Are interventions delivered by mobile phone effective in targeting behavioural risk factors for non-communicable diseases? A systematic review and meta-analysis.

Problem

An estimated 38 million deaths occur each year due to non-communicable diseases (NCDs). The World Health Organization has highlighted the potential of m-health technologies in the prevention and control of NCDs. Our systematic review aims to synthesise the evidence relating to the effectiveness of mobile phone-based interventions targeting NCD behavioural risk factors: physical activity, diet, tobacco use, and alcohol consumption.

Approach

We searched Medline, EMBASE, Global Health, and CINAHL (1990 - January 2016). Quality was assessed using Cochrane collaboration criteria. Primary outcomes were objective measures of outcomes related to the target behaviours. Secondary outcomes were self-reported measures. Two reviewers carried out data extraction.

Findings

76 studies met the inclusion criteria: Interventions targeted smoking cessation (n=18); physical activity (n=47), diet, or a combination of both; and harmful alcohol consumption.(n=8) Four studies were categorised as being at low risk of bias for all quality criteria. Interventions included 1-19 behaviour change techniques. SMS-based interventions increased continuous smoking abstinence (pooled RR: 2.19, 95% CI 1.80-2.67). SMS-based physical activity interventions had a borderline statistically significant effect on daily step count (pooled MD: 1256.9, 95% CI -159.7, 2673.6). Meta-analyses indicated that SMS-based interventions targeting diet and physical activity promoted greater weight-loss (pooled MD: -1.77 kg, 95% CI -2.95, -0.58), reduced the cumulative incidence of diabetes (pooled RR 0.67, 95% CI 0.49, 0.90) and had a borderline significant effect on BMI (pooled MD: -0.84, 95% CI -1.69, 0.01), however there was no evidence of effect on endpoint weight, cholesterol, or blood pressure. Pooled analysis of four trials of app-based interventions targeting diet and physical activity showed no effect on weight loss (pooled MD -0.94kg, 95% CI -2.47, 0.60). Four of six trials of interventions targeting alcohol use reported statistically significant self-reported outcomes.

Consequences

There is good evidence for the benefit of smoking cessation support delivered by text message on objective long term outcomes in high income countries. There is mixed evidence of benefit for interventions delivered by sms and no evidence of benefit from apps targeting diet and physical activity. High quality trials measuring objective long term outcomes for the most promising alcohol reduction, diet and physical activity interventions are needed. National smoking cessation interventions delivered by sms are already being implemented in England, USA, New Zealand, India, Phillipines and Tunisia. Further evaluation of the effects of interventions in low and middle income countries is needed.