TherapyMatch-D Trial
Research type
Research Study
Full title
A pilot cluster randomized controlled trial of psychological treatment selection for depression
IRAS ID
281430
Contact name
Jaime Delgadillo
Contact email
Sponsor organisation
Rotherham Doncaster and South Humber NHS Foundation Trust
Clinicaltrials.gov Identifier
applied; number to be confirmed, ISRCTN**
Duration of Study in the UK
3 years, 0 months, 1 days
Research summary
Reviews of clinical trials have concluded that different types of psychological interventions for depression are equally efficacious. Therefore, the largest public provider of psychological services in England, Improving Access to Psychological Therapies (IAPT), routinely offers Cognitive Behavioural Therapy (CBT) and Counselling for Depression (PCE-CfD) as front-line treatments. Despite efforts to improve recovery rates, 1 out of 2 patients accessing any evidence-based psychotherapy usually does not recover from depression.
Recent studies have suggested that identifying subgroups of patients that respond differently to diverse treatment modalities or intensity can improve clinical outcomes. One retrospective study used a large routine care dataset of patients who accessed either CBT or PCE-CfD. Researchers found that in about 30% of patients, if matched to their optimal treatment using artificial intelligence, patients were twice as likely to recover from depression.
The aim of this study is to pilot the effectiveness of using a treatment selection model based on that prior study. We will explore if more patients recover from depression by being provided a treatment recommendation compared to allocation as usual (which consists of shared decision making). IAPT sites will be randomised to TherapyMatch-D group or allocation as usual (control). During the initial assessment, clinicians in both groups will input suitable patients’ data into a computer programme. In TherapyMatch-D group, patients who according to prior research would benefit more from either CBT or PCE-CfD will be provided a treatment recommendation and then used shared decision making with the clinician to reach a final decision. Patients in the control sites will not be given a recommendation. We will then compare outcomes of both groups to explore if the treatment selection model is feasible and effective. Furthermore, interviews will be conducted with patients and clinicians to explore their views and experiences of using artificial intelligence to inform treatment selection.REC name
London - Riverside Research Ethics Committee
REC reference
23/LO/0487
Date of REC Opinion
21 Jul 2023
REC opinion
Further Information Favourable Opinion