Experimental Health Economics
Secondary Markets and the Opioid Crisis: Evidence from a Laboratory Experiment
Yue Deng, George Mason University
The opioid crisis is responsible for hundreds of thousands of deaths and trillions of dollars in costs. The secondary market for opioids contributes substantially to those numbers. While shutting down the secondary market can alleviate the health threat induced by the drug diversions, it could also trigger increased unnecessary prescriptions, leaving the welfare consequences of closing secondary market distribution ambiguous. Drawing on Schnell’s (2017) model of secondary markets and the opioid crisis, we design a laboratory experiment to investigate the effect of secondary markets on patient and physician behaviors. We find that, as compared to when the secondary market is absent, in the presence of a secondary market patients visit physicians more frequently and physicians provide more prescriptions. Consequently, we find that shutting down this distribution channel reduces total consumption of opioids, and positively impacts overall health outcomes. Our results provide clear evidence that policies aimed at restricting secondary markets would significantly contribute to mitigating the opioid crisis.
Yue Deng, George Mason University
Daniel Houser, George Mason University
The effects of acute stress on risk and time preferences. Can mindfulness meditation help?
Natalia Bulla, Universität Duisburg-Essen & CINCH - Forschungszentrum für Gesundheitsökonomik, Institut für Versorgungsforschung und Gesundheitsökonomie am Deutschen Diabetes-Zentrum
Stress influences decision-making processes and is one of the drivers of changes in economic preferences. Many techniques have been suggested to tackle stress, ranging from exercising to medical counselling. One of the novel techniques is mindfulness meditation, which aims to put the focus of the individual on the present moment and should be able to help manage stress. With this in mind, we investigate whether stress impacts individual risk and time preferences, and whether a brief mindfulness breathing exercise affects preferences as well. We conduct a controlled laboratory experiment with a student subject pool. We aim to reduce the level of individual stress with a mindfulness meditation task, while increasing stress via a cognitive load task. As a measure of stress, we track participants' heart rates in a continuous manner during the experimental session. Our preliminary results show that a brief mindfulness task reduces the average heart rate for participants who were exposed to stress by 3.8% and those who were not by 3.4%. So far, we do not find effects of the brief mindfulness breathing exercise on risk and time preferences.
Nadja Kairies-Schwarz, Universität Duisburg-Essen & CINCH - Forschungszentrum für Gesundheitsökonomik
Irene Mussio, Newcastle University (UK)
The Effects of Work Atmosphere on Physician Provision Behavior
Claudia Soucek, Universität Duisburg-Essen/CINCH
A detrimental question for health care policy makers is how to improve the quality of care provided to patients. Over the last decade there has been a focus on reforming financial incentives for physicians. Particularly, performance incentives have gained increasing relevance to raise the quality of care. Yet, the effects of these pay-for-performance incentives on the quality of care are, if at all present, rather modest and only temporary. Moreover, physicians tend to be more intrinsically or prosocially motivated towards their patients than the average employee. In contrast, there has been less focus on the non-monetary extrinsic factors in physicians’ daily practice that may affect their intrinsic or prosocial behavior and thus their provision behavior. Physicians are increasingly faced with bureaucracy in the form of documentation and time pressure when treating patients. These extrinsic factors may crucially affect their work motivation and treatment decisions.
In this study, we investigate how additional documentation tasks and time pressure affect hospital physicians’ provision behavior. For this, we use a controlled laboratory experiment with a medical framing. Medical students in the role of hospital physicians have to treat patients given a simplified DRG remuneration. We vary whether physicians are faced with documentation tasks, time pressure or both together. In the documentation treatments participants have to move sliders to a certain point while in the time pressure treatments a time limit was introduced by showing a timer that counted down the seconds to complete the tasks.
We observe significant differences in the quantity of medical services provided between the treatments. Preliminary results show that medical students discriminate between patients with high and mild disease severity given that the percentage of patient optimal decisions for high disease severities is higher in all treatments. Furthermore, time pressure and especially the combination of documentation and time pressure lead to significantly less patient optimal decisions. Lastly, this effect is time invariant since participants do experience a learning effect over time for the documentation, but not for the treatment decisions.
Physicians are increasingly faced with bureaucracy in the form of documentation and time pressure when treating patients. These extrinsic factors may crucially affect their work motivation and treatment decisions. To the best of our knowledge there exists no evidence on how documentation and time pressure affect physician provision behavior. Therefore, we aim at conducting a laboratory experiment with a medical framing. Preliminary results show a significant effect of documentation and time pressure on physician provision behavior.
Claudia Soucek, Universität Duisburg-Essen/CINCH
Nadja Kairies-Schwarz, Universität Duisburg-Essen/CINCH
Tommaso Reggiani, Cardiff University
Health Insurance Choice and Risk Preferences – Experimental Investigation of the Impact of a Preference-Based Insurance Finder
Benedicta Hermanns, HCHE/Universität Hamburg
Deciding on a health insurance contract can be a challenging task for consumers. In order to cope with complexity and demonstrating more consumer-friendliness, internet portals nowadays often provide health insurance finders where certain attributes can be selected. Based on the individual selection, the respective contracts then get displayed, ranked or highlighted. While there is several field evidence on the effects of such insurance finders on health insurance choice, the data is often prone to selection bias of the individuals using the finder. Moreover, it is very difficult to infer individuals’ risk preferences form choices made and thus evaluate their choices based on the latter. This study aims to complement the existing field evidence and examine experimentally which impact an insurance finder has on the choice of health insurance contracts when accounting for individually elicited risk preferences.
For this, we implemented a controlled laboratory experiment with a sequential design in which subjects (N=253) had to decide on health insurance contracts first and then went through a series of lotteries to elicit individual risk preferences. We varied the presence of an insurance finder. In the first part, subjects could suffer from six different illnesses with predetermined probabilities and had to select their preferred contract out of a set of six or twelve different contracts, in total twelve different choice sets. These designed contracts differed with regard to premium, deductible for the basic coverage and additional coverage of three other illnesses. An insurance finder was provided for the treatment group. This allows the subjects to select a maximum deductible and desired additional coverage such that contracts corresponding to their individual preferences are highlighted. In the second part, prospect theory risk preferences were elicited for all subjects through 70 lottery choices using the so-called trade-off method. This leads to individual utility and weighting functions. This enables us to classify the subjects based on their estimated parameter values and to evaluate the quality of health insurance choice in the first part of the experiment.
On average, the finder is used in more than half of the decisions and the greater the complexity of the decision the more it is used. Using the finder leads to decisions that differ in terms of preferences for contract attributes, e.g. it leads to a higher chance of choosing a contract that covers an illness with low risk but very high costs. However, our first results indicate that it does not necessarily result in a higher decision quality.
This study provides controlled laboratory evidence that a decision-making tool, while frequently used, may not necessarily improve decision quality.