Search

Objectives We aimed to apply natural language processing algorithms in routine healthcare records to identify reported somatic passivity (external control of sensations, actions and impulses) and thought interference symptoms (thought broadcasting, insertion, withdrawal), first-rank symptoms traditionally central to diagnosing schizophrenia, and determine associations with prognosis by analysing routine outcomes.Design Four algorithms were developed on deidentified mental healthcare data and applied to ascertain recorded symptoms over the 3 months following first presentation to a mental healthcare provider in a cohort of patients with a primary schizophreniform disorder (ICD-10 F20-F29) diagnosis.Setting and participants From the electronic health records of a large secondary mental healthcare provider in south London, 9323 patients were ascertained from 2007 to the data extraction date (25 February 2020).Outcomes The primary binary dependent variable for logistic regression analyses was any negative outcome (Mental Health Act section, >2 antipsychotics prescribed, >22 days spent in crisis care) over the subsequent 2 years.Results Final adjusted models indicated significant associations of this composite outcome with baseline somatic passivity (prevalence 4.9%; adjusted OR 1.61, 95% CI 1.37 to 1.88), thought insertion (10.7%; 1.24, 95% CI 1.15 to 1.55) and thought withdrawal (4.9%; 1.36, 95% CI 1.10 to 1.69), but not independently with thought broadcast (10.3%; 1.05, 95% CI 0.91 to 1.22).Conclusions Symptoms traditionally central to the diagnosis of schizophrenia, but under-represented in current diagnostic frameworks, were thus identified as important predictors of short-term to medium-term prognosis in schizophreniform disorders.

Introduction Of the many studies describing psychotic symptoms in schizophrenia, few have investigated their direct influence on prognosis. Objectives We aimed to apply natural language-processing (NLP) algorithms in routine healthcare records to identify reported somatic passivity and thought interference symptoms (thought broadcasting, insertion and withdrawal), and determine associations with prognosis by an analysis of routine outcomes. Methods Four algorithms were thus developed on de-identified mental healthcare data from a large south London provider and were applied to ascertain recorded symptoms over the three months following first presentation to that service in a cohort of patients with a primary schizophreniform disorder (ICD-10 F20-F29) diagnosis. The primary binary dependent variable for logistic regression analyses was any negative outcome (Mental Health Act section, >2 antipsychotics prescribed, >22 days spent in crisis care) over the subsequent 2 years, adjusted for age, gender, ethnic group, neighbourhood deprivation, diagnostic group, and recorded paranoia, persecutory delusions or auditory hallucinations. Results In 9,323 patients, final models indicated significant associations of this composite outcome with baseline somatic passivity (prevalence 4.9%; adjusted odds ratio 1.61, 95% CI 1.37-1.88), thought insertion (10.7%; 1.24, 1.15-1.55) and thought withdrawal (4.9%; 1.36, 1.10-1.69), but not independently with thought broadcast (10.3%; 1.05, 0.91-1.22). Conclusions Symptoms traditionally central to the diagnosis of schizophrenia, but under-represented in current diagnostic frameworks, were thus identified as important predictors of short- to medium-term prognosis. Disclosure No conflict of interest - past support from Janssen; GSK; Takeda; Induction Healthcare; Holmusk; the NIHR; SLaM NHS Trust; the MRC; KCL; the NIHR ARC; KCH NHS Trust; the Academy of Medical Sciences; The Wellcome Trust; BHF; Arthritis Research UK; the Roya

Objectives We 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.Design Development 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 participants Electronic 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).Outcomes Precision 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.Results Using 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.Conclusion This 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.

Abstract Background There is mounting evidence that people with severe mental illness have unhealthy lifestyles, high rates of cardiovascular and metabolic diseases, and greater risk of early mortality. This study aimed to assess the cost-effectiveness of a health promotion intervention seeking to improve physical health and reduce substance use in people with psychosis. Methods Participants with a psychotic disorder, aged 18-65 years old and registered on an enhanced care approach programme or equivalent were recruited from community mental health teams in six mental health trusts in England. Participants were randomisation to either standard community mental health team care (treatment as usual) or treatment as usual with an integrated health promotion intervention (IMPaCT). Cost-effectiveness and cost-utility analyses from health and social care and societal perspectives were conducted alongside a cluster randomised controlled trial. Total health and social care costs and total societal costs at 12 and 15 months were calculated as well as cost-effectiveness (incremental cost-effectiveness ratios and cost-effectiveness acceptability curves) at 15 months based on quality of life (SF-36 mental and physical health components, primary outcome measures) and quality adjusted life years (QALYs) using two measures, EQ-5D-3 L and SF-36. Data were analysed using bootstrapped regressions with covariates for relevant baseline variables. Results At 12-15 months 301 participants had full data needed to be included in the economic evaluation. There were no differences in adjusted health and social care costs (£95, 95% CI -£1410 to £1599) or societal costs (£675, 95% CI -£1039 to £2388) between the intervention and control arms. Similarly, there were no differences between the groups in the SF-36 mental component (−0.80, 95% CI -3.66 to 2.06), SF-36 physical component (−0.68, 95% CI -3.01 to 1.65), QALYs estimated from the SF-36 (−0.00, −0.01 to 0.00) or QALYs estimated from the EQ-5D-3 L (0.00, 95% CI -0.01 to 0.02). Cost-effectiveness acceptability curves for all four outcomes and from both cost perspectives indicate that the probability of the health promotion intervention being cost-effective does not exceed 0.4 for willingness to pay thresholds ranging from £0-£50,000. Conclusions Alongside no evidence of additional quality of life/clinical benefit, there is also no evidence of cost-effectiveness. Trial registration ISRCTN58667926 . Date retrospectively registered: 23/04/2010. Recruitment start date: 01/03/2010.

Objectives To investigate recorded poor insight in relation to mental health and service use outcomes in a cohort with first-episode psychosis.Design We developed a natural language processing algorithm to ascertain statements of poor or diminished insight and tested this in a cohort of patients with first-episode psychosis.Setting The clinical record text at the South London and Maudsley National Health Service Trust in the UK was used.Participants We applied the algorithm to characterise a cohort of 2026 patients with first-episode psychosis attending an early intervention service.Primary and secondary outcome measures Recorded poor insight within 1 month of registration was investigated in relation to (1) incidence of psychiatric hospitalisation, (2) odds of legally enforced hospitalisation, (3) number of days spent as a mental health inpatient and (4) number of different antipsychotic agents prescribed; outcomes were measured over varying follow-up periods from 12 months to 60 months, adjusting for a range of sociodemographic and clinical covariates.Results Recorded poor insight, present in 46% of the sample, was positively associated with ages 16-35, bipolar disorder and history of cannabis use and negatively associated with White ethnicity and depression. It was significantly associated with higher levels of all four outcomes over all five follow-up periods.Conclusions Recorded poor insight in people with recent onset psychosis predicted subsequent legally enforced hospitalisations and higher number of hospital admissions, number of unique antipsychotics prescribed and days spent hospitalised. Improving insight might benefit patients’ course of illness as well as reduce mental health service use.

Aims and method To support safe prescribing of antipsychotics in dementia, antipsychotic monitoring forms were embedded into our electronic health records. We present a review of the data collected on these forms to assess prescribing and identify areas for improvement in our practice and processes. Data were extracted from the structured fields of antipsychotic initiation and review forms completed between 1 January 2018 and 31 January 2020. Results We identified gaps in practice where improvements could be made, mainly with regard to physical health monitoring (and particularly electrocardiograms, performed in only 50% of patients) and the low (less than 50%) recorded use of non-pharmacological interventions for behavioural and psychological symptoms of dementia. In addition, antipsychotic treatment was continued despite lack of benefit in almost 10% of reviews. Clinical implications We advocate for recommendations on physical health monitoring of people with dementia taking antipsychotics to be added to the National Institute for Health and Care Excellence guidance on dementia and the Prescribing Observatory for Mental Health (POMH-UK) national audit. [ABSTRACT FROM AUTHOR]

This paper focuses on two types of definites in Greek--MONADICS and POLYDEFINITES--and provides a constraint-based account of their form, meaning and use. Specifically, I discuss three core issues that have not been addressed in previous work. First, the special pragmatic import of polydefinites. These are associated with contextual constraints that go beyond the uniqueness entailments of standard (monadic) definites. Their idiosyncratic morphosyntax achieves effects similar to those induced in other languages solely by prosodic means and illustrated by phenomena subsumed within the term DEACCENTING. Second, the morphosyntax of definites. I argue that the Greek definite article can be best analysed as a PHRASAL AFFIX, and provide a composition approach in the spirit of previous work couched in HPSG. Monadics and polydefinites are treated uniformly, without positing unmotivated complexity in the grammar for deriving the form of the latter. The definite concord and linear order facts that pose problems for previous analyses are directly derived and the morphosyntactic affinity between the Greek definite article and 'weak form' possessive is straightforwardly captured. Third, the semantics of definites. A quantificational semantics is provided that ensures that the semantic content of the definite article in polydefinites is integrated into the meaning of the sentence just once. Polydefinites are, therefore, semantically identical to monadics; the special import of the former originates from a contextual constraint on the anchoring of the index that interacts with the common morphosyntactic and semantic basis.

This paper focuses on the Information Packaging notion of linkhood and provides a structural definition of this notion for Greek. We show that a combination of structural resources--syntactic (left dislocation), morphological (clitic duplication) and phonological (absence of nuclear accent)--are simultaneously exploited to realize linkhood in Greek, a generalization that can be captured in a constraint-based grammar such as HPSG, which permits the expression of interface constraints. We assume Vallduvi's (1992) approach to Information Packaging, and Engdahl & Vallduvi's (1996) implementation of the latter in HPSG, but deviate from Vallduvi's work in adopting Hendriks & Dekker's (1996) revised definition of linkhood that relies on non-monotone anaphora. From an empirical point of view, our approach wide scope readings, as well as a number of systematic differences in felicity conditions between Clitic Left Dislocation and other apparently related phenomena (Topicalization and Clitic Doubling). From a theoretical perspective, our analysis departs from syntax-based notions of topichood or discourse-linking and supports a definition that unifies linkhood with other anaphora phenomena. As such, it arguably overcomes previously noted problems for Vallduvi's treatment of links as the current-locus-of-update in a Heim-style file-card system.