Implementing patient preferences in a shared decision-support tool to personalise antidepressant treatment for depression using real-time data
We aim to personalise treatment for depression using a living algorithm, which will incorporate multi-modal data (from demographic characteristics to imaging and genetics), individual preferences and real-time outcomes from patients. Embedded in a continuously learning environment, this clinical decision-support tool will produce more precise recommendations that will be more generalisable to people seen in routine care settings.
Part of our Mental health across the life course research theme
Research on pharmacological treatment of depression is rapidly evolving, implementing innovative methodologies to optimise patients care.
Recommendations from available evidence often target an “average” individual with depression, rather than addressing individual preferences and characteristics. Precision mental health is an emerging approach to personalise treatment that considers individual differences across several domains, from genes and lifestyle and other measurable characteristics.
We are currently working on a NIHR-funded project, called PETRUSHKA, aimed at combining data from observational studies and randomised controlled trials to predict individual responses to antidepressant treatments.
We are developing an evidence-based, digital shared decision-support tool using real-time data and patient preferences to personalise treatment, transform clinical outcomes and empower patients with major depression in the NHS.
Andrea Cipriani, Professor of Psychiatry
The PETRUSHKA prediction algorithm is based on a locked dataset, that has to be regularly updated in order to incorporate evidence from new clinical studies. Another important issue to address is how results from the clinical decision-support tool are presented to patients, and how to help them interpret the output in a clinically informative way.
During this DPhil project we will:
- systematically review the existing methods to elicit and incorporate individual preferences from patients, and, of possible, quantify their relative performance;
- explore the use of the multi-criteria decision analysis methodology to collect patient preferences in a clinical decision-support tool using artificial intelligence;
- integrate alternative types of data, such as imaging and genetic models, in the current PETRUSHKA tool;
- develop and test a new method to feedback into the algorithm outcome data from patients that are using the algorithm itself;
- build a user-friendly tool to facilitate communication of findings and their implementation in the shared decision-making process between individuals with depression and their clinicians.
How we are involving patients and the public
We will closely work with the OxPPL Patient and Public Involvement, Engagement and Participation (PPIEP) group that has already collaborated to develop the PETRUSHKA tool and is linked to the Patients & Research Strategy Group of the Oxford Health Biomedical Research Centre.
The OxPPL PPIEP group meet regularly (at least three times a year) and includes a diverse population of people with a living experience of mental health illnesses. We will leverage their expertise on precision psychiatry from the previous collaboration with the OxPPL group to improve collection and implementation of patient preferences, as well as to shape the framework for findings dissemination.
How we are planning to implement the research outputs
We aim at using the deliverables from this project to improve the care of people with depression, exploring the implementation of precision mental health in the current NHS primary care model. To do this, we will liaise with our local trust (Oxford Health NHS Foundation Trust) to evaluate a local integration into existing care pathway for people with depression.
We will also work with international collaborators to see if it is possible to export our clinical shared decision-support model beyond the UK (for instance, our existing collaboration with Toronto, which includes the Centre for Addiction and Mental Health and the Department of Psychiatry, University of Toronto).
Project Lead
Team Members
-
-
-
Huseyin Naci
Associate Professor of Health Policy, London School Economics and Political Science
Project end date
July 2024
Aims
To investigate:
Understand which methods better capture individual preferences and their implementation into clinical shared decision-making processes
Move from a “static” to a “dynamic” and “living” model, with outcome data continuously improving the clinical shared decision-support model
Expanding the existing care model for depression to effectively integrate findings from imaging and genetic data
Facilitate communication of findings in shared decision-support process
Deliverables
We will continue the ongoing discussion with experts in health decision making and behavioural science, local PPIEP group and representative from local GP surgeries to ensure the use of state-of-the-art models for patient preferences
We will transparently publish our findings in high impact scientific journals
We will explore the implementation of our clinical shared decision-support model locally with the Oxford Health NHS Foundation Trust
By engaging NIHR infrastructures like the NIHR Applied Research Collaboration Oxford and Thames Valley (ARC OxTV), we will explore how to successfully expand the implementation of our clinical shared decision-support model regionally, and then nationally
Expected Impact
We expect our work to have an impact on individual patients with depression with the introduction of precision mental health in the standard healthcare pathway. We expect our clinical shared decision-support model to be implemented into the primary care services under the NHS, starting locally at the Oxford Health NHS Foundation Trust.
Taken together, these will result in a reduction of direct costs by preventing prescription of only partially beneficial drugs compared to available alternatives, as well as indirect costs by affecting the economic consequences of work lost due to illness.