ABSTRACT
Environmental health studies are increasingly measuring multiple pollutants to characterize the joint health effects attributable to exposure mixtures. However, the underlying dose-response relationship between toxicants and health outcomes of interest may be highly nonlinear, with possible nonlinear interaction effects. Existing penalized regression methods that account for exposure interactions either cannot accommodate nonlinear interactions while maintaining strong heredity or are computationally unstable in applications with limited sample size. In this paper, we propose a general shrinkage and selection framework to identify noteworthy nonlinear main and interaction effects among a set of exposures. We design hierarchical integrative group least absolute shrinkage and selection operator (HiGLASSO) to (a) impose strong heredity constraints on two-way interaction effects (hierarchical), (b) incorporate adaptive weights without necessitating initial coefficient estimates (integrative), and (c) induce sparsity for variable selection while respecting group structure (group LASSO). We prove sparsistency of the proposed method and apply HiGLASSO to an environmental toxicants dataset from the LIFECODES birth cohort, where the investigators are interested in understanding the joint effects of 21 urinary toxicant biomarkers on urinary 8-isoprostane, a measure of oxidative stress. An implementation of HiGLASSO is available in the higlasso R package, accessible through the Comprehensive R Archive Network.
ABSTRACT
BACKGROUND: Anxiety and depression disorders (internalizing psychopathology) occur in approximately 50% of patients with alcohol use disorder (AUD) and mark a 2-fold increase in the rate of relapse in the months following treatment. In a previous study using network modeling, we found that perceived stress and drinking to cope (DTC) with negative affect were central to maintaining network associations between internalizing psychopathology INTP and drinking in comorbid individuals. Here, we extend this approach to a causal framework. METHODS: Measures of INTP, drinking urges/behavior, abstinence self-efficacy, and DTC were obtained from 362 adult AUD treatment patients who had a co-occurring anxiety disorder. Data were analyzed using a machine-learning algorithm ("Greedy Fast Causal Inference"[ GFCI]) that infers paths of causal influence while identifying potential influences associated with unmeasured ("latent") variables. RESULTS: DTC with negative affect served as a central hub for 2 distinct causal paths leading to drinking behavior, (i) a direct syndromic pathway originating with social anxiety and (ii) an indirect stress pathway originating with perceived stress. CONCLUSIONS: Findings expand the field's knowledge of the paths of influence that lead from internalizing disorder to drinking in AUD as shown by the first application in psychopathology of a powerful network analysis algorithm (GFCI) to model these causal relationships.
Subject(s)
Alcohol Drinking/psychology , Alcoholism/epidemiology , Anxiety Disorders/epidemiology , Models, Psychological , Adaptation, Psychological , Adult , Algorithms , Comorbidity , Craving , Female , Humans , Internal-External Control , Machine Learning , Male , Self Efficacy , Young AdultABSTRACT
CONTEXT: One fundamental way to honor patient autonomy is to establish and enact their wishes for end-of-life care. Limited research exists regarding adherence with code status. OBJECTIVES: This study aimed to characterize cardiopulmonary resuscitation (CPR) attempts discordant with documented code status at the time of death in the U.S. and to elucidate potential contributing factors. METHODS: The Cerner Acute Physiology and Chronic Health Evaluation (APACHE) outcomes database, which includes 237 U.S. hospitals that collect manually abstracted data from all critical care patients, was queried for adults admitted to intensive care units with a documented code status at the time of death from January 2008 to December 2016. The primary outcome was discordant CPR at death. Multivariable logistic regression models were used to identify patient-level and hospital-level associated factors after adjustment for age, hospital, and illness severity (APACHE III score). RESULTS: A total of 21,537 patients from 56 hospitals were included. Of patients with a do-not-resuscitate code status, 149 (0.8%) received CPR at death, and associated factors included black race, higher APACHE III score, or treatment in small or nonteaching hospitals. Of patients with a full code status, 203 (9.0%) did not receive CPR at death, and associated factors included higher APACHE III score, primary neurologic or trauma diagnosis, or admission in a more recent year. CONCLUSION: At the time of death, 1.6% of patients received or did not undergo CPR in a manner discordant with their documented code statuses. Race and institutional factors were associated with discordant resuscitation, and addressing these disparities may promote concordant end-of-life care in all patients.
Subject(s)
Cardiopulmonary Resuscitation , Terminal Care , APACHE , Adult , Hospitalization , Humans , Intensive Care Units , Resuscitation OrdersABSTRACT
Importance: Rib fractures are sustained by nearly 15% of patients who experience trauma and are associated with significant morbidity and mortality. Evidence-based practice (EBP) rib fracture management guidelines and treatment algorithms have been published. However, few studies have evaluated trauma center adherence to EBP or the clinical outcomes of each practice within a national cohort. Objective: To examine adherence to 6 EBPs for rib fractures across US trauma centers and the association with in-hospital mortality. Design, Setting, and Participants: A retrospective cohort study was conducted from January 1, 2007, to December 31, 2014, of 777 US trauma centers participating in the National Trauma Data Bank. A total of 625â¯617 patients (age, ≥16 years) were evaluated. Patients without rib fractures and those with no signs of life or institutions with poor data quality were excluded. Data analysis was performed from January 1, 2007, to December 31, 2014. Main Outcomes and Measures: Six EBPs were defined: (1) neuraxial blockade, (2) intensive care unit admission, (3) pneumatic stabilization, (4) chest computed tomographic scans for older adults (≥65 years) with 3 or more rib fractures, (5) surgical rib fixation for flail chest, and (6) tube thoracostomy placement for hemothorax and/or pneumothorax. Multiple imputation was used to account for missing data. Patients were propensity score matched in a 1:1 fashion based on demographic characteristics; injury severity parameters, including the Injury Severity Score (range, 0-75; higher scores indicate more severe injuries); and comorbidities. Logistic regression was used to determine the association of each practice with all-cause in-hospital mortality. Results: Of the 625â¯617 patients with rib fractures included in this analysis, 456 196 patients (73%) were white and 432 229 patients (69%) were male; the median age of the patients was 51 (interquartile range, 37-65) years, and the mean (SD) Injury Severity Score was 18.3 (11.1). The mean (SD) number of rib fractures was 4.2 (2.6). On univariate analysis, patients treated at verified level I trauma centers were more likely to receive 5 or 6 EBPs (all but pneumatic stabilization). Of those who met eligibility, only 4578 of 111 589 patients (4%) received neuraxial blockade, 46 456 of 111 589 patients (42%) were admitted to the intensive care unit, 3302 of 24 319 patients (14%) received surgical rib fixation, 1240 of 111 589 patients (1%) received pneumatic stabilization, 109 160 of 258 334 patients (42%) received tube thoracostomy, and 32 405 of 81 417 patients (40%) received chest computed tomographic scans. Three EBPs were associated with decreased mortality: neuraxial blockade (odds ratio [OR], 0.64; 95% CI, 0.51-0.79; P < .001) for patients aged 65 years or older with 3 or more rib fractures, surgical rib fixation (OR, 0.13; 95% CI, 0.01-0.18; P < .001), and intensive care unit admission (OR, 0.93; 95% CI, 0.86-1.00; P = .04) for patients aged 65 years or older with 3 or more rib fractures. Pneumatic stabilization (OR, 1.71; 95% CI, 1.25-2.35; P < .001) and chest tube placement (OR, 1.27; 95% CI, 1.21-1.33; P < .001) were associated with increased mortality in older patients with 3 or more rib fractures. On multivariable analysis, insurance status, race/ethnicity, injury severity, hospital bed size, and trauma center verification level were associated with receiving EBPs for rib fractures. Conclusions and Relevance: Significant variation appears to exist in the delivery of EBPs for rib fractures across US trauma centers. Three EBPs were associated with reduced mortality, but EBP adherence was poor. Multiple factors, including trauma center verification level, appear to be associated with patients receiving EBPs for rib fractures.
Subject(s)
Evidence-Based Practice/statistics & numerical data , Guideline Adherence/statistics & numerical data , Rib Fractures/mortality , Trauma Centers/statistics & numerical data , Adult , Aged , Databases, Factual , Evidence-Based Practice/standards , Female , Hospital Mortality , Hospitalization/statistics & numerical data , Humans , Injury Severity Score , Intensive Care Units/statistics & numerical data , Logistic Models , Male , Middle Aged , Odds Ratio , Practice Guidelines as Topic , Propensity Score , Retrospective Studies , Rib Fractures/therapy , United States/epidemiologyABSTRACT
With only 536 cases and 11 fatalities, India took the historic decision of a 21-day national lockdown on March 25. The lockdown was first extended to May 3 soon after the analysis of this paper was completed, and then to May 18 while this paper was being revised. In this paper, we use a Bayesian extension of the Susceptible-Infected-Removed (eSIR) model designed for intervention forecasting to study the short- and long-term impact of an initial 21-day lockdown on the total number of COVID-19 infections in India compared to other less severe non-pharmaceutical interventions. We compare effects of hypothetical durations of lockdown on reducing the number of active and new infections. We find that the lockdown, if implemented correctly, can reduce the total number of cases in the short term, and buy India invaluable time to prepare its healthcare and disease-monitoring system. Our analysis shows we need to have some measures of suppression in place after the lockdown for increased benefit (as measured by reduction in the number of cases). A longer lockdown between 42-56 days is preferable to substantially "flatten the curve" when compared to 21-28 days of lockdown. Our models focus solely on projecting the number of COVID-19 infections and, thus, inform policymakers about one aspect of this multi-faceted decision-making problem. We conclude with a discussion on the pivotal role of increased testing, reliable and transparent data, proper uncertainty quantification, accurate interpretation of forecasting models, reproducible data science methods and tools that can enable data-driven policymaking during a pandemic. Our software products are available at covind19.org.
ABSTRACT
OBJECTIVE: The objective of this study was to assess the potential of combining graph learning methods with latent variable estimation methods for mining clinically useful information from observational clinical data sets. MATERIALS AND METHODS: The data set contained self-reported measures of psychopathology symptoms from a clinical sample receiving treatment for alcohol use disorder. We used the traditional graph learning methods: Graphical Least Absolute Shrinkage and Selection Operator, and Friedman's hill climbing algorithm; traditional latent variable estimation method factor analysis; recently developed graph learning method Greedy Fast Causal Inference; and recently developed latent variable estimation method Find One Factor Clusters. Methods were assessed qualitatively by the content of their findings. RESULTS: Recently developed graphical methods identified potential latent variables (ie, not represented in the model) influencing particular scores. Recently developed latent effect estimation methods identified plausible cross-score loadings that were not found with factor analysis. A graphical analysis of individual items identified a mistake in wording on 1 questionnaire and provided further evidence that certain scores are not reflective of indirectly measured common causes. DISCUSSION AND CONCLUSION: Our findings suggest that a combination of Greedy Fast Causal Inference and Find One Factor Clusters can enhance the evidence-based information yield from psychopathological constructs and questionnaires. Traditional methods provided some of the same information but missed other important findings. These conclusions point the way toward more informative interrogations of existing and future data sets than are commonly employed at present.
Subject(s)
Alcoholism/psychology , Algorithms , Adult , Alcoholism/etiology , Alcoholism/therapy , Bayes Theorem , Causality , Datasets as Topic , Factor Analysis, Statistical , Humans , Latent Class Analysis , Self Report , Surveys and QuestionnairesABSTRACT
Research in the domain of psychopathology has been hindered by hidden variables-variables that are important to understanding and treating psychopathological illnesses but are unmeasured. Recent methodological advances in machine learning have culminated in the ability to discover and identify the influence of hidden variables that confound the observed relationships among measured variables. We apply a combination of traditional methods and more recent advances to a data set of alcohol use disorder patients with comorbid internalizing disorders, and find that the increasingly advanced methods produce increasingly informative and reliable results. These results include novel findings evaluated positively by our psychopathologists, as well as findings validated with knowledge from existing literature. We also find that advanced graph discovery methods can guide the use of latent variable modeling procedures, which can in turn explain the output of the graph discovery methods, resulting in a synergistic relationship between two seemingly distinct classes of methods.