The Value of Educational Messages Embedded in a Community-Based Approach to Combat Dengue Fever: A Systematic Review and Meta Regression Analysis

BACKGROUND: The effects of various dengue control measures have been investigated in previous studies. The aim of this review was to investigate the relative effectiveness (RE) of different educational messages embedded in a community-based approach on the incidence of Aedes aegypti larvae using entomological measures as outcomes. METHODS AND FINDINGS: A systematic electronic search using Medline, Embase, Web of Science and the Cochrane Library was carried out to March 2010. Previous systematic reviews were also assessed. Data concerning interventions, outcomes, effect size and study design were extracted. Basic meta-analyses were done for pooled effect size, heterogeneity and publication bias using Comprehensive Meta-analysis. Further analysis of heterogeneitity was done by multi-level modelling using MLwiN. 21 publications with 22 separate studies were included in this review. Meta-analysis of these 22 pooled studies showed an RE of 0.25 (95% CI 0.17–0.37), but with substantial heterogeneity (Cochran's Q = 1254, df = 21, p = <0.001,). Further analysis of this heterogeneity showed that over 60% of between study variance could be explained by just two variables; whether or not studies used historic or contemporary controls and time from intervention to assessment. When analyses were restricted to those studies using contemporary control, there was a polynomial relationship between effectiveness and time to assessment. Whether or not chemicals or other control measures were used did not appear have any effect on intervention effectiveness. CONCLUSION: The results suggest that such measures do appear to be effective at reducing entomological indices. However, those studies that use historical controls almost certainly overestimate the value of interventions. There is evidence that interventions are most effective some 18 to 24 months after the intervention but then subsequently decline.