This paper asks how hospitals improved the quality of care in response to a Medicare program that penalized excess readmissions of elderly heart attack patients. I find that hospitals reallocated heart attack patients toward cardiologists and toward doctors who, before the program, had larger heart attack caseloads, higher treatment intensity, and lower readmission and higher survival rates adjusted for patient health. Heart attack care also became more specialized, and these same doctors were increasingly assigned patients appropriate for intensive treatment, indicating better doctor-patient matches. Because many physicians are not hospital employees, I also examine the role of doctor-hospital affiliation. Both the quality improvements and the reallocation were larger in hospitals that employed or contracted with physicians than in hospitals with neither arrangement, suggesting that affiliation helps hospitals pass quality incentives on to physicians and control patient assignments.

We ask how urgent care centers (UCCs) impact healthcare costs and utilization among nearby Medicare beneficiaries. When residents of a zip code are first served by a UCC, total Medicare spending rises while mortality remains flat. In the sixth year after entry, 4.2% of the Medicare beneficiaries in a zip code that is served use a UCC, and the average per-capita annual Medicare spending in the zip code increases by $268, implying an incremental spending increase of $6,335 for each new UCC user. UCC entry is also associated with a significant increase in hospital stays and increased hospital spending accounts for half of the total increase in annual spending. These results raise the possibility that, on balance, UCCs increase costs by steering patients to hospitals. 

This paper develops a structural model of demand for illegal drugs to study how consumers substitute between drugs in response to government policies. We construct a unique dataset by scraping a darknet marketplace that dominated the retail illegal drug trade in Russia. We exploit novel micro-level moment conditions and estimate a median demand elasticity of -1.9. Using the estimated model, we show that after cannabis legalization, for every four additional doses of cannabis, one fewer dose of another drug is consumed. We also find that new synthetic drugs sharply increased total drug demand and that enforcement can raise drug-market revenue.

Research on healthcare provider quality focuses on average outcomes, but how providers perform when operations are suddenly disrupted is far less understood. We study how hospitals and nursing homes respond to one such disruption: major snowstorms. Using Medicare data, we compare the same facilities on the same calendar dates in snowstorm and non-snowstorm years. On the day of a snowstorm, nursing home staffing declines, and admissions and discharges fall sharply in both settings. In nursing homes, this transient shock delays discharges for weeks, extending stays with potential health and financial implications for residents. Delays are largest for short-stay residents, for whom reimbursement is typically highest. Hospitals, by contrast, accelerate some discharges before the storm and resolve delays within days. This pattern is consistent with differences in payment incentives: nursing homes are reimbursed per diem, whereas hospitals receive a fixed payment per admission and bear the cost of extra days. 

Snow routinely disrupts daily life, yet its effects on health and healthcare use have not been studied comprehensively. We link spatially granular daily snowfall data to decades of Medicare mortality and hospital records. Snow increases mortality on the day of a snowfall and for several days after. It also sharply reduces hospital visits on that day and the preceding day, consistent with anticipatory behavior. Although visits rebound afterward, the net effect is negative, suggesting that some care is forgone, with potentially adverse health consequences. In contrast, visits for fractures, cold-related conditions, and heart attacks increase, indicating direct negative health effects. The decline in hospital visits is larger in counties with less historical snowfall, consistent with adaptation, and smaller for patients living in the same zip code as a hospital, consistent with travel barriers. Finally, snow accounts for a substantial share of the well-established cold-mortality relationship and of cold's negative effect on hospital visits. 

Wide variation in hospital quality translates into substantial differences in health outcomes between patients admitted to high- and low-quality hospitals. This paper examines whether the Hospital Compare public reports shifted patients toward hospitals with high reported quality in terms of mortality and unplanned readmissions. Among elderly Medicare patients, the reports led to a reallocation of non-ambulance admissions but did not affect patients arriving by ambulance, who have limited choice of hospital. First, the effect is larger for transfers and referrals than for emergency room admissions, consistent with information being more useful when patients have the time and opportunity to choose. Because doctors typically guide patient choices in transfers and referrals, this pattern also suggests that doctors and patient surrogates are more responsive to quality information than patients themselves. The larger effect for patients with dementia, who often rely on others for decision-making, further supports this interpretation. Finally, quality reports are most valuable when patients have more options: the effect is concentrated in the most urbanized markets.