Artificial intelligence and clinical deterioration

Abstract

Purpose of reviewTo provide an overview of the systems being used to identify and predict clinical deterioration in hospitalised patients, with focus on the current and future role of artificial intelligence (AI). Recent findingsThere are five leading AI driven systems in this field: the Advanced Alert Monitor (AAM), the electronic Cardiac Arrest Risk Triage (eCART) score, Hospital wide Alert Via Electronic Noticeboard, the Mayo Clinic Early Warning Score, and the Rothman Index (RI). Each uses Electronic Patient Record (EPR) data and machine learning to predict adverse events. Less mature but relevant evolutions are occurring in the fields of Natural Language Processing, Time and Motion Studies, AI Sepsis and COVID-19 algorithms. SummaryResearch-based AI-driven systems to predict clinical deterioration are increasingly being developed, but few are being implemented into clinical workflows. Escobar et al. (AAM) provide the current gold standard for robust model development and implementation methodology. Multiple technologies show promise, however, the pathway to meaningfully affect patient outcomes remains challenging.

Incidence, outcomes and costs of severe sepsis and septic shock in geriatric trauma patients: Analysis of 2,563,463 hospitalizations at 3,284 hospitals

“…Incorporation of data on comorbidities and other patient specific characteristics into the modeling for automated "artificial intelligence" systems that alert clinicians to early signs of impending sepsis may prove fruitful for both prevention and early diagnosis but will require very large datasets to train and refine their machine learning algorithms. 27,28 This study has a number of limitations, most prominently, the retrospective nature of the analysis. 29,30 In addition, the study used the IPSAF dataset thereby excluding the approximately one-third of Medicare inpatients who had Medicare Advantage coverage during the study period.…”

“…It is unclear whether the incidence and progression of SS/SS in geriatric trauma patients is different than in other older adult inpatients, which is likewise an area of potential investigation. Incorporation of data on comorbidities and other patient specific characteristics into the modeling for automated “artificial intelligence” systems that alert clinicians to early signs of impending sepsis may prove fruitful for both prevention and early diagnosis but will require very large datasets to train and refine their machine learning algorithms 27,28 …”

Clinical Instability Is a Sign of Severity of Illness: A Cohort Study

“…Intuitive thinking is invaluable in dynamic settings, such as the ICU, where lack of time, multiple competing priorities, and heavy cognitive load are a part of this complex clinical environment (27). Recently, experts have proposed there is a “cognitive continuum” where both types of thinking operate in concert, and that little, if any, problem solving is performed by either cognitive system in isolation (27, 30). Depending on the situation, either intuitive judgment or analytic reasoning may be dominant.…”

Reframing the Morbidity and Mortality Conference: The Impact of a Just Culture

“…As already mentioned, AI has the potential to scan large amounts of patient data efficiently and automate the process of identifying adverse events. 60 We provide several theoretical applications of AI within M&M; while there are numerous studies of the applicability of AI within clinical medicine and the detection and prevention of adverse events, [80][81][82][83] there is no current published literature specifically on AI in M&M.…”