Search

Machine learning (ML) refers to computational algorithms that iteratively improve their ability to recognize patterns in data. The digitization of our healthcare infrastructure is generating an abundance of data from electronic health records, imaging, wearables, and sensors that can be analyzed by ML algorithms to generate personalized risk assessments and promote guideline-directed medical management. ML's strength in generating insights from complex medical data to guide clinical decisions must be balanced with the potential to adversely affect patient privacy, safety, health equity, and clinical interpretability. This review provides a primer on key advances in ML for cardiovascular disease prevention and how they may impact clinical practice.

Objective: Despite technological and treatment advancements over the past two decades, cardiogenic shock (CS) mortality has remained between 40-60%. A number of factors can lead to delayed diagnosis of CS, including gradual onset and nonspecific symptoms. Our objective was to develop an algorithm that can continuously monitor heart failure patients, and partition them into cohorts of high- and low-risk for CS.Methods: We retrospectively studied 24,461 patients hospitalized with acute decompensated heart failure, 265 of whom developed CS, in the Johns Hopkins Healthcare system. Our cohort identification approach is based on logistic regression, and makes use of vital signs, lab values, and medication administrations recorded during the normal course of care. Results: Our algorithm identified patients at high-risk of CS. Patients in the high-risk cohort had 10.2 times (95% confidence interval 6.1-17.2) higher prevalence of CS than those in the low-risk cohort. Patients who experienced cardiogenic shock while in the high-risk cohort were first deemed high-risk a median of 1.7 days (interquartile range 0.8 to 4.6) before cardiogenic shock diagnosis was made by their clinical team. Conclusions: This risk model was able to predict patients at higher risk of CS in a time frame that allowed a change in clinical care. Future studies need to evaluate if CS analysis of high-risk cohort identification may affect outcomes.

The term “big data” has gotten increasing popular attention, and there is growing focus on how such data can be used to measure and improve health and healthcare. Analytic techniques for extracting information from these data have grown vastly more powerful, and they are now broadly available. But for these approaches to be most useful, large amounts of data must be available, and barriers to use should be low. We discuss how “smart cities” are beginning to invest in this area to improve the health of their populations; provide examples around model approaches for making large quantities of data available to researchers and clinicians among other stakeholders; discuss the current state of big data approaches to improve clinical care including specific examples, and then discuss some of the policy issues around and examples of successful regulatory approaches, including deidentification and privacy protection.

A major goal of neonatal medicine is to identify neonates at highest risk for morbidity and mortality. Previously, we developed PhysiScore (Saria et al., 2010), a novel tool for preterm morbidity risk prediction. We now further define links between overall individual morbidity risk, specific neonatal morbidities, and placental pathologies.102 placentas, including 38 from multiple gestations, were available from the previously defined PhysiScore cohort (gestational age ≤ 34 weeks and birth weight ≤ 2000 g). Placentas were analyzed for gross and histologic variables including maternal malperfusion, amniotic fluid infection sequence, chronic inflammation, and fetal vascular obstruction. Risk as determined by PhysiScore and recorded neonatal morbidities were tested for statistical association with placental findings.In pair-wise correlations, respiratory distress syndrome, bronchopulmonary dysplasia, acute hemodynamic instability, post-hemorrhagic hydrocephalus, culture-positive sepsis, and necrotizing enterocolitis each significantly correlated with at least one placenta histology variable. Amniotic fluid infection sequence (p = 0.039), specifically the fetal inflammatory response (p = 0.017), correlated with higher PhysiScores (greater morbidity) but was not independent of gestational age and birth weight. In multivariate analyses correlating variables with all nine morbidities, gestational age (p 0.001), placental size10th percentile (p = 0.031), full thickness perivillous fibrin deposition (p = 0.001), and amniotic fluid infection sequence (umbilical arteritis, p = 0.031; ≥2 chorionic plate vessels with vasculitis, p = 0.0125), each were significant associations.Amniotic fluid infection sequence plays a significant role in neonatal morbidity. Less neonatal morbidity was observed in older and heavier infants and those with small placental size and full thickness perivillous fibrin deposition. The combined assessment of placental gross and histologic findings together with physiologic risk evaluation may allow more precise prediction of neonatal morbidity risk soon after delivery.