Three Essays on Medicare Quality Initiatives

My first chapter explores the relationship between readmission reduction efforts and hospital costs. A total hospital operating cost function is estimated using
over 5,000 observations from 2,129 US hospitals from the period 2012 - 2017. Using these cost estimates, I estimate a hospitals marginal cost of a 1% readmission
reduction for a single monitored disease. The average marginal cost of reducing risk-adjusted readmission rates for monitored diseases varies from $1,186,689 to $3,844,643. Significantly higher marginal cost are found for hospitals
with the highest number of dual-eligible patients, with hospitals spending up to an extra $839,027 to reduce readmission rates. These results contribute to
the growing literature on the burden of quality incentive programs on hospitals serving disproportionately low-income populations. The second chapter adds to the growing literature on the Hospital Readmission Reduction Program by describing the financial incentives faced by hospitals, estimating their magnitude and distribution, and testing whether hospitals facing larger financial incentives are more likely to improve performance. I estimate the magnitude of the expected future penalty for one additional readmission
across hospitals and procedures and find that on average hospitals can expect a penalty increase two periods in the future for one additional readmission today of: $27,906.74 for an additional AMI readmission, $39,161.94 for heart
failure, and $30,574.30 for an additional pneumonia admission. I find evidence that hospitals improve their readmission rates over time for the monitored conditions
for which they have the highest marginal incentives to improve. I also find evidence that approximately 30% of hospitals have no incentive to improve performance on any condition in a given year. In the third chapter, I investigate the effect of observed hospital quality measures
on patient demand for elective procedures. Using patient-level data from the state of Florida, I estimate a multinomial logit demand model using patient comorbidities and distance between patient zipcode and hospital zipcodes to
identify the effect of a marginal decrease in Hip and Knee Replacement complication rates on hospital demand. Previous literature has investigated the impact
of changes in readmission and mortality rates on hospital demand, but have not looked into complication rates. The findings indicate that patients have a significant
willingness to travel for improved quality measures, including lower complication rates for elective hip and/or knee replacement, lower 30-day readmission rates and lower in-hospital mortality rates for patients with serious treatable
conditions. Patient preference heterogeneity inputs older patients being less willing to travel further distances.