Incarceration is a significant social determinant of health, contributing to high morbidity, mortality, and racialized health inequities. However, incarceration status is largely invisible to health services research due to inadequate clinical electronic health record (EHR) capture. This study aims to develop, train, and validate natural language processing (NLP) techniques to more effectively identify incarceration status in the EHR.

The study population consisted of adult patients (≥ 18 y.o.) who presented to the emergency department between June 2013 and August 2021. The EHR database was filtered for notes for specific incarceration-related terms, and then a random selection of 1,000 notes was annotated for incarceration and further stratified into specific statuses of prior history, recent, and current incarceration. For NLP model development, 80% of the notes were used to train the Longformer-based and RoBERTa algorithms. The remaining 20% of the notes underwent analysis with GPT-4.

There were 849 unique patients across 989 visits in the 1000 annotated notes. Manual annotation revealed that 559 of 1000 notes (55.9%) contained evidence of incarceration history. ICD-10 code (sensitivity: 4.8%, specificity: 99.1%, F1-score: 0.09) demonstrated inferior performance to RoBERTa NLP (sensitivity: 78.6%, specificity: 73.3%, F1-score: 0.79), Longformer NLP (sensitivity: 94.6%, specificity: 87.5%, F1-score: 0.93), and GPT-4 (sensitivity: 100%, specificity: 61.1%, F1-score: 0.86).

Our advanced NLP models demonstrate a high degree of accuracy in identifying incarceration status from clinical notes. Further research is needed to explore their scaled implementation in population health initiatives and assess their potential to mitigate health disparities through tailored system interventions.

National health insurance (NHI) Taiwan has provided additional markups on dental service fees for people with specific disabilities, and the expenditure has increased significantly from TWD473 million (USD15 million) in 2016 to TWD722 million (USD24 million) in 2022. The purpose of this study was to determine oral health risk and to develop a risk assessment model for capitation outpatient dental payments in children with Autism.

Based on the literature and expert opinion, we developed a level of oral health risk model from the claim records of 2019. The model uses oral outpatient claim data to analyze: (i) the degree of caries disease; (ii) the level of dental fear or cooperation; and (iii) the level of tooth structure. Each factor was given a score from zero to four and a total score was calculated. Low-, medium-, and high-risk groups were formed based on the total points. The oral health risk capitation models are estimated by ordinary least squares using an individual’s annual outpatient dental expenditure in 2019 as the dependent variable. For subgroups based on age group and level of disability, expenditures predicted by the models are compared with actual outpatient dental expenditures. Predictive R-squared and predictive ratios were used to evaluate the model’s predictability.

The demographic variables, level of oral health risk, preventive dental care, and the type of dental health care predicted 30 percent of subsequent outpatient dental expenditure in children with autism. For subgroups (age group and disability level) of high-risk patients, the model substantially overpredicted the expenditure, whereas underprediction occurred in the low-risk group.

The risk-adjusted model based on principal oral health was more accurate in predicting an individual’s future expenditure than the relevant study in Taiwan. The finding provides insight into the important risk factor in the outpatient dental expenditure of children with autism and the fund planning of dental services for people with specific disabilities.

Somatic symptom disorders (SSD) and functional somatic syndromes (FSS) are often regarded as similar diagnostic constructs; however, whether they exhibit similar clinical outcomes, medical costs, and medication usage patterns has not been examined in nationwide data. Therefore, this study focused on analyzing SSD and four types of FSS (fibromyalgia, chronic fatigue syndrome, irritable bowel syndrome, functional dyspepsia).

This population-based matched cohort study utilized Taiwan's National Health Insurance (NHI) claims database to investigate the impact of SSD/FSS. The study included 2 615 477 newly diagnosed patients with SSD/FSS and matched comparisons from the NHI beneficiary registry. Healthcare utilization, mortality, medical expenditure, and medication usage were assessed as outcome measures. Statistical analysis involved Cox regression models for hazard ratios, generalized linear models for comparing differences, and adjustment for covariates.

All SSD/FSS showed significantly higher adjusted hazard ratios for psychiatric hospitalization and all-cause hospitalization compared to the control group. All SSD/FSS exhibited significantly higher adjusted hazard ratios for suicide, and SSD was particularly high. All-cause mortality was significantly higher in all SSD/FSS. Medical costs were significantly higher for all SSD/FSS compared to controls. The usage duration of all psychiatric medications and analgesics was significantly higher in SSD/FSS compared to the control group.

All SSD/FSS shared similar clinical outcomes and medical costs. The high hazard ratio for suicide in SSD deserves clinical attention.

We explored long-term employment status and income before and after depression diagnosis among men and women and at different working ages in Taiwan.

Data from 2006 to 2019 were obtained from the National Health Insurance Research Database (NHIRD). Individuals with newly diagnosed depressive disorder aged 15 to 64 years during the study period were identified. An equal number of individuals without depression were matched for their demographic and clinical characteristics. Employment outcomes included employment status, which was categorized into employed or unemployed, and annual income. Based on the occupation categories and monthly insurance salary recorded in the Registry for Beneficiaries of the NHIRD, a subject was defined as unemployed if he or she differed from the income earner or the occupation category was unemployed. Monthly income was defined as zero for unemployed subjects and proxied as monthly insurance salary for others. Annual income was the sum of monthly income in each observation year.

A total of 420,935 individuals with depressive disorder were included in the study, and an equal number of individuals with not diagnosed depression served as controls. Employment rate and income were lower in the depression group than in the control group before the year of diagnosis, with a difference of 5.7% in employment rate and USD 1,173 in annual income. This gap increased considerably after the year of diagnosis (7.3% in employment rate and USD 1,573 in annual incomes) and further widened in the subsequent years (8.1% in employment rate and USD 2,006 in annual incomes in the 5th following year). The drops in the employment rate and income caused by depression were more evident in men and older age groups than in women and younger age groups, respectively. However, the reduction in employment rate and income in the following years after the diagnosis was more considerable among younger age groups.

The effect of depression on employment status and income was significant during the year of diagnosis and continued afterwards. The effect on employment outcomes varied between genders and across all age groups.

Evidence has suggested that emotional dysregulation is a transdiagnostic feature in schizophrenia and major affective disorders. However, the relationship between emotional dysregulation and appetite hormone disturbance remains unknown in nonobese adolescents with first-episode schizophrenia, bipolar disorder, and major depressive disorder.

In total, 22 adolescents with schizophrenia; 31 with bipolar disorder; 33 with major depressive disorder; and 41 healthy age-, sex-, and body mass index (BMI)/BMI percentile-matched controls were enrolled for assessing levels of appetite hormones, namely leptin, ghrelin, insulin, and adiponectin. Emotional regulation symptoms were measured using the parent-reported Child Behavior Checklist―Dysregulation Profile.

Adolescents with first-episode schizophrenia, bipolar disorder, and major depressive disorder exhibited greater emotional dysregulation symptoms than the control group (P = .037). Adolescents with bipolar disorder demonstrated higher log-transformed levels of insulin (P = .029) and lower log-transformed levels of leptin (P = .018) compared with the control group. BMI (P < .05) and log-transformed ghrelin levels (P = .028) were positively correlated with emotional dysregulation symptoms.

Emotional dysregulation and appetite hormone disturbance may occur in the early stage of severe mental disorders. Further studies are required to clarify the unidirectional or bidirectional association of emotional dysregulation with BMI/BMI percentile and appetite hormones among patients with severe mental disorder.

Despite reports of an elevated risk of breast cancer associated with antipsychotic use in women, existing evidence remains inconclusive. We aimed to examine existing observational data in the literature and determine this hypothesised association.

We searched Embase, PubMed and Web of Science™ databases on 27 January 2022 for articles reporting relevant cohort or case-control studies published since inception, supplemented with hand searches of the reference lists of the included articles. Quality of studies was assessed using the Newcastle-Ottawa Scale. We generated the pooled odds ratio (OR) and pooled hazard ratio (HR) using a random-effects model to quantify the association. This study was registered with PROSPERO (CRD42022307913).

Nine observational studies, including five cohort and four case-control studies, were eventually included for review (N = 2 031 380) and seven for meta-analysis (N = 1 557 013). All included studies were rated as high-quality (seven to nine stars). Six studies reported a significant association of antipsychotic use with breast cancer, and a stronger association was reported when a greater extent of antipsychotic use, e.g. longer duration, was operationalised as the exposure. Pooled estimates of HRs extracted from cohort studies and ORs from case-control studies were 1.39 [95% confidence interval (CI) 1.11–1.73] and 1.37 (95% CI 0.90–2.09), suggesting a moderate association of antipsychotic use with breast cancer.

Antipsychotic use is moderately associated with breast cancer, possibly mediated by prolactin-elevating properties of certain medications. This risk should be weighed against the potential treatment effects for a balanced prescription decision.

Emotional dysregulation (ED) is a common characteristic of both attention deficit hyperactivity disorder (ADHD) and major depressive disorder (MDD), especially in adolescents. However, whether ADHD and MDD may share the specific ED-related neural networks remains unknown.

In total, 43 adolescents with clinical ED (22 adolescents with ADHD and 21 with MDD) were recruited; in addition, 29 sex- and age-matched healthy controls (HCs) were included. Resting-state functional connectivity (RSFC) analysis, voxel-based morphometry, and diffusion tensor imaging analysis were performed for each patient. In addition, we determined the significant regions of interest in patients with ED due to ADHD and MDD as compared with HCs and tested their correlations with clinical rating scale scores.

Compared with HCs, patients with ED had greater RSFC in the cerebellum and supramarginal gyrus (SMG), especially between vermis VI and the SMG in the attention networks, and lower RSFC between the right supplementary motor area and right lateral parietal area. Lower gray matter (GM) volume in the SMG was also found. RSFC was significantly correlated with clinical rating scale scores for all patients with ED due to ADHD or MDD. GM change was correlated with ED and MDD rating scale scores.

The cerebellum and attention networks might play major roles in ED pathophysiology in adolescents with ADHD and MDD. Increased connectivity of the vermis to the SMG serves as a possible underlying neural network.

Cognitive impairment is common in late-life depression, which may increase Alzheimer disease (AD) risk. Therefore, we aimed to investigate whether late-life major depressive disorder (MDD) has worse cognition and increases the characteristic AD neuropathology. Furthermore, we carried out a comparison between treatment-resistant depression (TRD) and non-TRD. We hypothesized that patients with late-life depression and TRD may have increased β-amyloid (Aβ) deposits in brain regions responsible for global cognition.

We recruited 81 subjects, including 54 MDD patients (27 TRD and 27 non-TRD) and 27 matched healthy controls (HCs). Neurocognitive tasks were examined, including Mini-Mental State Examination and Montreal Cognitive Assessment to detect global cognitive functions. PET with Pittsburgh compound-B and fluorodeoxyglucose were used to capture brain Aβ pathology and glucose use, respectively, in some patients.

MDD patients performed worse in Montreal Cognitive Assessment (p = 0.003) and had more Aβ deposits than HCs across the brain (family-wise error-corrected p < 0.001), with the most significant finding in the left middle frontal gyrus. Significant negative correlations between global cognition and prefrontal Aβ deposits existed in MDD patients, whereas positive correlations were noted in HCs. TRD patients had significantly more deposits in the left-sided brain regions (corrected p < 0.001). The findings were not explained by APOE genotypes. No between-group fluorodeoxyglucose difference was detected.

Late-life depression, particularly TRD, had increased brain Aβ deposits and showed vulnerability to Aβ deposits. A detrimental role of Aβ deposits in global cognition in patients with late-onset or non-late-onset MDD supported the theory that late-life MDD could be a risk factor for AD.

Few studies have explored the complex relationship of pro- and anti-inflammatory cytokines with cognitive function in adolescents with first-episode schizophrenia, bipolar disorder, or major depressive disorder.

In total, 26, 35, and 29 adolescents with first-episode schizophrenia, bipolar disorder, and major depressive disorder, respectively, and 22 age- and sex-matched controls were included in the current study. Cytokines, namely interleukin (IL)-2, IL-6, tumor necrosis factor (TNF)-α, and C-reactive protein (CRP), were assessed. The Wisconsin Card Sorting Test (WCST) and the working memory task were administered to assess cognitive function.

Using generalized linear models with adjustment for demographic data and clinical symptoms, patients with bipolar disorder were found to exhibit the highest levels of CRP (P = .023), IL-6 (P = .022), and TNF-α (P = .011), and had the lowest IL-2 levels (P = .034) among the four groups. According to the results of the WCST and working memory task, adolescents with schizophrenia exhibited the lowest performance in cognitive function. In addition, among the assessed cytokines, only CRP levels (P = .027) were negatively associated with WCST scores.

Dysregulated pro- and anti-inflammatory cytokines and impaired cognitive functioning were observed in first-episode adolescent-onset schizophrenia, bipolar disorder, and major depressive disorder. The altered cytokine profiles may play important roles in the pathophysiology of schizophrenia, bipolar disorder, and major depressive disorder.

The retention of patients under methadone maintenance treatment (MMT) is an indication for the effectiveness of the therapy. We aimed to explore the relation between mortality and the cumulative MMT duration.

A retrospective cohort analysis was performed using Taiwan Illicit Drug Issue Database (TIDID) and National Health Insurance Research Database (NHIRD) during 2012–2016. We included 9149 and 11 112 MMT patients as the short and long groups according to the length of their cumulative MMT duration, 1–364 and ⩾365 days, respectively. The risk of mortality was calculated by Cox proportional hazards regression model with time-dependent exposure to MMT, and the survival probability was plotted with the Kaplan-Meier curve.

The mortality rates were 2.51 and 1.51 per 100 person-years in the short and long cumulative MMT duration groups, respectively. After adjusting for on or off MMT, age, sex, marital status, education level, maximum methadone dose, and comorbidities (human immunodeficiency virus, depression, hepatitis C virus, hepatitis B virus, alcoholic liver disease, and cardiovascular disease), the long group had a lower risk of death (hazard ratio = 0.67; 95% confidence interval 0.60–0.75) than the short group. Increased risk was observed in patients with advanced age, being male, unmarried, infected by HIV, HCV, and HBV, and diagnosed with depression, ALD, and CVD. Causes of death were frequently related to drug and injury.

Longer cumulative MMT duration is associated with lower all-cause and drug-related mortality rate.