Embed Size (px)
Citation preview
Building the science of behaviour change:The Human Behaviour-Change Project
Susan MichieCentre for Behaviour Change
University College London, UK
@SusanMichie
Acknowledgments
• Funders including
• Many have contributed to my thinking and work
– especially …Robert West and Marie Johnston
• The Health Psychology Research Team
• The Centre for Behaviour Change
This talk
1. Some challenges in advancing behavioural science
2. Opportunities for advancing efficiently
3. The need and plans for developing a Behaviour Change
Intervention Ontology
A foundation on which to build
• We have a rich source of theories and methods for
intervention design and evaluation
• Considerable investment in interventions aimed at
individuals, communities and populations
– Trials: estimated 100’s behaviour change interventions per day
• Most have modest and variable effects– e.g. Cochrane database, National Institute for Health & Care Excellence (NICE)
How can we
improve this
situation?
Two challenges
1. Descriptions of behavioural interventions
– vague, partial and/or use terminology inconsistently
• 40–89% interventions non-replicable
• Recommendations include
– High quality and complete reporting demanded by
journals, authors and peer reviewers
• use reporting guidelines
Glasziou et al, Lancet, 2014
2nd challenge: Theory
We have a plethora of theories but many are
– overlapping & most are partial, underspecified and/or use
terminology inconsistently
• 83 theories identified in a review, 1700+ constructs
This means that …
• Our science underperforms
– replication is difficult
• We accumulate evidence slowly
– evidence synthesis is difficult
• It is difficult to implement effective interventions
Some solutions
1. Develop taxonomies of intervention characteristics
– Techniques (BCTs), modes of delivery, target behaviour,
mechanisms of action, target population, target context
2. Specify theories precisely in terms of constructs and
14 types of relationship
– Project with 83 theories aiming to identify ‘canonical’ theories
3. Develop an ‘ontology’ of behaviour change
interventions
Method for describing interventions: Behaviour change techniques (BCTs)
• “Active ingredients” within an intervention designed
to change behaviour
• They are
– discrete, low-level components of an intervention that on
their own have potential to change behaviour
– observable and replicable
Michie S, Johnston M, Carey R. (2016). Behavior change techniques. In
Turner, JR. (Ed.) Encyclopedia of Behavioral Medicine. Springer New York.
“Taxonomies” of BCTs
• Physical activity/healthy eating/mixed : 26 BCTs Abraham & Michie , 2008
• Physical activity & healthy eating: 40 BCTsMichie et al, Psychology & Health, 2011
• Smoking cessation: 53 BCTs Michie et al, Annals behavioural Medicine, 2010
• Reducing excessive alcohol use: 42 BCTsMichie et al, Addiction, 2012
• Condom use: 47 BCTsAbraham et al, 2012
• General behaviour change: 137 BCTsMichie et al, Applied Psychology: An International Review, 2008
• Competence framework: 89 BCTsDixon & Johnston, 2011
BCT Taxonomy v1
• Developed by 400 experts from 12
countries
• Clearly labelled, well defined, distinct,
precise; can be used with confidence
by a range of disciplines and countries
• Hierarchically organised to improve
ease of use
• Applies to an extensive range of
behaviour change interventions
BCT Taxonomy v1: 93 items in 16 groupings
Find by search term: BCTsor
The BCT smartphone app
• Search by BCT label, BCT category or
alphabetically
Feedback and plans for updating BCTTv1
http://www.ucl.ac.uk/behaviour-change-techniques/BCTTv1Feedback
To apply theory to developing interventions, need to know
how BCTs are linked to theory (mechanisms of action)
Developing a method for linking BCTs to
mechanisms: ‘Theory and Techniques’ project 2014-17
• International Advisory Board
41 experts from 11 countries
1. Systematic review: what does the literature (>300 articles) tell us?
2. Expert consensus: what do 96 experts from 18 countries think?
3. Triangulation
Protocol Paper
Two data sources:
1. Published reports of interventions
2. Expert consensus
BCTs
MoAs
Outputs:
Data are represented in heat
maps to indicate the relative
frequency with which each
BCT is hypothesised to link
to each MoA
Taxonomy of behaviours
• Led by Kai Larsen, University of Colorado
– with Robert West
• 5,461 articles from 3 leading journals in
– Psychology, Education, Behavioral Medicine, Business,
Management, Marketing, Information Systems, Nursing
• 2,375 behavioural variables
– Extending WHO’s International Classification
of Functioning, Disability and Health (ICF)
• Created 8 levels of hierarchy
Taxonomies of modes of delivery, populations, settings
• With Rachel Carey, Robert West (UCL) & Marie Johnston
(Aberdeen)
• Categories and structure
– inductively generated from published research and existing
ontologies
– Open peer review
• Presentations at
Results: Mode of Delivery Taxonomy (v1)
• 4-level hierarchical taxonomy with top-level categories:
1. Human: Delivery through human interaction in which the participant sees and/or
hears a person in real-time
2. Hard copy: Delivery through physical / hard-copy material (e.g. paper, acetate,
objects); given or sent to the participant); can include diagrams, pictures and/or text
3. Digital: Delivery through a form of digital technology, including computer,
smartphone, tablet, television, wearable & environmental devices
4. Somatic: Delivery through a device designed to act on the body
• 7 cross-cutting attributes identified: Grouping, Interaction, Tailoring,
Synchronicity, Gaming, Mobility, Format
Putting this together to answer the big question …
• Whether for researchers, policy-makers or practitioners:
‘What works, compared with what, how well, with what exposure, with what behaviours, for how long, for whom, in what settings and why?’
What is needed: an Ontology
• A systematic method for specifying interventions with
– a ‘‘controlled vocabulary’’ of agreed-upon terms & their inter-
relationships.
• 3 core elements1. a controlled vocabulary specifying and defining existing constructs
2. specification of the inter-relationships between constructs
3. codification in a computer-readable format to enable knowledge
generation, organisation, reuse, integration, and analysis
Larsen, Michie et al, J Beh Med, 2016
The Human Behaviour-Change Project
A Collaborative Award funded
by the
Participating organisations
@HBCProject
www.humanbehaviourchange.org
In a phrase …
• Brings together behavioural science to create an ontology …
• … with computer science to digitalise the structured evidence
• … and information science to make it accessible to users
The collaboration
Behavioural science Information science Computer science
Grant-holders Susan Michie (PI; UCL)Robert West (UCL)Marie Johnston (Aberdeen)Mike Kelly (Cambridge)
James Thomas (UCL) John Shawe-Taylor (UCL)Pol MacAonghusa (IBM)
Researchers Ailbhe Finnerty (UCL)Marta Marques (UCL)Emma Norris (UCL)Researcher: tba
Alison O’Mara-Eves (UCL)Software developer: tba
Lea Deleris (IBM)Debasis Ganguly (IBM)
Project manager Niccola Hutchinson Pascal
The Human Behaviour Change Project
The need Urgent need for more effective behaviour change interventions to improve health & wellbeing
The challenge
High volume of noisy data on effectiveness of interventions with large number of potential factors interacting to influence outcomes
The solution An Artificial Intelligence system that can extract relevant information from intervention evaluations, build a knowledge base that can be interrogated and continually improve as new information becomes available
The problem
Volume of research• Estimated >200 evaluations of behavioural interventions published each day
Reporting variability• Studies are reported very variably so difficult to synthesise or to draw
theoretical conclusions about mediation and moderation of effects
Context sensitivity• Much literature not directly relevant to specific contexts of users
Need for timeliness• Typical time for study results to be included in systematic reviews 2.5-6.5 years
Building the science of behaviour change
• The HBCP aims to revolutionise the ways in which we• Build knowledge and understanding about behaviour change
• Use that knowledge to answer real-world questions
‘What works, compared with what, how well, with what exposure, with what behaviours, for how long, for whom, in what settings and why?’
The 4-year plan
1. A Behaviour Change Intervention Ontology for organising relevant information from research reports
2. An automated feature extraction system to find and extract that information
3. Machine Learning and Reasoning Systems that integrate and extrapolate from that information to generate new knowledge and hypotheses about behaviour change
4. An Interface that answers users’ questions about behaviour change interventions and explains its conclusions
The Behaviour Change Intervention Ontology
A structured collection of terms and their inter-relationships and a
systematic method for defining them.
Three core elements:
1. a controlled vocabulary specifying and defining existing entities,
2. specification of the inter-relationships between entities, and
3. codification in a computer-readable format to enable knowledge generation, organisation, re-use, integration, and analysis
Larsen, Michie, Hekler et al (2016). Behavior change interventions: The potential of ontologies for advancing science and practice. Journal of Behavioral Medicine.
Use of Ontologies
• Tag data with the goal of making data deriving from heterogeneous sources more easily
• searchable,
• comparable or
• combinable
• Share information across communities of scientists with different sorts of expertise
Arp R, Smith ., & Spear AD (2015). Building ontologies with basic formal ontology. Cambridge: MIT Press.
The Preliminary Ontology of Behaviour Change Interventions
The task
1. Behavioural scientists annotate published reports using the BCI Ontology to train an Artificial Intelligence system to …
2. Systematically identify connections and patterns in data• Using Natural Language Processing and other feature extraction tools
3. Interpret evidence many times faster than a human• Using Machine Learning and automated reasoning
4. Generate reusable knowledge that both humans and machine can interpret• Accessed via a User Interface
Work streams in the Project
Examples of Users and Uses
E.g. what mechanisms of action are likely to account for the effect of x on y?
Public health policy-maker
E.g. what do I need to do to bring about this change in this population?
Behavioural scientist
The AI System
Detects patterns, makes inferences
The Promise
1. Remarkable success stories such as IBM Watson (playing Jeopardy), DeepMind (playing Atari games and Go) etc.
2. HBCP will leverage such approaches to teach the AI system to populate ontologies and suggest improvements to the structures
3. The AI system will be able to generate new insights and testable hypotheses about behaviour change
Challenges
1. Join knowledge from many sources• significant effort to identify connections
2. De-noise relevant knowledge• useful information represents small proportion of total content
3. Resolve content ambiguity• for precise semantics of ontology
4. Assign confidence to learned knowledge• assess evidence vs opinion
5. Connect rich semantic knowledge source to Machine Learning & AI• without combinatorial meltdown
The Interface
• Develop and evaluate an online open-access interface to • enable widespread use of the knowledge generated
• provide users with an easy method for intelligent searching of the behaviour change intervention literature and the inferences made from it
• interrogate and provide feedback from a wide perspective of views into the AI system – and the ontology developed by the AI System
• enable use by other computer programmes
Implementation
• Establish • Scientific Advisory Board
• International, interdisciplinary
• Users and Stakeholders Board• Co-design User Interface
• Partners e.g. • Cochrane and Campbell Collaborations
• Public sector and commercial users
Addressing the problem upstream
• Plan to publish template to encourage researchers to report studies using an ontological structure
• Will help• Clear, full reporting
• Data synthesis
• Interoperability i.e. link to other, related ontologies thus extending knowledge
Future collaborative projects
• HBCP will provide an infrastructure and tools that can enable collaborative projects e.g.
• Include cost-effectiveness evidence• Examine ethical aspects of how the HBCP can maximise
social benefit and minimise harm• Apply to individual-level datasets rather than aggregate-
level data as in published evaluation reports• Etc …
Conclusion
We need this research to:1. make rapid and efficient progress in advancing our
understanding of behaviour2. harness and develop the powers of AI for
effectively synthesising research evidence and generating new knowledge and hypotheses
3. make accessible up-to-date world literature on behavioural interventions ….
4. … for the benefit of:• Scientists• Policy-makers, practitioners and intervention designers• Public and commercial sectors, NGOs etc.
The Human Behaviour-Change Project
Questions and discussionwww.humanbehaviourchange.org
@HBCProject
For more information
• UCL Centre for Behaviour Change
– www.ucl.ac.uk/behaviour-change
• Susan Michie
All proceeds from CBC teaching, training, books and products go to further development
MSc in Behaviour Change
• Register now for September 2017
• Open to students from diverse backgrounds
• Full-time or part-time
• Cross-disciplinary
• Taught by world experts
• Links to placements
Course Directors:
Dr Leslie Gutman & Prof Susan Michie
www.ucl.ac.uk/behavior-change
