SUMMARY
Background:
Administrative billing codes for electrical cardioversion and ablation/maze procedures may be useful for atrial fibrillation (AF) research if the codes are accurate relative to medical record documentation.
Hypothesis:
We hypothesized that administrative billing codes accurately identify occurrence of electrical cardioversion and ablation/maze procedures in AF patients.
Methods:
We studied adults ages 30–84 who experienced new onset AF between 10/2001 and 12/2004 in Group Health Cooperative (acquired by Kaiser Permanente in 2017), an integrated health care system in Washington state and northern Idaho. Using medical record review as the gold standard, we calculated sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) for three administrative billing codes for electrical cardioversion and three codes for AF ablation/maze procedures.
Results:
Of 1,953 study participants, during a mean (SD) of 1.5 (0.7) years of follow-up after AF onset, 470 (24%) experienced electrical cardioversion and 44 (2%) experienced ablation/maze procedures according to medical record review. For electrical cardioversion, individual codes had 7.7% to 76.4% sensitivity, >99% specificity, 83.7% to 96.5% PPV, and 77.3% to 93.0% NPV. Considering any of three codes (code 1 or code 2 or code 3) improved sensitivity to 84.9%. For ablation/maze, individual codes had 18.2% to 47.7% sensitivity, >99% specificity, 66.7% to 95.5% PPV, and >98% NPV. Considering any of three codes improved sensitivity to 84.1%.
Conclusions:
Administrative billing data accurately identified electrical cardioversion and ablation/maze procedures, and can be used instead of medical record review. Our findings apply to health care settings with available administrative billing databases.