Our own Lorraine Landais got the opportunity to present parts of her PhD project at the last ISBNPA conference. Choice architecture (CA) is considered a promising approach to change health behaviors; however, its effectiveness in increasing physical activity and reducing sedentary behavior, especially after its removal, remains unclear.This systematic literature review aims to provide an overview of the effectiveness of CA interventions that promote physical activity or reduce sedentary behavior.
CA interventions were defined as interventions that aim to change behavior by altering the presentation of a choice in the information, physical or social micro-environment in which people make decisions. Following PRISMA guidelines, PubMed, Embase, PsycInfo and the Cochrane Library were searched for articles published up to February 2018. Additionally, backward and forward reference searches were conducted. Studies were eligible for inclusion if they (quasi)experimentally tested the effect of CA on either intention, behavior (physical activity or sedentary behavior) or health outcomes among an adult population. We excluded studies that did not meet our CA definition or combined CA with other behavior change techniques.After eligibility screening, articles were assessed for methodological quality (QualSyst tool), data was extracted and findings were narratively synthesized by type of choice architecture and type of outcome measure.
Of the 4888 unique records identified, 83 studies (81 on physical activity, 3 on sedentary behavior) met our inclusion criteria. In presence of CA, 57 studies found a significant effect, whereas 14 studies found no significant effect. When CA was no longer present, 20 studies found a significant effect, whereas 11 studies found no significant effect. Most studies (N=56) were conducted in the physical environment and mainly consisted of prompts to promote stair use, such as posters, signs, stair-riser banners and footprints on the floor. Studies in the information environment (N=34) mainly studied the effects of message framing, such as gain-framed messages versus loss-framed messages. Only 7 studies applied CA in the social environment. These studies consisted of intervention elements such as social norms, social comparison, modeling and social support.
Most CA interventions effectively increase physical activity or decrease sedentary behavior, although this is more pronounced in presence of CA than in absence. Suggestions for future research are to conduct more studies on sedentary behavior and in the social environment and to examine how the effectiveness of CA interventions can be sustained after their removal.