Using Natural Language Processing on Electronic Health Records to Enhance Detection and Prediction of Psychosis Risk

Using Electronic Health Records to Facilitate Precision Psychiatry

The use of clinical prediction models to produce individualized risk estimates can facilitate the implementation of precision psychiatry. As a source of data from large, clinically representative patient samples, electronic health records (EHRs) provide a platform to develop and validate clinical prediction models, as well as potentially implement them in routine clinical care. The current review describes promising use cases for the application of precision psychiatry to EHR data and considers their performance in terms of discrimination (ability to separate individuals with and without the outcome) and calibration (extent to which predicted risk estimates correspond to observed outcomes), as well as their potential clinical utility (weighing benefits and costs associated with the model compared to different approaches across different assumptions of the number needed to test). We review 4 externally validated clinical prediction models designed to predict psychosis onset, psychotic relapse, cardiometabolic morbidity, and suicide risk. We then discuss the prospects for clinically implementing these models and the potential added value of integrating data from evidence syntheses, standardized psychometric assessments, and biological data into EHRs. Clinical prediction models can utilize routinely collected EHR data in an innovative way, representing a unique opportunity to inform real-world clinical decision making. Combining data from other sources (e.g., meta-analyses) or enhancing EHR data with information from research studies (clinical and biomarker data) may enhance our abilities to improve the performance of clinical prediction models.

Extraction of sleep information from clinical notes of Alzheimer's disease patients using natural language processing

OBJECTIVES

Alzheimer's disease (AD) is the most common form of dementia in the United States. Sleep is one of the lifestyle-related factors that has been shown critical for optimal cognitive function in old age. However, there is a lack of research studying the association between sleep and AD incidence. A major bottleneck for conducting such research is that the traditional way to acquire sleep information is time-consuming, inefficient, non-scalable, and limited to patients' subjective experience. We aim to automate the extraction of specific sleep-related patterns, such as snoring, napping, poor sleep quality, daytime sleepiness, night wakings, other sleep problems, and sleep duration, from clinical notes of AD patients. These sleep patterns are hypothesized to play a role in the incidence of AD, providing insight into the relationship between sleep and AD onset and progression.

MATERIALS AND METHODS

A gold standard dataset is created from manual annotation of 570 randomly sampled clinical note documents from the adSLEEP, a corpus of 192 000 de-identified clinical notes of 7266 AD patients retrieved from the University of Pittsburgh Medical Center (UPMC). We developed a rule-based natural language processing (NLP) algorithm, machine learning models, and large language model (LLM)-based NLP algorithms to automate the extraction of sleep-related concepts, including snoring, napping, sleep problem, bad sleep quality, daytime sleepiness, night wakings, and sleep duration, from the gold standard dataset.

RESULTS

The annotated dataset of 482 patients comprised a predominantly White (89.2%), older adult population with an average age of 84.7 years, where females represented 64.1%, and a vast majority were non-Hispanic or Latino (94.6%). Rule-based NLP algorithm achieved the best performance of F1 across all sleep-related concepts. In terms of positive predictive value (PPV), the rule-based NLP algorithm achieved the highest PPV scores for daytime sleepiness (1.00) and sleep duration (1.00), while the machine learning models had the highest PPV for napping (0.95) and bad sleep quality (0.86), and LLAMA2 with finetuning had the highest PPV for night wakings (0.93) and sleep problem (0.89).

DISCUSSION

Although sleep information is infrequently documented in the clinical notes, the proposed rule-based NLP algorithm and LLM-based NLP algorithms still achieved promising results. In comparison, the machine learning-based approaches did not achieve good results, which is due to the small size of sleep information in the training data.

CONCLUSION

The results show that the rule-based NLP algorithm consistently achieved the best performance for all sleep concepts. This study focused on the clinical notes of patients with AD but could be extended to general sleep information extraction for other diseases.

Dynamic and Transdiagnostic Risk Calculator Based on Natural Language Processing for the Prediction of Psychosis in Secondary Mental Health Care: Development and Internal-External Validation Cohort Study

BACKGROUND

Automatic transdiagnostic risk calculators can improve the detection of individuals at risk of psychosis. However, they rely on assessment at a single point in time and can be refined with dynamic modeling techniques that account for changes in risk over time.

METHODS

We included 158,139 patients (5007 events) who received a first index diagnosis of a nonorganic and nonpsychotic mental disorder within electronic health records from the South London and Maudsley National Health Service Foundation Trust between January 1, 2008, and October 8, 2021. A dynamic Cox landmark model was developed to estimate the 2-year risk of developing psychosis according to the TRIPOD (Transparent Reporting of a multivariate prediction model for Individual Prognosis or Diagnosis) statement. The dynamic model included 24 predictors extracted at 9 landmark points (baseline, 0, 6, 12, 24, 30, 36, 42, and 48 months): 3 demographic, 1 clinical, and 20 natural language processing-based symptom and substance use predictors. Performance was compared with a static Cox regression model with all predictors assessed at baseline only and indexed via discrimination (C-index), calibration (calibration plots), and potential clinical utility (decision curves) in internal-external validation.

RESULTS

The dynamic model improved discrimination performance from baseline compared with the static model (dynamic: C-index = 0.9; static: C-index = 0.87) and the final landmark point (dynamic: C-index = 0.79; static: C-index = 0.76). The dynamic model was also significantly better calibrated (calibration slope = 0.97-1.1) than the static model at later landmark points (≥24 months). Net benefit was higher for the dynamic than for the static model at later landmark points (≥24 months).

CONCLUSIONS

These findings suggest that dynamic prediction models can improve the detection of individuals at risk for psychosis in secondary mental health care settings.

Protocol for Designing a Model to Predict the Likelihood of Psychosis From Electronic Health Records Using Natural Language Processing and Machine Learning

INTRODUCTION

Rapid identification of individuals developing a psychotic spectrum disorder (PSD) is crucial because untreated psychosis is associated with poor outcomes and decreased treatment response. Lack of recognition of early psychotic symptoms often delays diagnosis, further worsening these outcomes.

METHODS

The proposed study is a cross-sectional, retrospective analysis of electronic health record data including clinician documentation and patient-clinician secure messages for patients aged 15-29 years with ≥ 1 primary care encounter between 2017 and 2019 within 2 Kaiser Permanente regions. Patients with new-onset PSD will be distinguished from those without a diagnosis if they have ≥ 1 PSD diagnosis within 12 months following the primary care encounter. The prediction model will be trained using a trisourced natural language processing feature extraction design and validated both within each region separately and in a modified combined sample.

DISCUSSION

This proposed model leverages the strengths of the large volume of patient-specific data from an integrated electronic health record with natural language processing to identify patients at elevated chance of developing a PSD. This project carries the potential to reduce the duration of untreated psychosis and thereby improve long-term patient outcomes.

A real-world data analysis of electronic health records to investigate the associations of predominant negative symptoms with healthcare resource utilisation, costs and treatment patterns among patients with schizophrenia

OBJECTIVES

Negative symptoms in schizophrenia are associated with significant illness burden. We sought to investigate clinical outcomes for patients with schizophrenia who present with predominant negative symptoms (PNS) vs without PNS.

DESIGN

Retrospective analysis of electronic health record (EHR) data.

SETTING

25 US providers of mental healthcare.

PARTICIPANTS

4444 adults with schizophrenia receiving care between 1999 and 2020.

EXPOSURE

PNS defined as ≥3 negative symptoms and ≤3 positive symptoms recorded in EHR data at the time of the first recorded schizophrenia diagnosis (index date). Symptom data were ascertained using natural language processing applied to semistructured free text records documenting the mental state examination. A matched sample (1:1) of patients without PNS was used to compare outcomes. Follow-up data were obtained up to 12 months following the index date.

PRIMARY OUTCOME MEASURE

Mean number of psychiatric hospital admissions.

SECONDARY OUTCOME MEASURES

Mean number of outpatient visits, estimated treatment costs, Clinical Global Impression - Severity score and antipsychotic treatments (12 months before and after index date).

RESULTS

360 (8%) patients had PNS and 4084 (92%) did not have PNS. Patients with PNS were younger (36.4 vs 39.7 years, p<0.001) with a greater prevalence of psychiatric comorbidities (schizoaffective disorders: 25.0 vs 18.4%, p=0.003; major depressive disorder: 17.8 vs 9.8%, p<0.001). During follow-up, patients with PNS had fewer days with an antipsychotic prescription (mean=111.8 vs 140.9 days, p<0.001). Compared with matched patients without PNS, patients with PNS were more likely to have a psychiatric inpatient hospitalisation (76.1% vs 59.7%, p<0.001) and had greater estimated inpatient costs ($16 893 vs $13 732, p=0.04).

CONCLUSIONS

Patients with PNS were younger and presented with greater illness severity and more psychiatric comorbidities compared with patients without PNS. Our findings highlight an unmet need for novel therapeutic approaches to address negative symptoms to improve clinical outcomes.

Validation of an ICD-Code-Based Case Definition for Psychotic Illness Across Three Health Systems

BACKGROUND AND HYPOTHESIS

Psychosis-associated diagnostic codes are increasingly being utilized as case definitions for electronic health record (EHR)-based algorithms to predict and detect psychosis. However, data on the validity of psychosis-related diagnostic codes is limited. We evaluated the positive predictive value (PPV) of International Classification of Diseases (ICD) codes for psychosis.

STUDY DESIGN

Using EHRs at 3 health systems, ICD codes comprising primary psychotic disorders and mood disorders with psychosis were grouped into 5 higher-order groups. 1133 records were sampled for chart review using the full EHR. PPVs (the probability of chart-confirmed psychosis given ICD psychosis codes) were calculated across multiple treatment settings.

STUDY RESULTS

PPVs across all diagnostic groups and hospital systems exceeded 70%: Mass General Brigham 0.72 [95% CI 0.68-0.77], Boston Children's Hospital 0.80 [0.75-0.84], and Boston Medical Center 0.83 [0.79-0.86]. Schizoaffective disorder PPVs were consistently the highest across sites (0.80-0.92) and major depressive disorder with psychosis were the most variable (0.57-0.79). To determine if the first documented code captured first-episode psychosis (FEP), we excluded cases with prior chart evidence of a diagnosis of or treatment for a psychotic illness, yielding substantially lower PPVs (0.08-0.62).

CONCLUSIONS

We found that the first documented psychosis diagnostic code accurately captured true episodes of psychosis but was a poor index of FEP. These data have important implications for the case definitions used in the development of risk prediction models designed to predict or detect undiagnosed psychosis.

A transdiagnostic prodrome for severe mental disorders: an electronic health record study

Effective prevention of severe mental disorders (SMD), including non-psychotic unipolar mood disorders (UMD), non-psychotic bipolar mood disorders (BMD), and psychotic disorders (PSY), rely on accurate knowledge of the duration, first presentation, time course and transdiagnosticity of their prodromal stages. Here we present a retrospective, real-world, cohort study using electronic health records, adhering to RECORD guidelines. Natural language processing algorithms were used to extract monthly occurrences of 65 prodromal features (symptoms and substance use), grouped into eight prodromal clusters. The duration, first presentation, and transdiagnosticity of the prodrome were compared between SMD groups with one-way ANOVA, Cohen's f and d. The time course (mean occurrences) of prodromal clusters was compared between SMD groups with linear mixed-effects models. 26,975 individuals diagnosed with ICD-10 SMD were followed up for up to 12 years (UMD = 13,422; BMD = 2506; PSY = 11,047; median[IQR] age 39.8[23.7] years; 55% female; 52% white). The duration of the UMD prodrome (18[36] months) was shorter than BMD (26[35], d = 0.21) and PSY (24[38], d = 0.18). Most individuals presented with multiple first prodromal clusters, with the most common being non-specific ('other'; 88% UMD, 85% BMD, 78% PSY). The only first prodromal cluster that showed a medium-sized difference between the three SMD groups was positive symptoms (f = 0.30). Time course analysis showed an increase in prodromal cluster occurrences approaching SMD onset. Feature occurrence across the prodromal period showed small/negligible differences between SMD groups, suggesting that most features are transdiagnostic, except for positive symptoms (e.g. paranoia, f = 0.40). Taken together, our findings show minimal differences in the duration and first presentation of the SMD prodromes as recorded in secondary mental health care. All the prodromal clusters intensified as individuals approached SMD onset, and all the prodromal features other than positive symptoms are transdiagnostic. These results support proposals to develop transdiagnostic preventive services for affective and psychotic disorders detected in secondary mental healthcare.

Identifying Psychosis Episodes in Psychiatric Admission Notes via Rule-based Methods, Machine Learning, and Pre-Trained Language Models

Early and accurate diagnosis is crucial for effective treatment and improved outcomes, yet identifying psychotic episodes presents significant challenges due to its complex nature and the varied presentation of symptoms among individuals. One of the primary difficulties lies in the underreporting and underdiagnosis of psychosis, compounded by the stigma surrounding mental health and the individuals' often diminished insight into their condition. Existing efforts leveraging Electronic Health Records (EHRs) to retrospectively identify psychosis typically rely on structured data, such as medical codes and patient demographics, which frequently lack essential information. Addressing these challenges, our study leverages Natural Language Processing (NLP) algorithms to analyze psychiatric admission notes for the diagnosis of psychosis, providing a detailed evaluation of rule-based algorithms, machine learning models, and pre-trained language models. Additionally, the study investigates the effectiveness of employing keywords to streamline extensive note data before training and evaluating the models. Analyzing 4,617 initial psychiatric admission notes (1,196 cases of psychosis versus 3,433 controls) from 2005 to 2019, we discovered that the XGBoost classifier employing Term Frequency-Inverse Document Frequency (TF-IDF) features derived from notes pre-selected by expert-curated keywords, attained the highest performance with an F1 score of 0.8881 (AUROC [95% CI]: 0.9725 [0.9717, 0.9733]). BlueBERT demonstrated comparable efficacy an F1 score of 0.8841 (AUROC [95% CI]: 0.97 [0.9580,0.9820]) on the same set of notes. Both models markedly outperformed traditional International Classification of Diseases (ICD) code-based detection methods from discharge summaries, which had an F1 score of 0.7608, thus improving the margin by 0.12. Furthermore, our findings indicate that keyword pre-selection markedly enhances the performance of both machine learning and pre-trained language models. This study illustrates the potential of NLP techniques to improve psychosis detection within admission notes and aims to serve as a foundational reference for future research on applying NLP for psychosis identification in EHR notes.

Identifying Psychosis Episodes in Psychiatric Admission Notes via Rule-based Methods, Machine Learning, and Pre-Trained Language Models

Early and accurate diagnosis is crucial for effective treatment and improved outcomes, yet identifying psychotic episodes presents significant challenges due to its complex nature and the varied presentation of symptoms among individuals. One of the primary difficulties lies in the underreporting and underdiagnosis of psychosis, compounded by the stigma surrounding mental health and the individuals' often diminished insight into their condition. Existing efforts leveraging Electronic Health Records (EHRs) to retrospectively identify psychosis typically rely on structured data, such as medical codes and patient demographics, which frequently lack essential information. Addressing these challenges, our study leverages Natural Language Processing (NLP) algorithms to analyze psychiatric admission notes for the diagnosis of psychosis, providing a detailed evaluation of rule-based algorithms, machine learning models, and pre-trained language models. Additionally, the study investigates the effectiveness of employing keywords to streamline extensive note data before training and evaluating the models. Analyzing 4,617 initial psychiatric admission notes (1,196 cases of psychosis versus 3,433 controls) from 2005 to 2019, we discovered that the XGBoost classifier employing Term Frequency-Inverse Document Frequency (TF-IDF) features derived from notes pre-selected by expert-curated keywords, attained the highest performance with an F1 score of 0.8881 (AUROC [95% CI]: 0.9725 [0.9717, 0.9733]). BlueBERT demonstrated comparable efficacy an F1 score of 0.8841 (AUROC [95% CI]: 0.97 [0.9580, 0.9820]) on the same set of notes. Both models markedly outperformed traditional International Classification of Diseases (ICD) code-based detection methods from discharge summaries, which had an F1 score of 0.7608, thus improving the margin by 0.12. Furthermore, our findings indicate that keyword pre-selection markedly enhances the performance of both machine learning and pre-trained language models. This study illustrates the potential of NLP techniques to improve psychosis detection within admission notes and aims to serve as a foundational reference for future research on applying NLP for psychosis identification in EHR notes.

Validation of an ICD-code-based case definition for psychotic illness across three health systems

Background and Hypothesis

Early detection of psychosis is critical for improving outcomes. Algorithms to predict or detect psychosis using electronic health record (EHR) data depend on the validity of the case definitions used, typically based on diagnostic codes. Data on the validity of psychosis-related diagnostic codes is limited. We evaluated the positive predictive value (PPV) of International Classification of Diseases (ICD) codes for psychosis.

Study Design

Using EHRs at three health systems, ICD codes comprising primary psychotic disorders and mood disorders with psychosis were grouped into five higher-order groups. 1,133 records were sampled for chart review using the full EHR. PPVs (the probability of chart-confirmed psychosis given ICD psychosis codes) were calculated across multiple treatment settings.

Study Results

PPVs across all diagnostic groups and hospital systems exceeded 70%: Massachusetts General Brigham 0.72 [95% CI 0.68-0.77], Boston Children's Hospital 0.80 [0.75-0.84], and Boston Medical Center 0.83 [0.79-0.86]. Schizoaffective disorder PPVs were consistently the highest across sites (0.80-0.92) and major depressive disorder with psychosis were the most variable (0.57-0.79). To determine if the first documented code captured first-episode psychosis (FEP), we excluded cases with prior chart evidence of a diagnosis of or treatment for a psychotic illness, yielding substantially lower PPVs (0.08-0.62).

Conclusions

We found that the first documented psychosis diagnostic code accurately captured true episodes of psychosis but was a poor index of FEP. These data have important implications for the development of risk prediction models designed to predict or detect undiagnosed psychosis.

Language and communication rehabilitation in patients with schizophrenia: A narrative review

Language impairments often appear in patients with schizophrenia and are potential targets for rehabilitation. Clinical practice and research should be intimately connected. The aim was to perform a narrative review of the assessment and intervention tools that have been used for the rehabilitation of schizophrenia patients with language and communication impairments. Two types of tools, general and specific, were developed for both purposes. General tools include the Positive and Negative Syndrome Scale for assessment, and the Integrated Psychological Therapy for intervention. The specific tools used to evaluate language and communication impairments include the Scale for the Assessment of Thought, Language and Communication, the Formal Thought Disorder scales (for caregivers and patients), and the Thought and Language Disorder scale. The most recent language-specific intervention tools include the Cognitive Pragmatic Treatment, Conecta-2, Let's talk! Multimodal Speech-Gesture training, Speech Therapy Intervention Group, and PragmaCom. These tools primarily involve psychopathology/psychiatry, psychology, linguistics, speech and language therapy, and nursing. In conclusion, a wide range of assessment and intervention tools are available for the rehabilitation of language and communication impairments associated with schizophrenia. An integrative and interdisciplinary approach should always be considered for rehabilitation of language and communication in patients with schizophrenia throughout their lifetime.

Identification of risk factors for the onset of delirium associated with COVID-19 by mining nursing records

COVID-19 has a range of complications, from no symptoms to severe pneumonia. It can also affect multiple organs including the nervous system. COVID-19 affects the brain, leading to neurological symptoms such as delirium. Delirium, a sudden change in consciousness, can increase the risk of death and prolong the hospital stay. However, research on delirium prediction in patients with COVID-19 is insufficient. This study aimed to identify new risk factors that could predict the onset of delirium in patients with COVID-19 using machine learning (ML) applied to nursing records. This retrospective cohort study used natural language processing and ML to develop a model for classifying the nursing records of patients with delirium. We extracted the features of each word from the model and grouped similar words. To evaluate the usefulness of word groups in predicting the occurrence of delirium in patients with COVID-19, we analyzed the temporal changes in the frequency of occurrence of these word groups before and after the onset of delirium. Moreover, the sensitivity, specificity, and odds ratios were calculated. We identified (1) elimination-related behaviors and conditions and (2) abnormal patient behavior and conditions as risk factors for delirium. Group 1 had the highest sensitivity (0.603), whereas group 2 had the highest specificity and odds ratio (0.938 and 6.903, respectively). These results suggest that these parameters may be useful in predicting delirium in these patients. The risk factors for COVID-19-associated delirium identified in this study were more specific but less sensitive than the ICDSC (Intensive Care Delirium Screening Checklist) and CAM-ICU (Confusion Assessment Method for the Intensive Care Unit). However, they are superior to the ICDSC and CAM-ICU because they can predict delirium without medical staff and at no cost.

Applying contrastive pre-training for depression and anxiety risk prediction in type 2 diabetes patients based on heterogeneous electronic health records: a primary healthcare case study

OBJECTIVE

Due to heterogeneity and limited medical data in primary healthcare services (PHS), assessing the psychological risk of type 2 diabetes mellitus (T2DM) patients in PHS is difficult. Using unsupervised contrastive pre-training, we proposed a deep learning framework named depression and anxiety prediction (DAP) to predict depression and anxiety in T2DM patients.

MATERIALS AND METHODS

The DAP model consists of two sub-models. Firstly, the pre-trained model of DAP used unlabeled discharge records of 85 085 T2DM patients from the First Affiliated Hospital of Nanjing Medical University for unsupervised contrastive learning on heterogeneous electronic health records (EHRs). Secondly, the fine-tuned model of DAP used case-control cohorts (17 491 patients) selected from 149 596 T2DM patients' EHRs in the Nanjing Health Information Platform (NHIP). The DAP model was validated in 1028 patients from PHS in NHIP. Evaluation included receiver operating characteristic area under the curve (ROC-AUC) and precision-recall area under the curve (PR-AUC), and decision curve analysis (DCA).

RESULTS

The pre-training step allowed the DAP model to converge at a faster rate. The fine-tuned DAP model significantly outperformed the baseline models (logistic regression, extreme gradient boosting, and random forest) with ROC-AUC of 0.91±0.028 and PR-AUC of 0.80±0.067 in 10-fold internal validation, and with ROC-AUC of 0.75 ± 0.045 and PR-AUC of 0.47 ± 0.081 in external validation. The DCA indicate the clinical potential of the DAP model.

CONCLUSION

The DAP model effectively predicted post-discharge depression and anxiety in T2DM patients from PHS, reducing data fragmentation and limitations. This study highlights the DAP model's potential for early detection and intervention in depression and anxiety, improving outcomes for diabetes patients.

Combining clinical notes with structured electronic health records enhances the prediction of mental health crises

An automatic prediction of mental health crises can improve caseload prioritization and enable preventative interventions, improving patient outcomes and reducing costs. We combine structured electronic health records (EHRs) with clinical notes from 59,750 de-identified patients to predict the risk of mental health crisis relapse within the next 28 days. The results suggest that an ensemble machine learning model that relies on structured EHRs and clinical notes when available, and relying solely on structured data when the notes are unavailable, offers superior performance over models trained with either of the two data streams alone. Furthermore, the study provides key takeaways related to the required amount of clinical notes to add value in predictive analytics. This study sheds light on the untapped potential of clinical notes in the prediction of mental health crises and highlights the importance of choosing an appropriate machine learning method to combine structured and unstructured EHRs.

Machine learning techniques in diagnostics and prediction of the clinical features of schizophrenia: a narrative review

BACKGROUND

Schizophrenia is a severe psychiatric disorder associated with a significant negative impact. Early diagnosis and treatment of schizophrenia has a favorable effect on the clinical outcome and patients quality of life. In this context, machine learning techniques open up new opportunities for a more accurate diagnosis and prediction of the clinical features of this illness.

AIM

This literature review is aimed to search for information on the use of machine learning techniques in the prediction and diagnosis of schizophrenia and the determination of its clinical features.

METHODS

The Google Scholar, PubMed, and eLIBRARY.ru databases were used to search for relevant data. The review included articles that had been published not earlier than January 1, 2010, and not later than March 31, 2023. Combinations of the following keywords were applied for search queries: machine learning, deep learning, schizophrenia, neural network, predictors, artificial intelligence, diagnostics, suicide, depressive, insomnia, and cognitive. Original articles regardless of their design were included in the review. Descriptive analysis was used to summarize the retrieved data.

RESULTS

Machine learning techniques are widely used in the functional assessment of patients with schizophrenia. They are used for interpretation of MRI, EEG, and actigraphy findings. Also, models created using machine learning algorithms can analyze speech, behavior, and the creativity of people and these data can be used for the diagnosis of psychiatric disorders. It has been found that different machine learning-based models can help specialists predict and diagnose schizophrenia based on medical history and genetic data, as well as epigenetic information. Machine learning techniques can also be used to build effective models that can help specialists diagnose and predict clinical manifestations and complications of schizophrenia, such as insomnia, depressive symptoms, suicide risk, aggressive behavior, and changes in cognitive functions over time.

CONCLUSION

Machine learning techniques play an important role in psychiatry, as they have been used in models that help specialists in the diagnosis of schizophrenia and determination of its clinical features. The use of machine learning algorithms is one of the most promising direction in psychiatry, and it can significantly improve the effectiveness of the diagnosis and treatment of schizophrenia.

Natural Language Processing in Electronic Health Records in relation to healthcare decision-making: A systematic review

BACKGROUND

Natural Language Processing (NLP) is widely used to extract clinical insights from Electronic Health Records (EHRs). However, the lack of annotated data, automated tools, and other challenges hinder the full utilisation of NLP for EHRs. Various Machine Learning (ML), Deep Learning (DL) and NLP techniques are studied and compared to understand the limitations and opportunities in this space comprehensively.

METHODOLOGY

After screening 261 articles from 11 databases, we included 127 papers for full-text review covering seven categories of articles: (1) medical note classification, (2) clinical entity recognition, (3) text summarisation, (4) deep learning (DL) and transfer learning architecture, (5) information extraction, (6) Medical language translation and (7) other NLP applications. This study follows the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.

RESULT AND DISCUSSION

EHR was the most commonly used data type among the selected articles, and the datasets were primarily unstructured. Various ML and DL methods were used, with prediction or classification being the most common application of ML or DL. The most common use cases were: the International Classification of Diseases, Ninth Revision (ICD-9) classification, clinical note analysis, and named entity recognition (NER) for clinical descriptions and research on psychiatric disorders.

CONCLUSION

We find that the adopted ML models were not adequately assessed. In addition, the data imbalance problem is quite important, yet we must find techniques to address this underlining problem. Future studies should address key limitations in studies, primarily identifying Lupus Nephritis, Suicide Attempts, perinatal self-harmed and ICD-9 classification.

A survey on clinical natural language processing in the United Kingdom from 2007 to 2022

Much of the knowledge and information needed for enabling high-quality clinical research is stored in free-text format. Natural language processing (NLP) has been used to extract information from these sources at scale for several decades. This paper aims to present a comprehensive review of clinical NLP for the past 15 years in the UK to identify the community, depict its evolution, analyse methodologies and applications, and identify the main barriers. We collect a dataset of clinical NLP projects (n = 94; £ = 41.97 m) funded by UK funders or the European Union's funding programmes. Additionally, we extract details on 9 funders, 137 organisations, 139 persons and 431 research papers. Networks are created from timestamped data interlinking all entities, and network analysis is subsequently applied to generate insights. 431 publications are identified as part of a literature review, of which 107 are eligible for final analysis. Results show, not surprisingly, clinical NLP in the UK has increased substantially in the last 15 years: the total budget in the period of 2019-2022 was 80 times that of 2007-2010. However, the effort is required to deepen areas such as disease (sub-)phenotyping and broaden application domains. There is also a need to improve links between academia and industry and enable deployments in real-world settings for the realisation of clinical NLP's great potential in care delivery. The major barriers include research and development access to hospital data, lack of capable computational resources in the right places, the scarcity of labelled data and barriers to sharing of pretrained models.

Understanding psychiatric illness through natural language processing (UNDERPIN): Rationale, design, and methodology

Introduction

Psychiatric disorders are diagnosed through observations of psychiatrists according to diagnostic criteria such as the DSM-5. Such observations, however, are mainly based on each psychiatrist's level of experience and often lack objectivity, potentially leading to disagreements among psychiatrists. In contrast, specific linguistic features can be observed in some psychiatric disorders, such as a loosening of associations in schizophrenia. Some studies explored biomarkers, but biomarkers have yet to be used in clinical practice.

Aim

The purposes of this study are to create a large dataset of Japanese speech data labeled with detailed information on psychiatric disorders and neurocognitive disorders to quantify the linguistic features of those disorders using natural language processing and, finally, to develop objective and easy-to-use biomarkers for diagnosing and assessing the severity of them.

Methods

This study will have a multi-center prospective design. The DSM-5 or ICD-11 criteria for major depressive disorder, bipolar disorder, schizophrenia, and anxiety disorder and for major and minor neurocognitive disorders will be regarded as the inclusion criteria for the psychiatric disorder samples. For the healthy subjects, the absence of a history of psychiatric disorders will be confirmed using the Mini-International Neuropsychiatric Interview (M.I.N.I.). The absence of current cognitive decline will be confirmed using the Mini-Mental State Examination (MMSE). A psychiatrist or psychologist will conduct 30-to-60-min interviews with each participant; these interviews will include free conversation, picture-description task, and story-telling task, all of which will be recorded using a microphone headset. In addition, the severity of disorders will be assessed using clinical rating scales. Data will be collected from each participant at least twice during the study period and up to a maximum of five times at an interval of at least one month.

Discussion

This study is unique in its large sample size and the novelty of its method, and has potential for applications in many fields. We have some challenges regarding inter-rater reliability and the linguistic peculiarities of Japanese. As of September 2022, we have collected a total of >1000 records from >400 participants. To the best of our knowledge, this data sample is one of the largest in this field.

Clinical Trial Registration

Identifier: UMIN000032141.

Prognostic accuracy and clinical utility of psychometric instruments for individuals at clinical high-risk of psychosis: a systematic review and meta-analysis

Accurate prognostication of individuals at clinical high-risk for psychosis (CHR-P) is an essential initial step for effective primary indicated prevention. We aimed to summarise the prognostic accuracy and clinical utility of CHR-P assessments for primary indicated psychosis prevention. Web of Knowledge databases were searched until 1st January 2022 for longitudinal studies following-up individuals undergoing a psychometric or diagnostic CHR-P assessment, reporting transition to psychotic disorders in both those who meet CHR-P criteria (CHR-P + ) or not (CHR-P-). Prognostic accuracy meta-analysis was conducted following relevant guidelines. Primary outcome was prognostic accuracy, indexed by area-under-the-curve (AUC), sensitivity and specificity, estimated by the number of true positives, false positives, false negatives and true negatives at the longest available follow-up time. Clinical utility analyses included: likelihood ratios, Fagan's nomogram, and population-level preventive capacity (Population Attributable Fraction, PAF). A total of 22 studies (n = 4 966, 47.5% female, age range 12-40) were included. There were not enough meta-analysable studies on CHR-P diagnostic criteria (DSM-5 Attenuated Psychosis Syndrome) or non-clinical samples. Prognostic accuracy of CHR-P psychometric instruments in clinical samples (individuals referred to CHR-P services or diagnosed with 22q.11.2 deletion syndrome) was excellent: AUC = 0.85 (95% CI: 0.81-0.88) at a mean follow-up time of 34 months. This result was driven by outstanding sensitivity (0.93, 95% CI: 0.87-0.96) and poor specificity (0.58, 95% CI: 0.50-0.66). Being CHR-P + was associated with a small likelihood ratio LR + (2.17, 95% CI: 1.81-2.60) for developing psychosis. Being CHR-P- was associated with a large LR- (0.11, 95%CI: 0.06-0.21) for developing psychosis. Fagan's nomogram indicated a low positive (0.0017%) and negative (0.0001%) post-test risk in non-clinical general population samples. The PAF of the CHR-P state is 10.9% (95% CI: 4.1-25.5%). These findings consolidate the use of psychometric instruments for CHR-P in clinical samples for primary indicated prevention of psychosis. Future research should improve the ability to rule in psychosis risk.

Natural Language Processing and Psychosis: On the Need for Comprehensive Psychometric Evaluation

BACKGROUND AND HYPOTHESIS

Despite decades of "proof of concept" findings supporting the use of Natural Language Processing (NLP) in psychosis research, clinical implementation has been slow. One obstacle reflects the lack of comprehensive psychometric evaluation of these measures. There is overwhelming evidence that criterion and content validity can be achieved for many purposes, particularly using machine learning procedures. However, there has been very little evaluation of test-retest reliability, divergent validity (sufficient to address concerns of a "generalized deficit"), and potential biases from demographics and other individual differences.

STUDY DESIGN

This article highlights these concerns in development of an NLP measure for tracking clinically rated paranoia from video "selfies" recorded from smartphone devices. Patients with schizophrenia or bipolar disorder were recruited and tracked over a week-long epoch. A small NLP-based feature set from 499 language samples were modeled on clinically rated paranoia using regularized regression.

STUDY RESULTS

While test-retest reliability was high, criterion, and convergent/divergent validity were only achieved when considering moderating variables, notably whether a patient was away from home, around strangers, or alone at the time of the recording. Moreover, there were systematic racial and sex biases in the model, in part, reflecting whether patients submitted videos when they were away from home, around strangers, or alone.

CONCLUSIONS

Advancing NLP measures for psychosis will require deliberate consideration of test-retest reliability, divergent validity, systematic biases and the potential role of moderators. In our example, a comprehensive psychometric evaluation revealed clear strengths and weaknesses that can be systematically addressed in future research.

Language production impairments in patients with a first episode of psychosis

Language production has often been described as impaired in psychiatric diseases such as in psychosis. Nevertheless, little is known about the characteristics of linguistic difficulties and their relation with other cognitive domains in patients with a first episode of psychosis (FEP), either affective or non-affective. To deepen our comprehension of linguistic profile in FEP, 133 patients with FEP (95 non-affective, FEP-NA; 38 affective, FEP-A) and 133 healthy controls (HC) were assessed with a narrative discourse task. Speech samples were systematically analyzed with a well-established multilevel procedure investigating both micro- (lexicon, morphology, syntax) and macro-linguistic (discourse coherence, pragmatics) levels of linguistic processing. Executive functioning and IQ were also evaluated. Both linguistic and neuropsychological measures were secondarily implemented with a machine learning approach in order to explore their predictive accuracy in classifying participants as FEP or HC. Compared to HC, FEP patients showed language production difficulty at both micro- and macro-linguistic levels. As for the former, FEP produced shorter and simpler sentences and fewer words per minute, along with a reduced number of lexical fillers, compared to HC. At the macro-linguistic level, FEP performance was impaired in local coherence, which was paired with a higher percentage of utterances with semantic errors. Linguistic measures were not correlated with any neuropsychological variables. No significant differences emerged between FEP-NA and FEP-A (p≥0.02, after Bonferroni correction). Machine learning analysis showed an accuracy of group prediction of 76.36% using language features only, with semantic variables being the most impactful. Such a percentage was enhanced when paired with clinical and neuropsychological variables. Results confirm the presence of language production deficits already at the first episode of the illness, being such impairment not related to other cognitive domains. The high accuracy obtained by the linguistic set of features in classifying groups support the use of machine learning methods in neuroscience investigations.

Associations of presenting symptoms and subsequent adverse clinical outcomes in people with unipolar depression: a prospective natural language processing (NLP), transdiagnostic, network analysis of electronic health record (EHR) data

OBJECTIVE

To investigate the associations of symptoms of mania and depression with clinical outcomes in people with unipolar depression.

DESIGN

A natural language processing electronic health record study. We used network analysis to determine symptom network structure and multivariable Cox regression to investigate associations with clinical outcomes.

SETTING

The South London and Maudsley Clinical Record Interactive Search database.

PARTICIPANTS

All patients presenting with unipolar depression between 1 April 2006 and 31 March 2018.

EXPOSURE

(1) Symptoms of mania: Elation; Grandiosity; Flight of ideas; Irritability; Pressured speech. (2) Symptoms of depression: Disturbed mood; Anhedonia; Guilt; Hopelessness; Helplessness; Worthlessness; Tearfulness; Low energy; Reduced appetite; Weight loss. (3) Symptoms of mania or depression (overlapping symptoms): Poor concentration; Insomnia; Disturbed sleep; Agitation; Mood instability.

MAIN OUTCOMES

(1) Bipolar or psychotic disorder diagnosis. (2) Psychiatric hospital admission.

RESULTS

Out of 19 707 patients, at least 1 depression, overlapping or mania symptom was present in 18 998 (96.4%), 15 954 (81.0%) and 4671 (23.7%) patients, respectively. 2772 (14.1%) patients subsequently developed bipolar or psychotic disorder during the follow-up period. The presence of at least one mania (HR 2.00, 95% CI 1.85 to 2.16), overlapping symptom (HR 1.71, 95% CI 1.52 to 1.92) or symptom of depression (HR 1.31, 95% CI 1.07 to 1.61) were associated with significantly increased risk of onset of a bipolar or psychotic disorder. Mania (HR 1.95, 95% CI 1.77 to 2.15) and overlapping symptoms (HR 1.76, 95% CI 1.52 to 2.04) were associated with greater risk for psychiatric hospital admission than symptoms of depression (HR 1.41, 95% CI 1.06 to 1.88).

CONCLUSIONS

The presence of mania or overlapping symptoms in people with unipolar depression is associated with worse clinical outcomes. Symptom-based approaches to defining clinical phenotype may facilitate a more personalised treatment approach and better predict subsequent clinical outcomes than psychiatric diagnosis alone.

NeuroBlu, an electronic health record (EHR) trusted research environment (TRE) to support mental healthcare analytics with real-world data

PURPOSE

NeuroBlu is a real-world data (RWD) repository that contains deidentified electronic health record (EHR) data from US mental healthcare providers operating the MindLinc EHR system. NeuroBlu enables users to perform statistical analysis through a secure web-based interface. Structured data are available for sociodemographic characteristics, mental health service contacts, hospital admissions, International Classification of Diseases ICD-9/ICD-10 diagnosis, prescribed medications, family history of mental disorders, Clinical Global Impression-Severity and Improvement (CGI-S/CGI-I) and Global Assessment of Functioning (GAF). To further enhance the data set, natural language processing (NLP) tools have been applied to obtain mental state examination (MSE) and social/environmental data. This paper describes the development and implementation of NeuroBlu, the procedures to safeguard data integrity and security and how the data set supports the generation of real-world evidence (RWE) in mental health.

PARTICIPANTS

As of 31 July 2021, 562 940 individuals (48.9% men) were present in the data set with a mean age of 33.4 years (SD: 18.4 years). The most frequently recorded diagnoses were substance use disorders (1 52 790 patients), major depressive disorder (1 29 120 patients) and anxiety disorders (1 03 923 patients). The median duration of follow-up was 7 months (IQR: 1.3 to 24.4 months).

FINDINGS TO DATE

The data set has supported epidemiological studies demonstrating increased risk of psychiatric hospitalisation and reduced antidepressant treatment effectiveness among people with comorbid substance use disorders. It has also been used to develop data visualisation tools to support clinical decision-making, evaluate comparative effectiveness of medications, derive models to predict treatment response and develop NLP applications to obtain clinical information from unstructured EHR data.

FUTURE PLANS

The NeuroBlu data set will be further analysed to better understand factors related to poor clinical outcome, treatment responsiveness and the development of predictive analytic tools that may be incorporated into the source EHR system to support real-time clinical decision-making in the delivery of mental healthcare services.

Predicting Personality and Psychological Distress Using Natural Language Processing: A Study Protocol

Background

Self-report multiple choice questionnaires have been widely utilized to quantitatively measure one's personality and psychological constructs. Despite several strengths (e.g., brevity and utility), self-report multiple choice questionnaires have considerable limitations in nature. With the rise of machine learning (ML) and Natural language processing (NLP), researchers in the field of psychology are widely adopting NLP to assess psychological construct to predict human behaviors. However, there is a lack of connections between the work being performed in computer science and that of psychology due to small data sets and unvalidated modeling practices.

Aims

The current article introduces the study method and procedure of phase II which includes the interview questions for the five-factor model (FFM) of personality developed in phase I. This study aims to develop the interview (semi-structured) and open-ended questions for the FFM-based personality assessments, specifically designed with experts in the field of clinical and personality psychology (phase 1), and to collect the personality-related text data using the interview questions and self-report measures on personality and psychological distress (phase 2). The purpose of the study includes examining the relationship between natural language data obtained from the interview questions, measuring the FFM personality constructs, and psychological distress to demonstrate the validity of the natural language-based personality prediction.

Methods

Phase I (pilot) study was conducted to fifty-nine native Korean adults to acquire the personality-related text data from the interview (semi-structured) and open-ended questions based on the FFM of personality. The interview questions were revised and finalized with the feedback from the external expert committee, consisting of personality and clinical psychologists. Based on the established interview questions, a total of 300 Korean adults will be recruited using a convenience sampling method via online survey. The text data collected from interviews will be analyzed using the natural language processing. The results of the online survey including demographic data, depression, anxiety, and personality inventories will be analyzed together in the model to predict individuals' FFM of personality and the level of psychological distress (phase 2).

Psychosis Relapse Prediction Leveraging Electronic Health Records Data and Natural Language Processing Enrichment Methods

BACKGROUND

Identifying patients at a high risk of psychosis relapse is crucial for early interventions. A relevant psychiatric clinical context is often recorded in clinical notes; however, the utilization of unstructured data remains limited. This study aimed to develop psychosis-relapse prediction models using various types of clinical notes and structured data.

METHODS

Clinical data were extracted from the electronic health records of the Ajou University Medical Center in South Korea. The study population included patients with psychotic disorders, and outcome was psychosis relapse within 1 year. Using only structured data, we developed an initial prediction model, then three natural language processing (NLP)-enriched models using three types of clinical notes (psychological tests, admission notes, and initial nursing assessment) and one complete model. Latent Dirichlet Allocation was used to cluster the clinical context into similar topics. All models applied the least absolute shrinkage and selection operator logistic regression algorithm. We also performed an external validation using another hospital database.

RESULTS

A total of 330 patients were included, and 62 (18.8%) experienced psychosis relapse. Six predictors were used in the initial model and 10 additional topics from Latent Dirichlet Allocation processing were added in the enriched models. The model derived from all notes showed the highest value of the area under the receiver operating characteristic (AUROC = 0.946) in the internal validation, followed by models based on the psychological test notes, admission notes, initial nursing assessments, and structured data only (0.902, 0.855, 0.798, and 0.784, respectively). The external validation was performed using only the initial nursing assessment note, and the AUROC was 0.616.

CONCLUSIONS

We developed prediction models for psychosis relapse using the NLP-enrichment method. Models using clinical notes were more effective than models using only structured data, suggesting the importance of unstructured data in psychosis prediction.

Association between depressive symptoms and diagnosis of diabetes and its complications: A network analysis in electronic health records

OBJECTIVES

Diabetes and its complications are commonly associated with depressive symptoms, and few studies have investigated the diagnosis effect of depressive symptoms in patients with diabetes. The present study used a network-based approach to explore the association between depressive symptoms, which are annotated from electronic health record (EHR) notes by a deep learning model, and the diagnosis of type 2 diabetes mellitus (T2DM) and its complications.

METHODS

In this study, we used anonymous admission notes of 52,139 inpatients diagnosed with T2DM at the first affiliated hospital of Nanjing Medical University from 2008 to 2016 as input for a symptom annotation model named T5-depression based on transformer architecture which helps to annotate depressive symptoms from present illness. We measured the performance of the model by using the F1 score and the area under the receiver operating characteristic curve (AUROC). We constructed networks of depressive symptoms to examine the connectivity of these networks in patients diagnosed with T2DM, including those with certain complications.

RESULTS

The T5-depression model achieved the best performance with an F1-score of 91.71 and an AUROC of 96.25 compared with the benchmark models. The connectivity of depressive symptoms in patients diagnosed with T2DM (p = 0.025) and hypertension (p = 0.013) showed a statistically significant increase 2 years after the diagnosis, which is consistent with the number of patients diagnosed with depression.

CONCLUSION

The T5-depression model proposed in this study can effectively annotate depressive symptoms in EHR notes. The connectivity of annotated depressive symptoms is associated with the diagnosis of T2DM and hypertension. The changes in the network of depressive symptoms generated by the T5-depression model could be used as an indicator for screening depression.

Natural language processing in clinical neuroscience and psychiatry: A review

Natural language processing (NLP) is rapidly becoming an important topic in the medical community. The ability to automatically analyze any type of medical document could be the key factor to fully exploit the data it contains. Cutting-edge artificial intelligence (AI) architectures, particularly machine learning and deep learning, have begun to be applied to this topic and have yielded promising results. We conducted a literature search for 1,024 papers that used NLP technology in neuroscience and psychiatry from 2010 to early 2022. After a selection process, 115 papers were evaluated. Each publication was classified into one of three categories: information extraction, classification, and data inference. Automated understanding of clinical reports in electronic health records has the potential to improve healthcare delivery. Overall, the performance of NLP applications is high, with an average F1-score and AUC above 85%. We also derived a composite measure in the form of Z-scores to better compare the performance of NLP models and their different classes as a whole. No statistical differences were found in the unbiased comparison. Strong asymmetry between English and non-English models, difficulty in obtaining high-quality annotated data, and train biases causing low generalizability are the main limitations. This review suggests that NLP could be an effective tool to help clinicians gain insights from medical reports, clinical research forms, and more, making NLP an effective tool to improve the quality of healthcare services.

Timing of antipsychotics and benzodiazepine initiation during a first episode of psychosis impacts clinical outcomes: Electronic health record cohort study

The impact of timing of antipsychotics and benzodiazepine treatment during a first episode of psychosis on clinical outcomes is unknown. We present a RECORD-compliant electronic health record cohort study including patients (n = 4,483, aged 14-35) with a primary diagnosis of any non-organic ICD-10 first episode of psychosis at SLAM-NHS between 2007 and 2017. The impact of antipsychotic timing (prescription > 1 week after a first episode of psychosis) was assessed on the primary outcome (risk of any psychiatric inpatient admission over 6 years), and secondary outcomes (cumulative duration of any psychiatric/medical/accident/emergency [A&E] admission over 6 years). The impact of prescribing benzodiazepine before antipsychotic at any point and of treatment patterns (antipsychotic alone, benzodiazepine alone, combination of antipsychotic with benzodiazepine) within the first week after a first episode of psychosis were also assessed. Survival analyses and zero-inflated negative binomial regressions, adjusted for core covariates, and complementary analyses were employed. Antipsychotic prescribed >1 week after a first episode of psychosis did not affect the risk of any psychiatric admission (HR = 1.04, 95% CI = 0.92-1.17, p = 0.557), but increased the duration of any psychiatric (22-28%), medical (78-35%) and A&E (30-34%) admission (months 12-72). Prescribing benzodiazepine before antipsychotic at any point did not affect the risk of any psychiatric admission (HR = 1.03, 95% CI = 0.94-1.13, p = 0.535), but reduced the duration of any psychiatric admission (17-24%, months 12-72), and increased the duration of medical (71-45%, months 12-72) and A&E (26-18%, months 12-36) admission. Prescribing antipsychotic combined with benzodiazepine within the first week after a first episode of psychosis showed better overall clinical outcomes than antipsychotic or benzodiazepine alone. Overall, delaying antipsychotic 1 week after a first episode of psychosis may worsen some clinical outcomes. Early benzodiazepine treatment can be considered with concomitant antipsychotic but not as standalone intervention.

DASentimental: Detecting Depression, Anxiety, and Stress in Texts via Emotional Recall, Cognitive Networks, and Machine Learning

Most current affect scales and sentiment analysis on written text focus on quantifying valence/sentiment, the primary dimension of emotion. Distinguishing broader, more complex negative emotions of similar valence is key to evaluating mental health. We propose a semi-supervised machine learning model, DASentimental, to extract depression, anxiety, and stress from written text. We trained DASentimental to identify how N = 200 sequences of recalled emotional words correlate with recallers’ depression, anxiety, and stress from the Depression Anxiety Stress Scale (DASS-21). Using cognitive network science, we modeled every recall list as a bag-of-words (BOW) vector and as a walk over a network representation of semantic memory—in this case, free associations. This weights BOW entries according to their centrality (degree) in semantic memory and informs recalls using semantic network distances, thus embedding recalls in a cognitive representation. This embedding translated into state-of-the-art, cross-validated predictions for depression (R = 0.7), anxiety (R = 0.44), and stress (R = 0.52), equivalent to previous results employing additional human data. Powered by a multilayer perceptron neural network, DASentimental opens the door to probing the semantic organizations of emotional distress. We found that semantic distances between recalls (i.e., walk coverage), was key for estimating depression levels but redundant for anxiety and stress levels. Semantic distances from “fear” boosted anxiety predictions but were redundant when the “sad–happy” dyad was considered. We applied DASentimental to a clinical dataset of 142 suicide notes and found that the predicted depression and anxiety levels (high/low) corresponded to differences in valence and arousal as expected from a circumplex model of affect. We discuss key directions for future research enabled by artificial intelligence detecting stress, anxiety, and depression in texts.

Translating RDoC to Real-World Impact in Developmental Psychopathology: A Neurodevelopmental Framework for Application of Mental Health Risk Calculators

The National Institute of Mental Health Research Domain Criteria's (RDoC) has prompted a paradigm shift from categorical psychiatric disorders to considering multiple levels of vulnerability for probabilistic risk of disorder. However, the lack of neurodevelopmentally-based tools for clinical decision-making has limited RDoC's real-world impact. Integration with developmental psychopathology principles and statistical methods actualize the clinical implementation of RDoC to inform neurodevelopmental risk. In this conceptual paper, we introduce the probabilistic mental health risk calculator as an innovation for such translation and lay out a research agenda for generating an RDoC- and developmentally-informed paradigm that could be applied to predict a range of developmental psychopathologies from early childhood to young adulthood. We discuss methods that weigh the incremental utility for prediction based on intensity and burden of assessment, the addition of developmental change patterns, considerations for assessing outcomes, and integrative data approaches. Throughout, we illustrate the risk calculator approach with different neurodevelopmental pathways and phenotypes. Finally, we discuss real-world implementation of these methods for improving early identification and prevention of developmental psychopathology. We propose that mental health risk calculators can build a needed bridge between RDoC's multiple units of analysis and developmental science.

Developing and Validating an Individualized Clinical Prediction Model to Forecast Psychotic Recurrence in Acute and Transient Psychotic Disorders: Electronic Health Record Cohort Study

Acute and transient psychotic disorders (ATPDs) include short-lived psychotic episodes with a high probability of developing psychotic recurrences. Clinical care for ATPD is currently limited by the inability to predict outcomes. Real-world electronic health record (EHR)-based retrospective cohort study STROBE/RECORD compliant included all individuals accessing the South London and Maudsley NHS Trust between 2006 and 2017 and receiving a first diagnosis of ATPD (F23, ICD-10). After imputing missing data, stepwise and LASSO Cox regression methods employing a priori predictors (n = 23) were compared to develop and internally validate an individualized risk prediction model to forecast the risk of psychotic recurrences following TRIPOD guidelines. The primary outcome was prognostic accuracy (area under the curve [AUC]). 3018 ATPD individuals were included (average age = 33.75 years, 52.7% females). Over follow-up (average 1042 ± 1011 days, up to 8 years) there were 1160 psychotic recurrences (events). Stepwise (n = 12 predictors) and LASSO (n = 17 predictors) regression methods yielded comparable prognostic accuracy, with an events per variable ratio >100 for both models. Both models showed an internally validated adequate prognostic accuracy from 4 years follow-up (AUC 0.70 for both models) and good calibration. A refined model was adapted in view of the new ICD-11 criteria on 307 subjects with polymorphic ATPD, showing fair prognostic accuracy at 4 years (AUC: stepwise 0.68; LASSO 0.70). This study presents the first clinically based prediction model internally validated to adequately predict long-term psychotic recurrence in individuals with ATPD. The model can be automatable in EHRs, supporting further external validations and refinements to improve its prognostic accuracy.

Preventive psychiatry: a blueprint for improving the mental health of young people

Preventive approaches have latterly gained traction for improving mental health in young people. In this paper, we first appraise the conceptual foundations of preventive psychiatry, encompassing the public health, Gordon's, US Institute of Medicine, World Health Organization, and good mental health frameworks, and neurodevelopmentally-sensitive clinical staging models. We then review the evidence supporting primary prevention of psychotic, bipolar and common mental disorders and promotion of good mental health as potential transformative strategies to reduce the incidence of these disorders in young people. Within indicated approaches, the clinical high-risk for psychosis paradigm has received the most empirical validation, while clinical high-risk states for bipolar and common mental disorders are increasingly becoming a focus of attention. Selective approaches have mostly targeted familial vulnerability and non-genetic risk exposures. Selective screening and psychological/psychoeducational interventions in vulnerable subgroups may improve anxiety/depressive symptoms, but their efficacy in reducing the incidence of psychotic/bipolar/common mental disorders is unproven. Selective physical exercise may reduce the incidence of anxiety disorders. Universal psychological/psychoeducational interventions may improve anxiety symptoms but not prevent depressive/anxiety disorders, while universal physical exercise may reduce the incidence of anxiety disorders. Universal public health approaches targeting school climate or social determinants (demographic, economic, neighbourhood, environmental, social/cultural) of mental disorders hold the greatest potential for reducing the risk profile of the population as a whole. The approach to promotion of good mental health is currently fragmented. We leverage the knowledge gained from the review to develop a blueprint for future research and practice of preventive psychiatry in young people: integrating universal and targeted frameworks; advancing multivariable, transdiagnostic, multi-endpoint epidemiological knowledge; synergically preventing common and infrequent mental disorders; preventing physical and mental health burden together; implementing stratified/personalized prognosis; establishing evidence-based preventive interventions; developing an ethical framework, improving prevention through education/training; consolidating the cost-effectiveness of preventive psychiatry; and decreasing inequalities. These goals can only be achieved through an urgent individual, societal, and global level response, which promotes a vigorous collaboration across scientific, health care, societal and governmental sectors for implementing preventive psychiatry, as much is at stake for young people with or at risk for emerging mental disorders.

Gender differences in clinical presentation and illicit substance use during first episode psychosis: a natural language processing, electronic case register study

OBJECTIVE

To determine whether gender differences in symptom presentation at first episode psychosis (FEP) remain even when controlling for substance use, age and ethnicity, using natural language processing applied to electronic health records (EHRs).

DESIGN, SETTING AND PARTICIPANTS

Data were extracted from EHRs of 3350 people (62% male patients) who had presented to the South London and Maudsley NHS Trust with a FEP between 1 April 2007 and 31 March 2017. Logistic regression was used to examine gender differences in the presentation of positive, negative, depressive, mania and disorganisation symptoms.

EXPOSURES FOR OBSERVATIONAL STUDIES

Gender (male vs female).

MAIN OUTCOMES AND MEASURES

Presence of positive, negative, depressive, mania and disorganisation symptoms at initial clinical presentation.

RESULTS

Eight symptoms were significantly more prevalent in men (poverty of thought, negative symptoms, social withdrawal, poverty of speech, aggression, grandiosity, paranoia and agitation). Conversely, tearfulness, low energy, reduced appetite, low mood, pressured speech, mood instability, flight of ideas, guilt, mutism, insomnia, poor concentration, tangentiality and elation were more prevalent in women than men. Negative symptoms were more common among men (OR 1.85, 95% CI 1.33 to 2.62) and depressive and manic symptoms more common among women (OR 0.30, 95% CI 0.26 to 0.35). After adjustment for illicit substance use, the strength of associations between gender and negative, manic and depression symptoms increased, whereas gender differences in aggression, agitation, paranoia and grandiosity became insignificant.

CONCLUSIONS

There are clear gender differences in the clinical presentation of FEP. Our findings suggest that gender can have a substantial influence on the nature of clinical presentation in people with psychosis, and that this is only partly explained by exposure to illicit substance use.

"When they say weed causes depression, but it's your fav antidepressant": Knowledge-aware attention framework for relationship extraction

With the increasing legalization of medical and recreational use of cannabis, more research is needed to understand the association between depression and consumer behavior related to cannabis consumption. Big social media data has potential to provide deeper insights about these associations to public health analysts. In this interdisciplinary study, we demonstrate the value of incorporating domain-specific knowledge in the learning process to identify the relationships between cannabis use and depression. We develop an end-to-end knowledge infused deep learning framework (Gated-K-BERT) that leverages the pre-trained BERT language representation model and domain-specific declarative knowledge source (Drug Abuse Ontology) to jointly extract entities and their relationship using gated fusion sharing mechanism. Our model is further tailored to provide more focus to the entities mention in the sentence through entity-position aware attention layer, where ontology is used to locate the target entities position. Experimental results show that inclusion of the knowledge-aware attentive representation in association with BERT can extract the cannabis-depression relationship with better coverage in comparison to the state-of-the-art relation extractor.

Transdiagnostic individualized clinically-based risk calculator for the automatic detection of individuals at-risk and the prediction of psychosis: external replication in 2,430,333 US patients

The real-world impact of psychosis prevention is reliant on effective strategies for identifying individuals at risk. A transdiagnostic, individualized, clinically-based risk calculator to improve this has been developed and externally validated twice in two different UK healthcare trusts with convincing results. The prognostic performance of this risk calculator outside the UK is unknown. All individuals who accessed primary or secondary health care services belonging to the IBM® MarketScan® Commercial Database between January 2015 and December 2017, and received a first ICD-10 index diagnosis of nonorganic/nonpsychotic mental disorder, were included. According to the risk calculator, age, gender, ethnicity, age-by-gender, and ICD-10 cluster diagnosis at index date were used to predict development of any ICD-10 nonorganic psychotic disorder. Because patient-level ethnicity data were not available city-level ethnicity proportions were used as proxy. The study included 2,430,333 patients with a mean follow-up of 15.36 months and cumulative incidence of psychosis at two years of 1.43%. There were profound differences compared to the original development UK database in terms of case-mix, psychosis incidence, distribution of baseline predictors (ICD-10 cluster diagnoses), availability of patient-level ethnicity data, follow-up time and availability of specialized clinical services for at-risk individuals. Despite these important differences, the model retained accuracy significantly above chance (Harrell's C = 0.676, 95% CI: 0.672-0.679). To date, this is the largest international external replication of an individualized prognostic model in the field of psychiatry. This risk calculator is transportable on an international scale to improve the automatic detection of individuals at risk of psychosis.