A new study of a pay-for-performance program at 3 New York City hospitals finds costs for 8 kinds of procedures were reduced without harming outcomes or altering case mix, according to a report by Newswise.

The study, presented at the Annual Clinical Congress of the American College of Surgeons, is one of the first to examine patient outcomes under pay for performance. "There's always a concern that incentive programs may affect the quality of care," said Faiz Y. Bhora, MD, a general surgeon involved in the study. "We found no significant difference before and after implementation of pay for performance in studying the overall outcomes for these procedures."

Conducted at Continuum Health Partners hospitals, the study also found no evidence that physicians in the program avoided sicker, frailer patients to make their outcomes look better. "Our case mix index was no different before or after pay for performance," Dr. Bhora said. "There was no evidence of cherry-picking of cases."

The pay-for-performance program involved gainsharing, in which physicians meeting a variety of quality and efficiency measures can share cost savings on treating patients with the hospital. The study examined 1,768 patient cases at the hospitals from 2007 through 2010.

Researchers investigated gastric bypass, colorectal operations, removal of the appendix and gall bladder, bypass of occluded arteries, carotid artery endarterectomy and complete and partial removal of the lung.

With colorectal operations, for example, the complication rate went from 30 percent before the pay for performance started to 22 percent afterward. For lung operations, complication rates increased slightly while death rates declined slightly. None of the changes in outcomes were statistically significant.

Data for the study came from the American College of Surgeons National Surgical Quality Improvement Program and the Society of Thoracic Surgeons. The college's databank is the first nationally validated program to measure outcomes. It collects data on 136 variables, including risk factors before surgery, variables during surgery and survival and morbidity outcomes for up to 30 days after surgery.

Leigh Page