David Chan

Papers

Organizational Structure and Moral Hazard among Emergency Department Physicians (Job Market Paper)
Appendix A (Empirical Appendix) here
Appendix B (Theoretical Appendix) here
How does organizational structure affect physician behavior? I investigate this question by studying emergency department (ED) physicians who work in two organizational systems that differ in the extent of physician autonomy to manage work: a "nurse-managed" system in which physicians are assigned patients by a triage nurse "manager," and a "self-managed" system in which physicians decide among themselves which patients to treat. Taking advantage of several sources of quasi-random variation, I estimate that the self-managed system increases throughput productivity by 10-13%. Essentially all of this net effect can be accounted for by reducing a moral hazard I call "foot-dragging": Because of asymmetric information between physicians and the triage nurse, physicians prolong patient length of stay in order to appear busier and avoid getting new patients. I show that foot-dragging is sensitive to the presence of and relationship between peers. Finally, I show evidence consistent with theory that predicts more efficient assignment of new patients in the self-managed system.

Learning and Authority: The Dynamics of Housestaff Physician Practice Styles
Although a large literature documents variation in medical spending across areas, relatively little is known about the sources of underlying provider-level variation. I study physicians in training ("housestaff") at a single institution and measure the dynamics of their spending practice styles. Despite similar training experiences, housestaff physicians exhibit significant and growing variation in practice styles throughout their training. The increased variation doubles discontinuously as housestaff change informal roles at the end of the first year of training, from "interns" to "residents," suggesting that physician authority is important for the size of practice-style variation. Practice styles rapidly become more stable over time, especially with greater authority as residents, but somewhat less stable with less frequent practice. Consistent with highly individualized learning, practice styles are poorly explained by summary measures of training experiences. However, rotating to an affiliated community hospital decreases intern spending at the main hospital by more than half, reflecting an important and lasting effect of institutional norms. Pre-training characteristics as measures of intrinsic heterogeneity show little association with practice styles.

Publications

Chan DC, Gruber J. How Sensitive Are Low Income Families to Health Plan Prices? American Economic Review Papers and Proceedings 2010 May; 100(2):292-6.

Chan DC, Shrank WH, Cutler D, Fischer MA, Brookhart MA, Avorn J, Solomon D, Choudhry NK. Patient, Physician, and Payment Predictors of Statin Adherence. Medical Care 2010 Mar; 48(3):196-202.

Jha AK, Chan DC, Ridgway A, Franz C, Bates DW. Improving Safety and Eliminating Redundant Tests: Cutting Costs in US Hospitals. Health Affairs 2009; 28(5):1475-84.

Chan DC, Pollett PK, Weinstein MC. Quantitative Risk Stratification in Markov Chains with Limiting Conditional Distributions. Medical Decision Making 2009; 29:532-540.

Chan DC, Heidenreich PA, Weinstein MC, Fonarow GC. Heart Failure Disease Management Programs: A Cost-effectiveness Analysis. American Heart Journal 2008; 155(2):332-8.

Research in Progress

Cherry Picking and Matching in the Emergency Department
Do physicians choose patients that they are better at seeing, or do physicians all simply prefer and compete for the same types of patients? I estimate emergency department (ED) physician preferences by the likelihood that physicians choose a given patient type, when given the choice, and then determine whether patient outcomes are better for these patients when they are assigned patients of a similar type. I find that preferences are associated with improved efficiency in terms of length of stay when assigned patients with similar types. However, physician preferences are relatively homogeneous. As future research, I will account for rich information in “chief complaints,” which are highly informative free-text reasons of a patient’s visit and likely reflective of true heterogeneity. I will investigate whether outcomes are better when physicians with similar or different preferences are working together.

Clocking Out: Behavioral Effects Associated with Shift Work
Work is often performed in shifts, yet little is known about behavioral responses to the timing of work relative to the shift. I study emergency department (ED) physicians who work in shifts with patients arriving exogenously to their shift times and evaluate throughput, quality, and financial outcomes. Physicians are expected to stay until they complete seeing all patients, even if work lasts past the end of their shift. Physicians assigned patients at the end of their shifts discharge the patients 40% sooner, with a minor decrease in revenue and costs per patient. The distribution of discharges suggests that physicians time their discharges right before the end of their shifts. Physicians also avoid choosing patients near the end of their shift, if allowed to choose. As future research, I will attempt to identify mechanism(s) behind these behavioral effects, such as fatigue and moral hazard. I will explore whether behavioral effects are sensitive to social incentives and whether it is heterogeneous among physician types.

The Effect of the Doctor-patient Relationship on Birth Outcomes (with Daniela Carusi, Erin Johnson, and Marit Rehavi)
Access to an outpatient physician is thought to be important for quality care. This study tests whether this is because the doctor-patient relationship allows physicians to attain better outcomes for patients that they know. Using quasi-random variation in the timing of unscheduled deliveries, we ask whether patients who are delivered by their prenatal obstetricians have different outcomes than those who are delivered by another physician in the group with access to their patient records. We evaluate outcomes such as cesarean sections and complications from delivery. While non-prenatal obstetricians do not have a previous relationship with the patients they are delivering, prenatal obstetricians may face a pressure to deliver before their shift’s end.