In what may be a classic case of the fox guarding the hen house, physicians, through the American Medical Association and various specialty societies, generally set the work units associated with the relative value scale that is widely used for fee-for-service payments. Think there might be a temptation to overstate those units? A Health Affairs study explores just that question. (HA Article) The fee schedule assigns each procedure a number of relative value units, which are then multiplied by some dollar value to get reimbursement level. The RVUs are supposed to be based on actual data on how long the procedure takes and are periodically reviewed and updated. This work is largely done by physicians who are asked to estimate the amount of time for the procedure, the level of technical skill required and the intensity of the work. The number of respondents used for these surveys is astonishingly small, especially given the number of doctors in the country and in various specialties. The hip and knee replacement RVUs were based on 157 and 150 responses, respectively. CMS accepts over 90% of recommendations based on this physician survey and updates RVUs very infrequently. Prior research has shown that many RVUs are overestimated, based on actual time studies. The researchers in this study decided to do their own research on the knee and hip replacement RVUs.
Retroactive time stamp data from an electronic medical record was used to ascertain actual operating time for the procedures at two large academic hospitals. Over 2700 procedures on Medicare patients were used for the assessment. These actual procedure times were compared with the data used for the basis of the current RVUs, adjusting for patient characteristics and other factors. Both original and revision hip and knee replacement procedures had significantly lower actual procedure times than those used for the RVUs. For example, the average actual operating time for an original hip replacement was about 85 minutes, compared to 100 minutes used for the RVU. And for an original knee replacement, it was about 82 minutes versus 100 in the current RVU. The comparison was even worse for revision procedures–actual time for a hip revision was 149.5 minutes versus the 240 used for the RVU. For all four types of procedures, at least 83% of the cases were below the time estimate used for the RVU. Can you say massive overpayment. Shorter duration of procedure did not result in higher complication rates. These results should lead CMS to immediately stop using physician surveys as the basis for payment. The amounts of money involved would easily justify a comprehensive look at actual procedure times and a frequent revisiting of those times. For many procedures, like joint replacements, physicians should get more efficient over the years. Given Medicare’s budget issues, this should be attended to immediately, and there would be positive spillover effects for commercial insurers.