Differential effect of interventions in patients with prediabetes stratified by a machine learning‐based diabetes progression prediction model
Authors' conclusion
This study suggests personalized treatment for prediabetes according to their PR is necessary. ML-PR model with simple clinical variables may facilitate personal treatment strategies in participants with prediabetes.
Study context
- Population
- Participants with prediabetes: Pinggu Study (prospective population-based survey, suburban Beijing; n = 622) for model development; Beijing Prediabetes Reversion Program cohort (n = 1936) for stratificationplus 622 in Pinggu Study
- Participants
- 1,936 participants
- Comparison
- Four interventions within ML-PR risk strata
Methodological detailsmachine learning (XGBoost), development and internal validation in the Pinggu Study, external validation, risk-stratified subgroup analysis, multicentre
- Machine learning (XGBoost)
- Development and internal validation in the Pinggu Study
- External validation
- Risk-stratified subgroup analysis
- Multicentre
Primary result
As reported in the abstract
year 1 and end of trial
High-risk: pioglitazone plus intensive lifestyle therapy significantly reduced diabetes progression vs conventional lifestyle therapy with placebo, at year 1 and end of trial
about 50%
Further findings
ML-PR prediction of 1-year progression
- Internal AUC, Pinggu Study
- 0.80 (0.72-0.89)
- External AUC, Beijing Prediabetes Reversion Program
- 0.80 (0.74-0.86)
- Predictors: FPG, 2-h postprandial glucose after 75 g glucose, HbA1c, HDL-C, TG
Original publication
Differential effect of interventions in patients with prediabetes stratified by a machine learning‐based diabetes progression prediction model. Diabetes Obes Metab. Published 1 October 2023. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →