The Agency for Healthcare Research & Quality is plunging ahead on its comparative effectiveness research program, which produces excellent reviews on a variety of topics. Unfortunately, the Agency does not sponsor much new research on critical issues, but simply conducts meta-reviews which basically report on and analyze the current state of the evidence. Many of the reviews relate not to specific treatments for specific diseases, but to broader structural aspects of medical care. One such recent review examined bundled payment programs, ascertaining their effect on both costs and quality. (AHRQ Report) In the context of the report, bundled payment is a payment for services related to a specific disease, and thus differs from other “global payment” strategies, such as capitation, which pay for care for a particular patient.
The theory behind bundled payments is that they give providers a financial incentive to deliver efficient care and may promote better care coordination when the bundles cross multiple provider types. The biggest concern is that the financial incentives may lead to holding back necessary or appropriate care for financial reasons or to avoidance of high-cost, high-risk patients. In addition to ascertaining the quality and cost effects of existing bundled payment schemes, the authors sought to identify better or worse design features, but found insufficient evidence to do so with any confidence. The researchers found 53 relevant studies covering 18 different bundled payment systems, most of which were in public insurance settings. Fifteen of the systems involved bundling of services over a period of time by a single provider.
The headline finding is that switching from a fee-based or cost-based reimbursement method to bundled payments resulted in lowered spending and utilization, with small variations, positive and negative, in quality measures. The spending reduction appeared to average around 10% and the utilization reductions were on the order of 5-15%. In regard to quality measures, in the same study some measures improved and some declined and across studies, the same measure might show improvement under one bundled payment system and a decline in another. Overall the researchers rated the evidence as weak, but seemed to feel that the effects on spending and utilization were real and likely to be borne out in future experience and research.