ASHP 08: Evaluation of collaborative nurse-pharmacist intervention for improving medication reconciliation and patient safety

December 9, 2008

Edward G. Tessier, PharmD, reported on a study he conducted with colleagues in Massachusetts to create a nurse-friendly medication reconciliation process, with the goal of standardized data-gathering and improved patient safety. Dr. Tessier, clinical pharmacist, Baystate Franklin Medical Center, Greenfield, Massachusetts, said that pharmacists can help nurses with this task by focusing on the basics.

Edward G. Tessier, PharmD, reported on a study he conducted withcolleagues in Massachusetts to create a nurse-friendly medicationreconciliation process, with the goal of standardizeddata-gathering and improved patient safety. Dr. Tessier, clinicalpharmacist, Baystate Franklin Medical Center, Greenfield,Massachusetts, said that pharmacists can help nurses with thistask by focusing on the basics.

Start with the demographics, he advised. "There should be a focuson an allergy history, but we need more than just [an allergylisting]. What is the reaction? For example, the allergy might beto penicillin; however, depending on what the reaction is, youmight still be able to use a beta-lactam." In the studyintervention, the nurses were also instructed, "don't assume thatthe patient's list is true or up to date," he noted. Thisintervention also asked nursing staff to use open-ended questionsto facilitate patient and/or family dialogue. "The lastopen-ended question that the nurse needs to ask is, 'Is thereanything else you'd like to tell me about your medicationstoday?'"

This is not done in a vacuum, of course. Dr. Tessier noted thatthe success of the intervention is going to depend on how well itfits in with the way nurses work and think. "Nurses do systematichead-to-toe assessments. They think of the medication history asa separate task. So we wanted them to ask [when doing CNSevaluation, for example] ‘What medicines do you take when youhave a seizure?' Such pointed questions elicit much betterinformation."

Methods, results
The intervention was done first in a controlled setting(University of Massachusetts Amherst School of Nursing) and thenin two clinical centers (Baystate Franklin and Baystate MedicalCenter in Springfield, Massachusetts). The nursing students were16 second-degree BSN students speaking with 4 actors who wereusing a standardized script. The actors didn't release medicationinformation unless the students asked. There was improvement intwo of four possible scenarios. "We didn't focus on electrolytesor transdermal patches at first," he said. "We refined this andthen brought the intervention into the clinical setting."

In the clinical centers, the researchers set up a nursingeducation campaign at 4 nursing units (2 community and 2 largetertiary care centers). They gave nurses laminated cards withquestions on one side and the names of approximately 100brand-name and generic drugs on the other. "The nurses liked thedrug names better; they found them to be much more useful," hesaid.

Drug classes that were most often missed included antibiotics,cardiac drugs, corticosteroids, and agents for diabetes.Additionally, the researchers were most concerned with agentsthat aren't given every day, recreational drugs, alternativemedicines, and OTCs such as aspirin.

With regard to the outcome of medication discrepancies, a reviewof elements in the electronic health record (EHR) was carried outat three points: history and physical (H&P); medication lists;and discharge summary. With regard to categorization ofdiscrepancies, the implementation at the community hospital went"very well," said Dr. Tessier. "Electronic H&Ps were availablefor all patients save one. At the larger teaching hospital,electronic records were not consistently available so outcomemeasures were [variable]."The intervention did not appear to affect length of stay orallergy discrepancies. Top drug discrepancies includedantidepressants, vitamins, GI drugs, CNS drugs, hormones, andrespiratory drugs. Improvement was seen in the percentage ofpatients with no discrepancies: pre-intervention, 20%;post-intervention, 42% (p=.027).