LAS VEGAS – Tired of getting a patient medication list that isn’t reliable? A 3-year study on medication reconciliation could offer a roadmap for improving that list and reducing potentially harmful errors.
Preliminary results from MARQUIS (the Multi-Center Medication Reconciliation Quality Improvement Study) indicate that hospitals are able to reduce unintentional medication discrepancies by implementing a menu of interventions ranging from training providers to take a better medication history to stationing a designated provider in the emergency department to reconcile the patient’s information with pharmacies and primary care offices. But outside forces, such as problems with an electronic health record (EHR) system, can offset those effects.
Dr. Jason Stein
The 3-year study, which will wrap up in the next few months, is sponsored by the Society of Hospital Medicine, with $1.5 million in funding from the Agency for Healthcare Research and Quality.
In the first phase of the study, researchers identified evidence-based techniques for taking the best possible medication history from hospitalized patients and synthesized them into a toolkit for clinicians. The free toolkit includes how-to videos and pocket cards. The second phase of the study is the mentored implementation of the techniques at five sites: two academic medical centers, two community hospitals, and one Veterans Affairs medical center.
Initial results from two of the five sites show that training on how to take a better medication history and clarity about who should be working on reconciling the medication list can help improve medication discrepancies. But while the interventions are promising, the initial case studies show mixed results.
While one community hospital site was able to lower medication discrepancies significantly, a second community hospital site had a spike in its medication discrepancies as it underwent a problematic implementation of a new EHR system.
Eighteen minutes to win it
At the first site, a Medication Reconciliation Assistant (MRA) program was used to improve the reliability of medication lists. The MRA is stationed in the emergency department and interviews patients about their medications and then verifies that list with the pharmacy, the primary care physician, or the skilled nursing facility. That new medication list is then handed off to the admitting physician. The process usually takes about 18-20 minutes.
They MRA program is staffed by four full-time employees, all pharmacy technicians with experience in retail pharmacies, who work 8-hour shifts throughout the week.
"It doesn’t have to be a pharmacy technician," said Dr. Jason Stein, a MARQUIS coinvestigator and a professor of medicine at Emory University, Atlanta, at the annual meeting of the Society of Hospital Medicine. "But somebody needs to be spending that 18 minutes doing something that roughly looks like this."
At that first site, unintentional medication discrepancies from either history or reconciliation errors dropped from 4.5 per patient at the in the preintervention period to 3.4 per patient. And the total number of potentially harmful discrepancies was reduced from 0.25 to 0.09 per patient.
At the second site, a smaller community hospital, the quality improvement team provided training to front-line providers on medication history taking and counseling at discharge and created a new hospital policy that clarified expectations about who would perform medication reconciliation and when they would do it. And recently, they began stationing a provider in the emergency department 5 days a week to work on medication reconciliation.
But shortly after the second hospital began implementing the MARQUIS interventions, the hospital launched a new EHR system that created problems for medication reconciliation. For instance, with the paper system, the admitting physician was accountable for the initial medication history. But under the electronic system, that accountability was lost. Also, the new EHR was unable to group medications as "continued," "changed," "stopped," or "new," making it difficult to perform discharge medication counseling.
The result was that, after some initial success, unintentional medication discrepancies spiked. In the preintervention period, unintentional medication discrepancies were 2.0 per patient, but they rose to nearly 5 discrepancies per patient after the rollout of the new EHR system. They have since fallen back down to 3.8 per patient. Similarly, the total of potentially harmful discrepancies rose from 0.20 to 1.11 per patient.
The free MARQUIS toolkit is available online.
mschneider@frontlinemedcom.com
On Twitter @maryellenny