A psychiatrist reviewing a patient’s electronic health record will soon be able to see, at a glance, whether that person has fared better with remote or in-person appointments. The rapid switch to remote mental healthcare during the pandemic happened without a solid evidence base. Clinicians lack clear guidance on which patients and which types of consultation work well remotely, and which do not. This project will mine roughly 5 million clinical documents from 80,000 NHS patients, using natural language processing to extract information from free-text notes. The researcher will then build data visualisation tools that summarise safety and effectiveness outcomes for remote versus in-person care, designed to be used directly by NHS clinicians. If successful, the tools could help mental health services decide when to offer remote appointments and when to insist on face-to-face contact, reducing the risk of poor outcomes. The visualisation tools will be made publicly available, and the methods will be published as an open-access protocol so other NHS trusts can replicate the approach. The project also includes a pilot implementation in an NHS mental health setting to test whether clinicians actually find the tools useful in practice.
View original technical description
Research question How can electronic health records (EHRs) be used to optimise clinical outcomes of remote mental healthcare? Background The COVID-19 pandemic has been associated with the rapid adoption of remote technologies to deliver mental healthcare. While restrictions resulting from the pandemic are likely to ease as it is controlled, remote technology may remain a very useful adjunct to conventional in-person consultation, by reducing geographical barriers to the delivery of care by clinicians, and access to care by patients. However, the potential benefits of remote care must be balanced against the risk that its use might, in some situations, have adverse effects on quality of care, patient satisfaction, and clinical outcomes. Because the recent adoption of remote technology has been rapid and unplanned, the evidence base to support best practice is still sparse. EHR data could be used to address this by facilitating the analysis of factors associated with good remote care outcomes. Natural language processing (NLP) tools, which enable the automated extraction of clinical information from free text EHR documents, could be applied to establish which features of mental healthcare may be best suited to remote consultation. I will apply NLP to EHR data and develop data visualisation tools to support more effective implementation of remote mental healthcare technology. Aims and objectives EHR/NLP data study: to compare the safety and effectiveness of remote vs in-person assessments from a large sample of NHS patients with mental health disorders. Data visualisation: to summarise the clinical outcomes of patients receiving remote and in-person assessments in a format readily accessible to NHS clinicians. Implementation pilot: to evaluate the utility of the data visualisation tools developed in (ii) as part of the safe and effective use of remote technology in an NHS mental health setting. Methods Source of clinical data: De-identified EHR data comprising approximately 5 million documents from 80,000 patients in the South London and Maudsley NHS Foundation Trust (SLaM) NIHR Biomedical Research Centre (BRC) Case Register. Data extraction: Clinical Record Interactive Search tool (CRIS). NLP: CRIS Comprehensive Data Extraction (CRIS-CODE) framework based on the General Architecture for Text Engineering (GATE). Data analysis and visualisation: R, Exploratory.io (population level data) and CogStack (patient level data). Implementation pilot: Quantitative and qualitative survey data on usability (using NVivo software). Timelines for delivery Years 1 and 2: EHR data extraction and NLP analysis Years 3, 4 and 5: Data visualisation tool development and implementation pilot Anticipated impact and dissemination New evidence to support safe and effective use of remote technology for mental healthcare presented at international academic meetings and published in high-impact open access journal articles and clinical guidelines. Data visualisation tools will be made publicly available. Open-access protocol to enable researchers to replicate methods in other healthcare settings. Patient contribution to study design and dissemination of findings through video production and workshops. Demonstration of data visualisation tools to healthcare service managers and policy makers through the King's Health Partners Psychosis Clinical Academic Group and NIHR South London Applied Research Collaboration (ARC).
Plain English summaries and category classifications on this site are generated by AI and may not perfectly reflect the original research.
Is something wrong? Let us know