Designing internet interventions to test theories of behaviour change

1. Designing internet interventions to test theories of behaviour change Susan Michie University College London, UK EHPS Synergy, Bath, UK, 2008

2. This talk • What is theory • Testing theories of behaviour change • Advantages of internet for testing theory • Designing internet interventions to test theory

3. 1. What is theory • MRC-funded project to identify theories of behaviour change • Psychology, sociology, anthropology, economics • Shortlist of definitions • library catalogue searching • key reference sources e.g. American Psychological Association Dictionary and the Oxford Concise Dictionary of Sociology • Delphi exercise amongst study advisory group (n=12) • rated 12 definitions: • 0=not useful, 1=possibly useful, 2=useful, 3=very useful • definitions with a mean rating >1.5 selected for further analysis • Core constructs of resulting 7 definitions noted • definitions ranged from 1 to 6 constructs • constructs included by at least 50% of the definitions selected and synthesised as working definition

4. Our working definition “a set of concepts and / or statements with specification of how phenomena relate to each other. Theory provides an organising description of a system that accounts for what is known, and explains and predicts phenomena”

5. 2. Testing theories of behaviour change Describe phenomena – interventions and behaviour Map interventions onto theory and test theory by intervening Develop more effective interventions Refine theory

6. 2. Testing theories of behaviour change • Need to know how BC interventions relate to theoretical constructs that cause behaviour change • e.g. if want to test operant learning theory, need to know that an intervention (e.g. social support, money) is a positive reinforcer • Requires • Specifying interventions by component techniques • Mapping component techniques onto theoretical constructs causally associated with change

7. Specifying interventions into techniques • Interventions often “complex” • several, potentially interacting, techniques • Poorly described, difficult to replicate • Interventions often described vaguely e.g. “behavioural counselling” • Where protocols with more detail are available, terminology is variable

8. Example of the problem: Descriptions of “behavioural counselling” in two interventions

9. In biomedicine … “A big, oval red pill” would not be an acceptable intervention description …

10. Guidelines for specifying interventions • CONSORT guidelines for reporting RCTs • Evaluators should report “precise details of interventions [as] actually administered”Moher et al, 2001 • Which precise details? • the content or elements of the intervention • characteristics of those delivering the intervention • characteristics of the recipients, • characteristics of the setting (e.g., worksite) • the mode of delivery (e.g., face-to-face) • the intensity (e.g., contact time) • the duration (e.g., number sessions over a given period) • adherence to delivery protocols Davidson et al, Annals of Beh Med, 2003

11. The content: How to report? • Develop a parsimonious list (taxonomy) of conceptually distinct and defined techniques that can be reliably used in specifying interventions • An example • From 3 systematic reviews, identified techniques • inductively generated • Developed coding manual of definitions • Independent ratings to test reliability • 75% kappas >0.7 Abraham & Michie (2008) A taxonomy of behavior change techniques used in interventions. Health Psychology

12. Intervention techniques 14. Provide feedback 15. General encouragement 16. Contingent rewards 17. Teach to use cues 18. Follow up prompts 19. Social comparison 20. Social support/ change 21. Role model 22. Prompt self talk 23. Relapse prevention 24. Stress management 25. Motivational interviewing 26. Time management 1. General information 2. Information on consequences 3. Information about approval 4. Prompt intention formation 5. Specific goal setting 6. Graded tasks 7. Barrier identification 8. Behavioral contract 9. Review goals 10. Provide instruction 11. Model/ demonstrate 12. Prompt practice 13. Prompt monitoring

13. Further work …. • 35 techniques identified in reviews • 53 by brainstorming of 4 clinical/ health psychologists • 137 by consulting textbooks for training applied psychologists in BC techniques • Definitions extracted: 74.7% agreement between two raters • Need for work to map techniques and theories Michie, Johnston, Francis, Hardeman & Eccles (in press) Applied Psychology: An International Review

14. 3. Advantages of internet for testing theory • Characteristics • Precise specification of content and form • Control over what is delivered (fidelity) • Possibility of systematically and frequently changing intervention components • Large numbers of participants • Long-term follow-up • Automatic measures of contact time and patterns of interaction • Large, distal collaborations

15. These characteristics allow … • experimental testing • Of causal impact and interactions of constructs • Whole or parts of theory • systematic and rapid accumulation of evidence • assessment of • process as well as outcome • maintenance as well as change • powerful testing of effects, mediation and moderation • vary sequencing of intervention components • collaborative sharing/development of • intervention components and • designs for testing theory • e.g. testing one theory against another: focus on differences between theories, identify ones that are superior

16. 4. Designing internet interventions to test theory MOST: Multiphase Optimisation Strategy (Collins et al, 2007) • Principles drawn from engineering • More efficient than traditional cycle of • Intervention-RCT-post hoc analyses-revision of intervention-RCT • Very slow to optimise an intervention or test theory • Three phases • Screening • Refining • Confirming • Requires internet delivery re. rapidity of iterations and large amount of data generated

17. MOST: Multiphase Optimisation Strategy • Screening phase • Intervention components are efficiently identified for inclusion or rejection in an intervention, depending on their performance • Randomised experimentation via fractional factorial designs • Theory driven • Which components are having an effect, no effect or adverse effect? • Refining phase • Selected components are fine tuned and optimal levels ascertained • Randomised experimentation as above • Confirming phase • Optimised intervention evaluated in an RCT

18. Questions for discussion • Experiences people have had of testing theory using the internet? • Issues I haven’t thought about? • Other approaches?