An assessment of methods to measure handwashing behaviours

Significant global health attention and promotion has been focused on handwashing with soap due to the clear benefits observed in promoting and ensuring child health. However, the measurement and evaluation of handwashing behaviours remains complex. Demographic and Health Surveys (DHS) have measured handwashing using spot-check observation of the ownership of specific handwashing materials together with self-reported behaviour measures, although the validity of these approaches are questionable. Previous research has indicated that there is likely to be substantial over-reporting of handwashing behaviours using self-reports. Structured observation remains the most appropriate method to examine and understand handwashing behaviour, although issues also exist with this method.

Using data from a large-scale hygiene and sanitation project in Bangladesh (the UNICEF SHEWA-B project) this research assesses the validity of handwashing indicators through comparison of structured observation and responses to cross-sectional survey measures (spot-check observation and self-reported handwashing). Methods employed include simple bivariate analyses between observational and cross-sectional survey responses and logistic regression analyses to model response divergence of the different data collection methods for different handwashing exposures.

The results of this study indicate that the percentage of women handwashing with soap at various crucial times is low. Handwashing behaviours were over-reported compared with structured observation findings, indicating that current estimates of handwashing from the DHS are likely to also be overestimates. Multivariate analyses showed specific groups of women were more likely to give divergent answers. DHS should therefore reformulate their questions on handwashing behaviour and use other well-validated measures.