Application of construction and validation of logistic regression model in risk prediction of non‐proliferative diabetic retinopathy in type 2 diabetes mellitus
Authors' conclusion
The constructed Logistic regression model accurately predicts NPDR risk using four readily available clinical variables, offering high discriminative power, interpretability and clinical utility for stratifying high-risk T2DM patients in primary care settings.
Study context
- Population
- Patients with T2DM, single centre
- Participants
- 421 patients
- Exposure
- Training vs Validation
- Trainingn = 295
- Validationn = 126
- Outcome
- Non-proliferative DR (NPDR) risk prediction by logistic regression model
Methodological detailssingle-centre, prediction model development and validation, bootstrap internal validation, external validation
- Single-centre
- Prediction model development and validation
- Bootstrap internal validation
- External validation
Main finding
As reported in the abstract
Independent predictors of NPDR and model performance
| Group | AUC | Sensitivity | Specificity | H-L test |
|---|---|---|---|---|
| Training | 0.949 | 86.4% | 95.5% | p = 0.358 |
Training
AUC0.949
Sensitivity86.4%
Specificity95.5%
H-L testp = 0.358
- Macular oedema OR
- 3.247
- Fasting glucose OR
- 2.194
- HbA1c OR
- 2.799
- UACR OR
- 1.153
Observational data: associations, not effects.
Further findings
External validation
- AUC
- 0.918
- PPV
- 91.1%
- Calibration maintained
- P = 0.282
Bootstrap validation
- Bootstrap validation confirmed stability of HbA1c and UACR
- 1000 replicates
Original publication
Application of construction and validation of logistic regression model in risk prediction of non‐proliferative diabetic retinopathy in type 2 diabetes mellitus. Diabet Med. Published 29 May 2026. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →