Background: Increased workload, including workload related to electronic health record (EHR) documentation, is reported as a main contributor to nurse burnout, and adversely affects patient safety and nurse satisfaction. Although nurse burnout has been studied for decades, little has changed in the organization of clinical care. The measurement of nursing workload is not well understood. Traditional methods for workload analysis are either administrative measures (such as nurse-patient ratio) that do not represent actual nursing care, or are subjective and limited to snapshots of care (e.g., time-motion studies). Observing care, and testing workflow changes in real-time, can be obstructive to clinical care. An examination of EHR interactions through the use of EHR audit logs could provide a scalable, unobtrusive way to quantify nursing workload, at least to the extent that nursing work is represented in EHR documentation. EHR audit logs are extremely complex, however, and simple analytical methods can't discover complex temporal patterns, therefore requiring the use of state-of-the-art temporal data mining approaches. To effectively use these approaches it is necessary to structure the raw audit logs into a consistent and scalable data model that can be consumed by machine learning (ML) algorithms.
Objective:We aimed to conceptualize a data model for nurse-EHR interactions that would support the development of temporal ML models based on EHR audit log data.
Methods:We conducted a preliminary review of EHR audit logs to understand the types of nursing-specific data captured. Using concepts derived from literature and our previous experience studying temporal patterns in biomedical data, we formulated a data model that can describe the nurse-EHR interactions, nurse intrinsic and situational characteristics that may influence those interactions, and outcomes of relevance to nursing workload in a scalable and extensible manner.
Results:We described the data structure and concepts from EHR audit log data associated with nursing workload as a data model, that we name as RNteract. We conceptually demonstrated how using this data model could support temporal unsupervised machine learning and state-of-the-art artificial intelligence methods for predictive modeling.
Conclusions:The RNteract data model appears capable of supporting a variety AI-based systems (computational models), and should be generalizable to any type of EHR system or healthcare setting. Quantitatively identifying and analyzing temporal patterns of nurse-EHR interactions is foundational for developing interventions that support nursing documentation workload and address nurse burnout. Clinical
