Benchmarking operator performance in percutaneous coronary intervention: a novel approach using 30-day events

The infrequency of adverse in-hospital events limits scorecarding individual angioplasty operators. We assessed 30-day events for this purpose. Thirty-nine operators performed 1,950 coronary interventions from 1 April to 30 September 1998. Thirty-day follow-up was obtained in 1,896 (97.2%), who form the basis for this analysis. We recorded 16 baseline variables and chose an endpoint of mortality or target vessel revascularization. The endpoint occurred in 103 (5.4%) patients by 30 days. Independent predictors were identified by multivariate modeling. The expected event rate for the set of patients treated by each operator was determined. Two operators had significantly more adverse events by 30 days than predicted and two others had an event rate approaching significance. Two of these four performed at least 60 procedures during the 6-month period. With 30-day events, operators whose complication rates exceed laboratory standards can be identified with achievable sample sizes. Volume alone does not appear to be sufficient assurance of quality outcomes.