Equity in critical care: a review of artificial intelligence driven solutions

Abstract

The critical care unit provides complex care to patients with the highest risk of debility and death.While providers attempt to treat these patients in an objective and evidence-based manner, evidence suggests that patients of certain ethnicities, socio-economic backgrounds, age, and gender have worse outcomes in the ICU.While scoring systems have been employed to improve objectivity in this area, consistent application and generalizability remain a concern.One way to bridge this gap is to employ Machine Learning (ML) algorithms using patient data to obtain an objective assessment of the patient's condition to better make clinical care decisions.Prognostication tools such as ELDER-ICU use ML algorithms in conjunction with traditional scoring tools to control against bias.ML based protocols have also been employed in Unfractionated heparin dosing and other nomograms, developing higher quality calculations for eGFR and utilize heart rates and dermal temperature scores to model the effectiveness of pain control methods in the critical care unit.With the development of Chatbots and other interactive text-based applications, clinicians are also utilizing AI for informed consent and patient education in the critical care setting.With continuous advancements in the field of AI, each individual patient can have a tailor-made plan that incorporates real-time data into multiple ML protocols to provide the best outcomes.Care to be taken to ensure that these algorithms are designed in a controlled way that eliminates bias/disparities at the developmental stage.

Ähnliche Arbeiten