King's College London

Research portal

Illicit Substance Use in First Episode Psychosis (FEP): A Natural Language Processing (NLP) Electronic Health Record Study

Research output: Contribution to journalMeeting abstract

Original languageEnglish
Pages (from-to)99
Number of pages1
JournalEarly Intervention in Psychiatry
Volume12
Issue numberS1
Publication statusPublished - 17 Sep 2018

King's Authors

Abstract

Background: Cannabis use is associated with increased risk of developing psychosis and worse clinical outcomes. However, less is known about misuse of other illicit substances in first episode psychosis (FEP). We sought to compare the associations of cannabis, cocaine, amphetamines and MDMA with clinical outcomes in people with FEP using a large electronic health record (EHR) dataset.

Methods: Data were obtained from pseudonymised EHRs of 1,835 people with FEP in South London. Exposure to cannabis, cocaine, MDMA or amphetamines was identified using TextHunter NLP software. Data on subsequent hospital admission in the 5 years following presentation were obtained. Their relationship with illicit substance misuse was analysed using multivariable negative binomial regression with age, gender and ethnicity as covariates.

Results: Cannabis was the most frequently documented illicit substance (58.5% of patients) followed by cocaine (24.4%), amphetamines (3.7%) and MDMA (3.1%). Cannabis (IRR 1.58, 95% CI 1.37 to 1.83) and cocaine (1.36, 1.17 to 1.59) were significantly associated with increased number of hospital admissions while MDMA was associated with a reduction in number of hospital admissions (0.59, 0.40 to 0.86) and amphetamines with no significant difference (1.05, 0.77 to 1.45).

Conclusion: Cannabis and cocaine use are associated with a significant increase in number of hospital admissions following presentation to mental health services with psychosis. These findings highlight the importance of illicit substances in determining clinical outcomes in psychosis and the need to address their use when planning treatment.

View graph of relations

© 2018 King's College London | Strand | London WC2R 2LS | England | United Kingdom | Tel +44 (0)20 7836 5454