Explore the vast range of titles available.

Research into suicide prevention has been hampered by methodological limitations such as low sample size and recall bias. Recently, Natural Language Processing (NLP) strategies have been used with Electronic Health Records to increase information extraction from free text notes as well as structured fields concerning suicidality and this allows access to much larger cohorts than previously possible. This paper presents two novel NLP approaches – a rule-based approach to classify the presence of suicide ideation and a hybrid machine learning and rule-based approach to identify suicide attempts in a psychiatric clinical database. Good performance of the two classifiers in the evaluation study suggest they can be used to accurately detect mentions of suicide ideation and attempt within free-text documents in this psychiatric database. The novelty of the two approaches lies in the malleability of each classifier if a need to refine performance, or meet alternate classification requirements arises. The algorithms can also be adapted to fit infrastructures of other clinical datasets given sufficient clinical recording practice knowledge, without dependency on medical codes or additional data extraction of known risk factors to predict suicidal behaviour.

ObjectivesTo investigate factors associated with the recording of glycated haemoglobin (HbA1c) in people with first diagnoses of serious mental illness (SMI) in a large mental healthcare provider, and factors associated with HbA1c levels, when recorded. To our knowledge this is the first such investigation, although attention to dysglycaemia in SMI is an increasing priority in mental healthcare.DesignThe study was primarily descriptive in nature, seeking to ascertain the frequency of HbA1c recording in the mental healthcare sector for people following first SMI diagnosis.SettingsA large mental healthcare provider, the South London and Maudsley National Health Service Trust.ParticipantsUsing electronic mental health records data, we ascertained patients with first SMI diagnoses (schizophrenia, schizoaffective disorder, bipolar disorder) from 2008 to 2018.Outcome measuresRecording or not of HbA1c level was ascertained from routine local laboratory data and supplemented by a natural language processing (NLP) algorithm for extracting recorded values in text fields (precision 0.89%, recall 0.93%). Age, gender, ethnic group, year of diagnosis, and SMI diagnosis were investigated as covariates in relation to recording or not of HbA1c and first recorded levels.ResultsOf 21 462 patients in the sample (6546 bipolar disorder; 14 916 schizophrenia or schizoaffective disorder; mean age 38.8 years, 49% female), 4106 (19.1%) had at least one HbA1c result recorded from laboratory data, increasing to 6901 (32.2%) following NLP. HbA1c recording was independently more likely in non-white ethnic groups (black compared with white: OR 2.45, 95% CI 2.29 to 2.62), and was negatively associated with age (OR per year increase 0.93, 0.92–0.95), female gender (0.83, 0.78–0.88) and bipolar disorder (0.49, 0.45–0.52).ConclusionsOver a 10-year period, relatively low level of recording of HbA1c was observed, although this has increased over time and ascertainment was increased with text extraction. It remains important to improve the routine monitoring of dysglycaemia in these at-risk disorders.

ObjectivesTo examine whether depressive symptoms predict receipt of cognitive–behavioural therapy for psychosis (CBTp) in individuals with psychosis.DesignRetrospective cross-sectional analysis of electronic health records (EHRs) of a clinical cohort.SettingA secondary National Health Service mental healthcare service serving four boroughs of south London, UK.Participants20 078 patients diagnosed with an International Classification of Diseases, version 10 (ICD-10) code between F20 and 29 extracted from an EHR database.Primary and secondary outcome measuresPrimary: Whether recorded depressive symptoms predicted CBTp session receipt, defined as at least one session of CBTp identified from structured EHR fields supplemented by a natural language processing algorithm. Secondary: Whether age, gender, ethnicity, symptom profiles (positive, negative, manic and disorganisation symptoms), a comorbid diagnosis of depression, anxiety or bipolar disorder, general CBT receipt prior to the primary psychosis diagnosis date or type of psychosis diagnosis predicted CBTp receipt.ResultsOf patients with a psychotic disorder, only 8.2% received CBTp. Individuals with at least one depressive symptom recorded, depression symptom severity and 12 out of 15 of the individual depressive symptoms independently predicted CBTp receipt. Female gender, White ethnicity and presence of a comorbid affective disorder or primary schizoaffective diagnosis were independently positively associated with CBTp receipt within the whole sample and the top 25% of mentioned depressive symptoms.ConclusionsIndividuals with a psychotic disorder who had recorded depressive symptoms were significantly more likely to receive CBTp sessions, aligning with CBTp guidelines of managing depressive symptoms related to a psychotic experience. However, overall receipt of CBTp is low and more common in certain demographic groups, and needs to be increased.

Obsessive and Compulsive Symptoms (OCS) or Obsessive Compulsive Disorder (OCD) in the context of schizophrenia or related disorders are of clinical importance as these are associated with a range of adverse outcomes. Natural Language Processing (NLP) applied to Electronic Health Records (EHRs) presents an opportunity to create large datasets to facilitate research in this area. This is a challenging endeavour however, because of the wide range of ways in which these symptoms are recorded, and the overlap of terms used to describe OCS with those used to describe other conditions. We developed an NLP algorithm to extract OCS information from a large mental healthcare EHR data resource at the South London and Maudsley NHS Foundation Trust using its Clinical Record Interactive Search (CRIS) facility. We extracted documents from individuals who had received a diagnosis of schizophrenia, schizoaffective disorder, or bipolar disorder. These text documents, annotated by human coders, were used for developing and refining the NLP algorithm (600 documents) with an additional set reserved for final validation (300 documents). The developed NLP algorithm utilized a rules-based approach to identify each of symptoms associated with OCS, and then combined them to determine the overall number of instances of OCS. After its implementation, the algorithm was shown to identify OCS with a precision and recall (with 95% confidence intervals) of 0.77 (0.65–0.86) and 0.67 (0.55–0.77) respectively. The development of this application demonstrated the potential to extract complex symptomatic data from mental healthcare EHRs using NLP to facilitate further analyses of these clinical symptoms and their relevance for prognosis and intervention response.

ObjectivesWe set out to develop, evaluate and implement a novel application using natural language processing to text mine occupations from the free-text of psychiatric clinical notes.DesignDevelopment and validation of a natural language processing application using General Architecture for Text Engineering software to extract occupations from de-identified clinical records.Setting and participantsElectronic health records from a large secondary mental healthcare provider in south London, accessed through the Clinical Record Interactive Search platform. The text mining application was run over the free-text fields in the electronic health records of 341 720 patients (all aged ≥16 years).OutcomesPrecision and recall estimates of the application performance; occupation retrieval using the application compared with structured fields; most common patient occupations; and analysis of key sociodemographic and clinical indicators for occupation recording.ResultsUsing the structured fields alone, only 14% of patients had occupation recorded. By implementing the text mining application in addition to the structured fields, occupations were identified in 57% of patients. The application performed on gold-standard human-annotated clinical text at a precision level of 0.79 and recall level of 0.77. The most common patient occupations recorded were ‘student’ and ‘unemployed’. Patients with more service contact were more likely to have an occupation recorded, as were patients of a male gender, older age and those living in areas of lower deprivation.ConclusionThis is the first time a natural language processing application has been used to successfully derive patient-level occupations from the free-text of electronic mental health records, performing with good levels of precision and recall, and applied at scale. This may be used to inform clinical studies relating to the broader social determinants of health using electronic health records.

Lesbian, gay, bisexual, transgender, queer and related community (LGBTQ+) individuals have significantly increased risk for mental health problems. However, research on inequalities in LGBTQ+ mental healthcare is limited because LGBTQ+ status is usually only contained in unstructured, free-text sections of electronic health records. This study investigated whether natural language processing (NLP), specifically the large language model, Bi-directional Encoder Representations from Transformers (BERT), can identify LGBTQ+ status from this unstructured text in mental health records. Using electronic health records from a large mental healthcare provider in south London, UK, relevant search terms were identified and a random sample of 10 000 strings extracted. Each string contained 100 characters either side of a search term. A BERT model was trained to classify LGBTQ+ status. Among 10 000 annotations, 14% (1449) confirmed LGBTQ+ status while 86% (8551) did not. These other categories included LGBTQ+ negative status, irrelevant annotations and unclear cases. The final BERT model, tested on 2000 annotations, achieved a precision of 0.95 (95% CI 0.93-0.98), a recall of 0.93 (95% CI 0.91-0.96) and an F1 score of 0.94 (95% CI 0.92-0.97). LGBTQ+ status can be determined using this NLP application with a high success rate. The NLP application produced through this work has opened up mental health records to a variety of research questions involving LGBTQ+ status, and should be explored further. Additional work should aim to extend what has been done here by developing an application that can distinguish between different LGBTQ+ groups to examine inequalities between these groups.

We describe the incidence of suicidality (2007–2017) in people with depression treated by secondary mental healthcare services at South London and Maudsley NHS Trust ( n = 26 412). We estimated yearly incidence of ‘suicidal ideation’ and ‘high risk of suicide’ from structured and free-text fields of the Clinical Record Interactive Search system. The incidence of suicidal ideation increased from 0.6 (2007) to 1 cases (2017) per 1000 population. The incidence of high risk of suicide, based on risk forms, varied between 0.06 and 0.50 cases per 1000 adult population (2008–2017). Electronic health records provide the opportunity to examine suicidality on a large scale, but the impact of service-related changes in the use of structured risk assessment should be considered.

Activation of the innate immune system is commonly associated with depression. Immunomodulatory drugs may have efficacy for depressive symptoms that are co-morbidly associated with inflammatory disorders. We report a large-scale re-analysis by standardized procedures (mega-analysis) of patient-level data combined from 18 randomized clinical trials conducted by Janssen or GlaxoSmithKline for one of nine disorders (N = 10,743 participants). Core depressive symptoms (low mood, anhedonia) were measured by the Short Form Survey (SF-36) or the Hospital Anxiety and Depression Scale (HADS), and participants were stratified into high (N = 1921) versus low-depressive strata based on baseline ratings. Placebo-controlled change from baseline after 4–16 weeks of treatment was estimated by the standardized mean difference (SMD) over all trials and for each subgroup of trials targeting one of 7 mechanisms (IL-6, TNF-α, IL-12/23, CD20, COX2, BLγS, p38/MAPK14). Patients in the high depressive stratum showed modest but significant effects on core depressive symptoms (SMD = 0.29, 95% CI [0.12–0.45]) and related SF-36 measures of mental health and vitality. Anti-IL-6 antibodies (SMD = 0.8, 95% CI [0.20–1.41]) and an anti-IL-12/23 antibody (SMD = 0.48, 95% CI [0.26–0.70]) had larger effects on depressive symptoms than other drug classes. Adjustments for physical health outcome marginally attenuated the average treatment effect on depressive symptoms (SMD = 0.20, 95% CI: 0.06–0.35), but more strongly attenuated effects on mental health and vitality. Effects of anti-IL-12/23 remained significant and anti-IL-6 antibodies became a trend after controlling for physical response to treatment. Novel immune-therapeutics can produce antidepressant effects in depressed patients with primary inflammatory disorders that are not entirely explained by treatment-related changes in physical health.

Methylation of cytosine is a DNA modification associated with gene repression. Recently, a novel cytosine modification, 5-hydroxymethylcytosine （5-hmC） has been discovered. Here we examine 5-hmC distribution during mammalian development and in cellular systems, and show that the developmental dynamics of 5-hmC are different from those of 5-methylcytosine （5-mC）; in particular 5-hmC is enriched in embryonic contexts compared to adult tissues. A detectable 5-hmC signal appears in pre-implantation development starting at the zygote stage, where the paternal genome is subjected to a genome-wide hydroxylation of 5-mC, which precisely coincides with the loss of the 5-mC signal in the paternal pronucleus. Levels of 5-hmC are high in cells of the inner cell mass in blastocysts, and the modification colocalises with nestin-expressing cell populations in mouse post-implantation embryos. Compared to other adult mammalian organs, 5-hmC is strongly enriched in bone marrow and brain, wherein high 5-hmC content is a feature of both neuronal progenitors and post-mitotic neurons. We show that high levels of 5-hmC are not only present in mouse and human embryonic stem cells （ESCs） and lost during differentiation, as has been reported previously, but also reappear during the generation of induced pluripotent stem cells; thus 5-hmC enrichment cor- relates with a pluripotent cell state. Our findings suggest that apart from the cells of neuronal lineages, high levels of genomic 5-hmC are an epigenetic feature of embryonic cell populations and cellular pluri- and multi-lineage potency. To our knowledge, 5-hmC represents the first epigenetie modification of DNA discovered whose enrichment is so cell- type specific.

We aimed to investigate whether daily fluctuations in mental health-relevant Twitter posts are associated with daily fluctuations in mental health crisis episodes. We conducted a primary and replicated time-series analysis of retrospectively collected data from Twitter and two London mental healthcare providers. Daily numbers of ‘crisis episodes’ were defined as incident inpatient, home treatment team and crisis house referrals between 2010 and 2014. Higher volumes of depression and schizophrenia tweets were associated with higher numbers of same-day crisis episodes for both sites. After adjusting for temporal trends, seven-day lagged analyses showed significant positive associations on day 1, changing to negative associations by day 4 and reverting to positive associations by day 7. There was a 15% increase in crisis episodes on days with above-median schizophrenia-related Twitter posts. A temporal association was thus found between Twitter-wide mental health-related social media content and crisis episodes in mental healthcare replicated across two services. Seven-day associations are consistent with both precipitating and longer-term risk associations. Sizes of effects were large enough to have potential local and national relevance and further research is needed to evaluate how services might better anticipate times of higher risk and identify the most vulnerable groups.