Review of Risk Prediction Models for Type 2 Diabetes
Author Information
Author(s): Gary S Collins, Susan Mallett, Omar Omar, Ly-Mee Yu
Primary Institution: Centre for Statistics in Medicine, University of Oxford
Hypothesis
The study aims to systematically review and assess the methodology and reporting of risk prediction models for type 2 diabetes.
Conclusion
The review found widespread use of poor methods in developing risk prediction models for type 2 diabetes, which compromises their reliability and accuracy.
Supporting Evidence
- Thirty-nine studies comprising 43 risk prediction models were included in the review.
- 44% of studies reported models for predicting incident type 2 diabetes.
- 21% of models had fewer than 10 events per variable, indicating potential overfitting.
Takeaway
This study looked at how well different models predict type 2 diabetes and found that many of them are not very good because they use bad methods.
Methodology
A systematic search of PubMed and EMBASE databases was conducted to identify studies that developed risk prediction models for type 2 diabetes, focusing on aspects like study design, sample size, and model-building strategies.
Potential Biases
Many studies used poor design and reporting practices, which could introduce bias in the results.
Limitations
The review was limited to English-language articles and may have missed relevant studies published in other languages.
Participant Demographics
The studies included a diverse range of participants from various countries, primarily adults.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website