Using impairment and cognitions to predict walking in osteoarthritis: A series of n-of-1 studies with an individually tailored, data-driven intervention

Objectives First, this study compares the ability of an integrated model of activity and activity limitations, the International Classification of Functioning, Disability and Health (ICF), and the Theory of Planned Behaviour (TPB) to predict walking within individuals with osteoarthritis. Second, the effectiveness of a walking intervention in these individuals is determined. Design A series of n-of-1 studies with an AB intervention design was used. Methods Diary methods were used to study four community-dwelling individuals with lower-limb osteoarthritis. Data on impairment symptoms (pain, pain on movement, and joint stiffness), cognitions (intention, self-efficacy, and perceived controllability), and walking (step count) were collected twice daily for 12 weeks. At 6 weeks, an individually tailored, data-driven walking intervention using action planning or a control cognition manipulation was delivered. Simulation modelling analysis examined cross-correlations and differences in baseline and intervention phase means. Post-hoc mediation analyses examined theoretical relationships and multiple regression analyses compared theoretical models. Results Cognitions, intention in particular, were better and more consistent within individual predictors of walking than impairment. The walking intervention did not increase walking in any of the three participants receiving it. The integrated model and the TPB, which recognize a predictive role for cognitions, were significant predictors of walking variance in all participants, whilst the biomedical ICF model was only predictive for one participant. Conclusion Despite the lack of evidence for an individually tailored walking intervention, predictive data suggest that interventions for people with osteoarthritis that address cognitions are likely to be more effective than those that address impairment only. Further within-individual investigation, including testing mediational relationships, is warranted. Statement of contribution What is already known on this subject? N-of-1 methods have been used to study within-individual predictors of walking in healthy and chronic pain populations An integrated biomedical and behavioural model of activity and activity limitations recognizes the roles of impairment and psychology (cognitions) Interventions modifying cognitions can increase physical activity in people with mobility limitations What does this study add? N-of-1 methods are suitable to study within-individual predictors of walking and interventions in osteoarthritis An integrated and a psychological model are better predictors of walking in osteoarthritis than a biomedical model There was no support for an individually tailored, data-driven walking intervention