Effect of robotic automation process on research data abstraction in electronic health record system (Preprint)

BACKGROUND Despite the widespread adoption of electronic health records (EHR), extracting research data from EHRs depends on a time-consuming manual task. Robotic process automation (RPA) has the potential to solve this problem. OBJECTIVE Analyze the effectiveness of RPA in the abstraction of cancer registry data. METHODS The effect of RPA was analyzed by comparing the required duration of work before and after the implementation of RPA. Furthermore, an in-depth interview was conducted to investigate critical factors for implementing RPA for research data abstraction. RPA was implemented to extract 70 and 83 variables for gastric and breast cancer, respectively. RESULTS The data extraction time per patient was reduced by 74% and 30% for gastric and breast cancer, respectively. During in-depth interviews, participants who implemented RPA in the project emphasized the significance of explicitly defining the target process by identifying small, repeatable unit tasks. CONCLUSIONS In conclusion, RPA effectively extracted gastric and breast cancer registry data from EHRs, thus saving time.