This study set out to evaluate the Familia Nawiri community health education program on risk reduction for diarrhea and respiratory infections. Findings from the study revealed favorable results for this hygiene education program. Overall, after adjusting for potential confounders, participants exposed to the hygiene education program had 38% lower odds of having diarrhea or a respiratory infection compared to participants who were not exposed to the hygiene education program (adjusted odds ratio 0.62, 95% CI: 0.41–0.96). Our findings are consistent with estimates from other studies assessing effects of hand washing with soap and WASH educational intervention on reducing the incidence of diarrhea and respiratory infections [
11,
13,
14]. On the other hand, recent RCTs assessing interventions promoting healthy behaviour related to WASH, in improving child health outcomes like diarrhea and proper hand washing behaviour have yielded negative results [
15‐
18]. For instance, a trial conducted in Rwanda assessing the impact of community-led health clubs promoting WASH interventions reported no effect on care-giver reported diarrhea among children below 5 years across the three arms of the study (control group, eight community health club sessions group and 20 community health club sessions group) [
15]. A trial conducted in Kenya with seven arms (including: control arm, water intervention, sanitation intervention, hand washing intervention, combination of water, sanitation and hand washing intervention, nutrition intervention and combination of all the interventions [water, sanitation, hand washing and nutrition]) found no reduction in diarrhea in any of the intervention arms [
16]. However, a trial from Bangladesh – with similar intervention arms as the study in Kenya – reported reduction in diarrhea in all intervention arms, except the intervention arm receiving water treatment only [
17]. Although some of these recent studies have reported no effect of the interventions tested, the benefits of WASH for diarrheal diseases and other health outcomes should not be underestimated [
37,
38]. The findings may also not be generalizable across all contexts and therefore should be viewed in light of the specific interventions and setting [
37,
39,
40]. With regards to the intervention and the approach used to deliver the intervention, comparable examples in other low and middle income countries in Africa, have been provided by Sinharoy et al., Waterkeyn and Cairncross, and Lewycka et al. [
15,
41,
42]. However, findings from these interventions are varied. Cairncross, and Lewycka et al. [
41,
42] found that community health clubs and women’s group interventions can be effective in achieving high levels of health knowledge and hygiene behaviour change in Zimbabwe and also to improve maternal and child health in Malawi. On the other hand, Sinharoy et al., [
15] found that the use of community health clubs as implemented under Rwanda’s national Community-Based Environmental Health Promotion Program in western Rwanda had no effect on health outcomes such as diarrhea in children under 5 years old.
It cannot be excluded with certainty that unmeasured bias or confounding e.g. recall bias with respect to exposure to the Familia Nawiri education sessions, residing in areas with water improvement initiatives and attending other health education programs, may have impacted the results in either direction, i.e. either increasing or decreasing the odds ratio. However, since our findings were statistically significant (95% CI: 0.41–0.96, P = 0.03), they provide some support of an association between hygiene education and morbidity of diarrhea and respiratory tract infections.
Some of the strengths of this study include the fact that both cases and controls were recruited from the same health facilities, therefore, taking regional practice aspects into consideration. Selection of cases and controls was within 5 days of attending the clinic which helped to minimize potential seasonal influence on the outcomes.
The findings from this case-control study should be interpreted in light of some limitations. There is potential risk to overestimate the health impact as a result of relying on self-reported measurements. This could have been avoided by employing direct observation methods at the household level. However, due to logistical reasons and inconveniences that such a method would entail, we decided to rely on the self-reported measurements. We believe however, that the resulting misinformation would have been non-differential between cases and controls – in other words - any misinformation that occurred while collecting the data at the health facility may have occurred to the same extent for both groups and would therefore rather mask than exaggerate the impact of the intervention. Another limitation for our study is the fact that the intervention was delivered to a highly selective target audience - participants in group activities such as church groups or women’s groups. It can be expected that these people differ in many psychosocial parameters that may not reflect in the measured socio-economic indicators. Misclassification bias with respect to exposure is also very likely for this study. This is mainly because there was no objective way of verifying reported or non-reported exposure to the education sessions. This has a bearing on how the participants recall exposure to the intervention. However, for this study, both the cases and the controls were selected from participants seeking medical attention at health facilities; hence, both cases and controls would have similar concerns regarding the causes of their illness making them comparable and thus minimizing differential recall bias. Approximately 60% of the cases and the controls reported having attended some form of hygiene education session apart from the Familia Nawiri program while 45% of the controls and 34% of the cases reported residing in areas where there had been water improvement initiatives. The presence of these other initiatives may account for the finding of no difference in knowledge of the causes of diarrhea among cases and controls. Although the selected subjects for the study were statistically similar, the fact that most of them were children and young adults could perhaps reflect ability to access a health facility and therefore our findings should be interpreted in that context. Since the cases and controls in this study were selected from health facilities in the area where the health education program had purposefully been implemented, we cannot be certain that the findings from this study apply to people in other geographical locations or other health facilities. It is also important to note that recruitment at the health facility level probably shifts to more severe cases that require attendance of the health facility. Therefore, it is not clear whether the results would also apply to less severe cases of diarrhea and respiratory infections. Another limitation for this study is the fact that we did not consider the difference in population size of the three regions when enrolling participants to the study which could in turn affect the validity of the results. A final limitation for this study is the fact that we did not utilize theory of change in assessing the effect of the community health education program as the assessment was largely focused on clinical health outcomes and nor was an implementation framework utilized in the implementation of the community health education program. We are therefore limited in our understanding of how or why changes happened as a result of the implemented community health education program or which aspects of the program, if not all were effective.