Validation of claims-based algorithms for gout flares.

Area of Science:

• Rheumatology
• Health Informatics
• Epidemiology

Background:

• Gout is a prevalent inflammatory arthritis marked by recurrent acute flares.
• Accurate identification of gout flares is essential for evaluating gout treatment effectiveness.
• Claims databases are valuable for large-scale health research but require validated algorithms for specific conditions.

Purpose of the Study:

• To develop and assess the accuracy of a claims-based algorithm for identifying gout flares.
• To determine the utility of diagnosis codes combined with medication or procedure claims for gout flare detection.

Main Methods:

• A linked electronic medical record-Medicare claims dataset of patients with gout or hyperuricemia (2006-2010) was utilized.
• Gout flares were identified using algorithms combining gout diagnosis codes with medication claims or procedure codes (arthrocentesis/joint injection).
• Algorithm accuracy was evaluated by comparing with physician documentation (gold standard) for positive predictive value (PPV) and negative predictive value (NPV).

Main Results:

• Medication-based algorithms had a PPV of 50-54%, while procedure-based algorithms ranged from 59-68%.
• The combined medication and procedure claims algorithm demonstrated a high negative predictive value of 85.2%.
• A total of 3952 subjects were analyzed, identifying 503 flares via medication claims and 290 via procedure claims.

Conclusions:

• Claims-based algorithms combining gout diagnosis codes with medication or procedure data did not accurately capture gout flares.
• The developed algorithm may be more effective in identifying gout patients who have not experienced flares.
• Further refinement of claims-based algorithms is needed for precise gout flare identification in research settings.