Search

Mariam Heikal,1 Halim Saad,2 Pia Maria Ghanime,3 Tarek Bou Dargham,4 Maya Bizri,5 Firas Kobeissy,6 Wassim El Hajj,1 Farid Talih2 1Department of Computer Science, American University of Beirut, Beirut, Lebanon; 2Department of Psychiatry, Faculty of Medicine, American University of Beirut, Beirut, Lebanon; 3Department of Psychiatry, Massachusetts General Hospital, Boston, MA, USA; 4Department of Neurosurgery, Duke University Medical Center, Durham, NC, USA; 5Department of Psychiatry and Psychology, Cleveland Clinic, Cleveland, OH, USA; 6Department of Neurobiology, Morehouse School of Medicine, Atlanta, GA, USACorrespondence: Mariam Heikal, Department of Computer Science, American University of Beirut, Beirut, Lebanon, Email mas177@mail.aub.eduObjective: Delirium is a common and acute neuropsychiatric syndrome that requires timely intervention to prevent its associated morbidity and mortality. Yet, its diagnosis and symptoms are often overlooked due to its variable clinical presentation and fluctuating nature. Thus, in this study, we address the barriers to delirium diagnosis by utilizing a machine learning-based predictive algorithm for incident delirium that relies on archived electronic health records (EHRs) data.Methods: We used the Medical Information Mart for Intensive Care (MIMIC) database to create a detailed dataset for identifying delirium in intensive care unit (ICU) patients. Our approach involved training machine learning models on this dataset to pinpoint critical clinical features for delirium detection. These features were then refined and applied to non-ICU patients using EHRs from the American University of Beirut Medical Center (AUBMC).Results: Our study assessed machine learning models like Extreme Gradient Boosting (XGBoost), Categorical Boosting (CatBoost), Classification and Regression Trees (CART), Random Forest (RF), Neural Oblivious Decision Ensembles (NODE), and Logistic Regression (LR), highlighting superior delirium detection in diverse clinical settings. The CatBoost model excelled in ICU environments with an F1 Score of 89.2%, while XGBoost performed best in general hospital settings with a 75.4% F1 Score. Interpretations using Tabular Local Interpretable Model-agnostic Explanations (LIME) revealed critical indicators such as prothrombin time and hematocrit levels, enhancing model transparency and clinical applicability. These clinical insights help differentiate the delirium predictors between ICU patients, who are often sensitive to various factors.Conclusion: The proposed predictive algorithm improves delirium detection rates and streamlines efficiency in hospital electronic systems, thereby enabling prompt interventions to prevent delirium progression and associated complications. The clinical indicators for delirium that we identified in general hospital settings and ICU can greatly help healthcare professionals identify potential causes of delirium and reduce misdiagnosis.Keywords: Delirium, ICU delirium, Hospital-acquired delirium, electronic health records, machine learning, clinical indicators

Cancer and stroke are leading causes of global disability and mortality. With improvements in cancer-associated mortality and advancements in treatment of active malignancy, it is more common to encounter patients with ischemic stroke and active malignancy. Evidence suggests that cancer-associated ischemic stroke is a unique subtype of stroke; however, there is limited guidance when considering diagnostic workup, secondary prevention, rehabilitation, and future directions within this population. In this narrative review, we aim to describe the epidemiology, pathophysiological mechanisms, management, and future directions regarding understanding of cancer-associated ischemic stroke.

Background: Inclusion of diverse patient populations in stroke rehabilitation clinical trials is key for generalizability and detecting differences in subgroups. The aim of this study was to assess the reporting and inclusion of race, ethnicity, and sex in publications in post-stroke motor recovery clinical trials over the past 5 years., Methods: We conducted a systematic review of stroke motor recovery trials conducted only within the United States with at least one control group and published between 2019 and 2023. The percent of racial minorities, ethnicities, and women enrolled in the trials was extracted and calculated for those trials using available data found in the manuscript or on clinicaltrials.gov., Results: Sixty-eight trials (total of 2,801 participants) met the inclusion criteria and were included in the analysis. 100%, 35%, and 21% of the trials included reported enrollment by sex, race, and ethnicity in the manuscript. All publications reported sex and 38% of the subjects were female. Among the trials reporting race, 59% of the participants were White, followed by 34% Black, 0.4% Native American, 4% Asian, and 3% Other. Among the trials that reported ethnicity, Hispanic or Latino participants were represented as 13% of the total participants., Conclusions: In the past 5 years, while all stroke rehab trial publications reported data on sex, they underrepresented women. Most publications did not report race or ethnicity. Improvement in reporting of race/ethnicity in stroke motor recovery trials is needed for understanding of progress with inclusion, and improvement in inclusion of women is needed for better generalizability.