Every year, 64.7 million antidepressant prescriptions are dispensed in the UK, yet most patients stop taking them within 30 days—far too soon for the drugs to work. The core problem is that doctors have no evidence-based way to predict which antidepressant will work best for a specific patient, so prescribing remains a trial-and-error process that often fails. This project aims to build a clinical decision support tool—a treatment algorithm—that personalises antidepressant selection. The researcher will first combine data from thousands of patients in clinical trials and real-world NHS records to identify which patient characteristics (age, symptom severity, previous treatments) predict better outcomes with specific drugs. The algorithm will then incorporate each patient’s preferences about side effects, generating a tailored ranking of antidepressants. Finally, a randomised trial with 496 participants across five UK centres will test whether using the algorithm reduces the rate of treatment discontinuation compared to usual care. If successful, the tool could be deployed in NHS GP surgeries and outpatient clinics, replacing guesswork with personalised prescribing. This would mean fewer patients cycling through ineffective drugs, fewer adverse events, and more people completing a full course of treatment—directly improving outcomes for the millions of people treated for depression each year.
View original technical description
Background and research question Depression is the second leading causes of global burden and a significant challenge for health systems worldwide. Antidepressants are commonly used to treat moderate to severe depressive disorder. Only in the UK, 64.7m prescriptions of antidepressants were dispensed in 2016. Despite these figures, many patients are given antidepressants which do not work or cause adverse events. In fact, the majority of these prescriptions are for less than 30 days, which is much shorter than required to see a clinical effect. A key issue which drives this problem is that the perceived efficacy of the drugs is low and that, while there is a wide choice of antidepressants, no information is available to guide the personalised selection of the most efficacious and tolerable drug for a specific patient. Aims and objective I aim to develop and test an evidence-based clinical decision support tool to personalise treatment in depression in the NHS. Methods During Phase 1 (Evidence synthesis), I will: systematically review published and unpublished antidepressant trials in adults with depressive disorder; combine aggregate and individual-patient data in a network meta-analysis; rank the different antidepressants and stratify outcomes into subgroup of patients with different treatment indications based on patients' individual characteristics. During Phase 2 (Development of the treatment algorithm), I will: develop a clinical decision support tool (treatment algorithm) to combine the stratified recommendations from randomised evidence with the preferences and values of patients. Patients' characteristics (i.e. age, symptom severity, previous treatments) and preferences about side effects will be entered in the algorithm. This information will vary from one person to another and will change the final output, tailoring the ranking of antidepressants to each individual; incorporate routinely collected data derived from large observational datasets in primary and secondary care into the algorithm. Matching the patient's characteristics, the combination of information from randomised and observational data will increase the precision and reliability of the treatment recommendations; run a series focus groups with patients, carers and clinicians to inform the development of an easily used and acceptable platform by which the clinical value and practical utility of the algorithm can be maximised. During Phase 3 (Randomised trial), I will conduct a randomised, multi-centre trial. I will recruit 496 participants with depressive disorder who have been prescribed antidepressants from GP surgeries or outpatient clinics across 5 centres in the UK, comparing the use of the treatment algorithm versus usual care. The primary outcome will be the rate of discontinuation due to any cause, as a proxy of the clinical acceptability of the algorithm. Anticipated impact and dissemination Development and validation of a treatment algorithm for personalised prescription of antidepressants in depression, to be used in routine clinical settings within the NHS. Publication of the results from the three phases of the project in high-impact journals and dissemination of the findings to the general audience via patient and public engagement. Data and analyses will be made freely available, as appropriate, to support the future enhancement of classication algorithms.
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