Interactive voice response interventions targeting behaviour change: a systematic literature review with meta-analysis and meta-regression.

5 Repeated evidence

What this means for you: this is established.

Type of study: meta-analysis or systematic review
Studied in: people
Participants: not stated in the abstract
Published: 2018 in BMJ open
Cited: 26 times by other research
Who paid: Publicly funded, as far as was declared.
Source: PubMed 29478016 · doi 10.1136/bmjopen-2017-018974

The plain-language write-up of this study is still being written. Below is the abstract exactly as the researchers published it.

Abstract by the researchers — original text

OBJECTIVE: A number of promising automated behaviour change interventions have been developed using advanced phone technology. This paper reviewed the effectiveness of interactive voice response (IVR)-based interventions designed to promote changes in specific health behaviours. METHODS: A systematic literature review of papers published between January 1990 and September 2017 in MEDLINE, CINAHL, Embase, PsycINFO, SCOPUS and the Cochrane Central Register of Controlled Trials (CENTRAL) was conducted. From the total of 2546 papers identified, 15 randomised control trials (RCTs) met the eligibility criteria and were included in a random effects meta-analysis. Meta-regression analysis was used to explore whether behaviour change techniques (BCTs) that were used in the interventions were associated with intervention effectiveness. RESULTS: Meta-analysis of 15 RCTs showed that IVR-based interventions had small but significant effects on promoting medication adherence (OR=1.527, 95% CI 1.207 to 1.932, k=9, p=0.000) and physical activity (Hedges' g=0.254, 95% CI 0.068 to 0.439, k=3, p=0.007). No effects were found for alcohol (Hedges' g=-0.077, 95% CI -0.162 to 0.007, k=4, p=0.073) or diet (Hedges' g=0.130, 95% CI -0.088 to 0.347, k=2, p=0.242). In the medication adherence studies, multivariable meta-regression including six BCTs explained 100% of the observed variance in effect size, but only the BCT 'information about health consequences' was significantly associated with effect size (β=0.690, SE=0.199, 95% CI 0.29 to 1.08, p=0.000). CONCLUSION: IVR-based interventions appear promising in changing specific health behaviours, such as medication adherence and physical activity. However, more studies are needed to elucidate further the combination of active components of IVR interventions that make them effective and test their feasibility and effectiveness using robust designs and objective outcome measures.