A study at the University of Michigan Health System confirms something physicians have been complaining about for years: When a hospital, especially a large regional hospital, accepts transferred patients, outcome and benchmark measures can be adversely affected.
These findings point to what could be a doubly deceiving situation, says lead author Andrew Rosenberg, M.D., assistant professor of anesthesiology and internal medicine at UMHS. Not taking into account the condition of very sick transfer patients not only diminishes quality rankings at the receiving hospital, but it also can inflate the transferring hospital's scores.
"If a hospital that doesn't accept transfer patients has average outcomes, then the question is why--why aren't they doing better?" Rosenberg asks.
The study, published June 3 in the Annals of Internal Medicine, examined 4,579 admissions to the University of Michigan adult medical ICU; 25% were transfers to UMHS from other hospitals. The study's data was obtained from the university's clinical databases and from APACHE (Standard Acute Physiology Age and Chronic Health Evaluation) Medical Systems in Vienna, Va.
The APACHE data included an admission source correction. However, the study concluded that a hospital that gets 25% of its ICU patients as transfers from other hospitals would show an extra 14 deaths for every 1,000 admissions, as compared with a hospital that accepts no ICU transfer patients and provides exactly the same quality of care. The 1.4% difference is enough to drive down the hospital's score, Rosenberg says.
In a world where hospital rankings and quality surveys steer healthcare decisions, Rosenberg concludes that transfer rates ought to be routinely considered. Two companies that calculate ratings say they are.
"If you have a hospital that typically has large numbers of older, urban patients who have heart disease, diabetes and more than one bypass surgery, the hospital will be credited with the fact that many of its patients have higher mortality rates than a hospital where the average patient is a relatively healthy 65-year-old who walks in the door," says Sarah Loughran, senior vice president of provider services at Healthgrades, a Web-based hospital grading service in Lakewood, Colo.
Solucient, which annually ranks the top 100 hospitals, factors whether a patient is being admitted from an acute care facility or elsewhere into its ratings because this affects complications and mortality, says Janet Young, M.D., clinical support specialist for the Evanston, Ill.-based company.
She says there might be merit in factoring in whether a patient failed treatment elsewhere. "But practically speaking, the only place you can find that is in the patient history," not in the coded medical records on which the rating system algorithms rely.
Rosenberg disputes the contention of Healthgrades and Solucient that transfer data is now adequately considered. He points out that their evaluations are based on data provided by CMS. "There is a lot of information about the inadequacies of that kind of data service," Rosenberg says. "I can't find any use of transfer effect in the methodology they provide. I frankly doubt that they have good transfer data at all."
Jennie L. Phipps is an independent writer in Farmington Hills, Mich.