Achieving impact

DMI does not have a standard, one-size-fits-all method for designing and implementing mass media health campaigns. We do, however, build our campaigns on the basis of our Saturation+ approach for achieving impact. This has three elements: saturation, science and stories.
 


Saturation
 

Broadcasting messages several times a day for a sustained period

Intensity is key to any commercial advertising strategy, but it has been an underrated element of public health campaigning. There is a strong correlation between the frequency of broadcast of messages and their impact:

  • For our most successful campaign, in Ethiopia, we broadcast radio adverts 6-14 times per day for three years.
  • In Burkina Faso we are broadcasting 60-second radio adverts 10 times a day, seven days a week, in six languages. 
  • We are also broadcasting two 15-minute health-based dramas every weekday evening, as part of a live radio show. 

 

Devising radio and TV formats that can be produced quickly enough to enable frequent broadcasts

Producing enough messages to enable a high-intensity broadcasting schedule is challenging. We believe that, if the aim is to save lives, news and documentary formats are not very useful. These formats stress what is new and interesting, and are not based on repetition. Instead we place particular emphasis on two radio and TV formats:

  • Short advertising spots (based on drama, comedy or testimonial), which put across a basic message repeatedly. 
  • Longer dramas, which are useful for providing role models, demonstrating life-skills or allowing interactivity. These must use a format that can be produced cheaply and quickly enough to be broadcast every day, in many languages.

 

Working with media networks that reach the largest possible proportion of the target audience

Intensity of broadcasting needs to be matched by scale. This can be achieved by working with the radio or TV networks with the greatest numbers of listeners or viewers among the target audience:

  • Audiences in rural areas may listen to community radio stations that broadcast in local languages in preference to national networks, even if the signal of the national networks extends to those areas. 
  • A detailed economic and audience analysis will suggest the broadcasting strategy that generates the biggest return on investment, depending on the campaign’s objective and target audience.

 


Science
 

Predicting and maximising health impacts

We design and implement behaviour change campaigns to promote child and maternal survival. Our mathematical model predicts the impact on child and maternal mortality of each message in each country. For example, in DRC: 

  • A behaviour change radio campaign can save 14,400 under-five lives every year. 
  • This assumes that we can reach 5 million radio listeners over the age of 15 at 'saturation' intensity (or 15 million radio listeners at lower intensity).
  • We can predict impact by disease (eg for under-fives: 34% diarrhoea, 26% pneumonia, 25% malaria, 15% neonatal).
  • We can predict impact by behaviour (eg increasing rates of exclusive breastfeeding from 37% to 47%).
  • We can predict impact by province (eg most under-five lives will be saved in Orientale and Bandundu provinces).
  • We can predict the cost-effectiveness of our campaigns (a nationwide campaign in DRC would cost $6.40 per DALY).

 

Messaging on multiple behaviours

Wherever possible we message on multiple health issues and behaviours, because the economies of scale make it much more cost-effective than to set up and run a media campaign that is only focused on a vertical issue. One approach is to cover each behaviour for two weeks at a time, repeating key messages at certain points. The duration, frequency and relative weight attached to each message will vary depending on the following factors:

  • The predictions of our mathematical model (eg handwashing can save 5,900 under-five lives, ITNs can save 3,000).
  • The availability and quality of services and supplies, by province, that are needed to enable those behaviours (the mathematical model takes supply-side availability at national level into account when calculating lives saved).
  • Seasonal or other time-related factors, such as the malaria season, or co-ordination with other health initiatives.

 

Measuring and attributing health impacts

We can translate a series of behaviour change outcomes into health impacts (lives saved) using our model, but first we need to be able to measure the degree of behaviour change robustly, and to attribute those changes to our campaign rather than to any other initiatives. The ideal evaluation design is a randomised controlled trial, but this is not feasible or affordable in most cases. We have therefore developed a set of techniques for measuring and attributing the health impacts of our campaigns, including quasi-experimental evaluation designs. We use a combination of the following:

  • Baseline and endline cross-sectional surveys to measure trends in knowledge, attitudes and practice.
  • Dose-response relationships between behaviour change and target groups with low, medium and high exposure.
  • Triangulation of survey data with external data sources, including DHS, MICS and health system usage statistics. 
  • Smaller quarterly surveys of knowledge, attitudes and practice to enable a time-series analysis of impact.
  • Comparisons of behaviour outcomes between intervention areas and ‘natural’ control zones (eg media-dark areas).

 


Stories
 

A qualitative research process to help us to understand our audience

We cannot change behaviours if we do not understand the values, motivations and concerns of our target audience. We send teams of researchers to conduct detailed studies using focus groups and interviews, often in remote areas:

  • Formative research (identifying barriers to behaviour change) informs a ‘message brief’ for our scriptwriters.
  • Radio spots are pre-tested with focus groups representative of our target audience, to judge clarity and appeal.
  • Feedback research asks whether and why people who heard our messages have changed their behaviours (or not).

 

A creative process to generate content that engages audiences and changes behaviours

Our experience of producing radio and TV health behaviour change materials in developing countries suggests that:

  • They should be simple, funny, and engaging, convincing people to change their behaviours, rather than simply providing information.
  • Short, realistic dramas (in local languages) are the best format for changing behaviours: short enough that rural audiences can find the time to listen regularly (often while they work); realistic because people prefer stories that are rooted in their everyday lives, performed by local actors; and dramas, because the human brain is hardwired to respond to emotion better than it responds to intellectual reflection, and dramas play on that emotion, influencing our choices and behaviours.
  • A good creative approach miniaturises the Hollywood scriptwriting process, which structures films in three parts that replicate the process of behaviour change: Act I (where characters are given goals), Act II (where obstacles are thrown in front of those goals), and Act III (where characters change their goals or overcome the obstacles).
  • It is possible to create systems that encourage creativity, such as an editorial control process that reduces 30 script ideas per month to six that are selected for pre-testing, and four that are selected for production.
  • The most important, but often most neglected, element is to recruit a team with the necessary talent, motivation, and organisational, intellectual and creative abilities. Often, recruiting locally through an open scriptwriting competition rather than on the basis of previous experience is the most reliable way to find the top creative minds.

 


Find out more or get involved
 

         

Read about our Media Million Lives initiative

Download a PDF summary of DMI

Watch short videos
about our approach

Register to receive
 email updates

Contact our London office for a chat