Abstract
PURPOSE: ChemoPalRx is a novel provider order entry mobile application for chemotherapy. This study aims to evaluate the accuracy of prescribing chemotherapy using ChemoPalRx versus handwritten orders at a safety-net hospital in Los Angeles. METHODS: In a cross-sectional study from October 2019 to December 2019, we evaluated all outpatient chemotherapy orders for accuracy. Our primary predictor was type of prescription, dichotomized as handwritten or ChemoPalRx. Primary outcome was accuracy, dichotomized as accurate if no error was made on an order and as inaccurate if any error was made. Preplanned subgroup analyses were performed with covariates including provider experience, complexity of order, and day of order submission. We characterized error type and analyzed our data using univariate and multivariate logistic regression models. RESULTS: Among 288 orders (78.5% handwritten; 21.5% ChemoPalRx), prescription accuracy was higher among ChemoPalRx (93.5%) compared with handwritten orders (81.4%; P = .012). In multivariate analysis, prescription accuracy remained superior for ChemoPalRx after adjusting for provider experience, complexity of order, and day of order submission (adjusted odds ratio, 1.82; P = .012). Compared with handwritten orders, ChemoPalRx orders had less missing or incorrect information (1.6% v 13.7%; P = .0016). ChemoPalRx orders were also more accurate on prescriptions that contained two or fewer medications (92.2% v 80.2%; P = .032), submitted on the highest patient-volume clinic day of the week (96.7% v 83.2%; P = .035), and generated by a senior fellow or an attending (97.3% v 76.9%; P = .001). CONCLUSION: ChemoPalRx is associated with improved chemotherapy prescription accuracy over handwritten orders in the safety-net hospital setting and may serve as an alternative prescribing tool for oncology practices.