Social Determinants of Health and Limitation of Life-Sustaining Therapy in Neurocritical Care: A CHoRUS Pilot Project.

BACKGROUND: Social determinants of health (SDOH) have been linked to neurocritical care outcomes. We sought to examine the extent to which SDOH explain differences in decisions regarding life-sustaining therapy, a key outcome determinant. We specifically investigated the association of a patient's home geography, individual-level SDOH, and neighborhood-level SDOH with subsequent early limitation of life-sustaining therapy (eLLST) and early withdrawal of life-sustaining therapy (eWLST), adjusting for admission severity. METHODS: We developed unique methods within the Bridge to Artificial Intelligence for Clinical Care (Bridge2AI for Clinical Care) Collaborative Hospital Repository Uniting Standards for Equitable Artificial Intelligence (CHoRUS) program to extract individual-level SDOH from electronic health records and neighborhood-level SDOH from privacy-preserving geomapping. We piloted these methods to a 7 years retrospective cohort of consecutive neuroscience intensive care unit admissions (2016-2022) at two large academic medical centers within an eastern Massachusetts health care system, examining associations between home census tract and subsequent occurrence of eLLST and eWLST. We matched contextual neighborhood-level SDOH information to each census tract using public data sets, quantifying Social Vulnerability Index overall scores and subscores. We examined the association of individual-level SDOH and neighborhood-level SDOH with subsequent eLLST and eWLST through geographic, logistic, and machine learning models, adjusting for admission severity using admission Glasgow Coma Scale scores and disorders of consciousness grades. RESULTS: Among 20,660 neuroscience intensive care unit admissions (18,780 unique patients), eLLST and eWLST varied geographically and were independently associated with individual-level SDOH and neighborhood-level SDOH across diagnoses. Individual-level SDOH factors (age, marital status, and race) were strongly associated with eLLST, predicting eLLST more strongly than admission severity. Individual-level SDOH were more strongly predictive of eLLST than neighborhood-level SDOH. CONCLUSIONS: Across diagnoses, eLLST varied by home geography and was predicted by individual-level SDOH and neighborhood-level SDOH more so than by admission severity. Structured shared decision-making tools may therefore represent tools for health equity. Additionally, these findings provide a major warning: prognostic and artificial intelligence models seeking to predict outcomes such as mortality or emergence from disorders of consciousness may be encoded with self-fulfilling biases of geography and demographics.

A health equity monitoring framework based on process mining.

In the United States, there is a proposal to link hospital Medicare payments with health equity measures, signaling a need to precisely measure equity in healthcare delivery. Despite significant research demonstrating disparities in health care outcomes and access, there is a noticeable gap in tools available to assess health equity across various health conditions and treatments. The available tools often focus on a single area of patient care, such as medication delivery, but fail to examine the entire health care process. The objective of this study is to propose a process mining framework to provide a comprehensive view of health equity. Using event logs which track all actions during patient care, this method allows us to look at disparities in single and multiple treatment steps, but also in the broader strategy of treatment delivery. We have applied this framework to the management of patients with sepsis in the Intensive Care Unit (ICU), focusing on sex and English language proficiency. We found no significant differences between treatments of male and female patients. However, for patients who don't speak English, there was a notable delay in starting their treatment, even though their illness was just as severe and subsequent treatments were similar. This framework subsumes existing individual approaches to measure health inequities and offers a comprehensive approach to pinpoint and delve into healthcare disparities, providing a valuable tool for research and policy-making aiming at more equitable healthcare.

Interpretation and Use of Applied/Operational Machine Learning and Artificial Intelligence in Surgery.

Applications for artificial intelligence (AI) and machine learning in surgery include image interpretation, data summarization, automated narrative construction, trajectory and risk prediction, and operative navigation and robotics. The pace of development has been exponential, and some AI applications are working well. However, demonstrations of clinical utility, validity, and equity have lagged algorithm development and limited widespread adoption of AI into clinical practice. Outdated computing infrastructure and regulatory challenges which promote data silos are key barriers. Multidisciplinary teams will be needed to address these challenges and to build AI systems that are relevant, equitable, and dynamic.

Early COVID-19 respiratory risk stratification using machine learning.

Background: COVID-19 has strained healthcare systems globally. In this and future pandemics, providers with limited critical care experience must distinguish between moderately ill patients and those who will require aggressive care, particularly endotracheal intubation. We sought to develop a machine learning-informed Early COVID-19 Respiratory Risk Stratification (ECoRRS) score to assist in triage, by providing a prediction of intubation within the next 48 hours based on objective clinical parameters. Methods: Electronic health record data from 3447 COVID-19 hospitalizations, 20.7% including intubation, were extracted. 80% of these records were used as the derivation cohort. The validation cohort consisted of 20% of the total 3447 records. Multiple randomizations of the training and testing split were used to calculate confidence intervals. Data were binned into 4-hour blocks and labeled as cases of intubation or no intubation within the specified time frame. A LASSO (least absolute shrinkage and selection operator) regression model was tuned for sensitivity and sparsity. Results: Six highly predictive parameters were identified, the most significant being fraction of inspired oxygen. The model achieved an area under the receiver operating characteristic curve of 0.789 (95% CI 0.785 to 0.812). At 90% sensitivity, the negative predictive value was 0.997. Discussion: The ECoRRS score enables non-specialists to identify patients with COVID-19 at risk of intubation within 48 hours with minimal undertriage and enables health systems to forecast new COVID-19 ventilator needs up to 48 hours in advance. Level of evidence: IV.

Ethical, legal, and communication challenges in managing goals-of-care discussions in chronically critically ill patients.

Clinicians should expect controversial goals of care discussions in the surgical intensive care from time to time. Differing opinions about the likelihood of meaningful recovery in patients with chronic critical illness often exist between intensive care unit providers of different disciplines. Outcome predictions presented by health-care providers are often reflections of their own point of view that is influenced by provider experience, profession, and personal values, rather than the consequence of reliable scientific evaluation. In addition, family members of intensive care unit patients often develop acute cognitive, psychologic, and physical challenges. Providers in the surgical intensive care unit should approach goals-of-care discussions in a structured and interprofessional manner. This best practice paper highlights medical, legal and ethical implications of changing goals of care from prioritizing cure to prioritizing comfort and provides tools that help physicians become effective leaders in the multi-disciplinary management of patients with challenging prognostication.

Machine learning can accurately predict pre-admission baseline hemoglobin and creatinine in intensive care patients.

Patients admitted to the intensive care unit frequently have anemia and impaired renal function, but often lack historical blood results to contextualize the acuteness of these findings. Using data available within two hours of ICU admission, we developed machine learning models that accurately (AUC 0.86-0.89) classify an individual patient's baseline hemoglobin and creatinine levels. Compared to assuming the baseline to be the same as the admission lab value, machine learning performed significantly better at classifying acute kidney injury regardless of initial creatinine value, and significantly better at predicting baseline hemoglobin value in patients with admission hemoglobin of <10 g/dl.

Variability in Perioperative Fasting Practices Negatively Impacts Nutritional Support of Critically Ill Intubated Patients.

The aim of the study was to quantify nutritional losses related to pre- and postoperative fasts in critically ill intubated patients and to explore whether shorter fasts are safe and appropriate in this population. A retrospective review of mechanically ventilated adults undergoing surgery more than 24 hours after admission to a Level I trauma center over 15 months was done, which yielded 132 procedures and 81 unique patients. Ninety per cent of preoperative periods and 43 per cent of postoperative periods were affected by nonmedical barriers to feeding. Eighty-two per cent of gastrically fed nonemergent cases were fasted for longer than the 6-hour American Society of Anesthesiologists guideline, whereas 91 per cent of emergent cases had shorter fasts. There were no anesthetic complications, placing an upper limit of 6 per cent on the rate of aspiration for fasts shorter than six hours (95% confidence). Forty-three per cent of cases did not resume tube feeds within 90 minutes postoperatively, and only 37 per cent had a documented justification for delay. Intubated patients were frequently fasted preoperatively for longer than recommended and postoperatively for longer than medically indicated. No complications were observed with shorter-than-guideline fasts. This strengthens the evidence that "standard" preoperative fasting is unnecessary and deleterious in many critically ill intubated patients. New protocols and national guidelines are needed to ensure adequate nutrition.