Laboratory data from Electronic Health Records (EHR) are often used in prediction models where estimation bias and model performance from missingness can be mitigated using imputation methods. We demonstrate the utility of imputation in two real-world EHR-derived cohorts of ischemic stroke from Geisinger and of heart failure from Sutter Health to: (1) characterize the patterns of missingness in laboratory variables; (2) simulate two missing mechanisms, arbitrary and monotone; (3) compare cross-sectional and multi-level multivariate missing imputation algorithms applied to laboratory data; (4) assess whether incorporation of latent information, derived from comorbidity data, can improve the performance of the algorithms. The latter was based on a case study of hemoglobin A1c under a univariate missing imputation framework. Overall, the pattern of missingness in EHR laboratory variables wasnot at randomand was highly associated with patients’ comorbidity data; and the multi-level imputation algorithm showed smaller imputation error than the cross-sectional method.

Background: The long-term risk of recurrent ischemic stroke, estimated to be between 17% and 30%, cannot be reliably assessed at an individual level. Our goal was to study whether machine-learning can be trained to predict stroke recurrence and identify key clinical variables and assess whether performance metrics can be optimized. Methods: We used patient-level data from electronic health records, six interpretable algorithms (Logistic Regression, Extreme Gradient Boosting, Gradient Boosting Machine, Random Forest, Support Vector Machine, Decision Tree), four feature selection strategies, five prediction windows, and two sampling strategies to develop 288 models for up to 5-year stroke recurrence prediction. We further identified important clinical features and different optimization strategies. Results: We included 2091 ischemic stroke patients. Model area under the receiver operating characteristic (AUROC) curve was stable for prediction windows of 1, 2, 3, 4, and 5 years, with the highest score for the 1-year (0.79) and the lowest score for the 5-year prediction window (0.69). A total of 21 (7%) models reached an AUROC above 0.73 while 110 (38%) models reached an AUROC greater than 0.7. Among the 53 features analyzed, age, body mass index, and laboratory-based features (such as high-density lipoprotein, hemoglobin A1c, and creatinine) had the highest overall importance scores. The balance between specificity and sensitivity improved through sampling strategies. Conclusion: All of the selected six algorithms could be trained to predict the long-term stroke recurrence and laboratory-based variables were highly associated with stroke recurrence. The latter could be targeted for personalized interventions. Model performance metrics could be optimized, and models can be implemented in the same healthcare system as intelligent decision support for targeted intervention.

Background and Purpose: Stroke is reported as a consequence of severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) infection in several reports. However, data are sparse regarding the details of these patients in a multinational and large scale. Methods: We conducted a multinational observational study on features of consecutive acute ischemic stroke, intracranial hemorrhage, and cerebral venous or sinus thrombosis among SARS-CoV-2–infected patients. We further investigated the risk of large vessel occlusion, stroke severity as measured by the National Institutes of Health Stroke Scale, and stroke subtype as measured by the TOAST (Trial of ORG 10172 in Acute Stroke Treatment) criteria among patients with acute ischemic stroke. In addition, we explored the neuroimaging findings, features of patients who were asymptomatic for SARS-CoV-2 infection at stroke onset, and the impact of geographic regions and countries’ health expenditure on outcomes. Results: Among the 136 tertiary centers of 32 countries who participated in this study, 71 centers from 17 countries had at least 1 eligible stroke patient. Of 432 patients included, 323 (74.8%) had acute ischemic stroke, 91 (21.1%) intracranial hemorrhage, and 18 (4.2%) cerebral venous or sinus thrombosis. A total of 183 (42.4%) patients were women, 104 (24.1%) patients were <55 years of age, and 105 (24.4%) patients had no identifiable vascular risk factors. Among acute ischemic stroke patients, 44.5% (126 of 283 patients) had large vessel occlusion; 10% had small artery occlusion according to the TOAST criteria. We observed a lower median National Institutes of Health Stroke Scale (8 [3–17] versus 11 [5–17]; P =0.02) and higher rate of mechanical thrombectomy (12.4% versus 2%; P <0.001) in countries with middle-to-high health expenditure when compared with countries with lower health expenditure. Among 380 patients who had known interval onset of the SARS-CoV-2 and stroke, 144 (37.8%) were asymptomatic at the time of admission for SARS-CoV-2 infection. Conclusions: We observed a considerably higher rate of large vessel occlusions, a much lower rate of small vessel occlusion and lacunar infarction, and a considerable number of young stroke when compared with the population studies before the pandemic. The rate of mechanical thrombectomy was significantly lower in countries with lower health expenditures.

Background: Developing a decision support system based on advances in machine learning is one area for strategic innovation in healthcare. Predicting a patient’s progression to septic shock is an active field of translational research. The goal of this study was to develop a working model of a clinical decision support system for predicting septic shock in an acute care setting for up to 6 h from the time of admission in an integrated healthcare setting. Method: Clinical data from Electronic Health Record (EHR), at encounter level, were used to build a predictive model for progression from sepsis to septic shock up to 6 h from the time of admission; that is, T = 1, 3, and 6 h from admission. Eight different machine learning algorithms (Random Forest, XGBoost, C5.0, Decision Trees, Boosted Logistic Regression, Support Vector Machine, Logistic Regression, Regularized Logistic, and Bayes Generalized Linear Model) were used for model development. Two adaptive sampling strategies were used to address the class imbalance. Data from two sources (clinical and billing codes) were used to define the case definition (septic shock) using the Centers for Medicare & Medicaid Services (CMS) Sepsis criteria. The model assessment was performed using Area under Receiving Operator Characteristics (AUROC), sensitivity, and specificity. Model predictions for each feature window (1, 3 and 6 h from admission) were consolidated. Results: Retrospective data from April 2005 to September 2018 were extracted from the EHR, Insurance Claims, Billing, and Laboratory Systems to create a dataset for septic shock detection. The clinical criteria and billing information were used to label patients into two classes-septic shock patients and sepsis patients at three different time points from admission, creating two different case-control cohorts. Data from 45,425 unique in-patient visits were used to build 96 prediction models comparing clinical-based definition versus billing-based information as the gold standard. Of the 24 consolidated models (based on eight machine learning algorithms and three feature windows), four models reached an AUROC greater than 0.9. Overall, all the consolidated models reached an AUROC of at least 0.8820 or higher. Based on the AUROC of 0.9483, the best model was based on Random Forest, with a sensitivity of 83.9% and specificity of 88.1%. The sepsis detection window at 6 h outperformed the 1 and 3-h windows. The sepsis definition based on clinical variables had improved performance when compared to the sepsis definition based on only billing information. Conclusion: This study corroborated that machine learning models can be developed to predict septic shock using clinical and administrative data. However, the use of clinical information to define septic shock outperformed models developed based on only administrative data. Intelligent decision support tools can be developed and integrated into the EHR and improve clinical outcomes and facilitate the optimization of resources in real-time.

(1) Background: Stroke incidence and outcomes are influenced by socioeconomic status. There is a paucity of reported population-level studies regarding these determinants. The goal of this ecological analysis was to determine the county-level associations of social determinants of stroke hospitalization and death rates in the United States. (2) Methods: Publicly available data as of 9 April 2021, for the socioeconomic factors and outcomes, was extracted from the Centers for Disease Control and Prevention. The outcomes of interest were “all stroke hospitalization rates per 1000 Medicare beneficiaries” (SHR) and “all stroke death rates per 100,000 population” (SDR). We used a multivariate binomial generalized linear mixed model after converting the outcomes to binary based on their median values. (3) Results: A total of 3226 counties/county-equivalents of the states and territories in the US were analyzed. Heart disease prevalence (odds ratio, OR = 2.03, p < 0.001), blood pressure medication nonadherence (OR = 2.02, p < 0.001), age-adjusted obesity (OR = 1.24, p = 0.006), presence of hospitals with neurological services (OR = 1.9, p < 0.001), and female head of household (OR = 1.32, p = 0.021) were associated with high SHR while cost of care per capita for Medicare patients with heart disease (OR = 0.5, p < 0.01) and presence of hospitals (OR = 0.69, p < 0.025) were associated with low SHR. Median household income (OR = 0.6, p < 0.001) and park access (OR = 0.84, p = 0.016) were associated with low SDR while no college degree (OR = 1.21, p = 0.049) was associated with high SDR. (4) Conclusions: Several socioeconomic factors (e.g., education, income, female head of household) were found to be associated with stroke outcomes. Additional research is needed to investigate intermediate and potentially modifiable factors that can serve as targeted interventions.