Class of 2010 Abstract / OBJECTIVES: The objective of this study is to compare the number of medication refill errors that occur between the Pyxis Medstation 3500 and the Pyxis PARx automated dispensing systems. The accuracy of refilling Pyxis Medstation 3500 automated dispensing machines by pharmacy technicians at the University Medical Center (UMC) was assessed during six days in July 2009. The accuracy was then reassessed over the course of five days in September 2009, one month after implementation of the new Pyxis PARx barcode technology in August. All medications in both the morning and afternoon Pyxis refill reports generated by the UMC inpatient pharmacy were audited during the chosen days. METHODS: The accuracy of refilling Pyxis Medstation 3500 automated dispensing machines by pharmacy technicians at a 350-bed, tertiary-care, teaching hospital with a total of 50 automated dispensing systems was assessed during five days in July 2009. The accuracy was then reassessed over the course of five days in September 2009, one month after implementation of the new Pyxis PARx barcode technology in August. It was assumed that the following types of medication refill errors would be reduced: 1)Wrong drug, Wrong strength 2) Wrong drug, Right strength 3) Right drug, Wrong strength 4) Filling error/Wrong pocket 5) Overfill of pocket 6) Expired medication 7) Right drug, Wrong form. This study was a prospective evaluation of medication refill errors between an older automated dispensing system (Pyxis Medstation) and a newer system (Pyxis PARx). The addition of PARx barcode technology automates the pick and delivery method in order to enhance security during the medication refill process. When the appropriate drawer is opened, the medication must be scanned using the handheld scanner to make sure the correct medication is being refilled. All medications from each Pyxis refill report on the chosen days were audited during pre and post implementation of the new Pyxis PARx barcode technology. Medications to be audited were identified by collecting data from each automated dispensing system listed on the refill reports for the previous day. Audits were performed on the following day (i.e., Monday’s refill reports was audited on Tuesday) by study investigators. Each completed refill was audited for the above medication errors and a brief description of any errors found was noted. RESULTS: A total of 825 refilled items were audited prior to PARx installation (from dates 7/14/09-7/19/09) and a total of five errors were found. Post PARx installation, 959 items were audited (from dates 9/24/09-9/29/09) and a total of two errors were found. Overall, the types of errors encountered were Right Drug/Wrong Form (1/7, 14.3%), Filling Error/Wrong Pocket (2/7, 28.6%), Overfill (1/7, 14.3%), Expired Medication (2/7, 28.6%) and Right Drug/Wrong Strength (1/7, 14.3%). A Chi-square analysis was done to compare pre and post-PARx implementation errors found. The Chi square value was found to be 1.79 with a p-value of 0.18, meaning that there is an 18% probability that any deviation from expected is due to chance. Thus, per our investigation the installation of PARx did not significantly decrease refill error rates. CONCLUSIONS: In Progress

AN EXPLORATORY MODEL OF MEDICATION REFILL ADHERENCE BEHAVIOR by Gayle Holmes Payne Coronary heart disease (CHD) is the leading cause of death in the United States with over 15.8 million Americans suffering from the chronic disease (U.S. Department of Health and Human Services, 2007). Adherence to medication regimens has been identified as a key mediator between medical practice and patient outcomes (Kravitz & Melnikow, 2004). In this study, participants (N = 355) with CHD completed a questionnaire measuring their background characteristics, cognitive status, health literacy skills, self-efficacy levels, their perceived concerns and necessity beliefs about medication use, and enablers and barriers to their medication-taking behavior. Information regarding each participant’s number of medications and presence of disease was obtained from medical charts. Data regarding the dependent variable, cardiovascular medication refill adherence, were collected from pharmacy records. The data were used to see how the various variables work together in a model that explains cardiovascular medication refill adherence behavior. The study aimed to contribute to the body of adherence research by jointly examining all variables found to have an association with medication adherence through a path analysis to explain the determinants of medication refill adherence behavior. Analyses indicated that the hypothesized model did not fit the data. Additional analysis was conducted using a condensed revised model (age, self-efficacy, perceived concerns and necessity) and a self-reported measure of medication adherence (Adherence to Refills and Medication Scale) as the dependent variable. The revised model fit the data, X2(5, N = 355) = 6.71, p = .24. The revised model did not explain a statistically significant amount of the variance in medication adherence, suggesting that there may be other additional factors that may mediate the relationship between independent variables and medication refill adherence. Additional research is needed to reveal all the determinants of medication refill adherence behavior and to identify the most effective measure of adherence behavior. Given the number of people who suffer from CHD, and the often low rates of medication adherence, research that continues to explore and improve medication refill adherence will have a significant impact on morbidity and mortality rates.