Date: 
May 1, 2014

 

 

"There is a strong correlation between the intervention 'dose' (the number of weeks each message was broadcast) and the impact on behaviours..."

Development Media International (DMI) is conducting a 3-year randomised controlled trial to test the hypothesis that a radio campaign can reduce the large number of children dying before their fifth birthday in Burkina Faso. As detailed at the link below (see Related Summaries), the research involves the broadcast, beginning in March 2012, of health messages using radio spots (60-second adverts) and radio phone-in programmes. With the goal of engaging the public with health issues in order to change behaviours, this is a research initiative to evaluate a public health intervention delivered to 7 randomised geographic areas (clusters) across Burkina Faso with 7 additional clusters being used as controls.

Methodologies: 

Supervised by 2 people affiliated with the London School of Hygiene and Tropical Medicine (LSHTM), this independent survey is based on interviews with 5,000 mothers in the 7 intervention zones and 7 control zones halfway into the research intervention period (midline data).

Practices: 
  • Treatment-seeking
    • According to the researchers, the impact of the intervention on treatment-seeking behaviours is particularly important because of its contribution to the objective of reducing child mortality: the 3 main causes of under-5 deaths in Burkina Faso are: malaria (22% of under-5 deaths), pneumonia (13%), and diarrhoea (11%). Overall, the proportion of parents taking a sick child (with fever, cough, rapid breathing, or diarrhoea) to receive treatment at a health facility increased by 20.0 percentage points in intervention zones, compared to 11.1 in control zones (a "difference in difference" figure (based on cluster-level analysis) of 8.9). Adjusted for proximity to a health centre, the difference in difference is 12.9. ("Mortality levels were higher, and behaviour coverage levels lower, in our intervention zones; there are also twice as many health centres in the control zones as in the intervention zones.")
  • Diarrhoea:
    • The proportion of parents taking children with diarrhoea to a health facility for treatment increased by 20.1 percentage points in the intervention zones, compared to 8.6 in the control zones (a difference in difference of 11.5). Adjusted for proximity to a health centre, the difference in difference is 16.0. Parents in intervention areas are significantly more likely to have changed this behaviour compared to parents in control areas.
    • The number of parents providing oral rehydration therapy (ORT) or salts (ORS) or increasing the liquid intake of children who had diarrhoea increased by 25.5 percentage points in intervention zones, versus 1.5 in control zones (a difference in difference of 24.0). Analysed at cluster level, the difference in difference is 23.3. DMI focused less on promoting ORS in messages communicated during the first year of the initiative, as these have only recently become widely available in Burkina Faso. However, promoting ORS alone showed a difference in difference of 12 between the behaviour change in intervention zones and control zones.
  • Pneumonia:
    • The number of children who received antibiotics for pneumonia symptoms increased by 18.2 percentage points in the intervention zones, compared with 10.5 in the control zones (a difference in difference of 7.7). Adjusted for proximity to a health centre, the difference in difference is 14.8. (Researchers note that "[t]his is a very positive result despite the imbalance in access to health facilities between zones (at midline, 47% of women lived less than 2 km from a health facility in control zones, compared to just 16% in the intervention zones).") The proportion of parents seeking treatment for cough or fast/difficult breathing increased by 17.4 percentage points in intervention zones compared with 9.8 percentage points in control zones (an adjusted difference in difference of 10.3).
  • Malaria:
    • The proportion of parents taking children with fever to a health facility for treatment increased by 15.7 percentage points in intervention zones, compared to 8.7 in control zones (a difference in difference of 7.0). Adjusted for proximity to a health centre, the difference in difference is 9.1.
    • The proportion of women sleeping under an insecticide-treated bed net (ITN) during pregnancy increased by 17.5 percentage points in intervention zones, compared to 12.6 in control zones (a difference in difference of 4.9). Analysed at cluster level, the difference in difference is 3.4. (The researchers note that household ownership of ITNs was 95% at baseline in both intervention and control zones, "leaving little room for improvement.")
  • Sanitation and Hygiene:
    • The proportion of households with a latrine increased by 10.9 percentage points in the intervention zones compared to 9.6 in the control zones (a difference in difference of 1.3). Analysed at cluster level, the difference in difference is 2.3. "This result suggests that the considerable financial and practical barriers make it difficult to change this behaviour using a mass media intervention alone."
  • Nutrition:
    • The proportion of new mothers breastfeeding their babies within 2 hours of birth increased by 10.4 percentage points in the intervention zones, compared to a reduction of 2.5 percentage points in the control zones (a difference in difference of 12.9). Analysed at cluster level, the difference in difference is 10.7. "In some ethnic groups in Burkina Faso women discard their first milk, because they believe that colostrum is 'bad' or 'dirty', and do not know about its health benefits for their baby. The midline results suggest that our messaging is beginning to reduce these barriers to behaviour change."
    • The proportion of new mothers who were exclusively breastfeeding their babies aged 0 to 5 months increased by 12.2 percentage points in the intervention zones; surprisingly (for the researchers), it increased by 14.1 points in control zones (a difference in difference of -1.9). Analysed at cluster level, the difference in difference is -1.8. "Meanwhile, our qualitative research suggests that the impact of our ongoing campaign on exclusive breastfeeding may be subject to a time lag because of deep-rooted socio-cultural beliefs and practices, such as giving babies herbal 'tisanes'. We are conducting ongoing research to investigate this result."
  • Health Facility Delivery:
    • The proportion of mothers giving birth in a health facility increased by 9.3 points in intervention zones, compared to 11.5 in control zones (a difference in difference of -2.2). Adjusted for proximity to a health centre, the difference in difference is 0.2. "This was not a prominent message during the first 18 months of the campaign, and the low broadcast 'dose' may have reduced our impact."
Source: 

Burkina Faso RCT: Midline Results [PDF], accessed May 28 2014; and email from Cathryn Wood to The Communication Initiative on May 29 2014.