Putting pharmacists in charge of medication management as patients leave the hospital seems like a good idea. Does it work?

A great deal can go wrong after people have discharged from a hospital, especially if they are taking a lot of medications, says Joshua Pevnick, M.D., M.S.H.S., an associate professor of medicine at Cedars-Sinai Medical Center in Los Angeles.

Prescriptions may get started. Others may get stopped. Maybe the hospital pharmacy doesn’t dispense the particular brand of medication that a patient is taking. A switch is made, but it turns out that the patient’s insurance doesn’t cover that brand or covers it at a much higher out-of-pocket cost. Meanwhile, people who have been recently hospitalized are in a medically vulnerable state, recovering from an illness, trauma or surgery of some kind.

“One way of summing it up is that these are the sickest people we have that are not in the hospital, and they’re very prone to having an adverse drug event. They’re on a lot of medications, and we’ve just mucked with everything,” said Pevnick in an interview with Managed Healthcare Executive.

How prone is subject of some debate. Some research has put the rate of posthospitalization adverse drug events as high as 37%, but some recent research suggests that it is closer to 5% to 6%. Either way, it is a significant problem with cost implications for insurers and health systems if the adverse event leads to hospital readmission, to say nothing of the risk of harm to recovering patients’ health.

Some research has suggested that more deliberately inserting pharmacists into the many-moving-parts process of discharging patients from the hospital could help iron out many of the problems that arise with prescription drugs. Pevnick, who is co-director of the Division of Clinical Informatics at Cedars-Sinai, teamed up with Jeffrey L. Schnipper, M.D., M.P.H., M.H.M., a professor of medicine at Harvard Medical School and a leading health services research director at Harvard-affiliated Brigham and Women’s Hospital in Boston, to conduct a pragmatic trial of a pharmacist-led discharge process focused on medication management. Pevnick said that putting doctors or nurses in charge of medication management can lead to inconsistency. Pharmacists have the appropriate training, he said, and it helps that they are typically employed by the hospital, so the tasks can be made part of their responsibilities. “If you want to build the system, you need to have someone who not only has the time and the training, but they need to be accountable to do it,” he said, noting, though, that pharmacists don’t prescribe or stop prescriptions.

Pevnick and Schnipper’s was originally designed to have an age cutoff of 65 and older. It was lowered to 55 because of low enrollment, a consequence, at least in part, of the COVID-19 pandemic (the researchers started enrolling patients in December 2019). Some research has shown that posthospitalization adverse drug events occur mostly involving anticoagulants, antiplatelet agents, antihyperglycemics or opioids. Pevnick said he and co-investigators decided opioids were a separate problem. They also decided that the sheer number of prescriptions should be factored into the study, not just drugs from three classes of medications. As a result, the inclusion criteria included the use of 10 more medications; three or more prescriptions for anticoagulants, antiplatelet agents or antihyperglycemics; or both.

Pevnick and his colleagues enrolled approximately 6,300 patients into the study. Half received the pharmacist-led medication management intervention, and the other half, usual care. The intervention had multiple components, including three predischarge ones (medication review, patient education and discharge medication reconciliation), three postdischarge ones (communication with the outpatient pharmacy, communication with the primary care physician and a postdischarge follow-up telephone call with the patient), and optional ones, including motivational interviewing and drug-level monitoring. A sample of 40 hospitalizations suggested that interventions took the pharmacists just over an hour to complete and cost $101 per hospitalization. The primary end point Pevnick and his colleagues settled on was the proportion of patients with unplanned hospitalizations or emergency department utilization within 30 days after hospitalization. Their calculations before the study suggested that they would need a sample size of 9,776 patients to detect an absolute difference of 2.5% from the expected baseline of 27.5% of patients having unplanned hospitalizations or emergency department utilizations.

Results of the study, which was funded by the National Institute on Aging and other organizations, were published Monday in JAMA Network Open. In the introduction of the paper, Pevnick and his colleagues wrote that they were seeking evidence “to convince hospital and health system leaders to fund pharmacist-led TOC [transitions of care] interventions.” The results, as it turns out, may work against their goal because the primary end point was not met. Among all the patients whom the researchers randomly sorted into the usual and intervention groups, 19.5% (606 of 3,112) of those in the usual care group had an unplanned hospital utilization compared with 18.5% (579 of 3,126) of those in the intervention group, a difference that did not meet the standards of statistical difference. Another slice of the data that was limited to people in traditional Medicare but encompassed all hospital utilization showed a 26.4% (593 out of 2,243) rate in the usual care group compared with a 25.6% (570 out of 2,230) rate in the intervention, a difference that also did not meet the standards of statistical significance.

Pevnick and his colleagues offered five reasons for the results falling short on the primary end point, including under enrollment and “imperfect fidelity” to the intervention. In the interview with MHE, Pevnick noted that this was a pragmatic trial, designed in a way so the results can be translated more easily into real-world settings. “Some of the things that go along with that is we’re not ensuring that the intervention is done perfectly every time,” he said. Another factor that may have affected the outcome is that the patients in both the control and intervention groups received “best possible” medication history and reconciliation services upon admission. As a result, the difference between the control and intervention groups might have narrowed

Before the study was conducted, Pevnick and his colleagues had decided that their analysis of the results would include certain subgroups. One of those subgroups consisted of people who, according to a questionnaire they had filled out, had low medication adherence and literacy scores. In that group, which consisted of 589 patients, or roughly 10% of the study population, the intervention was associated with a 10 percentage point difference in hospitalization utilization (28.8% vs. 18.3%).

The authors in the JAMA Network Open paper and Pevnick in the interview with MHE wanted, understandably, the positive results in the MedAL subgroup highlighted as the study’s salient finding, not the shortcoming on the primary end point for the overall population.

“My takeaway for future researchers and people who are in operations would to be concentrate on this group,” Pevnick said. “If I were doing the trial again, I would say, ‘Let’s do it entirely in patients with low medication adherence and literacy’ and that’s where I would look for the effect.”