Informing behaviour change intervention design using Bayesian meta-analysis: physical activity in heart failure

Abstract

AbstractEmbracing the Bayesian approach, we aimed to synthesise evidence regarding barriers and enablers to physical activity in HF in a way that can inform behaviour change intervention development. This approach helps in concluding on the uncertainty in the evidence and facilitates the synthesis of qualitative and quantitative evidence. Qualitative and observational studies investigating barriers and enablers to physical activity in adults diagnosed with HF were included in a Bayesian meta-analysis. Evidence from three qualitative and 16 quantitative studies was synthesised. Qualitative evidence was annotated using Theoretical Domains Framework and represented as a prior distribution using an expert elicitation task. The maximum a posteriori probability (MAP) and Credible Intervals (CrI) was calculated as a summary statistic for the probability distribution of physical activity conditioned on each determinant, according to qualitative evidence alone and qualitative and quantitative evidence combined. Evidence concerning the modifiable barriers and enablers is highly uncertain: social support (MAP=0.11,CrI:[0.08;0.13]), negative attitude (MAP=0.22,CrI:[0.17;0.27]), positive attitude ( MAP=0.27,CrI:[0.23;0.31]), self-efficacy( MAP=0.31,CrI:[0.29;0.33]), symptom distress (MAP=021,CrI:[0.18;0.24]). The contextual barriers – low, moderate and high uncertainty respectively – are age (MAP=0.22,CrI:[0.22;0.23]), low Left Ventricular Ejection Fraction (MAP=0.20,CrI:[0.19;0.22]), and depression (MAP=0.14,CrI:[0.12;0.16]). This work extends the limited research on the modifiable barriers and enablers for physical activity by individuals living with HF.