Shedding light on 'black box' of inpatient opioid use

The findings, published today in the Annals of Internal Medicine, are among the first to shed light on the little-studied causes and consequences of inpatient opioid prescribing.

"I was surprised by the level of opioid prescribing to patients without a history of opioid use," said lead author Julie Donohue, Ph.D., professor in Pitt Public Health's Department of Health Policy and Management. "About half of the people admitted to the hospital for a wide variety of medical conditions were given opioids. The stability of this prescribing also was surprising. Nationally and regionally, as people have become more aware of how addictive opioids can be, we've seen declines in outpatient opioid prescribing. But we didn't see that in inpatient prescribing."

Previous studies have shown that some surgical and medical patients who fill opioid prescriptions immediately after leaving the hospital go on to have chronic opioid use. Until this study, however, little was known about how and if those patients were being introduced to the opioids while in the hospital.

Donohue and her colleagues reviewed the electronic health records of 191,249 hospital admissions of patients who had not been prescribed opioids in the prior year and were admitted to a community or academic hospital in Pennsylvania between 2010 and 2014.

Opioids were prescribed in 48% of the admissions, with those patients being given opioids for a little more than two-thirds of their hospital stay, on average.

Almost 6% of patients receiving opioids during their hospital stay were still being prescribed opioids three months later, compared with 3% of those without inpatient opioid use. And 7.5% of patients who received opioids less than 12 hours before discharge were still receiving opioids 90 days later, compared with 3.9% of their peers who were free of opioids for at least 24 hours prior to discharge.

Additionally, non-opioid painkillers and anti-inflammatory medications, such as ibuprofen, aspirin or naproxen, were rarely tried before an opioid was administered -- as little as 7.9% of the time for some conditions.

"Inpatient opioid use has been something of a black box," Donohue said. "And, while our study could not assess the appropriateness of opioid administration, we identified several practices -- low use of non-opioid painkillers, continuous use of opioids while hospitalized, opioid use shortly before discharge -- which may be opportunities to reduce risk of outpatient opioid use and warrant further study."

Kennedy, M.S., Christopher Seymour, M.D. M.Sc., Timothy Girard, M.D., M.S.C.I ., Oscar Marroquin, M.D., F.A.C.C., and Chung-Chou H. Chang, Ph.D., all of Pitt; Wei-Hsuan Lo-Ciganic, Ph.D., M.S.Pharm., of the University of Florida; Catherine H. Kim, Pharm.D., of UPMC; and Patience Moyo, Ph.D., of Brown University.

This research was funded by UPMC and Pitt.