSO-Cox, random survival forests and best subset selection
- Internal validation
- External validation
- LASSO-Cox, random survival forests and best subset selection
Performance
As reported in the abstract
Discrimination in ELSA external validation (C-index)
| Test or model | Risk-score C-index (95% CI) | Risk-group C-index (95% CI) |
|---|---|---|
| LASSO | 0.800 (0.773-0.827) | 0.737 (0.715-0.759) |
| RSF | 0.801 (0.774-0.827) | 0.727 (0.707-0.748) |
Risk-score C-index (95% CI)
LASSO0.800 (0.773-0.827)
RSF0.801 (0.774-0.827)
Risk-group C-index (95% CI)
LASSO0.737 (0.715-0.759)
RSF0.727 (0.707-0.748)
Further findings
Selected predictors
- LASSO and RSF both selected ethnicity, BMI, waist-to-height ratio, height, triglyceride-glucose index, fasting glucose, family history of diabetes and socioeconomic status; LASSO additionally smoking; RSF also age, TG, HDL-C, systolic and diastolic blood pressure
Original publication
Improving Short‐ and Long‐Term Type 2 Diabetes Risk Prediction for Middle‐Aged British Adults. Diabetes Obes Metab. Published 26 August 2026. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →