Understanding State-Level Variations in the US Infant Mortality: 2000 to 2015.

OBJECTIVE: To exploit state variations in infant mortality, identify diagnoses that contributed to reduction of the infant mortality rate (IMR), and examine factors associated with preterm-related mortality rate (PMR). STUDY DESIGN: Using linked birth-infant deaths files, we examined patterns in the leading causes of IMR. We compared these rates at both national and state levels to find reduction trends. Creating a cross-sectional time series of states' PMR and some explanatory variables, we implemented a fixed-effect regression model to examine factors associated with PMR at the state level. RESULTS: We found substantial state-level variations in changes of the IMR (range = - 2.87-2.08) and PMR (-1.77-0.67). Twenty-one states in which the IMR declined more than the national average of 0.99 (6.89-5.90) were labeled as successful. In the successful states, we found reduction in the PMR accounted for the largest decline in the IMR-0.90 fewer deaths. Changes in the other subgroups of leading causes did not differ significantly in successful and unsuccessful states. CONCLUSION: Trends in the causes of mortality are heterogeneous across states. Although its impact is not large, reducing the percentage of pregnant women with inadequate care is one of the mechanisms through which the PMR decrease.

Embracing Causal Complexity in Health Disparities: Metabolic Syndemics and Structural Prevention in Rural Minority Communities.

Chronic discrimination and associated socioeconomic inequalities have shaped the health and well-being of Black Americans. As a consequence of the intersection of these factors with rural deprivation, rural Black Americans live and work in particularly pathogenic environments that generate disproportionate and interacting chronic comorbidities (syndemics) compared to their White and/or urban counterparts. Traditional prevention research has been unable to fully capture the underlying complexity of rural minority health and has generated mostly low-leverage interventions that have failed to reverse adverse metabolic outcomes among rural Black Americans. In contrast, novel research approaches-such as system dynamics modeling-that seek to understand holistic system structure and determine complex health outcomes over time provide a robust framework to develop a more accurate understanding of the key factors contributing to type 2 diabetes. This framework can then be used to establish more efficacious interventions to address disparities among minorities in rural areas. This paper advocates for a unified complex systems epistemology and methodology in advancing rural minority health disparities research. Toward this goal, we (1) provide an overview of rural Black American metabolic health research, (2) introduce a complex systems framework in rural minority health disparities research, and (3) demonstrate how community-based system dynamics modeling and simulation can help us plow new ground in rural minority health disparities research and action. We anticipate that this paper can serve as a catalyst for a long-overdue discourse on the relevance of complex systems approaches in minority health research, with practical benefits for numerous disproportionately burdened communities.

Novel Screening Tool for Stroke Using Artificial Neural Network.

BACKGROUND AND PURPOSE: The timely diagnosis of stroke at the initial examination is extremely important given the disease morbidity and narrow time window for intervention. The goal of this study was to develop a supervised learning method to recognize acute cerebral ischemia (ACI) and differentiate that from stroke mimics in an emergency setting. METHODS: Consecutive patients presenting to the emergency department with stroke-like symptoms, within 4.5 hours of symptoms onset, in 2 tertiary care stroke centers were randomized for inclusion in the model. We developed an artificial neural network (ANN) model. The learning algorithm was based on backpropagation. To validate the model, we used a 10-fold cross-validation method. RESULTS: A total of 260 patients (equal number of stroke mimics and ACIs) were enrolled for the development and validation of our ANN model. Our analysis indicated that the average sensitivity and specificity of ANN for the diagnosis of ACI based on the 10-fold cross-validation analysis was 80.0% (95% confidence interval, 71.8-86.3) and 86.2% (95% confidence interval, 78.7-91.4), respectively. The median precision of ANN for the diagnosis of ACI was 92% (95% confidence interval, 88.7-95.3). CONCLUSIONS: Our results show that ANN can be an effective tool for the recognition of ACI and differentiation of ACI from stroke mimics at the initial examination.

Comparative analysis of machine learning versus traditional method for early detection of parental depression symptoms in the NICU.

Introduction: Neonatal intensive care unit (NICU) admission is a stressful experience for parents. NICU parents are twice at risk of depression symptoms compared to the general birthing population. Parental mental health problems have harmful long-term effects on both parents and infants. Timely screening and treatment can reduce these negative consequences. Objective: Our objective is to compare the performance of the traditional logistic regression with other machine learning (ML) models in identifying parents who are more likely to have depression symptoms to prioritize screening of at-risk parents. We used data obtained from parents of infants discharged from the NICU at Children's National Hospital (n = 300) from 2016 to 2017. This dataset includes a comprehensive list of demographic characteristics, depression and stress symptoms, social support, and parent/infant factors. Study design: Our study design optimized eight ML algorithms - Logistic Regression, Support Vector Machine, Decision Tree, Random Forest, XGBoost, Naïve Bayes, K-Nearest Neighbor, and Artificial Neural Network - to identify the main risk factors associated with parental depression. We compared models based on the area under the receiver operating characteristic curve (AUC), positive predicted value (PPV), sensitivity, and F-score. Results: The results showed that all eight models achieved an AUC above 0.8, suggesting that the logistic regression-based model's performance is comparable to other common ML models. Conclusion: Logistic regression is effective in identifying parents at risk of depression for targeted screening with a performance comparable to common ML-based models.

Modeling of drinking and driving behaviors among adolescents and young adults in the United States: Complexities and Intervention outcomes.

Alcohol-impaired driving is a formidable public health problem in the United States, claiming the lives of 37 individuals daily in alcohol-related crashes. Alcohol-impaired driving is affected by a multitude of interconnected factors, coupled with long delays between stakeholders' actions and their impacts, which not only complicate policy-making but also increase the likelihood of unintended consequences. We developed a system dynamics simulation model of drinking and driving behaviors among adolescents and young adults. This was achieved through group model building sessions with a team of multidisciplinary subject matter experts, and a focused literature review. The model was calibrated with data series from multiple sources and replicated the historical trends for male and female individuals aged 15 to 24 from 1982 to 2020. We simulated the model under different scenarios to examine the impact of a wide range of interventions on alcohol-related crash fatalities. We found that interventions vary in terms of their effectiveness in reducing alcohol-related crash fatalities. In addition, although some interventions reduce alcohol-related crash fatalities, some may increase the number of drinkers who drive after drinking. Based on insights from simulation experiments, we combined three interventions and found that the combined strategy may reduce alcohol-related crash fatalities significantly without increasing the number of alcohol-impaired drivers on US roads. Nevertheless, related fatalities plateau over time despite the combined interventions, underscoring the need for new interventions for a sustained decline in alcohol-related crash deaths beyond a few decades. Finally, through model calibration we estimated time delays between actions and their consequences in the system which provide insights for policymakers and activists when designing strategies to reduce alcohol-related crash fatalities.

Systems mapping of multilevel factors contributing to dental caries in adolescents.

Dental caries is a prevalent chronic disease among adolescents. Caries activity increases significantly during adolescence due to an increase in susceptible tooth surfaces, immature permanent tooth enamel, independence in pursuing self-care, and a tendency toward poor diet and oral hygiene. Dental caries in permanent teeth is more prevalent among adolescents in low-income families and racial/ethnic minority groups, and these disparities in adolescent dental caries experience have persisted for decades. Several conceptual and data-driven models have proposed unidirectional mechanisms that contribute to the extant disparities in adolescent dental caries experience. Our objective, using a literature review, is to provide an overview of risk factors contributing to adolescent dental caries. Specifically, we map the interactive relationships of multilevel factors that influence dental caries among adolescents. Such interactive multilevel relationships more closely reflect the complex nature of dental caries experience among the adolescent population. The methods that we use are two-fold: (1) a literature review using PubMed and Cochrane databases to find contributing factors; and (2) the system dynamics approach for mapping feedback mechanisms underlying adolescent dental caries through causal loop diagramming. The results of this study, based on the review of 138 articles, identified individual, family and community-level factors and their interactions contributing to dental caries experience in adolescents. Our results also provide hypotheses about the mechanisms underlying persistence of dental caries among adolescents. Conclusions: Our findings may contribute to a deeper understanding of the multilevel and interconnected factors that shape the persistence of dental caries experience among adolescents.

Exploring personalized psychotherapy for depression: A system dynamics approach.

Depressive disorders are the leading contributor to medical disability, yet only 22% of depressed patients receive adequate treatment in a given year. Response to treatment varies widely among individuals with depression, and poor response to one treatment does not signal poor response to others. In fact, half of patients who do not recover from a first-line psychotherapy will recover from a second option. Attempts to personalize psychotherapy to patient characteristics have produced better outcomes than usual care, but research on personalized psychotherapy is still in its infancy. The present study explores a new method for personalizing psychotherapy for depression through simulation modeling. In this study, we developed a system dynamics simulation model of depression based on one of the major mechanisms of depression in the literature and investigated the trend of depressive symptoms under different conditions and treatments. Our simulation outputs show the importance of individualized services with appropriate timing, and reveal a new method for personalizing psychotherapy to heterogeneous individuals. Future research is needed to expand the model to include additional mechanisms of depression.

What determines the success of states in reducing alcohol related crash fatalities? A longitudinal analysis of alcohol related crashes in the U.S. from 1985 to 2019.

In the United States, nearly 28 people die in alcohol-related motor vehicle crashes every day (1 fatality every 52 min). Over decades, states have enacted multiple laws to reduce such fatalities. From 1982 to 2019, the proportion of drivers in fatal crashes with a blood alcohol concentration (BAC) above 0.01 g/dl declined from 41% to 22%. States vary in terms of their success in reducing alcohol-related crash fatalities. The purpose of this study was to examine factors associated with changes in fatalities related to alcohol-impaired driving at the state level. We created a panel dataset of 50 states from 1985 to 2019 by merging different data sources and used fixed-effect linear regression models to analyze the data. Our two outcome variables were the ratio of drivers in fatal crashes with BAC ≥ 0.01 g/dl to those with BAC = 0.00, and the ratio of those with BAC ≥ 0.08 g/dl to those with BAC < 0.08 g/dl. Our independent variables included four laws (0.08 g/dl BAC per se law, administrative license revocation law, minimum legal drinking age law, and zero tolerance law), number of arrests due to impaired driving, alcohol consumption per capita, unemployment rate, and vehicle miles traveled. We found that the 0.08 g/dl per se law was significantly associated with lower alcohol-related crash fatalities while alcohol consumption per capita was significantly and positively associated with crash-related fatalities. Arrests due to driving under the influence (DUI) and crash fatalities were nonlinearly correlated. In addition, interaction of DUI arrests and two laws (0.08 g/dl BAC per se law, and zero tolerance) were significantly associated with lower crash-related fatalities. Our findings suggest that states which have more restrictive laws and enforce them are more likely to significantly reduce alcohol-related crash fatalities.

Trajectories and Outcomes of Adolescents that Ride With an Impaired Driver/Drive While Impaired.

Introduction: For young drivers, independent transportation has been noted to offer them opportunities that can be beneficial as they enter early adulthood. However, those that choose to engage in riding with an impaired driver (RWI) and drive while impaired (DWI) over time can face negative consequences reducing such opportunities. This study examined the prospective association of identified longitudinal trajectory classes among adolescents that RWI and DWI with their later health, education, and employment in emerging adulthood. Methods: We analyzed all seven annual assessments (Waves, W1-W7) of the NEXT Generation Health Study, a nationally representative longitudinal study starting with 10th grade (2009-2010 school year). Using all seven waves, trajectory classes were identified by latent class analysis with RWI (last 12 months) and DWI (last 30 days) dichotomized as ≥once = 1 vs. none = 0. Results: Four RWI trajectories and four DWI trajectories were identified: abstainer, escalator, decliner, and persister. For RWI and DWI trajectories respectively, 45.0% (N=647) and 76.2% (N=1,657) were abstainers, 15.6% (N=226) and 14.2% (N=337) were escalators, 25.0% (N=352) and 5.4% (N=99) were decliners, and 14.4% (N=197) and 3.8% (N=83) persisters. RWI trajectories were associated with W7 health status (χ2=13,20, p<.01) and education attainment (χ2=18.37, p<.01). Adolescent RWI abstainers reported better later health status than RWI escalators, decliners, and persisters; and decliners reported less favorable later education attainment than abstainers, escalators, and persisters. DWI trajectories showed no association with health status, education attainment, or employment. Conclusions: Our findings suggest the importance of later health outcomes of adolescent RWI. The mixed findings point to the need for more detailed understanding of contextual and time-dependent trajectory outcomes among adolescents engaging in RWI and DWI.

Mapping the complex causal mechanisms of drinking and driving behaviors among adolescents and young adults.

BACKGROUND: The proportion of motor vehicle crash fatalities involving alcohol-impaired drivers declined substantially between 1982 and 1997, but progress stopped after 1997. The systemic complexity of alcohol-impaired driving contributes to the persistence of this problem. This study aims to identify and map key feedback mechanisms that affect alcohol-impaired driving among adolescents and young adults in the U.S. METHODS: We apply the system dynamics approach to the problem of alcohol-impaired driving and bring a feedback perspective for understanding drivers and inhibitors of the problem. The causal loop diagram (i.e., map of dynamic hypotheses about the structure of the system producing observed behaviors over time) developed in this study is based on the output of two group model building sessions conducted with multidisciplinary subject-matter experts bolstered with extensive literature review. RESULTS: The causal loop diagram depicts diverse influences on youth impaired driving including parents, peers, policies, law enforcement, and the alcohol industry. Embedded in these feedback loops are the physical flow of youth between the categories of abstainers, drinkers who do not drive after drinking, and drinkers who drive after drinking. We identify key inertial factors, discuss how delay and feedback processes affect observed behaviors over time, and suggest strategies to reduce youth impaired driving. CONCLUSION: This review presents the first causal loop diagram of alcohol-impaired driving among adolescents and it is a vital first step toward quantitative simulation modeling of the problem. Through continued research, this model could provide a powerful tool for understanding the systemic complexity of impaired driving among adolescents, and identifying effective prevention practices and policies to reduce youth impaired driving.

Machine Learning-Enabled 30-Day Readmission Model for Stroke Patients.

Background and Purpose: Hospital readmissions impose a substantial burden on the healthcare system. Reducing readmissions after stroke could lead to improved quality of care especially since stroke is associated with a high rate of readmission. The goal of this study is to enhance our understanding of the predictors of 30-day readmission after ischemic stroke and develop models to identify high-risk individuals for targeted interventions. Methods: We used patient-level data from electronic health records (EHR), five machine learning algorithms (random forest, gradient boosting machine, extreme gradient boosting-XGBoost, support vector machine, and logistic regression-LR), data-driven feature selection strategy, and adaptive sampling to develop 15 models of 30-day readmission after ischemic stroke. We further identified important clinical variables. Results: We included 3,184 patients with ischemic stroke (mean age: 71 ± 13.90 years, men: 51.06%). Among the 61 clinical variables included in the model, the National Institutes of Health Stroke Scale score above 24, insert indwelling urinary catheter, hypercoagulable state, and percutaneous gastrostomy had the highest importance score. The Model's AUC (area under the curve) for predicting 30-day readmission was 0.74 (95%CI: 0.64-0.78) with PPV of 0.43 when the XGBoost algorithm was used with ROSE-sampling. The balance between specificity and sensitivity improved through the sampling strategy. The best sensitivity was achieved with LR when optimized with feature selection and ROSE-sampling (AUC: 0.64, sensitivity: 0.53, specificity: 0.69). Conclusions: Machine learning-based models can be designed to predict 30-day readmission after stroke using structured data from EHR. Among the algorithms analyzed, XGBoost with ROSE-sampling had the best performance in terms of AUC while LR with ROSE-sampling and feature selection had the best sensitivity. Clinical variables highly associated with 30-day readmission could be targeted for personalized interventions. Depending on healthcare systems' resources and criteria, models with optimized performance metrics can be implemented to improve outcomes.

Riding with an impaired driver and driving while impaired among adolescents: Longitudinal trajectories and their characteristics.

PURPOSE: The purpose of this study was to identify and characterize trajectory classes of adolescents who ride with an impaired driver (RWI) and drive while impaired (DWI). METHODS: We analyzed all 7 annual assessments (Waves W1-W7) of the NEXT Generation Health Study, a nationally representative longitudinal study starting with 10th grade (2009-2010 school year). Using all 7 waves, latent class analysis was used to identify trajectory classes with dichotomized RWI (last 12 months) and DWI (last 30 days; once or more = 1 vs. none = 0). Covariates were race/ethnicity, sex, parent education, urbanicity, and family affluence. RESULTS: Four RWI trajectories and 4 DWI trajectories were identified: abstainer, escalator, decliner, and persister. For RWI and DWI trajectories respectively, 45.0% (n = 647) and 76.2% (n = 1,657) were abstainers, 15.6% (n = 226) and 14.2% (n = 337) were escalators, 25.0% (n = 352) and 5.4% (n = 99) were decliners, and 14.4% (n = 197) and 3.8% (n = 83) persisters. Race/ethnicity (χ2 = 23.93, P = .004) was significantly associated with the RWI trajectory classes. Race/ethnicity (χ2 = 20.55, P = .02), sex (χ2 = 13.89, P = .003), parent highest education (χ2 = 12.49, P = .05), urbanicity (χ2 = 9.66, P = .02), and family affluence (χ2 = 12.88, P = .05) were significantly associated with DWI trajectory classes. CONCLUSIONS: Among adolescents transitioning into emerging adulthood, race/ethnicity is a common factor associated with RWI and DWI longitudinal trajectories. Our results suggest that adolescent RWI and DWI are complex behaviors warranting further detailed investigation of the respective trajectory classes. Our study findings can inform the tailoring of prevention and intervention efforts aimed at preventing illness/injury and preserving future opportunities for adolescents to thrive in emerging adulthood.

Modeling and estimating the feedback mechanisms among depression, rumination, and stressors in adolescents.

The systemic interactions among depressive symptoms, rumination, and stress are important to understanding depression but have not yet been quantified. In this article, we present a system dynamics simulation model of depression that captures the reciprocal relationships among stressors, rumination, and depression. Building on the response styles theory, this model formalizes three interdependent mechanisms: 1) Rumination contributes to 'keeping stressors alive'; 2) Rumination has a direct impact on depressive symptoms; and 3) Both 'stressors kept alive' and current depressive symptoms contribute to rumination. The strength of these mechanisms is estimated using data from 661 adolescents (353 girls and 308 boys) from two middle schools (grades 6-8). These estimates indicate that rumination contributes to depression by keeping stressors 'alive'-and the individual activated-even after the stressor has ended. This mechanism is stronger among girls than boys, increasing their vulnerability to a rumination reinforcing loop. Different profiles of depression emerge over time depending on initial levels of depressive symptoms, rumination, and stressors as well as the occurrence rate for stressors; levels of rumination and occurrence of stressors are stronger contributors to long-term depression. Our systems model is a steppingstone towards a more comprehensive understanding of depression in which reinforcing feedback mechanisms play a significant role. Future research is needed to expand this simulation model to incorporate other drivers of depression and provide a more holistic tool for studying depression.

Modeling the hypothalamus-pituitary-adrenal axis: A review and extension.

Multiple models of the hypothalamus-pituitary-adrenal (HPA) axis have been developed to characterize the oscillations seen in the hormone concentrations and to examine HPA axis dysfunction. We reviewed the existing models, then replicated and compared five of them by finding their correspondence to a dataset consisting of ACTH and cortisol concentrations of 17 healthy individuals. We found that existing models use different feedback mechanisms, vary in the level of details and complexities, and offer inconsistent conclusions. None of the models fit the validation dataset well. Therefore, we re-calibrated the best performing model using partial calibration and extended the model by adding individual fixed effects and an exogenous circadian function. Our estimated parameters reduced the mean absolute percent error significantly and offer a validated reference model that can be used in diverse applications. Our analysis suggests that the circadian and ultradian cycles are not created endogenously by the HPA axis feedbacks, which is consistent with the recent literature on the circadian clock and HPA axis.