Just What The Doctor Ordered: Tort Reforms and Hospital Imaging Utilization

Defensive medicine is considered to be a substantial driver of overutilization of medical services. As both doctors and hospitals can be held liable for injury and other adverse consequences, there are strong incentives to perform too much medical care. Many states have engaged in tort reforms in an effort to reduce the practice of defensive medicine, but it remains unclear how much tort reforms influence medical utilization, especially as findings vary across clinical areas and types of reforms. Further, identifying whether changes in utilization are related to tort reforms, and not physician decision-making about patient health, can be difficult to identify. This is particularly true for imaging orders, as imaging always provides additional information to providers even if this additional information was not needed to reach a diagnosis.

The Centers for Medicare and Medicaid Services (CMS) has identified a set of low value imaging services which are particularly susceptible to overutilization. One of these services, same day brain and sinus computed tomography (CT) scans, may provide an avenue to examine the effectiveness of tort reforms, because the clinical protocol requires a provider to decide in advance whether to perform just a brain CT or, alternatively, a brain CT followed by a same day sinus CT. By linking patient level Medicare claims data from 2011-2013 to the Database of State Tort Law Reforms, we are able to exploit variation in the passage of four major tort reforms (i.e. caps on total damages, collateral source reform, joint and several liability reform, and contingency fee reforms) to identify their effect on the utilization of low value same day sinus CT.

We employ two empirical approaches to identify the effect of tort reform: a patient level logistic difference-in-difference (DD) regression model and a synthetic control strategy. Our preliminary DD results suggest that the utilization of medical services is sensitive to tort reforms but that the direction of the effect varies by the type of reform. In particular, we find that caps on total damages reduces the likelihood of same day CT scans (-0.003; p = 0.01), while joint and several liability reform increases the likelihood of same day CT scans (0.003; p < 0.01). These results shed new light on how tort reforms can impact the utilization of medical services