Benchmarking to Perfection
In the last three years, hospital ownership of physician practices has increased by 86 percent. Coker’s financial consultants have been busy helping hospitals, health systems and medical practices align their services together and ensure that payments to physicians are appropriate and market-based. Further, as the growth of hospital–employed networks has proliferated Coker has been called upon to assist in the standardization of compensation and the development of compensation models that allow for success in both the fee-for-service and fee-for-value environments. One of the most powerful and robust tools that we have at our disposal is market benchmark data. Coker uses benchmark data, along with other resources, to ensure that compensation models are market-based and compliant with industry regulations.
To Blend or not to Blend?
As primary resources, Coker utilizes three industry benchmark surveys that are provided by the Medical Group Management Association (MGMA), the American Medical Group Association (AMGA), and Sullivan, Cotter and Associates, Inc. (SCA). The MGMA, AMGA, and SCA surveys are widely recognized industry surveys frequently used to evaluate physician compensation and productivity. Our approach is to blend the three surveys to develop a national average which we utilize in our analysis. We blend using a straight average of the respondents of the three surveys. By averaging three surveys, we create a much larger base of respondents, which lessens the potential of any one survey skewing the data for a specialty. Further, blending all three surveys creates more consistency in the numbers year over year based on the number of respondents. Further, while no longer a part of Stark law, the approach of using multiple surveys was highlighted in Stark II guidance. Even though it is no longer a part of Stark law, most consulting firms use this as a best practice, as opposed to relying on a single survey.
Where are you from?
All three surveys provide regional data, which is based on geographic sections of the country. Roughly a quarter of the nation’s respondents live in each geographic area. While many clients and physicians surface the idea of using regional data, Coker uses caution in applying these benchmarks because the surveys potentially do not reflect the situation in that particular state where they reside.As an example, in the 2016 survey of the MGMA southern region there are 13 states. 16,491 physician responses were tabulated, which is 25 percent of the MGMA national response (66,050 physician responses). Three states out of the thirteen generated 50 percent of the respondents, while six states (nearly half) represented five percent or less of the total respondents, with three of those six states representing two or even only one percent of the respondents. Thus, in this instance, if the hospital being bench-marked is in one of the states with low respondents, the regional data may not be any more useful/applicable than national data. In fact, it may be less reliable given its concentration on one specific geographic region that is not the region the hospital is in. Of course, the same risk can exist for certain specialties even at the national level if there is a low number of respondents in those specialties. Anytime there is a small pool of respondents, there is greater risk of the data being less useful.
Further, as we look at specific specialties in the southern regional data there are specialties which do not even have enough respondents for the survey companies to post a benchmark for that particular specialty. Out of 142 specialties, 50 specialties did not have enough respondents for MGMA to publish regional data. In addition, many of the regional respondents that MGMA does provide data for may have as few as 15 or less. Granted, these are specialties which have fewer overall physicians practicing in them but because of the small number of respondents, outliers in the data can greatly affect the overall quartile percentile rankings. With the lower physician counts in regional data there is more inconsistency in the data year over year in both the number of respondents as well as the survey data reported. The data can move up or down by larger amounts due to the heightened impact of outlier data points in a particular region if an organization that is the largest respondent in that region generates those data points. For all of these reason, in most instances Coker focuses on national data where more respondents reduce the issues with outliers in the data.
How many tools do you have in your tool shed?
When using the benchmarking “tools” in our analysis, there are several key benchmarks. Total Cash Compensation (TCC) is one benchmark, which encompasses all sources of physician remuneration such as professional fees, quality incentives, panel size payments, etc. We utilize TCC along with the physician’s productivity measurement, called the work relative units (wRVU). The wRVUs, which we can benchmark, shows how productive a physician is in practice in comparison to their peers within their specialty. A third key benchmark is the TCC per wRVU ratio which takes TCC divided by the wRVUs to create the TCC per wRVU ratio.
Each of these three components are reported separately by the surveys. Further, there is no correlation between the TCC per wRVU ratio and TCC or wRVUs, as it is a by-product of the latter two data points. As an example, TCC at the 75th percentile of internal medicine benchmarking divided by the 75th percentile of wRVUs results in a TCC per wRVU ratio of $52.62, which is just below the median of $53.09, not the corresponding 75th percentile ($66.00).
Additional benchmarking data points that we use in our analysis focus on factors such as collections and how they relate to other items. We benchmark collections, collections to wRVUs, along with compensation to collections. Benefits can be bench-marked as well and are often part of our analysis of a physician’s total compensation package, as benefits are considered another form of compensation.
A private practice’s operational costs can be bench-marked as well, both as a percentage of collections and on a per physician (FTE) basis. These can all be utilized to analyze the performance of a physician who is employed or in private practice yet has compensation which seems low on a percentile basis compared to their productivity or collections.The above just highlights key benchmarks, but only begin to touch on all the data available. In addition to these key resources, there are benchmarks for call pay, medical directorship pay, non-physician compensation, among others. Suffice it to say, market benchmark data is a robust tool which Coker uses daily in our work products to provide the most complete and accurate market analysis as possible for our clients.
- Modern Healthcare, September 7,2016 – Article by Maria Castellucci, Hospital Ownership of Medical Practices grows by 86% in three years
- 69 Fed. Reg. at 16128.
