Hospital Reimbursement in the Presence of Cherry Picking and Upcoding

• 成果类型: Article
• 署名作者: Savva, Nicos; Debo, Laurens; Shumsky, Robert A.
• 署名单位: University of London; London Business School; Dartmouth College
• 刊物名称: MANAGEMENT SCIENCE
• ISSN/ISSBN: 0025-1909
• DOI: 10.1287/mnsc.2023.4752
• 发表日期: 2023
• 页码: 6777-6799
• 关键词: yardstick competition; credence goods; hospital regulation; upcoding; cherry picking

摘要

Hospitals throughout the developed world are reimbursed based on diagnosis-related groups (DRGs). Under this scheme, patients are divided into clinically meaningful groups, and hospitals receive a fixed fee per patient episode tied to the patient DRG. The fee is based on the average cost of providing care to patients who belong to the same DRG across all hospitals. This scheme, sometimes referred to as yardstick competition, provides incentives for cost reduction, as no hospital wants to operate at a higher cost than average, and can be implemented using accounting data alone. Nevertheless, if costs within a DRG are heterogeneous, this scheme may give rise to cherry-picking incentives, where providers drop patients who are more expensive to treat than average. To address this problem, regulators have tried to reduce within-DRG cost heterogeneity by expanding the number of DRG classes. In this paper, we show that even if cost heterogeneity is eliminated, such expansion will fail to completely eliminate patient cherry picking. In equilibrium, the market will bifurcate into two groups, one of which will continue to cherry-pick patients and underinvest in cost reduction, whereas the other group treats all patients. Furthermore, we show that DRG expansion is particularly problematic if hospitals are also able to upcode patients, that is, intentionally assign patients to a more resource-intensive DRG than needed to increase income. Upcoding increases within-DRG cost heterogeneity and amplifies cherry-picking incentives. We examine potential solutions involving yardstick competition based on input statistics.