Prediction of Complications and Prognostication in Perioperative Medicine: A Systematic Review and PROBAST Assessment of Machine Learning Tools.

BACKGROUND: The utilization of artificial intelligence and machine learning as diagnostic and predictive tools in perioperative medicine holds great promise. Indeed, many studies have been performed in recent years to explore the potential. The purpose of this systematic review is to assess the current state of machine learning in perioperative medicine, its utility in prediction of complications and prognostication, and limitations related to bias and validation. METHODS: A multidisciplinary team of clinicians and engineers conducted a systematic review using the Preferred Reporting Items for Systematic Review and Meta-Analysis (PRISMA) protocol. Multiple databases were searched, including Scopus, Cumulative Index to Nursing and Allied Health Literature (CINAHL), the Cochrane Library, PubMed, Medline, Embase, and Web of Science. The systematic review focused on study design, type of machine learning model used, validation techniques applied, and reported model performance on prediction of complications and prognostication. This review further classified outcomes and machine learning applications using an ad hoc classification system. The Prediction model Risk Of Bias Assessment Tool (PROBAST) was used to assess risk of bias and applicability of the studies. RESULTS: A total of 103 studies were identified. The models reported in the literature were primarily based on single-center validations (75%), with only 13% being externally validated across multiple centers. Most of the mortality models demonstrated a limited ability to discriminate and classify effectively. The PROBAST assessment indicated a high risk of systematic errors in predicted outcomes and artificial intelligence or machine learning applications. CONCLUSIONS: The findings indicate that the development of this field is still in its early stages. This systematic review indicates that application of machine learning in perioperative medicine is still at an early stage. While many studies suggest potential utility, several key challenges must be first overcome before their introduction into clinical practice.

Reduced anticoagulation strategy is associated with a lower incidence of intracerebral hemorrhage in COVID-19 patients on extracorporeal membrane oxygenation.

BACKGROUND: Optimal anticoagulation strategies for COVID-19 patients with the acute respiratory distress syndrome (ARDS) on venovenous extracorporeal membrane oxygenation (VV ECMO) remain uncertain. A higher incidence of intracerebral hemorrhage (ICH) during VV ECMO support compared to non-COVID-19 viral ARDS patients has been reported, with increased bleeding rates in COVID-19 attributed to both intensified anticoagulation and a disease-specific endotheliopathy. We hypothesized that lower intensity of anticoagulation during VV ECMO would be associated with a lower risk of ICH. In a retrospective, multicenter study from three academic tertiary intensive care units, we included patients with confirmed COVID-19 ARDS requiring VV ECMO support from March 2020 to January 2022. Patients were grouped by anticoagulation exposure into higher intensity, targeting anti-factor Xa activity (anti-Xa) of 0.3-0.4 U/mL, versus lower intensity, targeting anti-Xa 0.15-0.3 U/mL, cohorts. Mean daily doses of unfractionated heparin (UFH) per kg bodyweight and effectively measured daily anti-factor Xa activities were compared between the groups over the first 7 days on ECMO support. The primary outcome was the rate of ICH during VV ECMO support. RESULTS: 141 critically ill COVID-19 patients were included in the study. Patients with lower anticoagulation targets had consistently lower anti-Xa activity values over the first 7 ECMO days (p < 0.001). ICH incidence was lower in patients in the lower anti-Xa group: 4 (8%) vs 32 (34%) events. Accounting for death as a competing event, the adjusted subhazard ratio for the occurrence of ICH was 0.295 (97.5% CI 0.1-0.9, p = 0.044) for the lower anti-Xa compared to the higher anti-Xa group. 90-day ICU survival was higher in patients in the lower anti-Xa group, and ICH was the strongest risk factor associated with mortality (odds ratio [OR] 6.8 [CI 2.1-22.1], p = 0.001). CONCLUSIONS: For COVID-19 patients on VV ECMO support anticoagulated with heparin, a lower anticoagulation target was associated with a significant reduction in ICH incidence and increased survival.