Process mining leveraging the analysis of patient journey and outcomes: Stroke assistance during the Covid-19 pandemic

Abstract

The patient journey had to be modified because of the Covid-19 pandemic, causing insecurity, especially in health conditions in a time-sensitive treatment. Identifying these changes and their consequences is essential to improving the healthcare process and guaranteeing patient safety. Process mining (PM) helps evaluate the patient journey discovering care delays, bottlenecks, and non-conformities. This paper aims to apply PM to discover and analyze the patient pathway during stroke care in two different contexts, before and after the Covid-19 outbreak, and to correlate these pathways to patient outcomes. It was a retrospective cross-sectional study including 509 analyzed event logs, employing the most relevant population-based stroke registry of Latin America. Two process models were uncovered to illustrate the patient journey before and during the pandemic. The main findings were the worsening of the patient's health status at their hospital admission, the reduction of hospitalization time, the increased delay for receiving reperfusion therapies after hospital admission, and the preference for the referral hospital instead of emergency services. PM assisted in identifying time-sensitive events and allowed the improvement of patient safety. This methodology can be replicated in other healthcare studies.