Identifying Patients with Treated Type 2 Diabetes Using Health Data
Author Information
Author(s): Laurence M Renard, Valery Bocquet, Gwenaelle Vidal-Trecan, Marie-Lise Lair, Sophie Couffignal, Claudine Blum-Boisgard
Primary Institution: Centre for Health Studies, Public Research Centre for Health, Luxembourg
Hypothesis
Can an algorithm effectively identify patients with treated type 2 diabetes using medico-administrative data?
Conclusion
The DIABECOLUX algorithm is effective in identifying treated type 2 diabetes patients and estimating their prevalence.
Supporting Evidence
- The algorithm showed a sensitivity greater than 98.2%.
- Prevalence of treated type 2 diabetes increased from 3.15% in 2000 to 3.79% in 2006.
- Validation against a national study confirmed the algorithm's accuracy.
- Data was collected from a representative health insurance claims database.
Takeaway
Researchers created a computer program to find people with type 2 diabetes using health records, which helps understand how many people are affected.
Methodology
The DIABECOLUX algorithm was developed using health insurance claims data from 2000 to 2006 to identify treated type 2 diabetes patients.
Potential Biases
Potential misclassification of type 1 diabetes patients as type 2 due to the algorithm's criteria.
Limitations
The study only considered diagnosed patients who were treated and reimbursed, excluding undiagnosed patients and those under lifestyle recommendations.
Participant Demographics
The population included 21,068 treated type 2 diabetes patients, with a median age of 66 and a gender distribution of 50.6% males in 2000.
Statistical Information
P-Value
p<0.05
Confidence Interval
[95% CI: 3.4-3.5]
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website