Physician Health Crime Deterrence

In 2016, over 4 billion health insurance claims were processed in the US. Although the majority of those claims likely represented health services that were administered, necessary, and effective, a small share of those claims were not only ineffective, but also fraudulent. According to the Federal Bureau of Investigations, these fraudulent claims—which include billing for services not rendered, billing for more expensive services than were actually performed, duplicate claims, unnecessary or excessive services—cost American tax payers nearly $80 billion a year. Of this amount, the Department of Justice only recovered $2.5 billion. While private insurers have implemented mechanisms to better monitor fraud, such as prior authorization and utilization management, publicly funded health insurance programs have less stringent monitoring and remain particularly susceptible to health fraud.

Since Becker (1968), a large literature on the determinants of crime has developed. Many of these studies highlight the difficulty of reducing crime through incapacitation, and a bulk of research shows that sanctions can deter crime (Freeman 1999). It may follow that health care sanctions create similar deterrence effects as police crackdowns, but unlike property or violent crimes, health care crimes are more difficult to detect. Health care delivery is complex, and the information asymmetry between physicians, patients, and insurers masks incidences of fraud. Moreover, medical malpractice insurance often shields a physician from having to face any individual out-of-pocket costs.

In this paper, we identify the impact of physician fraud deterrence by examining utilization and billing behaviors when a physician’s peer is barred from participating in all federal health care programs due to Medicare or Medicaid fraud, as well as offenses related to the delivery of services, patient abuse, and financial misconduct. We identify excluded physicians from the Office of the Inspector General’s List of Excluded Individuals and Entities (LEIE). Relying on a random 20% of Medicare claims data, we construct physician networks by identifying shared patient between fraudulent and non-fraudulent physicians within a given market (i.e., hospital referral region). We utilize a generalized difference-in-difference approach estimated at the physician-year-month level, where the control group consists of physicians who do not share patients with any fraudulent physician. While physician networks are far from random, our identification stems from the unpredictability of the date of exclusion.

We find that providers in a fraudulent physician’s network demonstrate increased total claims, total charges, and claims per admission in the twelve months leading up to the Medicare exclusion date. In the twelve months following the sanction, peer physicians reduce their utilization and billing, and trend toward the control group. Reponses are larger among exclusions that were not court mandated, and hence less predictable, and among tighter physician networks, defined by the number of shared patients a provider has with the fraudulent physician. Our results suggest that Medicare exclusions create a significant deterrence effect from overutilization or overbilling of care.