A prospective, observational cohort study was designed that included all adult non-ICU patients admitted to internal medicine teaching services from June 1, 2008, through January 31, 2009, who were receiving AST. Patients who had been transferred from the ICU were eligible for inclusion in the study. The study setting was a 650-bed academic medical center affiliated with schools of medicine and pharmacy. The Institutional Review Board approved the research methods.
Data were collected from electronic medical records. AST was defined as the prescription of at least one dose of PPI or H2RA at any point during hospitalization. AST was considered appropriate for all patients if at least 1 of the following indications could be identified in the patient's record: home AST at admission, history of gastroesophageal reflux disease, acute or suspected upper gastrointestinal (GI) bleeding or GI bleeding during the previous 3 months, erosive esophagitis or gastritis, dyspepsia, peptic ulcer disease, post bariatric surgery, or the administration of 2 or more of the following medications: nonsteroidal anti-inflammatory drugs, aspirin, clopidogrel, warfarin, heparin or low molecular weight heparin for therapeutic anticoagulation, or corticosteroids.2,10
AST was considered inappropriate if these conditions were not met. These criteria for AST appropriateness also were applied to patients transferred into the internal medicine service from the ICU.
Medical residents in the internal medicine residency program worked in teaching teams consisting of 1 third-year and 2 first-year medical residents supervised by an attending physician. From October to January, teaching teams also had 1 fourth-year advanced pharmacy practice experience (APPE) student participating in patient rounds with them for 5 days per week. The internal medicine teams without pharmacy students (June through September) served as the control group for this study.
Patients were selected for medication review from the internal medicine team's daily list of new admissions. Every third patient who had been prescribed at least one dose of an acid suppressive medication was selected for review until 20 patients from each week during the study period were identified, allowing for equal sampling from each pharmacy student's clinical APPE.
Each pharmacy student began the APPE by undergoing standard training with a pharmacist. Training included a 30-minute teaching session based on a key review article describing appropriate AST.6
Sessions emphasized evidence-based AST in internal medicine patients and the impact of unnecessary AST, including increased risk of the patient developing Clostridium difficile
colitis, pneumonia, osteoporosis, medication side effects, and drug-drug interactions, and increased costs. Strategies for student interventions for patients receiving inappropriate AST were discussed, including the pharmacy student questioning AST indication during rounds, presenting patient-specific implications of inappropriate AST, and following up with the patient's physician to determine whether the recommendation was accepted. Pharmacy students were evaluated on how successful their actions were in the discontinuation of inappropriate AST prescriptions.
The pharmacy students evaluated medication profiles daily for AST. When inappropriate AST was identified, students addressed AST therapeutic issues and determined the most effective method for communicating the information to the internal medicine team. Pharmacy students participated in patient rounds without a pharmacist preceptor present to encourage them to practice communicating directly with patients and physicians. Pharmacist preceptors met with students following rounds to review the recommendations the students had made. All pharmacy student recommendations to discontinue AST for a patient then were verified by a pharmacist preceptor who approved the recommendation or advised the student to communicate an alternative recommendation to the patient's physician.
An internal pilot study of AST appropriateness conducted in 50 patients prior to this study estimated that 40% of AST prescriptions were inappropriate. These internal data agreed with data from other studies which found that 22%-60% of AST was prescribed inappropriately.2,5
We estimated that pharmacy students would reduce the proportion of inappropriate AST prescriptions to 25%. Sampling required having 165 patients in the study group and 165 in the control group (330 patients), providing an expected effect size of 37.5%, with an alpha of 0.05 and 80% power. We calculated the proportion of patients receiving inappropriate AST prescriptions during hospitalization, proportion discharged on inappropriate AST, and median days of inappropriate AST. Proportions were compared using Pearson's chi-square or the Fisher exact test as appropriate. A Wilcoxon rank sum test was used to compare medians. Kruskal-Wallis ANOVA t
tests with unequal variances or F-tests for multiple comparisons were used to report significant differences in linear variables.