(B-169) Evaluation of completeness of patients’ follow up in a hospital healthcare database by matching with a closed claims database among atrial fibrillation patients

Background: Discontinuity in follow-up due to patients receiving care outside index hospitals may result in incomplete capture of study outcomes within hospital billing databases that can identify specific devices used.

Objectives: To compare the rate of AF recurrence after catheter ablation for AF patients between a large US hospital billing database and a closed claims database, which provides more complete follow-up.

Methods: A retrospective study used the Premier Healthcare Database (a hospital billing database) and the Optum closed claims database from 2018 to 2022. Premier includes inpatient and outpatient hospital visits; Optum captures all billed inpatient and outpatient encounters during the insurance enrollment period. Patients with AF diagnosis and catheter ablation conducted during inpatient or outpatient hospital visits were identified from the Optum and Premier databases The 1-year baseline Elixhauser comorbidities were calculated using data from index hospitals in both databases. Subclass propensity score matching was applied to balance patient demographic and clinic characteristics and provider characteristics between the two databases. AF recurrence was defined as repeat ablation, AF hospitalization, or direct cardioversion (DCCV). The difference in the cumulative incidence of these between the Optum and Premier databases was calculated for 30-, 90-, and 365-days post index procedure. The variance and 95% confidence intervals (CI) of the incidence were estimated using non-parametric bootstrapping.

Results: 22,789 in Optum and 135,549 in Premier patients met eligibility criteria. The difference in AF recurrence increased with longer follow-up. Within 30 days, repeat ablation rates were similar (Premier: 0.28%, Optum: 0.34%, difference: -0.06%, 95% CI: 0.03%, -0.15%), as were AF hospitalization rates (Premier: 0.98%, Optum: 0.82%, difference: 0.16%, 95% CI: 0.33%, -0.01%). Over 365 days, AF hospitalization rates showed a difference of -0.75% (95% CI: -0.31%, -1.19%, Premier: 5.41%, Optum: 6.15%) and repeat ablation rates -0.89% (95% CI: -0.44%, -1.35%, Premier: 5.31%, Optum: 6.20%). DCCV incidence was lower in Premier within 30 days (3.13% in Premier, 4.25% in Optum, difference: -1.12%, 95% CI: -0.69%, -1.56%) and within 365 days (9.90% vs 14.94%, difference: -4.35%, 95% CI: -4.35%, -5.74%).

Conclusions: Using the Optum closed claims database as more complete capture reference, the Premier hospital billing database had similar event rates of repeat ablation and AF hospitalization within 30 days. Events that occur in outpatient settings, like DCCV, are less captured in the hospital database than repeat procedures.