Academic medical centers, increasingly spurned by insurers for being more expensive than community hospitals, appear to have lower death rates for older adults than other facilities, a US study suggests.

Researchers reviewed millions of records for patients aged 65 and older and insured by Medicare, the US health program for the elderly. They found 8.3% of patients died within 30 days of admission at major teaching hospitals, compared with 9.2% at minor teaching hospitals and 9.5% at community hospitals. “We found, to our surprise, that across a wide range of medical and surgical conditions, patients at teaching hospitals did better – they were less likely to die,” said senior study author Dr. Ashish Jha of the Harvard T.H. Chan School of Public Health in Boston.

“While mortality may not be the only indicator that matters, it certainly is the most important one,” Jha said by email. “We know that short term mortality is driven largely by how well the hospital does in taking care of patients.”

Academic medical centers are often considered more expensive than community hospitals and some insurers have excluded teaching hospitals from their networks in an attempt to control costs, assuming that quality is comparable, Jha and colleagues noted online May 23 in JAMA.

For the study, researchers reviewed records from 21.4 million hospitalizations at 4,483 hospitals nationwide. This included 250 facilities designated as major teaching hospitals with membership in the Council of Teaching Hospitals, 894 hospitals with medical school affiliations designated as minor teaching hospitals, and 3,339 community hospitals.

When researchers looked at hospitals by size, they found teaching hospitals had lower death rates than non–teaching hospitals when the facilities were large, with at least 400 beds, and medium–sized, with 100 to 399 beds.

Among small hospitals with 99 or fewer beds, minor teaching hospitals had lower death rates than community hospitals, the study also found.

Teaching hospitals still had lower death rates after researchers accounted for differences in patients characteristics at the various hospital types, and for other characteristics of the hospitals themselves.

One limitation of the study is that it only included certain Medicare patients, and the results might not be representative of what death rates would look like for people at other ages or with other types of insurance, the authors note.

The lower death rates for teaching hospitals are surprising for two reasons, said Dr. Vineet Arora of the University of Chicago: they have more trainees who might make more mistakes than seasoned physicians and they often treat patients too sick or badly injured to get care at community hospitals who are more likely to die.

“The fact that we did not see this is reassuring,” Arora, who wasn’t involved in the study, said by email.

It’s possible, too, that the added supervision at teaching hospitals ensures that physicians are available around the clock when emergencies arise or patients deteriorate, improving mortality rates by saving patients who might otherwise have died, said Dr. Karl Bilimoria of Northwestern University Feinberg School of Medicine in Chicago.

“This study is critically important in that it shows lower mortality rates at teaching hospitals, thus demonstrating that care is quite safe at major academic centers even when trainees are involved,” Bilimoria, who wasn’t involved in the study, said by email.

“Mortality is the bottom line of health care,” Bilimoria added. “It is critically important and it is also a very fair and well–done metric, so patients should have more faith in that measure of hospital quality than many others.”

—Lisa Rapaport