Early detection alerts in sepsis treatment - An empirical study of patient treatment and outcomes

Sammanfattning: With rising antimicrobial resistance and an increasing number of sepsis incidence, it is essential to examine ways to improve sepsis patient outcomes. The purpose of this study is to evaluate the impact of an early detection sepsis alert on clinicians' behaviour and patient outcomes. During the observation period the alert system has an active and a silent running period. The active alert notifies clinicians when patients are potentially infected with sepsis. During the silent period clinicians are not notified but the system documents when the alert should have been triggered. Using repeated cross-sectional patient-level data, I gauge the effect of the alert system on clinicians' behaviour and patient outcomes employing three different regression forms. This allows me to estimate the differences in the outcome variables between patients that were exposed to the active alert and patients that were not. The results indicate a time trend leading to reduced mortality and length of stay (LOS) in hospital, but further research is needed to disentangle the effect of the alert from the time trend. Hence, this study provides suggestive evidence that the alert system has the potential to reduce mortality and LOS.