RCT Impact Data: Health, Media, Communication

[Please share your questions and analysis of the research outlined below. Either enter a comment below and/or open the Health Communication network network link for this discussion and share your perspectives].

One criticism of the communication and media for development, social and behaviour change field of work is that there is little high quality evidence for its impact. In particular those outside this field of work ask us for evidence based on that supposed gold standard of research designs - the Randomized Control Trial (RCT).

Of course we have a large amount of impact evidence using other quality research designs. But those with a strict science perspective will always look for the three "magic" letters "RCT" as the basis for assessing whether the evidence presented is credible and reliable.

Well, now we have an RCT! Its midline results (the final results will be available in 2015) facilitated the researchers to make this statement: "There is a strong correlation between the intervention 'dose' (the number of weeks each message was broadcast) and the impact on behaviours…"

Here are some examples of the ‘difference in difference’(the additional behaviour change in the intervention zones compared with the control zones) as measured at midline (and adjusted for confounding variables):
1. Parents taking children with diarrhoea to a health facility for treatment: 16.0 percentage points
2. Children receiving antibiotics for pneumonia symptoms (fast or difficult breathing): 14.8 percentage points;
3. Mothers initiating breastfeeding within two hours of birth: 10.7 percentage points;
For a full summary of the midline results please go to this link where there is also a link to the DMI paper

This independent research is funded by the Wellcome Trust and Planet Wheeler Foundation. The radio campaign is being implemented by Development Media International (DMI), and the evaluation is being supervised by the London School of Hygiene and Tropical Medicine (LSHTM).

It is a four-year randomised controlled trial in Burkina Faso, with three years of broadcasting. The research tests the hypothesis that a radio campaign can reduce the large number of children dying before their fifth birthday. The RCT is directly measuring the impact of the radio campaign on child mortality, with a 50,000 sample size at baseline and 100,000 at endline.

Because of the nature of the Burkina Faso media there are unique media footprints for each of the 14 zones. These are community media stations. There is weak national media penetration. Consequently it is possible to broadcast in intervention zones with minimal risk of ‘contaminating’ the control zones.

DMI wishes to stress that the research is a test of their strategy which is called "Saturation+". In Burkina Faso this involved, in the intervention zones, broadcasting 60-second radio adverts 10 times a day, seven days a week, in six languages on the major Burkina Faso health issues; and, two live radio dramas on health issues every weekday evening, as part of the primetime evening show on each station, since 2012.

The midline results for each of these issues include:

a. Treatment for diarrhoea at a clinic - 16% improvement difference in intervention zones compared to control zones.

b. Received ORT or increased liquids for diarrhoea - 23.3% improvement difference in intervention zones.

c. Received antibiotics for pneumonia (fast/difficult breathing) - 14.8% improvement difference

d. Sought treatment for fever at a clinic - 9.1% improvement difference

. Women sleeping under a bed net during pregnancy - 3.4% improvement difference

f. Household ownership of latrines - 2.3% improvement difference

g. Early initiation of breastfeeding (2 hours of birth) - 10.7% improvement difference

h. Exclusive breastfeeding aged 0 to 5 months - 1.8% negative difference in intervention zones compared to control zones

i. Gave birth in a health facility or with skilled attendant - no difference

j. Saved money for an emergency during pregnancy - 8.5% improvement difference in intervention zones compared to control zones

The DMI paper outlining these results can be accessed at the bottom of the online summary

The Board of DMI includes: Richard Horton - Editor-in-Chief, The Lancet; Joy Phumaphi - Executive Secretary, African Leaders Malaria Alliance; and, David Heymann - Chair, UK Health Protection Agency.

The Scientific Advisors include: Simon Cousens - Professor of Epidemiology, London School of Hygiene and Tropical Medicine (LHSTM); Cesar Victora - Professor of Epidemiology, Federal University of Pelotas; Bocar Kouyaté - Technical Advisor to the Minister of Health, Burkina Faso; and Anne Mills - Professor of Health Economics and Policy, LSHTM.

Please share your questions and analysis of the research outlined above. We would much welcome your comments and questions on this research. Please discuss this research in the Health Communication network at this network link.

What questions do you have on the results?

What questions do you have about the methodology?

DMI has kindly agreed to review and respond to questions and comments.

Thank you