ABSTRACT
Objectives: The risk factors for anastomotic leak (AL) after resection and primary anastomosis for traumatic bucket handle injury (BHI) have not been previously defined. This multicenter study was conducted to address this knowledge gap. Methods: This is a multicenter retrospective study on small intestine and colonic BHIs from blunt trauma between 2010 and 2021. Baseline patient characteristics, risk factors, presence of shock and transfusion, operative details, and clinical outcomes were compared using R. Results: Data on 395 subjects were submitted by 12 trauma centers, of whom 33 (8.1%) patients developed AL. Baseline details were similar, except for a higher proportion of patients in the AL group who had medical comorbidities such as diabetes, hypertension, and obesity (60.6% vs. 37.3%, p=0.015). AL had higher rates of surgical site infections (13.4% vs. 5.3%, p=0.004) and organ space infections (65.2% vs. 11.7%, p<0.001), along with higher readmission and reoperation rates (48.4% vs. 9.1%, p<0.001, and 39.4% vs. 11.6%, p<0.001, respectively). There was no difference in intensive care unit length of stay or mortality (p>0.05). More patients with AL were discharged with an ostomy (69.7% vs. 7.3%, p<0.001), and the mean duration until ostomy reversal was 5.85±3 months (range 2-12.4 months). The risk of AL significantly increased when the initial operation was a damage control procedure, after adjusting for age, sex, injury severity, presence of one or more comorbidities, shock, transfusion of >6 units of packed red blood cells, and site of injury (adjusted RR=2.32 (1.13, 5.17)), none of which were independent risk factors in themselves. Conclusion: Damage control surgery performed as the initial operation appears to double the risk of AL after intestinal BHI, even after controlling for other markers of injury severity. Level of evidence: III.
ABSTRACT
BACKGROUND: Primary aim was to assess the relative risk (RR) of anastomotic leak (AL) in intestinal bucket-handle (BH) compared to non-BH injury. METHODS: Multi-center study comparing AL in BH from blunt trauma 2010-2021 compared to non-BH intestinal injuries. RR was calculated for small bowel and colonic injury using R. RESULTS: AL occurred in 20/385 (5.2%) of BH vs. 4/225 (1.8%) of non-BH small intestine injury. AL was diagnosed 11.6 ± 5.6 days from index operation in small intestine BH and 9.7 ± 4.3 days in colonic BH. Adjusted RR for AL was 2.32 [0.77-6.95] for small intestinal and 4.83 [1.47-15.89] for colonic injuries. AL increased infections, ventilator days, ICU & total length of stay, reoperation, and readmission rates, although mortality was unchanged. CONCLUSION: BH carries a significantly higher risk of AL, particularly in the colon, than other blunt intestinal injuries.
Subject(s)
Abdominal Injuries , Wounds, Nonpenetrating , Humans , Anastomotic Leak/epidemiology , Anastomotic Leak/etiology , Retrospective Studies , Colon/surgery , Colon/injuries , Intestines/injuries , Wounds, Nonpenetrating/surgery , Abdominal Injuries/surgery , Anastomosis, SurgicalABSTRACT
In this work we demonstrate how to automate parts of the infectious disease-control policy-making process via performing inference in existing epidemiological models. The kind of inference tasks undertaken include computing the posterior distribution over controllable, via direct policy-making choices, simulation model parameters that give rise to acceptable disease progression outcomes. Among other things, we illustrate the use of a probabilistic programming language that automates inference in existing simulators. Neither the full capabilities of this tool for automating inference nor its utility for planning is widely disseminated at the current time. Timely gains in understanding about how such simulation-based models and inference automation tools applied in support of policy-making could lead to less economically damaging policy prescriptions, particularly during the current COVID-19 pandemic.
ABSTRACT
Making intelligent decisions from incomplete information is critical in many applications: for example, robots must choose actions based on imperfect sensors, and speech-based interfaces must infer a user's needs from noisy microphone inputs. What makes these tasks hard is that often we do not have a natural representation with which to model the domain and use for choosing actions; we must learn about the domain's properties while simultaneously performing the task. Learning a representation also involves trade-offs between modeling the data that we have seen previously and being able to make predictions about new data. This article explores learning representations of stochastic systems using Bayesian nonparametric statistics. Bayesian nonparametric methods allow the sophistication of a representation to scale gracefully with the complexity in the data. Our main contribution is a careful empirical evaluation of how representations learned using Bayesian nonparametric methods compare to other standard learning approaches, especially in support of planning and control. We show that the Bayesian aspects of the methods result in achieving state-of-the-art performance in decision making with relatively few samples, while the nonparametric aspects often result in fewer computations. These results hold across a variety of different techniques for choosing actions given a representation.
ABSTRACT
BACKGROUND AND OBJECTIVE: The volume of healthcare data is growing rapidly with the adoption of health information technology. We focus on automated ICD9 code assignment from discharge summary content and methods for evaluating such assignments. METHODS: We study ICD9 diagnosis codes and discharge summaries from the publicly available Multiparameter Intelligent Monitoring in Intensive Care II (MIMIC II) repository. We experiment with two coding approaches: one that treats each ICD9 code independently of each other (flat classifier), and one that leverages the hierarchical nature of ICD9 codes into its modeling (hierarchy-based classifier). We propose novel evaluation metrics, which reflect the distances among gold-standard and predicted codes and their locations in the ICD9 tree. Experimental setup, code for modeling, and evaluation scripts are made available to the research community. RESULTS: The hierarchy-based classifier outperforms the flat classifier with F-measures of 39.5% and 27.6%, respectively, when trained on 20,533 documents and tested on 2282 documents. While recall is improved at the expense of precision, our novel evaluation metrics show a more refined assessment: for instance, the hierarchy-based classifier identifies the correct sub-tree of gold-standard codes more often than the flat classifier. Error analysis reveals that gold-standard codes are not perfect, and as such the recall and precision are likely underestimated. CONCLUSIONS: Hierarchy-based classification yields better ICD9 coding than flat classification for MIMIC patients. Automated ICD9 coding is an example of a task for which data and tools can be shared and for which the research community can work together to build on shared models and advance the state of the art.
Subject(s)
Clinical Coding/methods , International Classification of Diseases , Support Vector Machine , Brain Ischemia/classification , Evaluation Studies as Topic , Humans , Patient DischargeABSTRACT
Improving activity recognition, with special focus on fall-detection, is the subject of this study. We show that Kalman smoothed in-painting of missing pose information and task-specific dimensionality reduction of activity feature vectors leads to significantly improved activity classification performance. We illustrate our findings by applying common classification algorithms to dimensionally reduced feature vectors, and compare our accuracy to previous work. In part two we investigate our methods on a small subset of the data, in order to ascertain what accuracy performance is achievable with the smallest amount of information available.
Subject(s)
Algorithms , Discriminant Analysis , Pattern Recognition, Automated , Humans , Time Factors , WalkingABSTRACT
INTRODUCTION: To our knowledge, the prevalence of Systemic Inflammatory Response Syndrome (SIRS) in pediatric patients with appendicitis has not been previously investigated. Our specific aim was to determine the prevalence of SIRS at the time of presentation of pediatric patients with appendicitis. Additionally, we sought to determine if the presence of SIRS had any value in predicting their clinical outcomes. METHODS: This retrospective cohort study included pediatric patients (age <17 years) presenting to a single hospital and being diagnosed with appendicitis between July 1, 2011, and June 30, 2012. The primary exposure variable of interest was SIRS, dichotomously defined as positive or negative. The primary outcome of interest was the presence/development of an intraabdominal abscess. The secondary outcome of interest was length of hospital stay (LOS). Chi-squared and t-tests were used to evaluate the association between presence of SIRS and development of abscess and LOS. RESULTS: This study consisted of 212 patients. The definition of SIRS was met in 66 patients (31.1%). Thirty of the 66 (45.6%) patients with SIRS had/developed an abscess versus 28 (19.2%) of those without SIRS (P<0.001). Patients with SIRS had a mean LOS of 4 days (+/-2.7), while those without SIRS stayed a mean of 2.5 days (+/-2.3) [p<0.0001]). Adjusting for age did not alter these associations. CONCLUSION: Our study found a 31.1% prevalence of SIRS in pediatric patients presenting with appendicitis. Our results suggest these patients with SIRS have a significantly higher risk of having/developing an intraabdominal abscess (RR, 2.4; 95% CI: 1.6-3.6) and significantly longer LOS.
Subject(s)
Abdominal Abscess/etiology , Appendectomy , Appendicitis/complications , Length of Stay/statistics & numerical data , Postoperative Complications/etiology , Systemic Inflammatory Response Syndrome/etiology , Abdominal Abscess/epidemiology , Adolescent , Appendicitis/diagnosis , Appendicitis/surgery , Child , Child, Preschool , Female , Humans , Logistic Models , Male , Postoperative Complications/epidemiology , Prevalence , Retrospective Studies , Risk Assessment , Risk Factors , Systemic Inflammatory Response Syndrome/epidemiology , Treatment OutcomeABSTRACT
We address the problem of analyzing sets of noisy time-varying signals that all report on the same process but confound straightforward analyses due to complex inter-signal heterogeneities and measurement artifacts. In particular we consider single-molecule experiments which indirectly measure the distinct steps in a biomolecular process via observations of noisy time-dependent signals such as a fluorescence intensity or bead position. Straightforward hidden Markov model (HMM) analyses attempt to characterize such processes in terms of a set of conformational states, the transitions that can occur between these states, and the associated rates at which those transitions occur; but require ad-hoc post-processing steps to combine multiple signals. Here we develop a hierarchically coupled HMM that allows experimentalists to deal with inter-signal variability in a principled and automatic way. Our approach is a generalized expectation maximization hyperparameter point estimation procedure with variational Bayes at the level of individual time series that learns an single interpretable representation of the overall data generating process.
ABSTRACT
The analysis of extra-cellular neural recordings typically begins with careful spike sorting and all analysis of the data then rests on the correctness of the resulting spike trains. In many situations this is unproblematic as experimental and spike sorting procedures often focus on well isolated units. There is evidence in the literature, however, that errors in spike sorting can occur even with carefully collected and selected data. Additionally, chronically implanted electrodes and arrays with fixed electrodes cannot be easily adjusted to provide well isolated units. In these situations, multiple units may be recorded and the assignment of waveforms to units may be ambiguous. At the same time, analysis of such data may be both scientifically important and clinically relevant. In this paper we address this issue using a novel probabilistic model that accounts for several important sources of uncertainty and error in spike sorting. In lieu of sorting neural data to produce a single best spike train, we estimate a probabilistic model of spike trains given the observed data. We show how such a distribution over spike sortings can support standard neuroscientific questions while providing a representation of uncertainty in the analysis. As a representative illustration of the approach, we analyzed primary motor cortical tuning with respect to hand movement in data recorded with a chronic multi-electrode array in non-human primates. We found that the probabilistic analysis generally agrees with human sorters but suggests the presence of tuned units not detected by humans.
Subject(s)
Action Potentials/physiology , Bayes Theorem , Models, Neurological , Neurons/physiology , Animals , Brain Mapping , Haplorhini , Humans , Motor Cortex/cytology , Neural Networks, Computer , Photic Stimulation/methods , Psychomotor Performance/physiology , Saccades/physiology , Signal Processing, Computer-AssistedABSTRACT
The genetic basis of reading disability (RD) has long been established through family and twin studies. More recently genetic linkage studies have identified genomic regions that appear to harbor susceptibility genes for RD. Association studies have been shown to have greater power for detecting genes of modest effect, particularly in genetically isolated populations. Hence, a case control study of RD was undertaken in the Afrikaner population in South Africa. Sixty-eight microsatellite markers in regions where linkages had been reported in previous studies were genotyped on 122 children with reading disability and 112 typically reading controls drawn from the same school population. A single allele of marker D6S299 showed a highly significant association with the RD phenotype (D6S299[229], P-value 0.000014). Other markers on other chromosomes also showed suggestive associations. Of particular interest were markers on chromosomes 1 and 15. These two regions have been implicated in studies of populations that formed the founding population in the Afrikaner population.
Subject(s)
Black People/genetics , Chromosomes, Human, Pair 6 , Dyslexia/genetics , Alleles , Case-Control Studies , Child , Chromosomes, Human, Pair 1 , Chromosomes, Human, Pair 15 , Dyslexia/epidemiology , Female , Genetic Predisposition to Disease , Genetics, Population , Genotype , Humans , Male , Microsatellite Repeats , South Africa/epidemiologyABSTRACT
OBJECTIVE: To examine psychiatric morbidity and functional impairment of adolescents with and without poor reading skills during mid- to late adolescence. METHOD: The sample consisted of 188 adolescents, 94 with poor reading skills and 94 with typical reading skills, screened from a larger sample in the public schools at age 15. To assess psychiatric disorders, participants were assessed annually with the Schedule for Affective Disorders and Schizophrenia for School-Age Children-Epidemiologic Version (up to 4.5 years; maximum age, 20 years). Functional impairment was assessed with the Child and Adolescent Functional Assessment Scale. RESULTS: Adolescents with poor reading skills evidenced higher rates of current attention-deficit/hyperactivity, affective, and anxiety disorders, particularly social phobia and generalized anxiety disorder. Anxiety disorders but not affective disorders were related to reading status after controlling for attention-deficit/hyperactivity disorder. Adolescents with poor reading evidenced more functional impairment across multiple areas than youths with typical reading skills, even after considering the presence of comorbid attention-deficit/hyperactivity disorder. CONCLUSIONS: The increased psychiatric morbidity and functional impairment of adolescents with reading problems highlight the importance of developing interventions that help these youths address reading deficits and associated vulnerabilities during the last years of secondary school.
Subject(s)
Anxiety Disorders/epidemiology , Attention Deficit Disorder with Hyperactivity/epidemiology , Dyslexia/epidemiology , Mood Disorders/epidemiology , Phobic Disorders/epidemiology , Schizophrenia/epidemiology , Adolescent , Adult , Anxiety Disorders/diagnosis , Attention Deficit Disorder with Hyperactivity/diagnosis , Comorbidity , Diagnostic and Statistical Manual of Mental Disorders , Female , Follow-Up Studies , Humans , Male , Mood Disorders/diagnosis , Phobic Disorders/diagnosis , Prevalence , Schizophrenia/diagnosis , Severity of Illness IndexABSTRACT
In recent years, multiple brain MR imaging modalities have emerged; however, analysis methodologies have mainly remained modality-specific. In addition, when comparing across imaging modalities, most researchers have been forced to rely on simple region-of-interest type analyses, which do not allow the voxel-by-voxel comparisons necessary to answer more sophisticated neuroscience questions. To overcome these limitations, we developed a toolbox for multimodal image analysis called biological parametric mapping (BPM), based on a voxel-wise use of the general linear model. The BPM toolbox incorporates information obtained from other modalities as regressors in a voxel-wise analysis, thereby permitting investigation of more sophisticated hypotheses. The BPM toolbox has been developed in Matlab with a user-friendly interface for performing analyses, including voxel-wise multimodal correlation, ANCOVA, and multiple regression. It has a high degree of integration with the SPM (statistical parametric mapping) software relying on it for visualization and statistical inference. Furthermore, statistical inference for a correlation field, rather than a widely used T-field, has been implemented in the correlation analysis for more accurate results. An example with in vivo data is presented, demonstrating the potential of the BPM methodology as a tool for multimodal image analysis.
Subject(s)
Brain Mapping/methods , Brain/anatomy & histology , Brain/physiology , Software , Humans , Statistics as TopicABSTRACT
The purpose of this study was to examine the risk of suicidal ideation and suicide attempts and school dropout among youth with poor reading in comparison to youth with typical reading (n = 188) recruited from public schools at the age of 15. In a prospective naturalistic study, youth and parents participated in repeated research assessments to obtain information about suicide ideation and attempts, psychiatric and sociodemographic variables, and school dropout. Youth with poor reading ability were more likely to experience suicidal ideation or attempts and more likely to drop out of school than youth with typical reading, even after controlling for sociodemographic and psychiatric variables. Suicidality and school dropout were strongly associated with each other. Prevention efforts should focus on better understanding the relationship between these outcomes, as well as on the developmental paths leading up to these behaviors among youth with reading difficulties.
Subject(s)
Dyslexia/epidemiology , Schools , Student Dropouts/statistics & numerical data , Suicide, Attempted/statistics & numerical data , Adolescent , Child , Child, Preschool , Dyslexia/diagnosis , Female , Humans , MaleABSTRACT
In this work we present and apply infinite Gaussian mixture modeling, a non-parametric Bayesian method, to the problem of spike sorting. As this approach is Bayesian, it allows us to integrate prior knowledge about the problem in a principled way. Because it is non-parametric we are able to avoid model selection, a difficult problem that most current spike sorting methods do not address. We compare this approach to using penalized log likelihood to select the best from multiple finite mixture models trained by expectation maximization. We show favorable offline sorting results on real data and discuss ways to extend our model to online applications.
Subject(s)
Action Potentials/physiology , Algorithms , Artificial Intelligence , Brain Mapping/methods , Nerve Net/physiology , Pattern Recognition, Automated/methods , Animals , Bayes Theorem , HaplorhiniABSTRACT
Recent studies have demonstrated that dyslexia is associated with deficits in the temporal encoding of sensory information. While most previous studies have focused on information processing within a single sensory modality, it is clear that the deficits seen in dyslexia span multiple sensory systems. Surprisingly, although the development of linguistic proficiency involves the rapid and accurate integration of auditory and visual cues, the capacity of dyslexic individuals to integrate information between the different senses has not been systematically examined. To test this, we studied the effects of task-irrelevant auditory information on the performance of a visual temporal-order-judgment (TOJ) task. Dyslexic subjects' performance differed significantly from that of control subjects, specifically in that they integrated the auditory and visual information over longer temporal intervals. Such a result suggests an extended temporal "window" for binding visual and auditory cues in dyslexic individuals. The potential deleterious effects of this finding for rapid multisensory processes such as reading are discussed.
Subject(s)
Auditory Perception/physiology , Dyslexia/psychology , Visual Perception/physiology , Acoustic Stimulation , Adult , Cues , Female , Humans , Male , Middle Aged , Photic Stimulation , Psychomotor Performance/physiology , Sensory Thresholds/physiologyABSTRACT
The purpose of this study was to examine the severity of behavioral and emotional problems among adolescents with poor and typical single word reading ability (N = 188) recruited from public schools and followed for a median of 2.4 years. Youth and parents were repeatedly assessed to obtain information regarding the severity and course of symptoms (depression, anxiety, somatic complaints. aggression, delinquent behaviors, inattention), controlling for demographic variables and diagnosis or ADHD. After adjustment for demographic variables and ADHD, poor readers reported higher levels of depression, trait anxiety, and somatic complaints than typical readers, but there were no difference, in reported self-reported delinquent or aggressive behaviors. Parent reports indicated no difference, in depression, anxiety or aggression between the two groups but indicated more inattention, somatic complaints, and delinquent behaviors for the poor readers. School and health professionals should carefully assess youth with poor reading for behavioral and emotional symptoms and provide services when indicated.
Subject(s)
Attention Deficit Disorder with Hyperactivity/epidemiology , Child Behavior Disorders/epidemiology , Dyslexia/epidemiology , Mood Disorders/epidemiology , Adolescent , Anxiety Disorders/diagnosis , Anxiety Disorders/epidemiology , Attention Deficit Disorder with Hyperactivity/diagnosis , Child , Child Behavior Disorders/diagnosis , Depression/diagnosis , Depression/epidemiology , Dyslexia/diagnosis , Female , Follow-Up Studies , Humans , Male , Mood Disorders/diagnosis , Psychological Tests , Severity of Illness Index , Socioeconomic FactorsABSTRACT
Study 1 retrospectively analyzed neuropsychological and psychoeducational tests given to N=220 first graders, with follow-up assessments in third and eighth grade. Four predictor constructs were derived: (1) Phonemic Awareness, (2) Picture Vocabulary, (3) Rapid Naming, and (4) Single Word Reading. Together, these accounted for 88%, 76%, 69%, and 69% of the variance, respectively, in first, third, and eighth grade Woodcock Johnson Broad Reading and eighth grade Gates-MacGinitie. When Single Word Reading was excluded from the predictors, the remaining predictors still accounted for 71%, 65%, 61%, and 65% of variance in the respective outcomes. Secondary analyses of risk of low outcome showed sensitivities/specificities of 93.0/91.0, and 86.4/84.9, respectively, for predicting which students would be in the bottom 15% and 30% of actual first grade WJBR. Sensitivities/specificities were 84.8/83.3 and 80.2/81.3, respectively, for predicting the bottom 15% and 30% of actual third grade WJBR outcomes; eighth grade outcomes had sensitivities/specificities of 80.0/80.0 and 85.7/83.1, respectively, for the bottom 15% and 30% of actual eighth grade WJBR scores. Study 2 cross-validated the concurrent predictive validities in an N=500 geographically diverse sample of late kindergartners through third graders, whose ethnic and racial composition closely approximated the national early elementary school population. New tests of the same four predictor domains were used, together taking only 15 minutes to administer by teachers; the new Woodcock-Johnson III Broad Reading standard score was the concurrent criterion, whose testers were blind to the predictor results. This cross-validation showed 86% of the variance accounted for, using the same regression weights as used in Study 1. With these weights, sensitivity/specificity values for the 15% and 30% thresholds were, respectively, 91.3/88.0 and 94.1/89.1. These validities and accuracies are stronger than others reported for similar intervals in the literature.
Subject(s)
Dyslexia/diagnosis , Mass Screening , Reading , Adolescent , Child , Child, Preschool , Cohort Studies , Dyslexia/prevention & control , Dyslexia/psychology , Early Intervention, Educational , Educational Status , Female , Follow-Up Studies , Humans , Male , Needs Assessment , Phonetics , Predictive Value of Tests , Psychosocial Deprivation , Semantics , VocabularyABSTRACT
Recent methods for motor cortical decoding have demonstrated relatively accurate reconstructions of hand trajectory from small populations of neurons in primary motor cortex. Decoding results are often reported only for periods when the subject is attending to the task. In a neural prosthetic interface, however, the subject must be able to switch between controlling a device or performing other mental functions. In this work we demonstrate a method for detecting whether or not a subject is attending to a motor control task. Using the firing activity of the same neural population used for decoding hand kinematics we demonstrate that a Fisher linear discriminant performs well in classifying the attentional state of a monkey. We use the output of this classifier to augment a hidden state in a first order Markov model and use particle filtering to recursively infer hand kinematics and attentional state conditioned on neural firing rates. We demonstrate high accuracy on test data where a monkey switches between attending to a task and not. By decoding a discrete "state" in addition to hand kinematics our proposed classification and estimation scheme may enable real-world neuroprosthetic functions such as "hold", "click", and "turn off/on".