Accreditation of hospitals for percutaneous coronary intervention on the basis of volume or clinical outcome using MEDPAR data sets: effect on patient mortality, cost and treatment accessibility.

BACKGROUND

The risk of major complications of percutaneous coronary revascularization (PCR) is modestly lower in high-volume as opposed to low-volume hospitals, but this is not a consistent finding for all hospitals. There are also limitations comparing risk-adjusted outcomes between hospitals. We sought to ascertain the effect of credentialing hospitals for PCR, either on the basis of procedural volume or outcome, on clinical outcome, cost and accessibility to treatment, in states of varied population density.

METHODS

We evaluated Medicare administrative data sets for all PCRs performed in 9 states during 1994Eth 1995. Based upon volume- and risk-adjusted in-hospital mortality during 1994, hospitals were OaccreditedO using varying volume and outcome thresholds, and the effect of accreditation using these thresholds was ascertained by analysis of 1995 outcomes. Sensitivity analyses were performed to assess the effect of altered assumptions.

RESULTS

During 1994, one hundred and thirty-three hospitals performed 34,879 PCRs in Medicare patients, with an overall mortality of 1.36%. If credentialing were performed based upon 1994 volumes, a sizable clinical benefit could be expected only if large numbers of catheterization laboratories were OclosedO, e.g., if laboratories with < 200Eth 300 Medicare cases/year (< 400Eth 900 total cases) were OclosedO, mortality would be expected to decrease to 0.17Eth 1.07% (maximum and minimum effect). Costs could be minimized by closing laboratories with < 100 Medicare cases/year (best case scenario, $512Eth $905/patient). Such laboratory closures would require transfer to hospitals > 50 miles distant in 6Eth 38% of patients, but as many as 18Eth 94% of patients in low-density states. If credentialing were done on the basis of 1994 adjusted mortality, a somewhat lesser reduction of risk of death (best case scenario, 0.93%), but little improvement in cost, could be expected.

CONCLUSIONS

If generalizable, these data suggest that to achieve a sizable reduction in procedure-related mortality by hospital-based credentialing, large numbers of catheterization laboratories would need to be closed and patient access to care would be adversely impacted. Cost savings of a very considerable magnitude may be more readily achieved.