Impact and determinants of community-level SODIS-interventions : experience from a community-randomised trial on solar water disinfection

Unsafe drinking water, insufficient sanitation and hygiene behaviour and indoor air pollution are some of the most important environmental risks, which are harmful to health. They cause diarrhoea, pneumonia and sepsis, and are with 65% the leading cause of death in children under 5 years of age. Lack of safe water perpetuates the cycle whereby poor populations become further disadvantaged and poverty entrenched.
Point-of-use household water treatment and safe storage systems (POU-HWT) represent one of the most promising strategies to block major transmission pathways associated with contracting diarrhoea, especially in low income countries. Home-based solar water disinfection (SODIS) is one of those POU-HWT systems being promoted worldwide. The SODIS-method consists of exposing water-filled, transparent PET bottles to sunlight for at least 6 hours. The evidence base of the health effectiveness of SODIS from population-based scientific evaluations is limited and experiences from large-scale roll outs of SODIS interventions are equally scarce. The community-randomised trial described in this thesis investigates the health effectiveness and the factors that lead to the uptake of the intervention.
The main goal of this thesis was to assess the effectiveness of a Latin American community-level SODIS dissemination programme in reducing child diarrhoea and the determination of factors related to the adoption of SODIS among the population. From 2004 to 2006, a community-randomised controlled trial was thus, conducted in 22 community-clusters situated in Bolivia. SODIS was implemented in the intervention communities by a local NGO over 15 months. The promotion of SODIS consisted of interactive, repeated, and standardised events held on community and household level. In order to comprehensively describe both, application and uptake, and the health effect of SODIS, a health surveillance system with community-based staff measuring compliance with SODIS as well as the occurrence of diarrhoea in children under 5 was established. Since no standards to classify households according to their SODIS-use exist, different indicators for use were measured by evaluators independent from the implementing NGO. Further, surveys assessed household determinants as well as SODIS promotion and -campaign factors, which were associated with the SODIS adoption.
Despite the extensive SODIS promotion campaign a possible health impact in this typical rural Bolivian setting was too low to be assessed by this study. The intention-to-treat analysis of the 1-year health monitoring of 725 children <5 (425 households) detected no significant difference in diarrhoea morbidity between the intervention and the control communities. This finding is neither in line with the results of former trials assessing the health impact of SODIS, nor with the results of trials testing a variety of different other POU-HWT technologies. This inconclusive finding might be explained by the moderate compliance or the fact that most of the endemic diarrhoeal disease is not exclusively transmitted through the consumption of contaminated drinking water only, but rather transmitted from person to person by hands, food and other fomites due to poor hygiene practices. Those risk factors could potentially have disguised the health effect of SODIS. Supposedly a more ‘holistic’ approach including community water improvements, sanitation and hygiene may produce better health outcomes than SODIS as a POU-HWT alone.
During the project and before the cessation of the SODIS promotional activities we measured an overall compliance with the intervention of 32% at any given day during the study. In contrast, around 80% of households reported using SODIS regularly after the first phase and again at the end of the implementation. The SODIS-implementing NGO observed an average SODIS-usage rate of 75%. The remarkable discrepancies of compliance with SODIS registered in our study when assessed by different staff raise questions of how to interpret the compliance rates of other published SODIS-effectiveness studies. The occurrence of bias due to self-reporting and courtesy of villagers resulting potentially in an over-estimation of SODIS-use could be prevented in future SODIS evaluations if assessed by independent agencies.
The continuous monitoring by our community-based staff of adoption, application, rejection and discontinuation of the SODIS intervention allowed the identification of household determinants and SODIS-campaign factors leading to adoption or rejection of the method among different SODIS-user groups. The analysis revealed that households that were more likely to use SODIS were those that participated more frequently in SODIS promotional events, included women, owned latrines, and had severely wasted children living in the home. Those household-level factors are easily assessable indicators that SODIS-programme managers could use to identify early adopters in SODIS promotion campaigns.
In summary, this thesis demonstrates that the SODIS-method promoted in a typical rural Bolivian setting was not effective in reducing diarrhoeal disease in children under 5 years, despite a comprehensive SODIS-dissemination of a local non-governmental organisation. Unless the overall adoption and acceptance of POU-HWT methods, especially across broad levels of the population most in need will be considerably increased, the public health benefit will be modest. This research suggests that additional work is needed to better understand how the well-established laboratory efficacy of this POU-HWT method translates into field effectiveness under various cultural settings and intervention intensities.