The Untapped Potential of NHS Talking Therapies Data: reflections during Mental Health Awareness Week

The untapped potential of NHS Talking Therapies data

Poor mental health is an urgent issue in England.

1 in 5 adults experience a common mental health problem. This impact is felt not only by individuals and families, but across society as a whole - in 2022, the Centre for Mental Health estimated that the economic and social costs of mental ill health in England was £300 billion a year - nearly double the annual NHS England budget.

When NHS Talking Therapies first launched in 2008, 40,000 people received treatment for depression and anxiety disorders, while over 670,000 people were treated this year. The service collects outcome data from 98% of people who complete a course of treatment, creating one of the largest and most complete mental health datasets in the world. Yet despite its scale and quality, this data has remained vastly underused in research.

This creates an enormous opportunity to better understand how mental health conditions are identified and treated in practice. For example, we know relatively little about how anxiety disorders and depression are recorded in primary care, which clinical codes are used, and how recording practices vary between GP practices, regions, and patient groups.

As demand for mental health support continues to rise, so does the need to better understand how NHS Talking Therapies are delivered - and which approaches are most effective for specific patient groups - so that we can better identify inequalities, improve care, and evaluate the effectiveness of services.

Building tools and services to better understand mental health care in primary care records

Thanks to a £17 million award from Wellcome across two separate grants, we’ve been working on incorporating NHS Talking Therapies data into OpenSAFELY, a platform that delivers safe and secure analyses across more than 58 million patients’ full pseudonymised primary care NHS records.

This will allow researchers to investigate many important questions about mental health treatment at a national scale, including:

• How talking therapies affect long-term health outcomes
• Which approaches work best for specific conditions and patient groups
• The best way to deliver services
• The relationship between mental health treatments and physical health

Large clinical datasets often contain hundreds of clinical codes and unstructured data, which can make them difficult to work with. To address this, we’ve developed two new tools, both built on top of aggregated, publicly available data published by the NHS: the nhstt R package, which makes it easier for researchers to make sense of NHS Talking Therapies data, and the opencodecounts R package, which helps researchers understand which clinical codes are used to record mental health assessments, diagnoses, and treatments - and how these recordings have changed over time.

But, tools alone are not enough. Researchers also need practical support to navigate and use these complex datasets effectively.

To help them make the most of this data, we’ll be launching an Electronic Health Record Support Unit - once the NHS Talking Therapies data becomes available through OpenSAFELY - to help researchers design and deliver studies using linked NHS Talking Therapies and GP data.

Expanding mental health research in OpenSAFELY

Before we can support other researchers to use NHS Talking Therapies and GP data effectively, we need to understand how mental health conditions are currently recorded in GP records, which codes are used, and where recording is inconsistent or incomplete. These insights will directly inform the tools, guidance, and support we develop for researchers.

To build this foundation, we’ve recently submitted an OpenSAFELY research application to access individual-level GP electronic health records. This work will help us better understand how common mental health conditions are assessed, diagnosed, and treated in GP practices across England.

Using GP records from 2013 to 2024, we’ll analyse trends and variation in the recording of patient-reported outcome measures for anxiety disorders and depression - including PHQ-9 and GAD-7 scores - standard questionnaires used in healthcare to measure the severity of depression and anxiety symptoms - as well as antidepressant prescribing and referrals to NHS Talking Therapies.

Comparing data in the GP records with recent results from the Adult Psychiatric Morbidity Survey, we’ll be able to identify gaps in recognition and treatment, and explore how recording of care varies by age, sex, ethnicity, deprivation, and geography.

Laying the groundwork for the future of mental health research in England

Mental Health Awareness Week is a timely reminder for us to prioritise learning about how mental health care is delivered, which treatments work best for different people, and where gaps in access and outcomes remain.

As more people continue to seek support for anxiety, depression, and other common mental health conditions, large-scale mental health data analysis will be key to improving mental health care in England, so that everyone can receive timely and effective support, when and where they need it.

To learn more about our work with big mental health data, visit the Bennett Institute’s website, follow the Bennett Institute on LinkedIn and BlueSky, and sign up to receive notifications of the latest blog posts and newsletter updates.