Price Transparency in the Health Care Sector

This dissertation consists of three essays on price and insurance coverage transparency in the health care market. In these essays, I empirically examine how providing patients with information about health care prices and insurance coverage characteristics affects both the utilization and prices of health care and then, in turn, affects the physician's decision to disseminate this information. To this end, I study the consequences of a specific Medicare document that Part B physicians use to warn patients about expected out-of-pocket costs prior to receipt of medical care. In the first chapter, "The Effect of Health Care Price Transparency on Utilization and Prices'', I exploit a March 2008 reform to the Medicare document that required providers to provide additional charge price information to patients at the time of service to estimate the effects of price transparency on utilization and prices themselves. I argue that this reform affected certain providers and not others and use this insight to implement a difference-in-differences empirical strategy with Medicare administrative claims data. I find that, for preventive gynecological procedures, this type of price information given directly to patients reduces the number of claims submitted to Medicare for reimbursement and reduces the prices providers charge. These results uncover a potential downside of increased price information that should be considered along with the previously established benefit of incentivizing patients to switch to low-cost providers. Price transparency has the potential to reduce utilization of medical care, and, depending on the value of the medical care that is reduced, the welfare consequences are unclear. In the second chapter, "Information Disclosure in the Presence of General Uncertainty: Evidence from Medicare Part B'', I document surprising patterns of physician use of the Medicare document. The strategies can be explained by the presence of general uncertainty over the accuracy of the information and the optimal response of the distributor to the demand consequences of making an ex-post "mistake". I show that a substantial number of physicians make ex-post mistakes when giving this document; they do not give this document to patients who experience an ex-post non-coverage and vice versa. In March 2008, the informational content of the document was changed in such a way that shifted the demand consequences of ex-post mistakes differently across physicians. I show that the response of physician strategies are consistent with the presence of high costs in discerning which patients will be denied. Using Medicare claims data from before and after this reform, I find that physicians restrain their distribution strategies accordingly, but the probability of making an ex-post mistake remains unchanged. These results imply that the presence of general uncertainty plays a nontrivial role in explaining physician disclosure behavior, and that physicians strategically respond to the demand consequences of revealing price information to patients. In the third and final chapter, "The Effect of Insurance Coverage Information on Utilization of Health Care for New Medicare Enrollees '', I investigate the association between receipt of salient Medicare coverage information soon after enrollment on future utilization of medical care. In particular, I use the Medicare document to examine the effect of being warned within the first six months of enrollment that Medicare does not cover all procedures on total charges and procedure counts in the subsequent two years. The results indicate that this warning is associated with fewer submitted charges and fewer procedures performed. For immediate policy relevance, the results suggest that early patient experience with new health care plans can be important in determining future health care spending. However, I find suggestive evidence that vulnerable groups, the poor and less-educated, refrain from receiving covered care in addition to non-covered care after receipt of this information, arguably due to the high cost of disentangling the two. This indicates that the information can be effective in reducing spending, but this benefit is mitigated by reductions in necessary care by disadvantaged populations