Identifying cases of congestive heart failure from administrative data: a validation study using primary care patient...

Area of Science:

• Health Informatics
• Epidemiology
• Cardiology

Background:

• Accurate identification of congestive heart failure (CHF) patients is crucial for population health management.
• Existing methods for identifying CHF patients may not fully leverage administrative data sources.

Purpose of the Study:

• To evaluate the accuracy of 9 algorithms in identifying patients with congestive heart failure (CHF) using combined administrative data.
• To determine the optimal algorithm for CHF patient identification from administrative sources.

Main Methods:

• Tested 9 algorithms using a validation cohort combining electronic medical records and physician billing data from Ontario, Canada.
• Algorithms were assessed using sensitivity, specificity, positive predictive value, ROC AUC, and likelihood ratios.

Main Results:

• The most effective algorithm combined one hospital record or physician billing with a second record from either source within one year.
• This optimal algorithm achieved a sensitivity of 84.8% and a specificity of 97.0% for CHF identification.

Conclusions:

• Combined administrative data from hospitalization and ambulatory care can accurately identify patients with congestive heart failure (CHF).
• This approach enables precise measurement of CHF population prevalence.