Recognising, quantifying and accounting for classification uncertainty in type 2 diabetes subtypes
Authors' conclusion
The NRE provides a means to quantify and compare individual classification uncertainty in type 2 diabetes subtypes. Classification uncertainty varied between subtypes and individuals with type 2 diabetes, and accounting for it improved the ability of the subtypes to predict 10 year CVD risk.
Study context
- Population
- Individuals with recent-onset T2D from the German Diabetes Study (GDS); subtypes defined by age, BMI, HbA1c, HOMA-IR and HOMA-BGDS cohort
- Participants
- 859 individuals
- Exposure
- Classification uncertainty (normalised relative entropy (NRE)) across T2D subtypes; SCORE2-Diabetes with vs without weighting by classification certainty
- Mild age-related diabetesn = 395
- Mild obesity-related diabetesn = 316
- Severe insulin-resistant diabetesn = 130
- Severe insulin-deficient diabetesn = 18
- Outcome
- Classification uncertainty quantified by normalised relative entropy (NRE) from distances to cluster centroids; lower NRE = greater uncertainty
Methodological detailscluster-based subtype classification
- Cluster-based subtype classification
Main finding
As reported in the abstract
Mild age-related diabetes
0.155
median · 95% CI 0.142–0.177
Mild obesity-related diabetes
0.119
median · 95% CI 0.107–0.131
Severe insulin-resistant diabetes
0.086
median · 95% CI 0.075–0.108
Severe insulin-deficient diabetes
0.082
median · 95% CI 0.071–0.109
Observational data: associations, not effects.
Further findings
Proportion of variation in SCORE2-Diabetes explained by subtypes (R2), unweighted vs weighted by classification certainty
- R2 before weighting
- 17.4% (95% CI 12.8, 23.0)
- R2 after weighting
- 31.5% (95% CI 26.4, 37.1)
Predicted 10 year CVD risk (SCORE2-Diabetes), mild age-related diabetes, unweighted vs weighted
- Before weighting
- 10.3% (95% CI 9.8, 10.7)
- After weighting
- 11.6% (95% CI 11.2, 12.0)
Original publication
Composed by AI from the published abstract and checked against it. How Briefs are made →