Thirty-day readmissions--truth and consequences.

tals would then have a direct incentive to invest in reducing readmissions, and Medicare would benefit from immediate savings and progressively lower rates. Our approach would measure a hospital’s readmissions performance against its own historical record and population, avoiding the need for complex risk adjustment or adjustments for readmissions to nonindex hospitals, since they would already be included in each hospital’s baseline rate. Medicare would provide extra payment in accordance with a hospital’s historical (perhaps 3-year rolling average) readmission rate, for either all cases or a large subgroup of diagnoses in which readmissions are potentially avoidable. Britain, Germany, and Maryland, which runs its own all-payer hospital rate-setting program, have all adapted the DRG payment approach in this way, so they pay nothing for many or all readmissions within a specified period. We acknowledge that this approach has the perverse effect of further rewarding all admissions other than readmissions that fall within the 15or 30-day window; a recent study showed a strong association between rehospitalization rates and overall admission rates.5 Only fundamentally different payment methods, such as population-based, global payments, would avoid the volumegenerating incentives of episodebased approaches. We doubt that the marginally higher payment rates would substantially alter the already strong incentive for hospitals to fill beds. Increasing penalties for readmissions, but not all admissions, provides a clearer incentive for hospitals to pursue readmissions-reduction programs and enhance quality and efficiency, while more fundamental payment reform is being explored. Disclosure forms provided by the authors are available with the full text of this article at NEJM.org.