Gestational diabetes and preterm birth are perinatal morbidities that significantly impact women and infants' health. While clinical factors like cesarean delivery, multiple gestation, preeclampsia, and hypertensive disorder are associated with these conditions, it is increasingly recognized that social determinants of health play a crucial role. This study aims to measure the associations between the social vulnerability index (SVI) and these perinatal morbidities using multivariate logistic regression models. The results indicate that factors across all four themes in SVI are significantly associated with these conditions. These findings suggest that interventions targeting these areas are needed to achieve better reproductive health.

Telehealth is an alternative care delivery model to in-person care. It uses electronic information and telecommunication technologies to provide remote clinical care to patients, especially those living in rural areas that lack sufficient access to health care services. Like other areas of care affected by the COVID-19 pandemic, the prevalence of telehealth has increased in prenatal care. This study reports on telehealth use in prenatal care at a large academic medical center in Middle Tennessee, USA. We examine the electronic health records of over 2500 women to characterize 1) the volume of prenatal visits participating in telehealth, 2) disparities in obstetric patients using telehealth, and 3) the impact of telehealth use on obstetric outcomes, including duration of intrapartum hospital stays, preterm birth, Cesarean rate, and newborn birthweight. Our results show that telehealth mainly was used in the second and third trimesters, especially for consulting services. In addition, we found that certain demographics correlated with lower telehealth utilization, including patients who were under 26 years old, were Black and/or Hispanic, were on a state-sponsored health insurance program, and those who lived in urban areas. Furthermore, no significant differences were found on preterm birth and Cesarean between the patients who used telehealth in their prenatal care and those who did not.

Telehealth is designed to provide health services through the use of electronic information and telecommunication technologies. It has quickly become an important tool to ensure continued care in response to the COVID-19 pandemic while mitigating the risk of viral exposure for patients and providers. This study compared the number of monthly telehealth visits in primary care settings at a large academic medical center from 2019 and 2020. To investigate what health conditions are suitable for telehealth visits, we report on the ten ICD-10 codes with the largest number of telehealth visits.

Background Few intensive care unit (ICU) staffing studies have examined the collaboration structures of health care workers (HCWs). Knowledge about how HCWs are connected to the care of critically ill patients with COVID-19 is important for characterizing the relationships among team structures, care quality, and patient safety. Objective We aimed to discover differences in the teamwork structures of COVID-19 critical care by comparing HCW collaborations in the management of critically ill patients with and without COVID-19. Methods In this retrospective study, we used network analysis methods to analyze the electronic health records (EHRs) of 76 critically ill patients (with COVID-19: n=38; without COVID-19: n=38) who were admitted to a large academic medical center, and to learn about HCW collaboration. We used the EHRs of adult patients who were admitted to the COVID-19 ICU at the Vanderbilt University Medical Center (Nashville, Tennessee, United States) between March 17, 2020, and May 31, 2020. We matched each patient according to age, gender, and their length of stay. Patients without COVID-19 were admitted to the medical ICU between December 1, 2019, and February 29, 2020. We used two sociometrics—eigencentrality and betweenness—to quantify HCWs’ statuses in networks. Eigencentrality characterizes the degree to which an HCW is a core person in collaboration structures. Betweenness centrality refers to whether an HCW lies on the path of other HCWs who are not directly connected. This sociometric was used to characterize HCWs’ broad skill sets. We measured patient staffing intensity in terms of the number of HCWs who interacted with patients’ EHRs. We assessed the statistical differences in the core and betweenness statuses of HCWs and the patient staffing intensities of COVID-19 and non–COVID-19 critical care, by using Mann-Whitney U tests and reporting 95% CIs. Results HCWs in COVID-19 critical care were more likely to frequently work with each other (eigencentrality: median 0.096) than those in non–COVID-19 critical care (eigencentrality: median 0.057; P<.001). Internal medicine physicians in COVID-19 critical care had higher core statuses than those in non–COVID-19 critical care (P=.001). Nurse practitioners in COVID-19 care had higher betweenness statuses than those in non–COVID-19 care (P<.001). Compared to HCWs in non–COVID-19 settings, the EHRs of critically ill patients with COVID-19 were used by a larger number of internal medicine nurse practitioners (P<.001), cardiovascular nurses (P<.001), and surgical ICU nurses (P=.002) and a smaller number of resident physicians (P<.001). Conclusions Network analysis methodologies and data on EHR use provide a novel method for learning about differences in collaboration structures between COVID-19 and non–COVID-19 critical care. Health care organizations can use this information to learn about the novel changes that the COVID-19 pandemic has imposed on collaboration structures in urgent care.

Objective: In intensive care units (ICUs), a patient's brain function status can shift from a state of acute brain dysfunction (ABD) to one that is ABD-free and vice versa, which is challenging to forecast and, in turn, hampers the allocation of hospital resources. We aim to develop a machine learning model to predict next-day brain function status changes. Materials and Methods: Using multicenter prospective adult cohorts involving medical and surgical ICU patients from 2 civilian and 3 Veteran Affairs hospitals, we trained and externally validated a light gradient boosting machine to predict brain function status changes. We compared the performances of the boosting model against state-of-the-art models-an ABD predictive model and its variants. We applied Shapley additive explanations to identify influential factors to develop a compact model. Results: There were 1026 critically ill patients without evidence of prior major dementia, or structural brain diseases, from whom 12 295 daily transitions (ABD: 5847 days; ABD-free: 6448 days) were observed. The boosting model achieved an area under the receiver-operating characteristic curve (AUROC) of 0.824 (95% confidence interval [CI], 0.821-0.827), compared with the state-of-the-art models of 0.697 (95% CI, 0.693-0.701) with P< .001. Using 13 identified top influential factors, the compact model achieved 99.4% of the boosting model on AUROC. The boosting and the compact models demonstrated high generalizability in external validation by achieving an AUROC of 0.812 (95% CI, 0.812-0.813). Conclusion: The inputs of the compact model are based on several simple questions that clinicians can quickly answer in practice, which demonstrates the model has direct prospective deployment potential into clinical practice, aiding in critical hospital resource allocation.

Objectives: Like other areas of care affected by the COVID-19 pandemic, telehealth (both audio and video) was rapidly adopted in the obstetric setting. We performed a retrospective analysis of electronic health record (EHR) data to characterize the sociodemographic and clinical factors associated with telehealth utilization among patients who received prenatal care. Materials and Methods: The study period covered March 23rd, 2020 to July 2nd, 2020, during which time 2,521 patients received prenatal care at a large academic medical center. We applied a generalized logistic regression to measure the relationship between the patients’ sociodemographic factors (in terms of age, race, ethnicity, urbanization level, and insurance type), pregnancy complications (namely, type 2 diabetes, chronic hypertension, and fetal growth restriction), and telehealth usage, as documented in the EHR. Results: During the study period, 2,521 patients had 16,516 prenatal care visits. 938 (37.2%) of the patients participated in at least one of 1,934 virtual prenatal care visits. Prenatal visits were more likely to be conducted through telehealth for patients who were older than 25 years old and lived in rural areas. In addition, patients who were with type 2 diabetes were more likely to use telehealth in their prenatal care (adjusted Odds Ratio (aOR) 7.247 [95% Confidence Interval (95% CI) 4.244 – 12.933]). By contrast, patients from racial and ethnic minority groups were less likely to have a telehealth encounter compared to white or non-Hispanic patients (aOR 0.603 [95% CI 0.465 – 0.778] and aOR 0.663 [95% CI 0.471 – 0.927], respectively). Additionally, patients who were on state-level Medicaid were less likely to use telehealth (aOR 0.495 [95% CI 0.402 – 0.608]). Discussion: Disparities in telehealth use for prenatal care suggest further investigations into access barriers. Hispanic patients who had low English language proficiency may not willing to see doctors via virtual care. Availability of high-speed internet and/or hardware may hold these patients who were insured through state-level Medicaid back due to poverty. Future work is advised to minimize access barriers to telehealth in its implementation. Conclusions: While telehealth expanded prenatal care access for childbearing women during the COVID-19 pandemic, this study suggested that there were non-trivial differences in the demographics of patients who utilized such settings.

OBJECTIVE:The characteristics of clinician activities while interacting with electronic health record (EHR) systems can influence the time spent in EHRs and workload. This study aims to characterize EHR activities as tasks and define novel, data-driven metrics.MATERIALS AND METHODS:We leveraged unsupervised learning approaches to learn tasks from sequences of events in EHR audit logs. We developed metrics characterizing the prevalence of unique events and event repetition and applied them to categorize tasks into 4 complexity profiles. Between these profiles, Mann-Whitney U tests were applied to measure the differences in performance time, event type, and clinician prevalence, or the number of unique clinicians who were observed performing these tasks. In addition, we apply process mining frameworks paired with clinical annotations to support the validity of a sample of our identified tasks. We apply our approaches to learn tasks performed by nurses in the Vanderbilt University Medical Center neonatal intensive care unit.RESULTS:We examined EHR audit logs generated by 33 neonatal intensive care unit nurses resulting in 57 234 sessions and 81 tasks. Our results indicated significant differences in performance time for each observed task complexity profile. There were no significant differences in clinician prevalence or in the frequency of viewing and modifying event types between tasks of different complexities. We presented a sample of expert-reviewed, annotated task workflows supporting the interpretation of their clinical meaningfulness.CONCLUSIONS:The use of the audit log provides an opportunity to assist hospitals in further investigating clinician activities to optimize EHR workflows.

Objective: There is a wide range of risk factors predisposing to the onset of neonatal encephalopathy (NE), including maternal antepartum/intrapartum comorbidities or events. However, few studies have investigated the difference in the impact of acute and chronic diseases on forecasting NE, which could assist clinicians in choosing the best course of action to prevent NE or reduce its severity and complications. In this study, we aimed to engineer features based on acute and chronic diseases and assess the differences of the impact of acute and chronic diseases on NE prediction using machine learning models. Materials and methods: We used ten years of electronic health records of mothers from a large academic medical center to develop three types of features: chronic disease, recurrence of an acute disease, and temporal relationships between acute diseases. Two types of NE prediction models, based on acute and chronic diseases, respectively, were trained with feature selection. We further compared the prediction performance of the models with two state-of-the-art NE forecasting models. The machine learning models ranked the three types of engineered features based on their contributions to the NE prediction. Results: The NE model trained on acute disease features showed significantly higher AUC than the model relying on chronic disease features (AUC difference: 0.161, p-value < 0.001). The NE model trained on both acute and chronic disease features achieved the highest average AUC (0.889), with a significant improvement over the best existing model (0.854) with p = 0.0129. Recurrence of "known or suspected fetal abnormality affecting management of mother (655) " was assigned the highest weights in predicting NE. Conclusions: Machine learning models based on the three types of engineered features significantly improve NE prediction. Our results specifically suggest that acute disease-associated features play a more important role in predicting NE. (c) 2021 Elsevier B.V. All rights reserved.

Deep learning architectures have an extremely high-capacity for modeling complex data in a wide variety of domains. However, these architectures have been limited in their ability to support complex prediction problems using insurance claims data, such as readmission at 30 days, mainly due to data sparsity issue. Consequently, classical machine learning methods, especially those that embed domain knowledge in handcrafted features, are often on par with, and sometimes outperform, deep learning approaches. In this paper, we illustrate how the potential of deep learning can be achieved by blending domain knowledge within deep learning architectures to predict adverse events at hospital discharge, including readmissions. More specifically, we introduce a learning architecture that fuses a representation of patient data computed by a self-attention based recurrent neural network, with clinically relevant features. We conduct extensive experiments on a large claims dataset and show that the blended method outperforms the standard machine learning approaches.

Adopting non-pharmaceutical interventions (NPIs) can affect COVID-19 growing trends, decrease the number of infected cases, and thus reduce mortality and healthcare demand. Almost all countries in the world have adopted non-pharmaceutical interventions (NPIs) to control the spread rate of COVID-19; however, it is unclear what are differences in the effectiveness of NPIs among these countries. We hypothesize that COVID-19 case growth data reveals the efficacy of NPIs. In this study, we conduct a secondary analysis of COVID-19 case growth data to compare the differences in the effectiveness of NPIs among 16 representative countries in the world. This study leverages publicly available data to learn patterns of dynamic changes in the reproduction rate for sixteen countries covering Asia, Europe, North America, South America, Australia, and Africa. Furthermore, we model the relationships between the cumulative number of cases and the dynamic reproduction rate to characterize the effectiveness of the NPIs. We learn four levels of NPIs according to their effects in the control of COVID-19 growth and categorize the 16 countries into the corresponding groups. The dynamic changes of the reproduction rate are learned via linear regression models for all of the studied countries, with the average adjusted R-squared at 0.96 and the 95% confidence interval as [0.94 0.98]. China, South Korea, Argentina, and Australia are at the first level of NPIs, which are the most effective. Japan and Egypt are at the second level of NPIs, and Italy, Germany, France, Netherlands, and Spain, are at the third level. The US and UK have the most inefficient NPIs, and they are at the fourth level of NPIs. COVID-19 case growth data provides evidence to demonstrate the effectiveness of the NPIs. Understanding the differences in the efficacy of the NPIs among countries in the world can give guidance for emergent public health events. NA

The data in a patient’s laboratory test result is a notable resource to support clinical investigation and enhance medical research. However, for a variety of reasons, this type of data often contains a non-trivial number of missing values. For example, physicians may neglect to order tests or document the results. Such a phenomenon reduces the degree to which this data can be utilized to learn efficient and effective predictive models. To address this problem, various approaches have been developed to impute missing laboratory values; however, their performance has been limited. This is due, in part, to the fact no approaches effectively leverage the contextual information (1) in individual or (2) between laboratory test variables. We introduce an approach to combine an unsupervised prefilling strategy with a supervised machine learning approach, in the form of extreme gradient boosting (XGBoost), to leverage both types of context for imputation purposes. We evaluated the methodology through a series of experiments on approximately 8200 patients’ records in the MIMIC-III dataset. The results demonstrate that the new model outperforms baseline and state-of-the-art models on 13 commonly collected laboratory test variables. In terms of the normalized root mean square derivation (nRMSD), our model exhibits an imputation improvement by over 20%, on average. Missing data imputation on the temporal variables can be largely improved via prefilling strategy and the supervised training technique, which leverages both the longitudinal and cross-sectional context simultaneously.

Objective It has been recognized that efficient resource allocation can improve patient care quality and reduce health care expenditure, especially in a trauma setting, where inpatient resource consumption is extremely high. Predicting patient discharge disposition early can help healthcare organizations allocate an appropriate amount of resources for an estimated number of inpatient stays. Concurrently, early prediction of discharge disposition can help providers to schedule subsequent post-trauma care to ensure continuity of care. With patient overcrowding being a widespread issue in hospitals across the US, proper allocation of resources and early coordination of post-trauma care can alleviate this issue. The objective of our study is to apply graph convolutional networks (GCN) on electronic health records (EHRs) to predict discharge dispositions one day after patient admissions.

Severe maternal morbidity (SMM) encompasses a wide range of serious health complications that would likely result in death without in-time medical attention. It has been recognized that various demographic factors (e.g., age and race) and medical conditions (e.g., preeclampsia and organ failure) are associated with SMM However, how medical conditions develop into SMM is seldom investigated. We hypothesize that SMM has a progression path, which is associated with a sequence of risk factors rather than a set of independent individual factors. We implemented a data-driven framework that leverages electronic health records (EHRs) in the antepartum period to learn the temporal patterns and measure their relationships with SMM during the delivery hospitalization. We evaluate the framework with two years of data from 6,184 women who had delivery hospitalizations at Vanderbilt University Medical Center. We discovered 69 temporal patterns, 12 of which were confirmed to be significantly associated with SMM

Objective: Models for predicting preterm birth generally have focused on very preterm (28-32 weeks) and moderate to late preterm (32-37 weeks) settings. However, extreme preterm birth (EPB), before the 28th week of gestational age, accounts for the majority of newborn deaths. We investigated the extent to which deep learning models that consider temporal relations documented in electronic health records (EHRs) can predict EPB. Study design: EHR data were subject to word embedding and a temporal deep learning model, in the form of recurrent neural networks (RNNs) to predict EPB. Due to the low prevalence of EPB, the models were trained on datasets where controls were undersampled to balance the case-control ratio. We then applied an ensemble approach to group the trained models to predict EPB in an evaluation setting with a nature EPB ratio. We evaluated the RNN ensemble models with 10 years of EHR data from 25,689 deliveries at Vanderbilt University Medical Center. We compared their performance with traditional machine learning models (logistical regression, support vector machine, gradient boosting) trained on the datasets with balanced and natural EPB ratio. Risk factors associated with EPB were identified using an adjusted odds ratio. Results: The RNN ensemble models trained on artificially balanced data achieved a higher AUC (0.827 vs. 0.744) and sensitivity (0.965 vs. 0.682) than those RNN models trained on the datasets with naturally imbalanced EPB ratio. In addition, the AUC (0.827) and sensitivity (0.965) of the RNN ensemble models were better than the AUC (0.777) and sensitivity (0.819) of the best baseline models trained on balanced data. Also, risk factors, including twin pregnancy, short cervical length, hypertensive disorder, systemic lupus erythematosus, and hydroxychloroquine sulfate, were found to be associated with EPB at a significant level. Conclusion: Temporal deep learning can predict EPB up to 8 weeks earlier than its occurrence. Accurate prediction of EPB may allow healthcare organizations to allocate resources effectively and ensure patients receive appropriate care.

Severe maternal morbidity (SMM) is broadly defined as significant complications in pregnancy that have an adverse effect on women's health. Identifying women who experience SMM and reviewing their obstetric care can assist healthcare organizations in recognizing risk factors and best practices for management. Various definitions of SMM have been posited, but there is no consensus. Existing definitions are further limited in that they 1) are often rooted in existing clinical knowledge (which is problematic as many risk factors remain unknown), leading to poor positive predictive performance (PPV), and 2) have limited scalability as they often require substantial chart review. Thus, in this paper, a machine learning framework was introduced to automatically identify SMM and relevant risk factors from electronic health records (EHRs). We evaluated this framework with EHR data from 45,858 deliveries at a large academic medical center. The framework outperformed a state-of-the-art model from the U.S. Centers for Disease Control and Prevention (AUC of 0.94 vs. 0.80). Specially, it improved upon PPV by 59% (CDC: 0.22 vs. our model: 0.35). In the process, we revealed several novel SMM indicators, including disorders of fluid or electrolytes, systemic inflammatory response syndrome, and acidosis.

Data quality plays an important role in the data-driven based biomedical informatics research because the effectiveness of these research heavily relies on the completeness of data being collected. The problem of missing values, however, is commonly encountered in research practice, which impedes researchers to build accurate models and then make reasonable decisions. Simply removing the data instances having missing value(s) is a candidate strategy, but risks incorporating biases or even yielding incorrect models. [1]. Thus, imputing accurate missing values is a prerequisite to training good machine learning models.

There is a general belief that the workflow of surrounding location transfers between locations documented in electronic health record (EHR) during hospitalization is associated with a patient's length of stay (LOS). However, this belief has had little formal investigation in a data-driven manner. Location transfers in patients' hospitalization are hypothesized to be related to LOS. The objective of this study is to assess this relationship, using data derived from the EHR system of a large hospital system, with a focus on the obstetric setting - a clinical environment that exhibits wide swing in resource utilization. We designed a data-driven framework to infer patterns of location transfers and developed a zero-truncated negative binomial model, adjusting for demographics and billed diagnoses, to learn the association between patterns of location transfers and LOS. Indicative factors found to be of indicative of location transfer patterns were further investigated via their odds ratios, Pearson Correlation Coefficients, and Chi-squared test. We evaluated our approach with two years of data on from 5,774 obstetric inpatient encounters from the EHR system of Northwestern Memorial Hospital. The results indicated that the average LOS for patients with patterns of repetitious location transfers (RLTs) was 4.25 days (95% confidence interval [4.02, 4.47]) longer than patients with no RLT. This difference reduced to 3.62 days (95% confidence interval [3.61, 3.64]) after adjusting for age, race and billed diagnoses. We further discovered 21 indicative factors of RLT (statistically significant with a significance level of 0.05), in the form of billed diagnosis codes, each of which exhibited an odds ratio larger than 4. This study suggests that RLT patterns are associated with a prolonged LOS in the obstetric setting. As such, healthcare organizations may need to pay more attention to patients with RLTs to refine location transfers workflow and to boost efficiency in obstetric care.