Medication errors and missing doses are common problems that can occur during a patient's hospital visit. However, patient safety can be improved by utilizing automated dispensing cabinets (ADCs) and automation in the medication order fulfillment process to reduce the number of medication errors and missing doses. ADCs create a decentralized inventory which increases the inventory cost while automation is costly and has limited use due to throughput capacity. We have created a model which improves patient safety subject to a budget constraint by improving hospital pharmacy operations. Our model determines a primary pathway to deliver medications from the hospital pharmacy to the patient based on the inpatient unit, the medication order type, and the forecasted demand. Using data from a local hospital to determine model parameters, we solve the problem to attain the optimal delivery paths that reduce missing doses and medication errors. In order to determine which parameters have the most influence on the optimal pathways, we perform a sensitivity analysis on the parameters in model. Specifically, we consider the implications of using different service level policies on the model results.