Trends in Physician Exclusion: Characteristics for Consideration to Complement Analyses

By Amy K., Training & Development Manager

January 16, 2019

What is predictive profiling? How does predictive profiling help the healthcare industry? How do statistics and identifying traits inform decision-making about healthcare providers?

Predictive profiling is often criticized for its use by law enforcement as potentially being biased or perpetuating inequality.  However, I have spent the last 24 years investigating financial fraud with data analytics.  In the investigative process, “cumulative similarities” can assist in further identifying red flags and eliminating false-positives.  While its value may be incremental, it brings a quantitative element to decision analysis and optimization.

Over the past ten years (2007-2017), the number of physicians excluded from participation in public healthcare programs has increased 200%. During that period, roughly 0.29% of U.S. physicians, or 2,222, were temporarily or permanently excluded due to a variety of infractions to include: healthcare fraud schemes, health crime convictions, and/ or the unlawful prescription of controlled substances.

Using characteristics of those excluded, a “profile” was created and found they were more likely to be: male, older in age, an IMG (international medical graduate), to have osteopathic vs. allopathic training, or not have a faculty appointment at a U.S. medical school. A greater proportion of those excluded tended to work in the following specialties: family and/or internal medicine, psychiatry, anesthesiology, surgery, or obgyn. They were also more likely to practice in the West and Southeast census regions, with West Virginia having the highest exclusion rate.  While CA, NY, FL, and TX all had the highest counts of excluded populations, they also have the highest physician populations.  Those same states all have Medicare Fraud Strike Forces in place to combat over-payments due to high levels of Medicare waste per beneficiary and a top 20% national ranking. Exclusions were least common in the specialties of cardiology and radiology and in graduates of top 20 medical schools.

Several different explanations could be attributed to such a large increase over time.  It may be that both policies and increased funding to assist in the reduction of fraud, waste, and abuse have served to better identify perpetrators.  However, it could also be due to the exponential increase in the number of physicians participating in public insurance in the U.S.  After the passage of the ACA (Affordable Care Act), enrollment increased by 12.6% between 2013 and 2017, higher than the 7.9% increase in private insurance participation.

Adding a layer of study to include outliers that possess characteristics of those that may be more likely to engage in fraudulent activities may assist with reducing false-positives when looking at Fee-for-Service medical claims data alone. This would allow agencies to not only analyze anomalies and properly allocate investigative resources, but to best identify those patterns that warrant enforcement action.

It is obvious that individuals matching a single criterion, as well as those that match each and every one of the criteria listed, may have no aberrant billing patterns and be models of ethical behavior.  However, the value derived from the identification of patterns used by those that have previously committed fraud cannot be understated.  This is true not only in law enforcement for gathering leads to support criminal prosecution, but also in identifying traits belonging to those more likely to engage in healthcare schemes and violations.  It is then up to the individual investigator and their respective agency to follow-up and either prove or disprove the alleged behavior.