Introduction

This paper is a report on an experimental intervention that was carried out by the authors in 2001 in Lima, Peru. It describes the entire process, including the design, implementation, and evaluation of an intervention that was intended to improve upon hygiene practices, with a specific goal of encouraging behaviours for prompt and sanitary removal of child feces. The authors developed a health communications strategy that included use of pamphlets, video presentations, and counselling within the framework of existing routine health services. While the intervention materials and approach were well received by both the community and participating health workers, the limited time frame of the intervention prevented the authors from developing statistically significant data and conclusions on the effectiveness of the intervention.

Key Points

Diarrhoeal diseases remain a major contributor to childhood mortality and morbidity, and increasing focus has been given to the promotion of improved domestic hygiene as a preventative strategy. This strategy is two-track, and involves the improvement of sanitary hardware facilities as well as encouraging changes in behaviour such as hand washing. The authors note that even in cases where the infrastructure is lacking, behaviour modification can have positive impacts. Yet, relatively little is known about feces disposal practices and the relevant determinants and feasibility for change. The authors sought to answer some of these questions by designing and implementing an experimental intervention in the densely populated shanty-town areas of Lima, Peru, which are characterised by limited piper water and sewage systems, leading to widespread surface defecation in the immediate area and in hastily built and typically shallow latrines.

When designing the intervention the authors sought to answer questions such as: 1) which practices put children at risk; 2) which practices are a priority for intervention; 3) who the intervention audience should be; 4) how can the link between hygiene and diarrhoea be built upon and used to motivate behaviour changes; and 5) what communication channels would be most effective. The formative research that was carried out to answer these questions is only briefly summarised in this article and is mostly contained in another article by the same author. The research revealed that demonstrating proper toilet training techniques to parents by exposing them to messages through existing health services would be the most effective and feasible option.

The authors developed a 20-minute, soap-opera style video, pamphlets, and a training package for community nurses who were then to introduce mothers to the information each time they attended clinics or health posts over a period of 6 months. The intervention was introduced into 4 communities while 4 more where retained as control communities, all of which were monitored using a variety of survey techniques. These included exit interviews with mothers leaving health clinics, consultations with health personnel and a widespread pre- and post-implementation survey that was carried out in both the intervention and control groups. Two types of evaluation indicators were assessed: impact indicators, such as the use of potties among children aged 15-47 months, and process indicators, such as the level of exposure to the intervention messages. Other data was collected to help inform the impact assessment including household characteristics, knowledge of hygiene issues, and information on other non-target hygiene behaviours. The behaviours were measured through structured observations within the home, while other information was conducted using structured questionnaires. The specific hypothesis (whether potty usage increased or not relative to exposure) was not revealed to participants in the structured observations and the questionnaires relating to exposure to the intervention materials were asked after the observations.

The results of the study provided insight into several areas. Both the audiences and the health workers involved in the project liked the soap-opera style of the video, with many mothers suggesting that it be used for other health messages rather than the standard lecture format. Yet, the intervention and the evaluation were plagued by relatively low exposure rates to the materials. Only about 10% of eligible mothers had seen either the video or the pamphlet at mid-term, and at the post-intervention assessment the number had only risen to 19%, though the percent exposed varied from 13-27% between the 4 communities within the intervention group. Exit interviews revealed that only about 40% of mothers who had been consulted at the health centres had been exposed to the materials. There was also a difference in exposure between mothers of different educational levels, ranging from only 11% of those with primary or less, to 26% of those with secondary or more. The authors attribute these low and variable coverage rates to several factors. Despite the training provided, the intervention remained a relatively low priority for nurses because the intervention was not on a long list of activities that the Ministry of Health tracks and evaluates the nurses on, evaluations which can have a substantial effect on nurse job prospects. In addition it was discovered that the nurses are potentially less likely to expose women with lower educational levels to more information.

With these relatively low levels of exposure in hand the authors faced the difficult challenge of assessing the impact of the intervention. The authors noted that message retention (measured on 5 point scale, overall mean = 1.59) was higher in areas where exposure was better and was higher amongst women with more education (1.89). The result of the hygiene practices observation was however, inconclusive. While potty usage increased somewhat over the course of the study, it increased in both the intervention and control groups, with the control group actually demonstrating a higher rate of potty usage (46% to 41%) than the intervention group in the follow-up evaluation. This phenomenon, combined with the low exposure numbers makes it impossible for the authors to make any definitive statements as to the effectiveness of their intervention.

The authors suggest that the short time period of the study was a major barrier to greater coverage, as was the reliance on existing routine health delivery structures for interventions rather than specific agents. They admit that this study actually provides more evidence as to the difficulty of carrying out such interventions within routine systems than it does to the intervention itself. Nevertheless, the authors note several other lessons learned from this essentially failed exercise, including the importance of collecting process indicators which would have otherwise left them unable to explain the low outcomes. The authors conclude with a methodological discussion on how they might better be able to test the health communication package rather than the efficacy (or lack thereof) of the delivery mechanism, as was the case in this study.