Using data mining to Do things better

– By Billiejoe Charlton, KTP Associate, Do Something Different

As part of my KTP work with Do Something Different Ltd and University of Brighton, I recently attended and presented a paper at the 7th International Digital Health Conference in London [link: http://www.acm-digitalhealth.org/ ]. Held at 30 Euston Square – the headquarters of the Royal College of General Practitioners – the conference offered the opportunity to disseminate the scientific findings from my KTP project to an interested and knowledgeable audience, and raise awareness of the vital work that Do Something Different [http://www.dsd.me/] is doing to help people change their lives for the better. It was also a wonderful chance to network with like-minded professionals from a wide variety of backgrounds, and learn about exciting new developments in the field of Digital Health.

The theme of the conference was “Global Public Health, Personalised Medicine, and Emergency Medicine in the Age of Big Data”, and it attracted experts from many fields, including medicine, disaster management, sensor technologies, data mining and social marketing. Kicking things off was Dr Oliver Morgan from the World Health Organisation, whose fascinating keynote asked, “How can we make better use of data to protect people’s health and save lives?”. Dr Morgan described the WHO’s development of a new global surveillance system for disease outbreaks, which will bring together data from national public health systems as well as from less structured sources such as news and social media reports. This set the tone for an event where the focus was firmly on using emerging technologies not for the accumulation of profit but for the benefit of all.

The research I presented is about how we have used data mining techniques to improve the behaviour change programmes delivered by Do Something Different. Each such programme consists of a series of personalised “Dos” – small recommended activities to help people practice behaving in new ways and break their habits. These behavioural prompts are delivered by a smartphone app, or by SMS or email. The approach is based on decades of psychological research, and programmes have been designed to address many personal development goals, such as smoking cessation, stress reduction, better diabetes self-management, leadership skills and so on.

A slide from my presentation: Dos are prompts for small actions, delivered by smartphone, designed to help people change their behaviour.

 

In our research we have applied data mining techniques to interaction data and psychological questionnaires from a sample of Do Something Different’s users. Our data set included information about 15,550 people who have taken part in a Do Something Different programme. Using correlation networks and regression models, we were able to construct a new, more precise model of the connections between the behaviours promoted by Do Something Different and a person’s wellbeing and happiness. This has led us to refine the contents of the programmes. The paper, titled Using Data Mining to Refine Digital Behaviour Change Interventions [link: http://dl.acm.org/citation.cfm?id=3079468 ], is co-authored with John Kingston, Miltos Petridis and Ben (C) Fletcher. Interested readers can try out Do Something Different today by going to https://dsd.me/get-started/.

My presentation was in a session on Study Metholodogies, which also featured a fantastic talk by Emily Keane from charity Save The Children, about the use of a smartphone app in the treatment of malnourished children in Africa. The app replaces paper-based child registers for recording data, and guides the health worker step-by-step through the treatment protocol, so that no steps are inadvertently missed. Video illustrations show how to correctly measure a child’s mid upper arm circumference, a key indicator in assessing nutrition, which is frequently performed incorrectly.

Away from the talks, some exciting technology was on display. I was particularly impressed by the Advies.chat system (http://www.soaaids.nl/en/advieschat), a “chatbot” advice tool that offers free and anonymous advice about sexual health and STI testing to young people, based on clinical guidelines. The developers note that “over 20 years of experience in one-on-one counseling via telephone, e-mail and direct message chat about STIs, HIV, testing and prevention were manually translated into structured responses to common questions”. This approach was thought-provoking for me, as Do Something Different is currently exploring ways to improve its user interaction experience.

I returned from the conference having made useful contacts, and more motivated than ever to work on my KTP project, where we are continually looking for new ways to use technology to help people make the changes they want in their lives, and be happier and more fulfilled. I hope to attend the conference again next year, and present further discoveries from analysing data about behaviour change.

– Billiejoe Charlton, KTP Associate, Do Something Different

5 Reasons to be ‘Appy

KTP company Do Something Different (DSD) seek to make the world a better place by inspiring millions of people to do something different through behaviour change programmes. Clients are sent personalised ‘dos’ which aim to take the individual out of their comfort zone, breaking the habits which prevent them reaching their full potential and developing behavioural flexibility.

KTP Associates Billiejoe Charlton recently posted on the blog for his company DSD to tell us about their new app which you can log into every day to log how you feel on the ‘happy tracker’. Why use the app? App users give on average 4.07 stars out of 5, vs 3.67 for non-app users, suggesting they have a better overall experience. You can read Billiejoe’s full blog post here: https://dsd.me/2017/03/15/5-reasons-to-be-appy/

DSD have amassed a wealth of data from delivering programmes over the years – this KTP project seeks to apply machine learning to optimise their programmes to help the company expand and grow.