Unpacking the black box: theory-based design & evaluation of a multiple behaviour change intervention to control diarrhoeal disease.

Handwashing with soap, exclusive breastfeeding and the use of oral rehydration salts and zinc to manage diarrhoea episodes are apparently simple practices through which many child deaths from diarrhoea could be prevented. However, their success depends on caregiver behaviour, which has proved difficult to influence. This thesis describes the development and evaluation of a theory-based, multiple behaviour change intervention to improve uptake of these diarrhoea control practices in Lusaka Province, Zambia. The research was guided by a conceptual framework based on an ecological theory of behaviour and comprised three studies: formative research, outcome evaluation and process evaluation. Formative research focussed on actual practices of target behaviours and their determinants and informed the design of an intervention based on the motive affiliation. Outcome evaluation used a cluster-randomised trial to assess the effect of the intervention on handwashing with soap, exclusive breastfeeding, and use of oral rehydration salts and zinc to manage childhood diarrhoeal episodes. The intervention resulted in a small improvement in reported practice of exclusive breastfeeding. Changes in the other target behaviours were observed in areas where higher reach was achieved. The process evaluation used mixed methods and was guided by the intervention’s theory of change to explore how intervention content and delivery influenced behavioural outcomes. The process evaluation revealed that the intervention was popular and memorable, but it achieved low and variable levels of reach and did not change psychological mediators of behaviour. This research adds to the currently small evidence base in the field of multiple behaviour change for diarrhoea control. In light of the findings of low reach, further work is needed to ascertain whether delivery alone is the issue, or whether the intervention content also requires revision. The findings underscore the importance of modifying intervention delivery strategies to account for the local context.