Physician compensation models based on work relative-value units should consider not only industry median benchmarks but also variations across different productivity levels.
Healthcare organizations have long relied on compensation models that reward physicians for generating work relative-value units (wRVUs). In many if not most cases, the rate paid per wRVU is tied to industry median benchmarks that are published in a variety of physician compensation and productivity surveys. Although there are good reasons for using benchmarks, the misapplication of median compensation per wRVU benchmarks, and the subsequent reporting of results in industry surveys, creates a feedback loop that affects the survey data. Over time, the benchmarks become more and more divorced from the actual economics of the practice, and they are progressively inflated to the point that, in many cases, they are unsustainable.
Another issue, which is perhaps more subtle but equally important, is the potentially counter-intuitive relationship between physician compensation and productivity. Although it is true that compensation generally tends to rise with increases in productivity, an analysis of how much physicians earn on a per-unit basis at varying levels of productivity shows this is not always the case, particularly for physicians on the higher end of the productivity spectrum.
The Default Assumption
A widely held view among people involved with physician compensation planning is that a physician’s compensation per wRVU should increase as his or her productivity increases, and many observers might assume that this pattern generally occurs. This perception is based on the economics of the medical practice; because of the significant fixed costs, a certain level of clinical activity is required to generate enough revenue to cover the practice’s overhead. But because variable costs are relatively few, a major portion of any incremental revenue typically will be available for physician compensation. Not surprisingly, many organizations have structured their physician compensation models with several tiers, such that a lower rate is paid for all wRVUs up to a specified threshold, with higher rates at subsequent levels of production, as shown in the exhibit below.
Tiered Compensation Models
The existence of these common tiered models, together with the practice economics just described, might lead one to expect that market data would follow a pattern similar to the orange line in the exhibit. But that may not actually be the case.
The (Potentially) Surprising Truth
There is evidence that instead of higher-producing physicians earning more per wRVU, the opposite is actually the case. The exhibit below shows that physicians at the lower deciles of wRVU productivity tend to have much higher rates of compensation per wRVU, and the very highest-producing physicians have the lowest rates of compensation per wRVU.
Physician Productivity and Compensation Per wRVU
For many readers, it will come as no surprise that the lowest-producing physicians earn the most on a per-WRVU basis—the reason being that physicians, particularly those who are early in their career and building their practices, commonly receive income guarantees. In such instances, productivity often does not match compensation levels, resulting in high compensation-to-productivity ratios.
What may be surprising is how this relationship changes for the middle to upper portions of the productivity spectrum. One might expect the curve to at least flatten if not climb again at higher levels of productivity. Instead, the income curve continues to decline steadily so that the most productive physicians tend to earn the least on a per-unit basis.
A geographic component (not shown in the exhibit) also comes into play. In some markets, physicians tend to earn more per wRVU for a variety of factors, such as favorable payment by commercial insurers, local competition for physicians among health systems, and geographies that require a premium for recruitment purposes. In such markets, physicians also tend to produce less, with the result that total compensation levels are more consistent across markets than compensation per wRVU rates.
A Theoretical Explanation
Microeconomic theory offers some clues as to why physician compensation and activity relate to each other in this way. As it relates to an individual’s earnings, the theory is that workers decide how hard to work based in part on the wage rate they are offered. If economic conditions are such that workers’ wages are fixed at low levels (lower compensation per wRVU in this context), the workers may determine that it is hardly worth their while to do much work, and so they will be inclined to live simply and focus more on enjoying leisure time instead. Because medicine is a relatively high-paying profession, we would not expect to observe this behavior very often.
However, if a higher wage is being offered (i.e., more lucrative compensation per wRVU ratios), the value of work increases, and workers will respond by spending more of their time working and taking less time off. There is a tipping point, however, when workers’ wages are high enough to give them the luxury of actually working less while still maintaining an attractive compensation level. In effect, workers now have sufficient income to “spend” on more leisure time.
The result is that when workers are offered higher wage rates (i.e., a higher rate per wRVU), they may choose to work less (because they can afford to do so). Conversely, if they are offered lower wage rates, they may end up working more (because they believe they have to). This scenario corresponds to the relationship between compensation per wRVU and wRVU productivity depicted in the exhibit above and may help to explain why that phenomenon exists at the higher end of the productivity scale.
Implications for Compensation Planning
For the reasons described above, the common practice of basing physician compensation on median compensation per wRVU may produce results that are at variance with the market. Median compensation per wRVU represents all physicians in the data set and includes dissimilar practices, as well as many that are true outliers.
Median Compensation Per wRVU Model Results: Noninvasive Cardiology
To illustrate this point, the exhibit directly above superimposes a median compensation per wRVU plan (portrayed by the sloped orange line) on actual survey data in which each data point in the graph represents an individual physician in the survey. Although based on real-life data, the model depicted is hypothetical.
These data show that a physician who produces above the median would be overpaid if he or she were paid on a median compensation per wRVU basis. This overpayment occurs because the majority of higher-producing physicians are actually receiving compensation that is less than the median compensation per wRVU shown in this hypothetical model. However, the majority of the lower-producing physicians appear above the orange line, which suggests that such a model would actually underpay them.
These observations have several implications for physician compensation planning efforts, as detailed below.
Realize that high rates for high producers may not produce the desired effect. As shown, it is not guaranteed that higher rates of pay per wRVU will drive higher productivity. Furthermore, in a value-based environment, extremely high productivity may not be clinically appropriate or financially advantageous. In terms of influencing physician behavior, it is important to recognize that physicians are motivated by many factors—for example, achieving professional satisfaction or simply responding to the demand for their services—and that maximizing earnings may not be their top priority.
Be attentive to fair market value (FMV). Regulators tend to zero in on physicians with the highest compensation levels, which usually means those with high productivity. This information suggests that, for these physicians, even a slightly elevated rate per wRVU may push the compensation package out of FMV range.
Consider using a “fully productive” benchmark. Rather than simply using median compensation per wRVU as the starting point for compensation formulas, it may be helpful to use a rate that reflects the productivity level the organization expects its physicians to achieve. For example, a rate that reflects the performance of physicians whose wRVU productivity is between the 50th (the median) and 75th percentile may be appropriate for many organizations because these physicians tend to be the ones whose practices are mature and productive. Survey data suggest that this approach may result in a rate per wRVU that is about 6 percent below the published median rate, as shown in the exhibit below.
Variance to Median Compensation Per wRVU, by Productivity Quartile
Be creative in giving lower producers incentives to improve productivity. Addressing incentives for lower-producing physicians is important because this area is where most of the opportunity for improvement will be. The point is not necessarily to pay lower producers more; rather, it is to structure compensation to make sure the reward for increased productivity is large enough to influence behavior. One method would be to reduce or eliminate the base component and to increase the marginal rate per wRVU offered to lower producers. Incremental improvement is the key; someone who has historically been at the 25th percentile is unlikely to reach the 75th percentile but might be encouraged to reach the 30th or 35th percentile. Of course, even this percentile may not reflect sufficient improvement, and in some cases, administrative intervention may be required.
Address operational and other issues. Offering financial incentives is one means of influencing physician productivity, but there are limits to the effectiveness of such incentives, and they can be costly and controversial as well. Sometimes the problem has nothing at all to do with the physicians’ willingness to put in effort. Instead, there could be issues with patient access, clinic operational efficiency, marketing, health plan contracting, and a host of other factors related to generating and responding to the demand for clinical services. It may be best to ensure that other areas are functioning properly before tackling the difficult issue of physician compensation redesign.
A Better Understanding of the Data
Using industry data in the design of physician compensation plans is a necessary practice, but achieving the desired results requires both knowledge of and insight into what the data actually represent, why the benchmarks behave as they do, and what can be achieved through the compensation arrangement. Such understanding provides the essential foundation for any health system’s efforts to design effective incentive structures.
Dave Wofford is associate principal, ECG Management Consultants, San Diego..
Maria Hayduk is associate principal, ECG Management Consultants, St. Louis.
Kenneth Robbins, MD, is executive vice president and chief medical officer, Hawaii Pacific Health, Honolulu, Hawaii.