A few weeks ago I attended the European Federation for Medical Informatics and the Farr Institute of Health Informatics Research’s Manchester-based conference – Informatics for Health 2017. The conference was a vibrant mix of academic thought topped off with a generous helping of public collaboration, showing that the field of health and medical informatics takes collaboration and public involvement very seriously.
Since health informatics covers all aspects of health-data collection, storage and processing it would be impossible to do justice to the sheer breadth of research presented at this conference in a single article. Therefore, here I will focus on a couple of my personal highlights.
On Tuesday the 25th, Susan Michie from University College London gave a keynote talk about the Human Behavioural Change Project:
With environmental, social and health concerns appearing endemic in our society, Suzan noted that one of the best ways to address these issues would be through targeted behavioural change interventions. These take a huge array of forms from subtle nudges implemented by many governments and large organisations (encouraging everything from litter reduction to targeted urinal use – see here for examples), to less than subtle public health campaigns. These interventions are widely documented across academic literature and show a range of outcomes and successes. Susan outlined a vision where this literature could be used to answer the big question:
‘What behaviour change interventions work, how well, for whom, in what setting, for what behaviours and why’
This is undoubtedly a pretty ambitious question to answer and it is made harder by the fact that the literature on this subject, although vast, is often fragmented, inconsistent and sometimes incomplete. So how do Susan’s team propose to tackle this big data problem?
The Human Behaviour-Change Project, funded by the Wellcome Trust, draws together some of the best minds in behavioural, computer and information science. Their output will depend on the close working relationships and interplay between all disciplines involved.
Behaviour scientists have been tasked with developing an ‘ontology’, basically a standardised method of categorising different behavioural change interventions. It is then hoped that this standardised ontology can be used to both sort existing literature and as a template on which new studies can be based. It is hoped that this will add some much needed order to the current fragmented literature and pave the way for further analysis. Specifically, computer scientists on this team will use Natural Language Processing (a branch of computer science which employs artificial intelligence and computational linguistics to sort and process large bodies of text) to extract and organise information from these studies, whilst also learning as they process this information.
Finally information scientists, the big data miners, will develop effective user interfaces which allow researchers to delve into this data and to untangle it in a way that reveals answers to many important research questions.
This is undoubtedly a huge task but with the combined input of so many specialists it certainly seems tractable.
On Wednesday the 26th the conference was drawn to a close with a compelling talk from Sally Okun, Vice President for Advocacy, Policy and Patient Safety at PatientsLikeMe, an online patient powered research network. The PatientsLikeMe network partners with 500,000+ patients living with 2700+ conditions and offers a platform for patients to share experiences and where researchers can learn more about treatments directly from those undergoing them. Indeed, more than 90 peer reviewed papers have already stemmed from data collected through the PatientsLikeMe network.
The theory behind this work is compelling and almost begs the question as to why such networks are not yet commonplace. Indeed, it’s no secret that online marketers spend billions analysing our search histories and purchase data in an attempt to feed us highly personalised targeted marketing, so why shouldn’t patient experiences be used to tailor personalised medicine? Although there are undoubtedly greater complications linked to the use of patient data, not to mention the perils of misinformation, this is no excuse not to try and work towards a digital ideal.
Sally also discussed the launch of their new platform, the Digital Me. This platform will combine a plethora of personal health data including genetic data, medical histories, activity tracking – basically if you can collect it you can include it. Their hope is that this data can be used to personalise medical treatments, tailoring them to your own individual requirements. Indeed, advances in statistical methods could take us beyond blanket prescribing and into a world where your digital profile can be compared to those similar to you (similarity being based on a large number of patient characteristics) and recommendations made based on successes and failure of treatments for you nearest digital neighbours (those sharing most of your traits).
As my first experience of an informatics-based conference, I was struck by both the breadth and depth of knowledge in the field and the ethos of working together to optimise our outputs – a skill which is often found lacking in other fields. It was also plain that researchers in this area value patient input and many elements of this conference were tailored to be accessible and engaging for a lay audience. Indeed, representatives from HeRC’s own patient public forum who attended the event enjoyed the opportunity to engage further with researchers and learn about engagement and involvement work being conducted across the field.
Post by: Sarah Fox
Save
Save