Can Mass Media Interventions Reduce Child Mortality?

 "...[H]ow can we persuade national governments to spend money on promoting demand [for health services], rather than on midwives or medicines, when none of us know how important demand is? The challenge here is to quantify the importance of human knowledge and motivation relative to other spending imperatives."

In addressing this challenge, the authors of this Lancet Viewpoint piece outline an approach to implementing and evaluating mass media campaigns that have been developed to combat child mortality - as well as to garnering funding to support these campaigns. It is Development Media International (DMI's) Saturation+ approach, which is based on "saturation, science, and stories."

As the authors explain: "Our own experience, based on 20 years of delivering mass media campaigns, is that substantial behaviour change is possible when campaigns are broadcast with sufficient intensity." For example, they cite a study carried out in Ethiopia, where hygiene promotion messages were broadcast up to 14 times per day for 3 years. This campaign was notable for great reductions in observed dirty hands, falling from 74% to 26% (p<0·001) and for a 20% reduction in trachoma prevalence without the use of antibiotics. Saturation coverage, they argue, is the most important ingredient in a successful mass media campaign.

However, there are methodological problems associated with linking this success to the mass media intervention, such as the fact that studies often rely on cross-sectional before-and-after comparisons, so other potentially confounding factors could have affected the results. Also, many evaluations rely on self-reported behaviours: People may know the "correct" answers after an intensive campaign without actually changing their behaviour. Thus, a randomised controlled trial (RCT) of a media campaign is needed but is challenging when the national media has a strong and broad reach. Furthermore, "the effect of any single-issue campaign on all-cause mortality is modest and therefore hard to identify statistically."

The authors are, thus, approaching these challenges via a 35-month trial in Burkina Faso, which is an effort to "measure rigorously, using a cluster-randomised design, how many lives mass media can save in a low-income country, and at what cost." They explain the scientific reasoning behind the trial (summarised above yet elaborated in the article) and outline the research methodology. To begin, in order to estimate the potential effect of a comprehensive campaign, they gathered evidence from previous multiple-issue media campaigns to predict how much they could increase coverage of key interventions (such as breastfeeding or seeking treatment for pneumonia). They adjusted their predictions for service provision and media penetration in each country and then analysed the effect on mortality of these increases in intervention coverage in a range of sub-Saharan African countries using the Lives Saved Tool (LiST). "LiST predicts that a sustained, comprehensive campaign could reduce under-5 mortality by between 16% and 23% during the third and subsequent years of a campaign. If these predictions are correct, mass media campaigns would be among the most cost effective of all currently available health interventions."

To test these predictions, the authors developed a media methodology, which involved, first, addressing the challenge of randominising a mass media intervention by choosing a country (Burkina Faso) where local media is very strong and national media is relatively weak. Burkina Faso's private FM broadcast in local languages, whereas the national station broadcasts most of its outputs in French. The result is that, in most locations, national radio achieves only a 10% market share or less. Thus, they can broadcast messages in 7 intervention areas using local FM radio stations with limited range without leaking into the 7 control areas and without "sacrificing the factor crucial to the power of mass media (ie, that most people are listening)." Specifically:

• The saturation component entails broadcasting messages 6-12 times per day on market-leading radio stations or at least 3 times per day on market-leading television stations, using 60-second spots in local languages as the foundation of the campaign.
• The science component entails: (i) quantifying the geographic coverage and market share of media channels in different parts of the country and with different demographic groups and (ii) targeting behaviours that are predicted to save the most lives per dollar spent.
• The stories component involves building emotional identification between the audience and the characters as they advance through the early stages of the narrative (characterisation, their goals, the obstacles they face). "It is also important that the emotional climax of the story (the moment of decision, in which protagonists must either overcome the obstacles or revise their goals) addresses the crucial barrier to behaviour change, as identified by formative research. All stories are pre-tested for clarity and cultural sensitivity."

The evaluation, led by The London School of Hygiene & Tropical Medicine (LSHTM) and Centre Muraz, includes baseline and endline mortality surveys, as well as baseline (March 2012), midline [see Related Summaries, below, for a summary of, and access to these results], and endline (November 2014) behavioural surveys. Full trial results are expected in 2016. In the meantime: "Our model predicts that if comprehensive campaigns are implemented in ten African countries for 5 years, one million lives of children younger than 5 years should be saved. If this claim is sustained by the trial, saturation-based media campaigns should belong in the mainstream of public health interventions and a priority for governments."