Monitoring Institutions in Health Care Markets: Experimental Evidence

A key characteristic of health care markets is the information asymmetry between patients and physicians. Physicians know more about the disease and the appropriate treatment than patients. This may result in three forms of physician misbehavior: providing more treatments than necessary, i.e. overtreatment; providing less treatment than necessary, i.e. undertreatment and charging more treatments than provided, i.e. overcharging. Patients have to trust in physicians that they receive appropriate treatment. This is why health services are often referred to as credence goods (Darby and Karni, 1973; Dulleck and Kerschbamer, 2006).

Monitoring physicians’ treatment decisions has gained more and more attention in the past two decades. The German government, for example, introduced the “Medical Service of Health Insurers“ (Medizinischer Dienst der Krankenversicherung) that reviews and enforces physicians’ treatment and charging behavior. Similarly, the importance of patient protection agencies has risen throughout the industrialized nations. We capture these recent developments and investigate whether monitoring of physicians can help reduce mistreatment in health care with the use of a laboratory experiment.

The experimental design is based on the credence goods framework established by Dulleck and Kerschbamer (2006) and Dulleck et al. (2011). In total, 6 conditions of market interactions with 280 undergraduate students either in the role of physicians (N=140) or patients (N=140) were conducted. In the baseline condition (B) we assume that patients can neither verify the type of treatment received nor is the physician liable for providing the appropriate treatment. In the other conditions a monitoring institution is introduced either endogenously (M-End) where the patient decides whether to exert costly monitoring or exogenously where a third party exerts monitoring for a random sample of physicians with three different probabilities of monitoring and one variation in information provision: (i) 65% (M-Ex65), (ii) 35% (M-Ex35), (iii) 10% (M-Ex10) and (iv) 10% without information on monitoring probability to participants (M-ExUnk).

In line with our predictions, we find that monitoring is an effective institution to reduce physician misbehavior. Compared to the baseline condition, the level of undertreatment significantly reduces by 31, 28, respectively 24 percentage points for M-End, M-Ex65 respectively M-Ex35. The level of overcharging is significantly lower for M-End, M-Ex65 and M-ExUnk than for B. There is virtually no overtreatment in any of the conditions. Efficiency measured as the sum of patient and physician surplus increases significantly for endogenous and marginally significantly for exogenous monitoring compared to no monitoring. A monitoring frequency of 10% with common knowledge of the monitoring probability has no impact on physician behavior compared to the baseline condition, whereas an unknown monitoring probability of 10% reduces undertreatment in the first periods and overcharging overall periods.