Active Mental Health

OpenSAFELY Talking Therapies: Using large datasets to advance psychological therapy research and policy

In plain English

AI plain-English summary

The NHS Talking Therapies dataset—containing records from over a million patients who have received psychological treatment—is being added to the OpenSAFELY research platform, giving mental health researchers secure access to data they have long struggled to use. Mental health research has lagged behind other medical fields partly because patient data from psychological therapy services has been difficult to access and analyse. Many psychotherapy researchers lack the technical skills or infrastructure to work with large, anonymised electronic health records. This project builds a dedicated support unit—staffed by software developers and psychotherapy researchers—that handles the data management and privacy safeguards, freeing researchers to focus on clinical questions. If successful, this framework could accelerate the development of better treatments for anxiety and depression, and clarify how mental and physical health interact. The same model could later extend to child mental health and psychosis services. The work does not itself test a new therapy; it creates the infrastructure that makes such testing faster, cheaper, and more rigorous. For patients, the payoff is indirect but real: faster evidence for what works, and treatments shaped by data from millions of real-world therapy sessions.

View original technical description
OpenSAFELY is an established national research resource funded by the Wellcome Trust and NHS England, providing an effective support service for new researchers. We propose to extend this service by (1) adding the NHS Talking Therapies (NHSTT) dataset and (2) establishing an Electronic Health Record Support Unit (ehrSU) that provides comprehensive support for mental health researchers. Integration of the NHSTT dataset would turbo-charge psychological therapy research, unlocking the full potential of this invaluable dataset, while preserving patients’ privacy. We will hire skilled software developers and researchers with a background in psychotherapy research, while enhancing our public and patient involvement (PPI) efforts by embedding mental health representatives to guide our projects. This will allow us to establish a fully operational EHR Support Unit that supports studies that facilitate Wellcome’s strategy of creating a step change in interventions for anxiety & depression, and illuminate the links between mental and physical health. The service is designed for mental health researchers, particularly those with expertise in psychotherapy research but limited experience in data management. Once successfully established, this collaborative framework can be extended to other areas of mental health beyond anxiety & depression in adults, such as child mental health and psychosis services.

View the original record at the funder ↗

Researchers

Alex Walker (EPMC Awardee)Amelia Green (EPMC Awardee)Amir Mehrkar (EPMC Awardee)Andrea Schaffer (EPMC Awardee)Ben Goldacre (EPMC Awardee)David Clark (EPMC Awardee)David Evans (EPMC Awardee)Helen Curtis (EPMC Awardee)Louis Fisher (EPMC Awardee)Milan Wiedemann (EPMC Awardee)Pete Stokes (EPMC Awardee)Peter Inglesby (EPMC Awardee)Rose Higgins (EPMC Awardee)Sebastian Bacon (EPMC Awardee)WIlliam Hulme (EPMC Awardee)

Related Research

Grants with similar aims, by meaning.

Addressing COVID-19 Surge and professional staff shortages in NHS Mental Health Services
OpenSAFELY Phase Two
The Open Psychiatry Project: a co-produced web interface and platform for mental health omic data and federated data analysis
The development and research testing of digital mental health therapy homework utilising the Dr Julian Platform to improve compliance of homework and outcomes of therapy
FAIR TREATMENT: Federated analytics and AI Research across TREs for AdolescenT MENTal health

Original classification

Discretionary Award

Plain English summaries and category classifications on this site are generated by AI and may not perfectly reflect the original research.