Posted on: 20 July 2020
Could the impact of digital mental health interventions be improved by more personalised care?
This is the subject of a recently published JAMA article (https://jamanetwork.com/journals/jamanetworkopen/fullarticle/2768347) based on work by a team of researchers at SilverCloud Health, the world’s largest provider of digital mental health services, Microsoft Research Cambridge, Trinity College Dublin’s School of Psychology [e-mental health group] and the Trinity College School of Computer Science and Statistics.
Extensive research has shown that iCBT (cognitive behavioural therapy, delivered digitally) interventions help increase access to mental health treatment whilst achieving clinical outcomes comparable to traditional face-to-face therapy.
The research team explored de-identified longitudinal behavioural engagement data and clinical measures from 54,604 patients who used SilverCloud Health. It used probabilistic latent variable modelling, a machine–learning technique which can be used to infer distinct patient subtypes based on longitudinal patterns of patient engagement with iCBT. These patient subtypes are inferred based on patterns in how individuals interacted with different sections of the treatment platform over a 14-week period.
A machine learning framework was established to
identify subtypes of patients based on their use of the iCBT intervention over time
identify possible ways of triaging these different subtypes based on early patterns of behaviour
investigate how these subtypes predict clinical outcomes for patients with comorbid symptoms of depression and anxiety.
The work shows the potential for triaging patients for more personalised mental healthcare, demonstrated on the largest dataset of its kind.
Dr Derek Richards, Chief Science Officer, SilverCloud Health and co-director E-mental Health Research Group, School of Psychology, Trinity College Dublin, said:
The insights from this study allow us to tailor interventions for specific subtypes of engagement. For example, by identifying different subtypes of patients early on, we may be able to front load specific recommendations as to what types of content or tools may be associated with improved therapy engagement and clinical outcomes for patients.
We hope that the findings in this study may in the future facilitate tailoring interventions according to specific subtypes of engagement.
The sensitive and fluctuating nature of mental health symptomology, evidenced especially now during the COVID-19 pandemic, leads to a necessity in increasing access to interventions which can be personalised as per the behavioral patterns of the individual. It is therefore also our hope that this publication helps to increase the awareness of the importance of mental health and online interventions during the time of the COVID pandemic and beyond.
Dr Gavin Doherty, Associate Professor at the School of Computer Science and Statistics, said:
User engagement is central to the design of effective digital health interventions and has been a key pillar of the design of the SilverCloud platform from the outset. While broadly speaking, more engagement leads to better outcomes, this research reveals a more nuanced picture. This analysis uncovers a number of different classes of user who differ in terms of their engagement, and also their outcomes. This research brings a new set of techniques to bear on understanding user engagement in digital health, and is a first step towards leveraging the power of machine learning techniques in the design and delivery of online mental health interventions.
User engagement is central to the design of effective digital health interventions and has been a key pillar of the design of the SilverCloud platform from the outset. While broadly speaking, more engagement leads to better outcomes, this research reveals a more nuanced picture.
This analysis uncovers a number of different classes of user who differ in terms of their engagement, and also their outcomes. This research brings a new set of techniques to bear on understanding user engagement in digital health, and is a first step towards leveraging the power of machine learning techniques in the design and delivery of online mental health interventions.
Project Talia is a continuing research collaboration between Microsoft Research Cambridge and SilverCloud Health.
SilverCloud Health provides a suite of internet-delivered Cognitive Behavioural Therapy (iCBT) interventions for the treatment of symptoms of depression, anxiety, and other functional impairments (e.g. stress). The research collaborators include Dr Derek Richards, Chief Science Officer, SilverCloud Health and co-director E-mental Health Research Group, School of Psychology, Trinity College; Dr Angel Enrique and Dr Jorge Palacios, Digital Health Scientists at SilverCloud Health and Research Fellows at the e-mental health group, Trinity College Dublin; and Dr Gavin Doherty from the School of Computer Science and Statistics.
