Tools for sharing evidence for effectiveness from a systematic review of complex interventions to improve food security: Which do policy makers prefer?

Article type
Authors
Kristjansson E1, Dubois A1, Svensson K1, Szijarto B2, Welch V1, Thompaon H3, Liberato S4, Lawrence M5, Armstrong R6, Burns C7, Barnett BM1, Donoso J1, Houssain A1, Platts J1, Labelle P1, Salewski E8
1University of Ottawa
2 University of Ottawa
3NHS Health Scotland, MRC Social & Public Health Sciences Unit
4Menzies School of Health Research, Darwin
5Deakin University
6Cochrane Public Health, Melbourne School of Public Health
7Asylum Seekers
8Ottawa Public Health
Abstract
Background: We are completing an equity-focused review of community interventions to improve food security. At study onset, we assembled seven policy makers at regional, national and international levels; and they provided input into several aspects of the review. The review and the interventions within it are complex; we need to share results in ways that are clear, attention-grabbing and relevant.

Objectives: 1)To compare ways of presenting results to policy makers; and, 2) to understand the impact of involving policy makers in data interpretation.

Methods: We followed standard Cochrane procedures for the review. Synthesis methods included Random Effects meta-analysis and narrative synthesis with Effect Direction plots to visualise data. To ensure that our Advisory Group has input into data interpretation and knowledge translation, we are holding a ‘Data Dive’ workshop in May 2017.

Results: Searches retrieved 26 178 articles. After culling by title, we screened 2456 abstracts; 507 articles were reviewed in depth and 40 studies were included. Ten assessed the impact of the US WIC program, another concerned income supplementation for pregnant mothers in Canada. Seven studies assessed Meals on Wheels, nine assessed vouchers or rebates for fruits and vegetables. The remainder assessed the impact of Food Stamps, healthy corner stores, new supermarkets in deprived areas, a novel food pantry and the Good Food Box. Effectiveness varied by intervention, process factors and outcome. For example, we found a significant effect of rebates on purchase of fruits and vegetables; those who received them purchased 818 grams more of fruits and vegetables per week (95% CI 147.5, 1488.1).

At the Data Dive, we will study reactions to different ways of presenting evidence, including data placemats, infographics, forest plots and Effect Direction plots. We will incorporate Advisory Group interpretations of the evidence and policy relevance into the review.

Conclusions: This presentation will highlight data-presentation tools that were most attractive to policy makers; we will also discuss the ways in which we incorporated Advisory Group Input.