Given the increasing availability of electronic health records, it has become increasingly feasible to adopt data-driven approaches to capture a deep understanding of the patient journeys. Nevertheless, simply using data-driven techniques to depict the patient journeys without an integrated modeling and analysis approach is proving to be of little benefit for improving patients’ experiences. Indeed, a model of the journey patterns is necessary to support the improvement process.
We presented a 3-phase methodology that integrates a process mining–based understanding of patient journeys with a stochastic graphical modeling approach to derive and analyze the analytical expressions of some important performance indicators of an emergency department including mean and variance of patients’ length of stay (LOS).
Analytical expressions were derived and discussed for mean and variance of LOS times and discharge and admission probabilities. LOS differed significantly depending on whether a patient was admitted to the hospital or discharged. Moreover, multiparameter sensitivity equations are obtained to identify which activities contribute the most in reducing the LOS at given operating conditions so decision makers can prioritize their improvement initiatives.
Data-driven based approaches for understanding the patient journeys coupled with appropriate modeling techniques yield a promising tool to support improving patients’ experiences. The modeling techniques should be easy to implement and not only should be capable of deriving some key performance indicators of interest but also guide decision makers in their improvement initiatives.
Farhood Rismanchian Sara Hosseinzadeh Kassani Seyed Mahdi Shavarani Young Hoon Lee