Key issues: Coding, definitions, fairness
Payers know how many patients each physician sees, how intensive those visits are, and how productive they are. But most medical groups don't have a clue about anything but their overall patient loads. That is why medical groups increasingly are studying their RVUs, relative value units as defined by Medicare, to determine productivity.
The Medical Group Management Association (MGMA) in Englewood, CO, has just completed a four-year Physician Profiling Project that enabled 3,900 physicians in 71 participating practices to compare their clinical work profiles. "All of us like to know how we do compared to our peers," says Donna Burman MA, research database administrator for the MGMA project. "What is unique about this project is that we do it across all payers."
Burman and her colleagues calculated RVUs from CPT codes, which were provided by the participants. The data were reported according to smaller and larger practices, above or below 100 full-time equivalent (FTE) providers. Nurse practitioners, family practice physicians, and other specialists received reports comparing their RVU values to each other. In addition, the practices could compare overall values to those in each of the four participating states and the overall study population.
Within CPT ranges, a practice could learn whether its physicians' coding was weighted toward low- or high-end procedures. Within any given CPT range, reimbursement levels can vary significantly. "For example, family practice doctors would be able to see how many RVUs they had per FTE and how other practices compared," says Burman. The study did not indicate what the preferred level would be, she adds.
Accurate coding essential
MGMA hopes to develop physician profiling by RVU into a product available to practices with monthly or quarterly reports, says Burman. However, there is no target date for such a system.
At a February conference in Denver, physician profiling members discussed the lessons learned in the demonstration project. Some practices have implemented their own physician profiling based on various indicators and joined the project primarily to obtain national comparisons.
Gregory Angstman, MD, medical director of the Wabasha (MN) Clinic in the Mayo Health System, found that his clinic's physicians are highly productive. "The family physicians in the groups I'm responsible for are quite active clinically. In comparison to their peers, their productivity seems to be higher than typical. Yet you couldn't see anything that indicates they have a markedly different practice from their peer groups."
The profiling project made practices aware of what is necessary to gather complete and accurate information, as well as how to use the data. Burman and Angstman shared the following observations:
· Coding quality must improve to ensure reliable comparisons of physician work and adequate reimbursement.
In the MGMA project, several practices implemented coding training for providers, Burman says. "The providers would look at the data and say, 'That's not what I do.' We would say, 'this is what you're reporting,'" she says. For example, one practice coded office visits for established patients at the middle of the range: CPT 99213. Yet the five codes in the range represent different types or patients and visits, from one that doesn't require the presence of a physician to a comprehensive, 40-minute history-taking and exam appointment that involves a medical problem of moderate to high severity.
Disparities in coding can affect reimbursement as well as productivity measures, Burman notes. "If you want accurate profiles, you have to properly code what you do," she says.
· Participants in a profiling study must use the same definitions.
What is a FTE provider? Practices may define that term differently, particularly when determining the value of a part-time provider. Is he or she working at .5 FTE or .8 FTE? For the purposes of the MGMA study, a FTE provider saw patients for 35 to 40 hours per week. Providers who worked more than that could not be counted as more than one FTE. If practices use different definitions, clearly their comparisons won't be valid, Angstman says.
· Physicians should examine their variations to determine differences in practice.
Using data effectively
Variations aren't necessarily bad, notes Burman. For example, one physician might subspecialize in a certain type of care or may have an older patient population. The MGMA project didn't include case-mix adjustment to account for differences in severity of illness among physician populations. "In order to find out about what causes the variation, you would need to sit down with physicians and talk about what they are doing differently," she says.
At Wabasha Clinic, while data collection is centralized, the different clinical sites analyze the information to set priorities, Angstman says. Individual physicians and teams of physicians and nurses develop improvement projects based on the profiling information. The indicators include flu vaccines, tetanus shots, mammograms, Pap smears, and cholesterol screening.
· Productivity profiling can be used to create an objective method of determining compensation.
Wabasha Clinic establishes a compensation pool each year and divides it among providers based on RVUs. Those units can be monitored electronically and provide an objective way to determine who has performed the most work, Angstman says. "Our physicians accept that [measure] as long as it's applied equally," says Angstman, who notes that RVUs aren't used to make judgments about the quality or effectiveness of physician performance.