Introduction The determinants of health and disease can be classified as proximal and distal depending on their position in the causal pathway. This study attempts find out whether behaviour change communication interventions targeted at the proximal determinants (knowledge and infrastructure) of handwashing and water treatment change the relative importance of distal socio-economic status determinants.

Methods The analysis uses data from an evaluation of a Safe Water System intervention in urban slums of Haridwar and Dehradun districts of Uttarakhand, India. As a part of evaluation, two independent surveys were conducted in the SWS intervention area in 2004 (n=1125) and 2005 (n=1128). Analysis was based on framework of hierarchical modelling. Logistic regression was applied and the change in “population attributable fraction” (PAF) from baseline to endline models was calculated for arriving at the relative importance of the determinants.

Results At baseline, distal determinants were important for water treatment. Introduction of software variables to the model resulted in decline in the PAF of distal determinants from 89% to 74%. At endline, the importance of distal determinants was further reduced; (PAF of 23%). For handwashing at baseline, the distal variables were less important for handwashing (PAF of 36%). After intervention, there was smaller decline in their importance (PAF of 22.7%). Proximal determinants assumed importance only after intervention. The models in the study could only partially explain the variance in two behaviours under study.

Conclusions Behaviour change communication interventions can help overcome influence of low socio-economic status or illiteracy by changing proximal determinants.