N-of-1 methods: A practical guide to exploring trajectories of behaviour change and designing precision behaviour change interventions

Objectives: 
(1) To introduce N-of-1 methods and how they can help the researchers identify predictors of behavioural outcomes, (2) to provide examples of studies that test individual theory-based predictions of physical activity and/or exercise; (3) to provide a practical example dataset to illustrate how to design and undertake a basic analysis for an N-of-1 study; and (4) to suggest a future agenda for N-of-1 physical activity and exercise research.
Design: 
Factors for consideration when designing an N-of-1 study include variability of predictors and outcomes, assessment frequency and appropriate analysis methods. Existing literature and piloting can help inform these aspects.
Methods
We use a dataset of 24 individuals who collected data over 28 days to illustrate example analysis procedures. Data, guidance and associated SPSS and R syntax are made available to provide researchers with tools to learn about and practice N-of-1 analysis.
Results: 
Guidance on dealing with missing data, looking at graphical representations of N-of-1 data, managing autocorrelation using the prewhitening method and analysing N-of-1 datasets is provided. Using the example dataset, we demonstrate how to identify antecedents of physical activity (steps) to assess directionality of associations. We also include an overview of aggregating N-of-1 datasets using multilevel modelling.
Conclusions: 
N-of-1 methodology provides a means of tracking individual patterns of behaviour and identifying potential antecedents of physical activity and exercise to help determine causality. Assisted by mobile technologies, there is great potential to enrich our understanding of movement behaviour using this approach to inform interventions.