Higher hospital spending linked to better outcomes

Bucking a widely accepted correlation between health spending and patient outcomes, a recent study finds hospitals that spend more than average may be getting better results.

Heart failure patients' survival rates in the 180 days after hospitalization improved significantly in facilities with higher hospital spending, according to a study by California academic researchers published online Oct. 13 in the journal Circulation: Cardiovascular Quality and Outcomes.

The study also found significant variation in both hospital spending and mortality rates for the six California teaching hospitals in the study, based on their care of 3,999 Medicare patients who were hospitalized with a principal heart failure diagnosis between Jan. 1, 2001, and June 30, 2005. The highest-spending hospital used nearly double the resources of the lowest-spending hospital, and mortality varied by more than 50%.

"The more difficult thing is, what do you do about that variation?" asked lead author Michael Ong, MD, PhD, an assistant professor of medicine at the David Geffen School of Medicine at the University of California, Los Angeles. "We can't say definitively that spending more is going to lead to better outcomes."

The study focused on the five University of California teaching hospitals and Cedars-Sinai Medical Center in Los Angeles, which is academically affiliated with UCLA. The authors did not identify how individual hospitals performed.

The study was prompted in part by Dartmouth Atlas Project conclusions that Medicare patients in higher-spending regions did not receive significantly better care than did patients in lower-spending regions. Leaders at UC wanted to understand differences between some of its higher-spending and lower-spending facilities as tracked by the Dartmouth research, Dr. Ong said.

The new study's authors also wanted to take a wider look at hospital care, according to Dr. Ong. Some of the better-known Dartmouth Atlas research on hospitals focused primarily on patient care during the last six months of life, he said. The California researchers said that time frame does not provide a complete picture of hospitals' performance, especially on mortality.

Attempts to reconcile the two studies

Differing risk-adjustment methods may help explain the varying results, suggested Jonathan Skinner, PhD, a Dartmouth Atlas author and professor of economics at the Dartmouth University Dept. of Economics.

Dr. Ong said more research is needed on spending and quality. "There's always issues about risk adjustment that come up."

Skinner also said the University of California study's six-hospital sample is not large enough to draw national conclusions. "While this paper is a promising first step, only further research can address that issue."

Dr. Ong said the Dartmouth researchers had produced great work on variation over the years.

The newest study provides interesting insights at a local level, said Nancy Foster, vice president for quality and patient safety policy at the American Hospital Assn. She sees the articles as companion pieces. The Dartmouth work, to use her analogy, shows how schools' test scores compare with the national average, while the California investigation details how specific students are performing.

Recent criticism of the Dartmouth Atlas work has prompted its team to be more conservative on its estimates of waste in the nation's health system. Their previous research concluded that up to 30% of all U.S. health spending may not produce better outcomes, but Skinner said 20% may be a more accurate estimate.

Medical liability reform and advance care directives would cut some of the variation and excess spending in emergency medicine, said Sandra Schneider, MD, president-elect of the American College of Emergency Physicians.

Only about 10% of her patients have advance directives, about 30% to 40% have talked about it, and the rest haven't discussed it. "It's clearly a difficult discussion to have," she said. "It's not something you want to bring up at the holiday party."

Improving care guidelines also would reduce some of the variation, Dr. Schneider said. She noted, however, that no protocol will work for every patient and situation.