Use this URL to cite or link to this record in EThOS:
Title: Investigating clinical outcomes in psychotic disorders using an electronic case register
Author: Patel, Rashmi
Awarding Body: King's College London
Current Institution: King's College London (University of London)
Date of Award: 2016
Availability of Full Text:
Access from EThOS:
Access from Institution:
Background: Psychotic disorders have a lifetime prevalence of around 3% and cost the UK around £10 billion per year. One of the key challenges faced by clinicians in managing these disorders is that it is not possible to predict clinical outcomes or the course of illness. To date, research that has investigated clinical outcomes in psychosis has generally involved patient samples that are relatively modest in size and may be unrepresentative of the patients seen in everyday clinical practice. In many centres, routine clinical information is now recorded in the form of electronic health records (EHRs) rather than in paper case notes. While some of these data comprise responses to specific questions, at present most of the clinical data is still recorded in the form of unstructured free text entries. The large volume of free text means that it is not feasible to manually read through records to identify data of interest in a large sample of patients. However, automated information extraction methods such as natural language processing (NLP) offer the opportunity to quickly extract large volumes of meaningful data from free text EHRs and perform observational research studies in much larger samples than would be possible through direct participant recruitment. Aims of this thesis: To extract and analyse clinical data using NLP from a large electronic case register of patients with psychotic disorders to test the following hypotheses: (i) In people with first episode psychosis (FEP), those who initially present to a service for people at high risk of psychosis will have better outcomes than those who present to conventional services. (ii) Concurrent cannabis use in people with FEP is associated with poor clinical outcomes which are partly mediated by it reducing the effectiveness of treatment with antipsychotic medication (iii) Negative symptoms are common in people with schizophrenia and are associated with worse clinical outcomes. Methods (the same overall approach was used to test all hypotheses): Dataset: South London and Maudsley NHS Trust (SLaM) Biomedical Research Centre (BRC) Case Register. The dataset comprises anonymised EHRs of over 250,000 people who have received mental healthcare from SLaM. NLP development: The software package TextHunter was used. All sentences containing keywords relevant to the constructs investigated were extracted and used to develop NLP applications using a support vector machine learning (SVM) approach. Outcomes: number of days spent in hospital and frequency of hospital admission. Covariates: age, gender, ethnicity, marital status and diagnosis. Statistical analysis: multivariable logistic, negative binomial, linear regression and mediation analysis using STATA. Results: FEP patients who initially presented to high‐risk services (n=2,943): Presentation to a high‐risk service was associated with 17 fewer days spent in hospital (95% CI ‐33.7, ‐0.3) and a lower frequency of admission (incidence rate ratio: 0.49, 0.39‐0.61) in the 24 months following referral, as compared to patients who presented to conventional services. Cannabis in FEP (n=2,026): Concurrent cannabis use was associated with increased frequency of hospital admission (incidence rate ratio 1.50, 1.25‐1.80) and a greater number of days spent in hospital (B coefficient 35.1 days, 12.1‐58.1). An increase in the number of unique antipsychotics prescribed to cannabis users mediated both an increased frequency of hospital admission (natural indirect effect: 1.09, 1.01‐1.18; total effect: 1.50, 1.21‐1.87) and a greater number of days spent in hospital (NIE: 17.9, 2.4‐33.4; TE: 34.8, 11.6‐58.1). Negative symptoms in chronic schizophrenia (n=7,678): 55.7% of people with schizophrenia had at least one negative symptom documented. Among patients with schizophrenia, negative symptoms were associated with increased likelihood of hospital admission (odds ratio 1.24, 95% CI 1.10‐1.39), re‐admission (1.58, 1.28‐1.95) and length of stay (B coefficient 20.5, 7.6‐ 33.5). Conclusions: EHR data can be used to investigate associations between variables assessed during routine care and clinical outcomes in patient samples that are much larger than can be recruited to conventional research studies. Moreover, the specific findings obtained using this approach have a number of implications for healthcare service delivery. First, the finding that engaging first episode patients in the prodromal phase is associated with better outcomes indicates that contact at this stage may not only reduce the risk of developing psychosis, but also improve outcomes in those at high risk who subsequently become psychotic. Secondly, the finding that both cannabis use and negative symptoms in patients with psychosis are independently linked to significantly poorer clinical outcomes highlights the need for the development of effective treatments to reduce cannabis use and ameliorate negative symptoms.
Supervisor: McGuire, Philip ; Stewart, Robert James Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available