Physical Fitness and Risk of Mental Disorders in Children and Adolescents

Importance

With the rising prevalence of mental disorders among children and adolescents, identifying modifiable associations is critical.

Objective

To examine the association between physical fitness and mental disorder risks.

Design, Setting, and Participants

This nationwide cohort study used data from the Taiwan National Student Fitness Tests and National Health Insurance Research Databases from January 1, 2009 to December 31, 2019. Participants were divided into 2 cohorts targeting anxiety and depression (1 996 633 participants) and attention-deficit/hyperactivity disorder (ADHD; 1 920 596 participants). Participants were aged 10 to 11 years at study entry and followed up for at least 3 years, had a nearly equal gender distribution, and an average follow-up of 6 years. Data were analyzed from October 2022 to February 2024.

Exposures

Assessments of physical fitness included cardiorespiratory fitness (CF), muscular endurance (ME), muscular power (MP), and flexibility, measured through an 800-m run time, bent-leg curl-ups, standing broad jump, and sit-and-reach test, respectively.

Main Outcomes and Measures

Kaplan-Meier method calculated the cumulative incidence of anxiety, depression, and ADHD across fitness quartiles. Additionally, multivariable Cox proportional hazards models were used that included all 4 fitness components and explored sex and income as modifiers.

Results

The anxiety and depression cohort had 1 996 633 participants (1 035 411 participants were male [51.9%], and the median [IQR] age was 10.6 [10.3-11.0] years), while the ADHD cohort had 1 920 596 (975 568 participants were male [51.9%], and the median [IQR] age was 10.6 [10.3-11.0] years). Cumulative incidence of mental disorders was lower among participants in better-performing fitness quartiles, suggesting a dose-dependent association. Gender-specific analyses, controlling for confounders, revealed that improved CF, indicated by a 30-second decrease in run times, was associated with reduced risks of anxiety, depression, and ADHD in female participants, and lower risks of anxiety and ADHD in male participants (adjusted hazard ratio [aHR] for ADHD risk for female participants, 0.92; 95% CI, 0.90-0.94; P < .001; for male participants, 0.93; 95% CI, 0.92-0.94; P < .001). Enhanced ME, marked by an increase of 5 curl-ups per minute, was associated with decreased risks of depression and ADHD in female participants, and lower anxiety and ADHD risks in male participants (aHR for ADHD risk for female participants, 0.94; 95% CI, 0.92-0.97; P < .001; for male participants, 0.96; 95% CI, 0.95-0.97; P < .001). Improved MP, reflected by a 20-cm increase in jump distance, was associated with reduced risks of anxiety and ADHD in female participants and reduced anxiety, depression, and ADHD in male participants (aHR for ADHD risk for female participants, 0.95; 95% CI, 0.91-1.00; P = .04; for male participants, 0.96; 95% CI, 0.94-0.99; P = .001).

Conclusions and Relevance

This study highlights the potential protective role of cardiorespiratory fitness, muscular endurance, and muscular power in preventing the onset of mental disorders. It warrants further investigation of the effectiveness of physical fitness programs as a preventive measure for mental disorders among children and adolescents.

Development and multi-site external validation of a generalizable risk prediction model for bipolar disorder

Bipolar disorder is a leading contributor to disability, premature mortality, and suicide. Early identification of risk for bipolar disorder using generalizable predictive models trained on diverse cohorts around the United States could improve targeted assessment of high risk individuals, reduce misdiagnosis, and improve the allocation of limited mental health resources. This observational case-control study intended to develop and validate generalizable predictive models of bipolar disorder as part of the multisite, multinational PsycheMERGE Network across diverse and large biobanks with linked electronic health records (EHRs) from three academic medical centers: in the Northeast (Massachusetts General Brigham), the Mid-Atlantic (Geisinger) and the Mid-South (Vanderbilt University Medical Center). Predictive models were developed and valid with multiple algorithms at each study site: random forests, gradient boosting machines, penalized regression, including stacked ensemble learning algorithms combining them. Predictors were limited to widely available EHR-based features agnostic to a common data model including demographics, diagnostic codes, and medications. The main study outcome was bipolar disorder diagnosis as defined by the International Cohort Collection for Bipolar Disorder, 2015. In total, the study included records for 3,529,569 patients including 12,533 cases (0.3%) of bipolar disorder. After internal and external validation, algorithms demonstrated optimal performance in their respective development sites. The stacked ensemble achieved the best combination of overall discrimination (AUC = 0.82-0.87) and calibration performance with positive predictive values above 5% in the highest risk quantiles at all three study sites. In conclusion, generalizable predictive models of risk for bipolar disorder can be feasibly developed across diverse sites to enable precision medicine. Comparison of a range of machine learning methods indicated that an ensemble approach provides the best performance overall but required local retraining. These models will be disseminated via the PsycheMERGE Network website.

Impact of selection bias on polygenic risk score estimates in healthcare settings

BACKGROUND

Hospital-based biobanks are being increasingly considered as a resource for translating polygenic risk scores (PRS) into clinical practice. However, since these biobanks originate from patient populations, there is a possibility of bias in polygenic risk estimation due to overrepresentation of patients with higher frequency of healthcare interactions.

METHODS

PRS for schizophrenia, bipolar disorder, and depression were calculated using summary statistics from the largest available genomic studies for a sample of 24 153 European ancestry participants in the Mass General Brigham (MGB) Biobank. To correct for selection bias, we fitted logistic regression models with inverse probability (IP) weights, which were estimated using 1839 sociodemographic, clinical, and healthcare utilization features extracted from electronic health records of 1 546 440 non-Hispanic White patients eligible to participate in the Biobank study at their first visit to the MGB-affiliated hospitals.

RESULTS

Case prevalence of bipolar disorder among participants in the top decile of bipolar disorder PRS was 10.0% (95% CI 8.8-11.2%) in the unweighted analysis but only 6.2% (5.0-7.5%) when selection bias was accounted for using IP weights. Similarly, case prevalence of depression among those in the top decile of depression PRS was reduced from 33.5% (31.7-35.4%) to 28.9% (25.8-31.9%) after IP weighting.

CONCLUSIONS

Non-random selection of participants into volunteer biobanks may induce clinically relevant selection bias that could impact implementation of PRS in research and clinical settings. As efforts to integrate PRS in medical practice expand, recognition and mitigation of these biases should be considered and may need to be optimized in a context-specific manner.

An efficient landmark model for prediction of suicide attempts in multiple clinical settings

Growing evidence has shown that applying machine learning models to large clinical data sources may exceed clinician performance in suicide risk stratification. However, many existing prediction models either suffer from "temporal bias" (a bias that stems from using case-control sampling) or require training on all available patient visit data. Here, we adopt a "landmark model" framework that aligns with clinical practice for prediction of suicide-related behaviors (SRBs) using a large electronic health record database. Using the landmark approach, we developed models for SRB prediction (regularized Cox regression and random survival forest) that establish a time-point (e.g., clinical visit) from which predictions are made over user-specified prediction windows using historical information up to that point. We applied this approach to cohorts from three clinical settings: general outpatient, psychiatric emergency department, and psychiatric inpatients, for varying prediction windows and lengths of historical data. Models achieved high discriminative performance (area under the Receiver Operating Characteristic curve 0.74-0.93 for the Cox model) across different prediction windows and settings, even with relatively short periods of historical data. In short, we developed accurate, dynamic SRB risk prediction models with the landmark approach that reduce bias and enhance the reliability and portability of suicide risk prediction models.

AI-assisted prediction of differential response to antidepressant classes using electronic health records

Antidepressant selection is largely a trial-and-error process. We used electronic health record (EHR) data and artificial intelligence (AI) to predict response to four antidepressants classes (SSRI, SNRI, bupropion, and mirtazapine) 4 to 12 weeks after antidepressant initiation. The final data set comprised 17,556 patients. Predictors were derived from both structured and unstructured EHR data and models accounted for features predictive of treatment selection to minimize confounding by indication. Outcome labels were derived through expert chart review and AI-automated imputation. Regularized generalized linear model (GLM), random forest, gradient boosting machine (GBM), and deep neural network (DNN) models were trained and their performance compared. Predictor importance scores were derived using SHapley Additive exPlanations (SHAP). All models demonstrated similarly good prediction performance (AUROCs ≥ 0.70, AUPRCs ≥ 0.68). The models can estimate differential treatment response probabilities both between patients and between antidepressant classes for the same patient. In addition, patient-specific factors driving response probabilities for each antidepressant class can be generated. We show that antidepressant response can be accurately predicted from real-world EHR data with AI modeling, and our approach could inform further development of clinical decision support systems for more effective treatment selection.

Initial antidepressant choice by non-psychiatrists: Learning from large-scale electronic health records

OBJECTIVES

Pharmacological treatment of depression mostly occurs in non-psychiatric settings, but the determinants of initial choice of antidepressant treatment in these settings are unclear. We investigate how non-psychiatrists choose among four antidepressant classes at first prescription (selective serotonin reuptake inhibitors [SSRI], bupropion, mirtazapine, or serotonin-norepinephrine reuptake inhibitors [SNRI]).

METHOD

Using electronic health records (EHRs), we included adult patients at the time of first antidepressant prescription with a co-occurring diagnosis code for a depressive disorder. We selected 64 variables based on a literature search and expert consultation, constructed the variables from either structured codes or through applying natural language processing (NLP), and modeled antidepressant choice using multinomial logistic regression, using SSRI as the reference class.

RESULTS

With 47,528 patients, we observed significant associations for 36 of 64 variables. Many of these associations suggested antidepressants' known pharmacological properties/actions guided choice. For example, there was a decreased likelihood of bupropion prescription among patients with epilepsy (adjusted OR 0.49, 95%CI: 0.41-0.57, p < 0.001), and an increased likelihood of mirtazapine prescription among patients with insomnia (adjusted OR 1.59, 95%CI: 1.40-1.80, p < 0.001).

CONCLUSIONS

Broadly speaking, non-psychiatrists' selection of antidepressant class appears to be at least in part guided by clinically relevant pharmacological considerations.

Development and Multi-Site External Validation of a Generalizable Risk Prediction Model for Bipolar Disorder

Bipolar disorder is a leading contributor to disability, premature mortality, and suicide. Early identification of risk for bipolar disorder using generalizable predictive models trained on diverse cohorts around the United States could improve targeted assessment of high risk individuals, reduce misdiagnosis, and improve the allocation of limited mental health resources. This observational case-control study intended to develop and validate generalizable predictive models of bipolar disorder as part of the multisite, multinational PsycheMERGE Consortium across diverse and large biobanks with linked electronic health records (EHRs) from three academic medical centers: in the Northeast (Massachusetts General Brigham), the Mid-Atlantic (Geisinger) and the Mid-South (Vanderbilt University Medical Center). Predictive models were developed and validated with multiple algorithms at each study site: random forests, gradient boosting machines, penalized regression, including stacked ensemble learning algorithms combining them. Predictors were limited to widely available EHR-based features agnostic to a common data model including demographics, diagnostic codes, and medications. The main study outcome was bipolar disorder diagnosis as defined by the International Cohort Collection for Bipolar Disorder, 2015. In total, the study included records for 3,529,569 patients including 12,533 cases (0.3%) of bipolar disorder. After internal and external validation, algorithms demonstrated optimal performance in their respective development sites. The stacked ensemble achieved the best combination of overall discrimination (AUC = 0.82 - 0.87) and calibration performance with positive predictive values above 5% in the highest risk quantiles at all three study sites. In conclusion, generalizable predictive models of risk for bipolar disorder can be feasibly developed across diverse sites to enable precision medicine. Comparison of a range of machine learning methods indicated that an ensemble approach provides the best performance overall but required local retraining. These models will be disseminated via the PsycheMERGE Consortium website.

Causal inference in medical records and complementary systems pharmacology for metformin drug repurposing towards dementia

Metformin, a diabetes drug with anti-aging cellular responses, has complex actions that may alter dementia onset. Mixed results are emerging from prior observational studies. To address this complexity, we deploy a causal inference approach accounting for the competing risk of death in emulated clinical trials using two distinct electronic health record systems. In intention-to-treat analyses, metformin use associates with lower hazard of all-cause mortality and lower cause-specific hazard of dementia onset, after accounting for prolonged survival, relative to sulfonylureas. In parallel systems pharmacology studies, the expression of two AD-related proteins, APOE and SPP1, was suppressed by pharmacologic concentrations of metformin in differentiated human neural cells, relative to a sulfonylurea. Together, our findings suggest that metformin might reduce the risk of dementia in diabetes patients through mechanisms beyond glycemic control, and that SPP1 is a candidate biomarker for metformin's action in the brain.

Active deep learning to detect cognitive concerns in electronic health records

BACKGROUND

Timely diagnosis of dementia is important to patients and their caregivers for advanced planning, yet dementia is under-diagnosed by healthcare professionals and under-coded in claims data. Sensitive and specific tools to detect cognitive concerns in diverse clinical settings could prompt referral for cognitive evaluation and specialist care.

METHOD

We developed a deep learning natural language processing (NLP) method to detect cognitive concerns in unstructured clinician notes from electronic health records (EHR). We leveraged a gold-standard set of ∼1000 patients sampled randomly from three strata: patients with diagnosis codes, patients with specialist visits but no code, or patients with neither. The physician performed a detailed chart review and adjudication of cognitive status, noting "cognitive concern" (i.e., any evidence of cognitive difficulties) and rating patients on a 5-point scale: normal, normal vs. MCI, MCI, MCI vs. dementia, and dementia. We used 10% of the labeled data as a test set and the remaining 90% for training and validation of the model to classify patients with any cognitive concerns (normal vs. other). We also built a web-based chart review annotation tool that facilitates labeling and enables an active learning loop to scale up labeling to thousands of charts.

RESULT

In a random sample from the gold-standard dataset, 30 out of 80 patients with cognitive concerns had no diagnosis code or medication related to dementia; we hypothesized that our deep learning tool could leverage clinical text to improve detection of cognitive concerns. Indeed, a model with codes and medications had an area under the receiver operating characteristic (AUROC) curve of 0.79, sensitivity of 0.59, and specificity of 1.00 for the binary classification task. The deep learning model improved the AUROC to 0.90, increased sensitivity to 0.79, and maintained specificity of 0.98. Notes from primary care, specialties such as neurology and psychiatry, and social workers had the highest likelihood of containing information.

CONCLUSION

The deep learning model was successful in detecting cases without a dementia-related diagnosis code or medication. Automatic processing of electronic medical records with a deep learning tool can be used for early detection of cognitive concern to optimize patient care and predict hospital readmission.

Firearm access and adolescent suicide risk: toward a clearer understanding of effect size

BACKGROUND

Strong and consistent associations between access to firearms and suicide have been found in ecologic and individual-level observational studies. For adolescents, a seminal case-control study estimated that living in a home with (vs without) a firearm was associated with a fourfold increase in the risk of death by suicide.

METHODS

We use data from a nationally representative study of 10 123 US adolescents aged 13-18 years to (1) measure how much adolescents who live in a home with a firearm differ from those who do not in ways related to their risk of suicide, and (2) incorporate these differences into an updated effect estimate of the risk of adolescent suicide attributable to living in a home with firearms.

RESULTS

Almost one-third (30.7%) of adolescents reported living in a home with firearms. Relative to those who did not, adolescents reporting living in a home with a firearm were slightly more likely to be male, older and reside in the South and rural areas, but few differences were identified for mental health characteristics. The effect size found by Brent and colleagues appeared robust to sources of possible residual confounding: updated relative risks remained above 4.0 across most sensitivity analyses and at least 3.1 in even the most conservative estimates.

CONCLUSIONS

Although unmeasured confounding and other biases may nonetheless remain, our updated estimates reinforce the suggestion that adolescents' risk of suicide was increased threefold to fourfold if they had lived in homes with a firearm compared with if they had not.

Illuminating the Black Box: Interpreting Deep Neural Network Models for Psychiatric Research

Psychiatric research is often confronted with complex abstractions and dynamics that are not readily accessible or well-defined to our perception and measurements, making data-driven methods an appealing approach. Deep neural networks (DNNs) are capable of automatically learning abstractions in the data that can be entirely novel and have demonstrated superior performance over classical machine learning models across a range of tasks and, therefore, serve as a promising tool for making new discoveries in psychiatry. A key concern for the wider application of DNNs is their reputation as a "black box" approach-i.e., they are said to lack transparency or interpretability of how input data are transformed to model outputs. In fact, several existing and emerging tools are providing improvements in interpretability. However, most reviews of interpretability for DNNs focus on theoretical and/or engineering perspectives. This article reviews approaches to DNN interpretability issues that may be relevant to their application in psychiatric research and practice. It describes a framework for understanding these methods, reviews the conceptual basis of specific methods and their potential limitations, and discusses prospects for their implementation and future directions.

Racial differences in suicide deaths after cancer diagnosis: A SEER-based analysis of 2,336,949 patients

244

Background: Risk of suicide is increased after cancer diagnosis. Our study aims to investigate the racial differences on risk of suicide after cancer diagnosis in a nationwide cohort of U.S. patients. Methods: Patients ≥ 18 years and diagnosed with breast (n = 616,099; 26.4%), lung (n = 585,978; 25.0%), colorectal (n = 429,060; 18.4%), or prostate cancer (n = 705,812; 30.2%) from 1988-2010 in Surveillance, Epidemiology and End Results Program (SEER) were identified. A Cox-proportional hazard model was used to compare the suicide mortality of the races (Hispanic white, African American, Asian/Pacific Islander, American Indian/Alaska Native) to non-Hispanic white patients adjusting for age, sex, marital status, household income, education level and cancer sites. Results: A total of 2,336,949 patients were identified, and there were 3,406 suicide death events. Fifty percent of suicide deaths were within 32 months after cancer diagnosis (interquartile range: 66 months). After a median follow-up of 49 months, the suicide mortality was lower in African American (HR: 0.29, 95% CI: 0.25-0.35, p value <.0001), Hispanic white (HR: 0.49, 95% CI: 0.41-0.58, p value <.0001), Asian/Pacific Islander (HR: 0.67, 95% CI: 0.57-0.79, p value<.0001). The suicide mortality was the same comparing American Indian/Alaska Native to non-Hispanic white patients (HR: 1.01, 95% CI: 0.62-1.65, p value: 0.97). Conclusions: After adjusting for sociodemographic factors, non-Hispanic white cancer patients still had higher risk of suicide death after cancer diagnosis. Future studies should address the racial differences of psychological stress after cancer diagnosis. [Table: see text]

Dosage and duration of antipsychotic treatment in demented outpatients with agitation or psychosis

BACKGROUND/PURPOSE

The USA Food and Drug Administration (FDA) issued warnings regarding the use of antipsychotics in patients with dementia in 2003 and 2005. We aimed to study the dose and duration of antipsychotic treatment in dementia, and to examine whether physicians' prescription behaviors changed after the FDA warnings.

METHODS

Medical charts of outpatients who had Alzheimer's disease, vascular dementia, or mixed dementia were reviewed. Patients must have achieved a clinically stable state for at least 4 weeks after receiving antipsychotic treatment for agitation or psychosis. Demographics, clinical correlates, and duration of antipsychotic treatment were compared among different antipsychotic groups. Because the quetiapine group had the largest sample size, the optimal dose and duration of quetiapine treatment were compared among three time periods (before 2003, 2003-2005, after 2005).

RESULTS

Stable state was achieved in 215 patients (80 had Alzheimer's disease, 117 vascular dementia, and 18 mixed dementia). Most patients (177) took quetiapine, 25 took risperidone, and 13 took sulpiride. The whole sample had a long total duration of antipsychotic treatment (median 525 days, mean 707 days). The median dose and total duration of antipsychotic treatment were 1.0mg/day and 238 days for risperidone, 100mg/day and 390 days for sulpiride, and 25mg/day and 611 days for quetiapine, respectively. The optimal dose and total duration of quetiapine treatment decreased significantly after FDA warning in 2005, although the duration remained long.

CONCLUSION

The optimal doses of antipsychotics were not higher than those of western reports, but the total duration of antipsychotic treatment was quite long. Although our study suggests the prescription dosage and duration of antipsychotic treatment decreased significantly after FDA warning in 2005, the duration of treatment was still long. Given the serious safety concerns, more effort should be made to avoid unnecessary and prolonged prescription.

Rapid response to antipsychotic treatment on psychotic prodrome: implications from a case series

The feasible intervention strategy at the prodromal state of psychosis is under debate. We report nine subjects clinically in a putative prodromal state of psychosis who responded to low-dose aripiprazole within the first week of medication. We conjecture that the pathophysiological processes might be easier to modify by antipsychotics in the prodromal state and we believe a short-term low-dose trial of antipsychotic agents is a convenient option for subjects at ultra high risk of psychosis. We urge specific attention to monitor the dissolution of psychotic-like symptoms carefully in order to have a better understanding of the pathogenesis and pharmacotherapy in the inception of psychosis.