Do automated digital health behaviour change interventions have a positive effect on self-efficacy? A systematic review and meta-analysis

ABSTRACT

Self-efficacy is an important determinant of health behaviour. Digital interventions are a potentially acceptable and cost-effective way of delivering programmes of health behaviour change at scale. Whether behaviour change interventions work to increase self-efficacy in this context is unknown. This systematic review and meta-analysis sought to identify whether automated digital interventions are associated with positive changes in self-efficacy amongst non-clinical populations for five major health behaviours, and which BCTs are associated with that change. A systematic literature search identified 20 studies (n = 5624) that assessed changes in self-efficacy and were included in a random-effects meta-analysis. Interventions targeted: healthy eating (k = 4), physical activity (k = 9), sexual behaviour (k = 3) and smoking (k = 4). No interventions targeting alcohol use were identified. Overall, interventions had a small, positive effect on self-efficacy $(\overline{g}=0.190,CI[0.078;0.303\left]\right)$. The effect of interventions on self-efficacy did not differ as a function of health behaviour type (Q-between = 7.3704, p = .061, df = 3). Inclusion of the BCT ‘information about social and environmental consequences’ had a small, negative effect on self-efficacy $(\mathrm{\Delta }\overline{g}=-0.297,Q=7.072,p=.008)$. Whilst this review indicates that digital interventions can be used to change self-efficacy, which techniques work best in this context is not clear.

KEYWORDS:

• Self-efficacy
• digital
• health behaviour
• behaviour change techniques

Disclosure statement

No potential conflict of interest was reported by the authors.

Data availability statement

The data that support the findings of this study are openly available in OSF at https://osf.io/vf52h[DOI10.17605/OSF.IO/VF52H].

Notes

1 These reviews both used predecessors of the 93-item taxonomy used in the present review. A 26-item version was used by Prestwich et al. (2014) and a 40-item version used by Williams and French (2011).

2 These reviews both used predecessors of the 93-item taxonomy used in the present review. The 26-item version used by Prestwich et al. (2014) labelled this BCT as ‘provide information on consequences’, whereas the 40-item version used by Williams and French (2011) labelled it as ‘provide information on the consequences of behaviour in general’.

Additional information

Funding

This work was supported by a Coventry University Early Career Researcher Pump Priming Award.