Each spring, the University of College London Centre for Behaviour Change (CBC) hosts a conference, whose theme is “Behaviour Change for Health: Current & Emerging Science & Technologies”. This annual event has been a highlight for me since 2015 because of its dual focus on use of technology and the latest in behaviour change science, which reflects my dual role as head of behavioural science as well as member of the HRW innovation team. For the first time, we had been invited to present, both as part of a symposium on virtual reality and chatbots, as well as a poster sharing our latest research on mobile phone apps for health behaviour change. so as the pandemic ramped up, it was disappointing to see the April conference get postponed, but happily it was rescheduled for September 16th-18th as a virtual conference. The format of the virtual conference did add some challenges; namely all presentations were pre-recorded, so the live sessions only involved Discussion and Q&A. This system required delegates to pre-watch sessions and come prepared with questions rather than just immersing in the sessions as they occurred. As a presenter, this format made it much more relaxing on the day as your session had been pre-recorded. Over 400 delegates from a mix of academia, commercial, and third sectors attended and created a vivacious community of discussion during presentations and on the Whova conference app.
At HRW, especially within the behavioural science team HRW shift, we value touchpoints with the academic community, especially with the UCL CBC which is a world-leading institution for applied behavioural science. Academia’s robust investigations into the multiple dimensions of behaviour change in controlled trials and systematic reviews give us a great evidence basis to draw from in our work with HRW clients. And as in previous years, the conference exceeded expectations in the quality of content and the learnings and implications we can take away and apply to our work.
Overall, three key themes stood out:
Behavioural science has often reiterated the importance of context; the physical, social, and psychological state in which behaviour takes place has a huge non-conscious role. However, because of the enormous variability of context, it is one of the more difficult components to account for in market research. Here the CBC conference had some great examples. Perhaps the best was a poster about a context mapping tool from Dr Weston Baxter and colleagues at London School of Hygiene and Tropical Medicine, a templated framework and series of questions designed to take a more methodical account of the contextual factors.
There were also great examples of how taking account of context works in active interventions – Dr. John Torous of Beth Israel Medical Center, Harvard Medical School who talked about understanding patients affected by health inequity and realising how digital interventions struggle in these populations because of lack of access to even mobile phones. Similarly, Dr Rob Mooney spoke about the implementation of the COVID tracker app in Ireland and how critical that the chief medical officer was on television each day to talk about the app to increase uptake.
Another component of context is our habits, which Professor Wendy Wood (author of ‘Good Habits, Bad Habits’ and business professor at my undergraduate alma mater, University of Southern California) stated in her keynote account for 43% of behaviours each day. And even that figure was generated using self-report (based on whether respondents were thinking about the task or thinking about something else) so I’d hazard a guess that the proportion is higher. Taking that further, Eamon Colvin of the University of Ottawa questioned whether there are mental habits as well as physical habits: most work on habits has focussed on physical/observable behaviours but there might be patterns of thinking that might be considered ‘mental habits’.
Finally, I was glad to contribute again, this time in relation to the discussion of context with a poster session on Friday 18th. Here we shared the results of our self-funded study alongside technology company RealityMine that allowed us to passively track the mobile phone use with people using apps for health behaviour change (medication adherence or alcohol reduction). Seeing this app use in its full context allowed us to put our observations in perspective and share with the conference just how small the focus on health is in relation to the other components of life:less than 1% of activity on the phone related to health, and use of social media was 8-11 times greater and for up to 38 times longer than use of the behaviour change apps.
Utilising technology where appropriate
Given the focus of this conference is around emerging technology, it is no surprise thatthere is lots of discussion about cool new tools. But one of my favourite things about this conference is that each technology is thoroughly questioned about whether using the technology would change or negatively impact the intervention. A classic example of this came from the keynote speaker Heleen Riper, professor of mental health, who talked about the use of iCBT (internet delivered cognitive behavioural therapy): because of the serious nature of mental healthcare, there had been few trials comparing iCBT with traditional in-person visits. However amidst the pandemic the dramatic need for iCBT created dramatic uptake, and her recommendations both recognised the validity of iCBT efficacy from these case studies, but also recommended a blended approach where iCBT is used alongside other delivery mediums including face to face.
In the symposium on Virtual Reality and Chatbots we presented alongside 4 academic institutions as we shared a case study of a virtual reality training module that helped our client refine applications of behavioural science. Much of the discussion centred not only on the benefits of these delivery mechanisms such as engagement, memorability, and ability to simulate worlds beyond what’s possible, but also the important downsides: the technical requirements, cost, age-based difficulty, unproven longevity of effect. These nuanced discussions introduce such important considerations to make sure the delivery medium doesn’t overshadow the impact.
There were, however, some excellent examples of using new technology to its fullest extent. Within our session, Dr. Olga Perski shared a great example of a chatbot that increased engagement with a stop-smoking app. Lean Kramer from Wageningen University gave a paper presentation about the effectiveness of embodied conversational agents (think Clippy the paperclip in Microsoft word) as an intervention to nudge older adults engage towards healthy eating. And in possibly the most astonishing example of a good use of technology came from Peter Piroli, Senior Research Scientist at the Institute for Human and Machine Cognition, who was using AI informed by behavioural science to more reliably estimate how people would comply (or not) with psychological interventions related to COVID-19 and then using this to inform epidemiological models.
Behavioural Evidence in design
Centre for Behaviour change director (and behavioural science Rockstar) Susan Michie said during the conference ‘popularity is not the same as effectiveness’. And this is a consistent theme that comes through at this conference and is another reason why we’re such regular delegates. The academic community does incredible long-term and thorough work on understanding the effectiveness of different interventions, and we are incredibly passionate about using this evidence basis in our commercial work. This year was no exception, with several sessions providing useful learnings and sources to improve the efficacy of behavioural interventions.
Several sessions talked about the challenges of behavioural consultancy in high-speed environments. Day 1 keynote presenter Rachel Carey from Zinc talked about the role of evidence in a startup company environment and provided a nice roadmap for how she and her team are able to ensure adoption of expertise in early set up stages as well as throughout the duration of a project. Building on that theme, Sander Hermsen of One Planet research centre gave a case study of how behavioural consultancy can be delivered even alongside an agile product development process as long as the experts are present. These examples support our belief in the role of integrating behavioural science expertise within a company rather than keeping it separate.
Some case studies on how companies have used particular behaviour change techniques in technological offerings and their efficacy were also compelling: Paulina Bondaronek of the UCL E-health unit described the behavioural components present in the popular ‘Couch to 5k’ and ‘7 min workout’ apps. Thomas Rouyard of Hitotsubashi University gave a paper presentation about the behavioural components designed to communicate risk to type II diabetes patients (including communicating their personalised ‘heart age’ and how many days of life they were losing each month). Madalina Jager of Denmark and her global team presented a poster outlining the results of her scoping review of patient facing apps for multimorbid patients including 74 studies, giving a very robust picture of what techniques are used in patient-facing apps. These academic systematic reviews and meta-analyses are the best resource for us to continue to build a strong evidence basis for the behaviour change interventions we discuss and test with clients.
All in all, it was a delightful and busy three days, and we were honoured to be recognised as presenter calibre. For a fuller summary or for links to any of the papers referenced please get in touch: email@example.com
By Katy Irving