June 28, 2019
4 min read

STS News, Summer 2019 — The concept of “fair market value” is a centerpiece in physician employment compensation agreements as mandated by federal law. In this edition of STS News, Dr. G. Randall Green explores the challenges associated with the real-world application of this construct.

Frank L. Fazzalari, MD, MBA, Chair, Workforce on Practice Management


G. Randall Green, MD, JD, MBA
Division Chief of Cardiac Surgery, Upstate Medical University, Syracuse, NY

If physicians used clinical research the way that the health care industry uses physician compensation surveys, patient outcomes would be as unpredictable as practice valuations. Most market participants believe that compensation surveys establish a fair market value of physician compensation. Defining ranges of fair market value of physician compensation using these surveys, however, reveals a systemic misunderstanding of the data and leads to an indefensible valuation practice.

Compensation market research lacks the rigor of the STS National Database to be used with the same confidence for valuation purposes. Survey data are drawn from voluntary samples, and specialty sample sizes reported can be comparatively small. Many commercial studies also are biased toward large, multispecialty groups providing data. By using a single questionnaire for every specialty, current surveys fail to capture specialty-specific and even subspecialty-specific drivers of value in each of these very different businesses. In our own specialty, consider the significant differences between adult cardiac, pediatric cardiac, and general thoracic surgery. Such incomplete data collection ultimately limits comparability.

To use market research for physician compensation valuation, comparability of the survey data to the subject transaction is a threshold issue. Market participants routinely turn to physician productivity as the sole measure of comparability. Here, measures of productivity such as the much-derided wRVU percentile are used to identify a corresponding percentile of compensation. This practice assumes a relatedness between survey data tables and a linear correlation between productivity and compensation. Tim Smith of TS Healthcare Consulting, however, has shown using Medical Group Management Association data that productivity fails to account for as much as 60%-70% of physician compensation. Physician compensation follows a multifactor economic model, and survey instruments that fail to collect a comprehensive dataset limit comparability. The lack of scholarship on the physician compensation data itself limits our understanding and, therefore, the utility of the surveys.

The importance of these studies should be clear to anyone in clinical practice. Although a few holdouts remain, the majority of STS members are now employed, leased, or in management positions and, therefore, have financial relationships with hospitals. The requirement that such relationships be at fair market value means that existing physician productivity and compensation market research will be used to establish levels of compensation. What may not be so clear to high-earning physicians is how survey data are used in regulatory compliance cases where hospitals and health systems are alleged to have paid above fair market value for physician services.

In two notable cases, Tuomey and Halifax, the compensation valuation expert for the Department of Justice took the position that fair market value should be based exclusively on physician compensation survey data. Due to the nature of participation in voluntary surveys, it is possible that high-earning physicians compose a very small proportion of the sample used to construct currently available compensation data. As such, those same physicians fail to establish the true market for their professional services with resulting increased compliance risk.

Although physicians cannot control how the data are used, the power to create a truly representative database of physician productivity and compensation lies with each one of us. As one of the first societies to collect our clinical data to improve clinical outcomes, perhaps it is time to lead once again by using our own practice data to serve our profession.

Note: STS will conduct its quinquennial practice survey this fall. More information about this important survey will be provided in the coming months.