The objectives of this study were to show which determinants commonly explain handwashing with soap and how well they do so and to identify population-specific determinants to be targeted in different contexts.
The study collates data from more than 5000 households in total, from nine surveys conducted in eight different countries: one study each in Haiti, Ethiopia, Chad, Gambia, Guinea Bissau, Senegal, and Zimbabwe and two studies in different provinces of Burundi. Questionnaires with predefined questions were used in face-to-face interviews to measure the psychosocial factors of the Risks, Attitudes, Norms, Abilities, and Self-regulation (RANAS) model (find the questionnaire here).
Handwashing Questionnaire (pdf)
These psychosocial factors were correlated with handwashing with soap by regression analysis. This statistical analysis allows a calculation to define how well the difference between high-frequency handwashers and low-frequency handwashers can be explained (explained variance). It also shows the relative degree of influence that a factor exerts on handwashing. By merging the results from several studies, it is possible both to determine how well the RANAS factors can explain handwashing and to extract the psychosocial factors that seem to influence handwashing universally, those that might be case and country specific, and those that do not seem to play a role.
From the table below, you can learn the following:
The first two columns on the left show the factor blocks and behavioral factors of the RANAS model. They are followed by columns for the countries in which handwashing surveys were conducted. At the bottom, we have rows with ‘Sign. Add. Factors’ which denotes which additional factors significantly influenced the behavior. ‘N’ is the number of households interviewed, and ‘Adj R’ is the explained variance (a measure of how well the factors can explain the behavior). In the cells of the table, ‘+’ indicates that the factor had a positive influence on handwashing, ‘-’ means that it had a negative influence, ‘0’ indicates that it had no significant influence, and an empty cell means that this factor was not measured in this study.
The most interesting column is the last one on the right-hand side of the table. ‘Relevance in percentage of cases’ means that we first count the number of studies in which the factor had a significant influence, and second count the total number of studies in which the factor was measured. We then divide the first number by the second one (see the term in brackets) and multiply it with 100, which gives us the percentage.
The results reveal that handwashing with soap can be explained very well using the RANAS model, with a mean of 47% explained variance. The behavioral factors that most commonly influence handwashing are Others’ behavior, Others’ approval, and Feelings. This means that in most cases observing handwashing by relatives, friends, and neighbors is a strong motivator for an individual’s handwashing. Likewise, the affirmation by others that they support and approve of handwashing is also a driver. In addition, feelings associated with handwashing, such as liking and pleasantness or disgust when not performing handwashing, were a main driver of handwashing. The factors of Severity, Beliefs about costs and benefits, Remembering, and Confidence in performance are influential in half of the cases, and Health knowledge, Action planning, and Commitment are significant in only a third of the cases.
The first important conclusion from this review is that promotion activities should mostly target feelings and social norms relating to handwashing. However, other factors are also relevant in specific socio-cultural contexts. This means that, to design population-tailored interventions, all the factors of the RANAS model should be tested for their significance for handwashing in each socio-cultural context. The review gives clear recommendations which behavioral factors to address as priorities and the corresponding behavior change techniques to be used in contexts where formative research might not be possible.
Poster of the review (pdf)