Physician Time in the Medicare Physician Fee Schedule: Implications for Alternative Payment Models

Objectives: Recent legislation ended the Medicare Sustainable Growth Rate (SGR) policy and moves physicians in the direction of Alternative Payment Models (APMs) beginning in 2019.  Despite these payment innovations, Medicare’s resource-based relative value scale (RBRVS), currently the basis for most public and private payment systems, is likely to be the foundation for many APMs or, at least, a significant point of comparison for them.  However, concerns have been raised about systematically distorted fees resulting from discrepancies between the amount of physician time used to establish fees in the RBRVS and the actual time physicians spend delivering a given service. This study examines the relationship between these two measures of physician time at the service level and across specialties and their potential for distorting fees.

Design: Using administrative data from large integrated health systems and a survey of physicians, we examine the time input to the RBRVS at the physician and service level. For the physician analysis, we acquired data on individual physicians' service volume and administrative measures of time worked for a year. Fee schedule time values were multiplied by each service's volume and summed across services for each physician to examine the ratio of fee schedule time to "actual time” worked. We assessed how this ratio varies by physician specialty and different types of services. In the service-level component, we fielded a survey asking two samples of physicians – a convenience sample of physicians affiliated with several large systems and a random sample of physicians - to estimate service times for 6 services in each of 5 specialties. We examine the correlation between survey responses and fee schedule times for all physicians, by specialty, and between physician samples.

Population: Administrative data analysis: all physicians in three large multi-specialty practices. Survey: one random sample of physicians from the AMA Masterfile and a sample of physicians from several large integrated systems.

Findings:  The administrative data from the multi-specialty practices suggest that the ratio of fee schedule time to actual time worked by individual physicians exceeds 1 and differs across physician specialty.  The survey data also show that fee schedule times are high relative to times reported for the surveyed services and that the distortions are greatest to Ophthalmologists and Radiologists.  Both sets of findings show that distortions are less pronounced for primary care services than for speciality care.

Implications:  Many have questioned the accuracy of the time data used in the RBRVS and have proposed that overstated time values for some groups of services have led to misvalued fees and, by extension, unjustified discrepancies in physician incomes.  The physician-level and service-level time estimates presented here show that the ratio of fee schedule time to actual time worked varies by specialty and implies there are likely distortions in the RBRVS used in the Medicare fee schedule.  The study results should guide policymakers to seek remedies before these misvaluations get embedded within future APMs.  This study also tests new strategies for obtaining physician time data that could be incorporated into RBRVS updates.