My third week here at the Banner Medical Group Corporate Office was fantastic. I thought it would not be as good as my other weeks because my mentor was busy for a large portion of my time at the office. However, as a result of his time away from me, I got to do some of the work that actual financial analysts do. I was given access to the information listing the level of productivity of physicians, measured in relative value units and information regarding their compensation as a result of their efficiency. For each service, a payment formula contains three RVUs, one for physician work, one for practice expense, and one for malpractice expense. On average, the proportion of costs for Medicare are 52%, 44% and 4%, respectively. The three RVUs for a given service are each multiplied by a unique geographic practice cost index, referred to as the GPCI adjustment. The GPCI adjustment has been implemented to account for differences in wages and overhead costs across regions of the country. The sum of the three geographically weighted RVU values is then multiplied by the Medicare conversion factor to obtain a final price. Historically, a private group of 29 (mostly specialist) physicians—the American Medical Association's Specialty Society Relative Value Scale Update Committee (RUC)—have largely determined Medicare's RVU physician work values. For every relative value unit that these doctors produce over the median, they receive a certain amount of money based on their specialty. Someone had made a mistake and listed some of the doctors’ productivity incorrectly, resulting in Banner Medical Group paying these doctors the wrong amounts. My job was to comb through over one hundred thousand lines of data and determine the reason behind this incorrect measurement. Initially, this process was painstaking and meticulous, but I later discovered a more efficient method of analyzing this data.
First, I exported the data to Microsoft Excel. Next, I wrote a function using a vlookup, which essentially lets you export a data point from one set of data to another set of data. One set of data had the accurate relative value unit number, and the other had the inaccurate number that Banner Medical Group used. Unfortunately, despite spending 15 hours looking at the data, I couldn’t come to any real concrete cause for the inaccurate measurements of RVUs.