Hospital mortality rates analyzed
By LOIS BAKER
Contributing Editor
A study of data from more than 26,000 U.S. hospitals covering outcomes of 38 million patients has shown that people treated in private for-profit hospitals in the U.S. have a greater risk of dying than those cared for in private not-for-profit hospitals.
The study, carried out by researchers from UB, McMaster University in Hamilton, Ontario, and the University of Toronto, appeared last month in the Canadian Medical Association Journal (CMAJ).
The researchers undertook the meta-analysis of mortality statistics to help inform the debate under way in Canada over whether to move into the for-profit health-care-delivery arena.
"Most of the debate so far has focused on economics," said P. J. Devereaux, research fellow in the departments of Medicine and Clinical Epidemiology and Biostatistics at McMaster University and lead author on the study. "The emphasis has been on determining if for-profit hospitals can contain costs and run more efficiently; if having for-profits would create 'two-tier medicine,' and on the potential for foreign investors to become involved and influence Canadian health policy in light of NAFTA (North American Free Trade Agreement).
"What has been missing from this debate is how expansion of private for-profit hospitals would affect patients," he said. "We undertook the study to find out the relative impact of private for-profit versus private not-for-profit delivery of hospital care on patient mortality."
Private not-for-profit hospitals are owned by religious organizations, communities, regional health authorities or hospital boards. For-profit hospitals are owned by shareholders or investors.
To conduct their review, the researchers systematically identified all relevant studies that compared private for-profit with private not-for-profit hospital mortality. They ended up analyzing 15 studies containing data from approximately 38 million patients hospitalized in 26,000 U.S. hospitals between 1982 and 1995. The data were adjusted for confounders such as teaching status of the hospital, the patients' severity of illness and hospitals' case mix.
Results showed that for-profit hospitals were associated with significantly higher mortality.
"Although the relative increase in risk amounts to 2 percent in our analysis, which may seem small, the absolute number is frightening," said Holger Schunemann, assistant professor of medicine and social and preventive medicine, and UB author on the study.
"In Canada, this percent is equivalent to 2,200 deaths yearly, which equals Canadian death rates for suicide, colon cancer or motor -ehicle accidents. The total numbers will be much higher in the U.S.," he said. "They speak for themselves."
The researchers point to pressure to achieve the expected profit for investors as the probable cause of this increase in mortality.
"Heads of for-profit hospitals are rewarded based on meeting or surpassing the profit margin," said Devereaux. When reimbursement comes from the same source as not-for-profits, such as from Medicare, for-profits are faced with trying to achieve the same outcomes while having less to spend on patient care.
"These issues raise concerns that the profit motive causes hospitals to limit care in ways that affect patient outcomes, and our findings suggest such concerns are justified."
The meta-analysis results have significant implications for health-care delivery.
"All data are derived from U.S. studies," said Schunemann. "The results are directly applicable to the American public. Being treated in a private for-profit hospital puts patients at increased risk, and the number of private for-profit hospitals in the U.S. is growing."
Additional researchers from McMaster University were Peter T-L Choi, Christina Lacchetti, Bruce Weaver, Ted Haines, John Lavis, David R.S. Halsam, Mohit Bhandari, Deborah J. Cook, Stephen D. Walter, Maureen Meade, Humaira Khan, Neera Bhatnagar and Gordon H. Guyatt.
Brydon J.B. Grant from the UB Department of Medicine and Terrence Sullivan from the Department of Health Policy Management and Evaluation at the University of Toronto also participated.
http://www.buffalo.edu/reporter/vol33/vol33n29/n2.html
