Innovative methods for observing and changing complex health behaviors: four propositions

Abstract

Precision health initiatives aim to progressively move from traditional, group-level approaches to health diagnostics and treatments toward ones that are individualized, contextualized, and timely. This article aims to provide an overview of key methods and approaches that can help facilitate this transition in the health behavior change domain. This article is a narrative review of the methods used to observe and change complex health behaviors. On the basis of the available literature, we argue that health behavior change researchers should progressively transition from (i) low- to high-resolution behavioral assessments, (ii) group-only to group- and individual-level statistical inference, (iii) narrative theoretical models to dynamic computational models, and (iv) static to adaptive and continuous tuning interventions. Rather than providing an exhaustive and technical presentation of each method and approach, this article articulates why and how researchers interested in health behavior change can apply these innovative methods. Practical examples contributing to these efforts are presented. If successfully adopted and implemented, the four propositions in this article have the potential to greatly improve our public health and behavior change practices in the near future.

Characterizing and predicting person-specific, day-to-day, fluctuations in walking behavior

“…The first set of results (i.e., the identification of sudden gains and losses), is in accordance with both (i) previous empirical results, showing that psychological, physiological, social and behavioral systems, when observed at a high-resolution, might changes in a non-linear fashion (e.g., [26,32]), and (ii) complex systems theory assumptions, which postulate that any natural system is "bubbling" with change and tends to evolve in a non-stable manner [53]. Our observations thus provide support for several recent position papers arguing that changes in health behaviors can be fruitfully studied with research designs, methods, and statistics that reflect these non-linearities [8,9,54]. Further, at the theoretical level, results from this study support the adoption of complex systems theory to better understand the dynamics of health behaviors over time and across contexts [6,16].…”

“…Previous assumptions about gradual behavior change following adoption and maintenance phases should also be refined, or at least completed, to better account for rapidly changing phase transitions. At the methodological level, results from this study suggest that future studies interested in the dynamics of behavior changes should transition from (i) lowto high-resolution behavioral assessments to better capture potential non-linearities in those processes; (ii) group-only to group-and individual-level statistical inference to accurately modelled individual variability; and (iii) static to adaptive and continuous tuning interventional designs to better account for rapidly changing behavioral states and contexts (see [9] for further justifications). If successfully applied, that type of adaptive interventions would also argue against the utilization of traditional randomized control trials to test interventions in the physical activity context, at least as they are often implemented, i.e., with low-resolution behavioral measures, exclusively using a nomothetic approach and fixed intervention components.…”

“…Evaluation of physical activity interventions typically incorporate measurement bursts, i.e., regular measurement of physical activity for 3 to 7 days, that are dispersed across time, such as immediately before taking part in an intervention, immediately after an "intervention phase", and, occasionally, a few weeks or months after the intervention is completed (e.g., [12]). Using this measurement approach, physical activity levels within each measurement burst are often aggregated at the weekly level (e.g., minutes/week of moderate-to-vigorous intensity physical activity), and insight into the high-resolution day-to-day variations in physical activity removed through aggregation [9]. As both a cause and consequence to this methodological paradigm, physical activity changes have been mostly conceptualized as a relatively linear and deterministic process (e.g., "the adoption and maintenance paradigm"; [13][14][15][16]).…”

“…If strong non-linear patterns of change are observed, then the current measurement approach, which largely relies on snapshots of behavior across time, would be demonstrably insufficient to understand behavioral dynamics. Consequently, there would be a need to move from low-to highresolution physical activity assessments [9]. If results illustrate non-linear patterns of change, this would also establish the need for behavioral theories of physical activity that better account for the dynamic, non-linear and potentially complex nature of physical activity [6,7].…”

“…Although several explanations for this relative failure have been proposed (see for example [4,5]), little prior research has been conducted to understand the day-to-day, "high-resolution", and long-term changes in physical activity behavior, despite clear theoretical and practical benefits to do so. Indeed, gaining a better understanding of the resolution of change could provide important information to refine current physical activity guidelines and theories to reflect a more dynamic perspective and, ultimately, inform a new forms of quasi-real time physical activity interventions [6][7][8][9]. The present study aims to improve our understanding of day-to-day changes in physical activity within individuals over months.…”