Sustainability of Solar Disinfection to Provide Safe Drinking Water in Rural Peru

The World Health Organization estimates that 1.8 million people die yearly from diarrheal diseases, with 90% of deaths in children younger than 5 years of age.¹ Disease results from the consumption of untreated or improperly treated drinking water contaminated with viral, bacterial, or parasitic agents.² In Peru, diarrhea is the second most common cause of death in those younger than 5 years of age.³ Solar disinfection (SODIS) is a simple, low-cost technology for disinfecting drinking water.^4,5 SODIS involves filling plastic bottles with untreated water and then exposing them to the sun for six hours. Turbid water must first be filtered, and the bottles are placed on roofs exposed to sunlight. A combination of increased temperature and the ultraviolet A band radiation destroys or inactivates waterborne pathogens.⁶

In the laboratory^7–16 and the field,^17–23 SODIS inactivates viruses, protozoa, and bacteria that cause diarrheal diseases. The reduction of the microbiological content of drinking water is similar to other methods such as chlorination and filtration.²⁴ Due to its microbial efficacy, affordability, accessibility, and ease of use, SODIS has been widely adopted initially.^25–27 To our knowledge, however, no published research exists on the continued use of SODIS in communities over extended periods of time.

In contrast, various studies have examined factors associated with the initial adoption of SODIS. In Bolivia, Moser et al. found that early adopters were involved in the topic of safe drinking water and water treatment. These were often people who worked as health promoters or aqueduct and water technicians. Late adopters did so only when a majority of their peers supported the technology.²⁵ In Nicaragua, Altherr et al. found that intention to use and actual use were associated with a positive attitude about SODIS, including knowledge of contaminated drinking water’s association with diarrheal diseases.²⁸ A study in Bolivia found that current and intended SODIS use was associated with reduced costs and perceived better taste of SODIS water.²⁶ A study in Zimbabwe cited disease prevention and lack of safety of untreated drinking water as reasons for using SODIS.²⁹

The primary objectives of this pilot study were to (1) estimate the current prevalence of SODIS use among families in 2008 that were targeted in 2001 for SODIS intervention, and (2) describe the -association of -motivating factors and impediments to current SODIS use. We identified an early SODIS intervention in Peru (in which one member of this research team participated), where SODIS was instituted in 11 rural communities near Arequipa following an earthquake. Seventy-five percent of the population reported using SODIS at the end of a 10-month introductory campaign.²⁷

Given its practicality and efficacy, we hypothesized that SODIS use was sustained in rural Arequipa even though there had been no reinforcement or organized efforts in health education since the initial SODIS project of 2001. We further hypothesized that factors such as the presence of small children and recent episodes of diarrheal illness would be associated with increased use of SODIS.

Methods

Study design

A study was designed to determine current prevalence of SODIS use (Objective 1) and motivators and impediments (Objective 2). The first phase involved two community meetings and informal interviews (qualitative phase). The second phase involved a household survey that comprised a questionnaire and observational component in five communities (quantitative phase). We analyzed the second phase using both descriptive statistics and case-control analyses.

Study setting

Study participants were recruited from five of the 11 small communities in the Polobaya district of Arequipa, Peru, that were targeted for SODIS use in 2001. The 11 communities were similar in the following respects: water sources, size, region, and language. We selected five of the communities based on personal contacts and proximity. While there are many Quechua speakers in Peru, the study population speaks Spanish. This population was described in 2001 as living in conditions of extreme poverty and having a low level of education. Livelihoods were described as primarily subsistence agriculture, unskilled labor, and informal business.²⁷

Local Peruvian Red Cross volunteers were recruited to facilitate community meetings and administer questionnaires. Interviewers had backgrounds in health promotion and recent experience with focus groups and the administration of questionnaires (one member of the team [Victor Quispe] was involved in the prior implementation of SODIS in 2001). Obtaining written and oral permission from local authorities preceded all phases of this study.

Ethics board approval

We obtained Institutional Review Board (IRB) approval from Asociación Benéfica Proyectos en Informática, Salud, Medicina y Agricultura (PRISMA), a nongovernmental organization based in Lima, Peru, as well as Columbia University in New York City (Protocol IRB-AAAD6801). All study participants gave written consent prior to participation.

Phase one: community meetings

In phase one, we facilitated a one-hour community meeting in each of two communities (out of the 11 total) targeted for SODIS in 2001, during which motivators and impediments to SODIS use were identified. Local health promoters who had participated in the original intervention invited participants to attend. Approximately 20 participants attended each meeting. Information obtained from the community meetings was used to inform the development of the questionnaire.

Phase two: survey implementation (questionnaire and observation)

The structured questionnaire was administered to 69 out of the 90 total households in five communities (two of which had participated in phase one—the community meetings) targeted for SODIS in 2001 to quantify current SODIS use and assess motivators and barriers to use. There was also an observational component (bottle sightings), which aimed to further confirm current SODIS use (Objective 1). Questions were based upon those used by Heri et al.²⁶ and adapted with information obtained during the community meetings. Respondents were asked 10 questions about trips to the local health clinic because of diarrhea, perceived safety of untreated drinking water, usage of in-home water purification techniques (e.g., boiling and chlorination), knowledge of SODIS, past and present SODIS use, positive aspects and obstacles to SODIS use, and household demographics. SODIS users were asked why they liked the method, and nonusers (based on self-report) were asked why they disliked the method. While these questions were open-ended, the answers fit into predetermined categories identified from both the literature review and community meetings.

Sampling and recruitment

Female respondents were purposefully interviewed due to the fact that they had been targeted for participation during the intervention in 2001. If adult females were not available, males were asked to participate.

Sample size estimation

Based upon a power calculation for an unmatched case-control study with α=0.05, a prevalence of 25% for SODIS use, and precision of estimates ±10%, we calculated that we needed a sample size of 50. To calculate the sample size for assessment of the relationship of each variable with the outcome of interest, we did the power calculation for a case-control study. Cases were families currently using SODIS based on self-report; controls were families not using SODIS. Using OpenEpi version 2.2.1, and assuming exposure of 80% of case families, a total sample of 50 families had 80% power to detect a prevalence of exposure of 40% in the control population.³⁰

Data management and analysis

We conducted Chi-square analyses of variables we hypothesized might be associated with SODIS use by using SAS®.³¹ All variables associated with use at p≤0.2 were considered for inclusion in a multivariate logistic regression. However, only place of residence met this criterion, so we stopped at the bivariate stage of analysis. The comparison of proportions was made between use in Arequipa in 2001,²⁷ when 312/416 (75%) families reported SODIS use, and the 29/69 (42%) families who reported SODIS usage in 2008.

Results

Community meetings

Motivators for continued SODIS use identified were (1) availability of plastic bottles; (2) belief that SODIS reduces incidence of diarrheal disease; (3) belief that SODIS is less of an economic burden than boiling water, which requires fuel; and (4) belief that SODIS water tastes better than boiled water (respondents reported a smoky taste in boiled water).

Impediments identified included (1) belief that SODIS water causes stomachaches, (2) belief that SODIS is too time intensive, (3) lack of information regarding SODIS (some people reported never having learned the method), (4) lack of interest in using SODIS, and (5) belief that newly installed piped water was potable.

Survey (questionnaire and observational component)

Of the 90 households in the five communities, sixty-nine families were available on interview day. For each household, an adult female (25 in Agua Buena, nine in Candelaria, 15 in Bella Vista, 12 in Totorani, and eight in San Jose) was invited to interview (Table 1). Nine households (13%) elected to have a male interview. The 69 respondents represented 297 household residents. Households had an average of 4.3 people.

Table 1.

Characteristics of residents living in five communities in the rural Polobaya district in the Province of Arequipa, Peru, 2008

^aColumns add up to >100% as some households use multiple water treatment methods.

To answer Objective 1, we found that 42% of respondent families reported continued SODIS use (Table 1). The validity of this estimate is unclear, however, as only 32% of the respondents that reported continued SODIS use were observed to have bottles placed on their roofs on the day that the questionnaires were administered. Each family home was visited once to administer the questionnaire and make observations for rooftop SODIS bottles.

All respondents reported treating their drinking water by boiling it (100%), while others, to a lesser extent, used SODIS; few people reported chlorination (7%). Despite the prevalence of in-home water purification practices, trips to the local health clinic for treatment of diarrhea were common, with 54% of respondents reporting a trip to the local health clinic within the last year (Table 1).

Drinking water sources varied in the five communities. Untreated water from public taps was the most common source of water in all communities except for San Jose, which only had access to river water (Table 1).

Using Chi-square analyses, we examined the association between SODIS use and variables related to water sources, knowledge and practices, and family demographics; however, we did not find any significant associations (Table 2).

Table 2.

Characteristics of SODIS users and nonusers residing in five communities in the Polobaya district in the Province of Arequipa, Peru, 2008

SODIS = solar disinfection

OR = odds ratio

CI = confidence interval

Ref. = referent group

Comparison of SODIS users and nonusers

The mean age of SODIS users was 45 years and for nonusers was 42 years. SODIS users on average lived in the community longer than nonusers: a mean of 32 years for SODIS users vs. 28 years for nonusers. Table 2 shows odds ratios and associated p-values of eight variables, identified through community discussions to perhaps be associated with SODIS use.

Among the users, 27% responded that they liked SODIS because of a perceived lessening of cases of diarrhea, 28% replied that SODIS was more economical (didn’t have to burn firewood to boil water), 28% noted the improved taste of SODIS water in comparison to the smoky taste of boiled water, and 18% cited the ease of use of SODIS, which only requires plastic bottles and sunlight. Nonusers gave five reasons for disliking SODIS: 9% of respondents said that SODIS water gave them a stomachache, 18% commented that SODIS takes too much time, 9% replied that they did not use SODIS because they never learned how, 9% of people said they were not interested in using SODIS, and 56% responded that they didn’t use SODIS because they believed the newly installed piped water (tap water) was potable (data not shown).

Discussion

Hypothesis testing

The aims of this study were to estimate the prevalence of SODIS use seven years after its introduction following an emergency (earthquake) and to identify and evaluate motivators of and barriers to SODIS use among respondents who currently used and did not use SODIS.

Forty-two percent of respondents indicated that their family currently used SODIS. Compared with the 75% reporting SODIS use after the original intervention (2001), and considering the seven-year interval, this finding suggests that SODIS is a sustainable method. These results, while lower, are believed to be consistent (considering the seven-year interval) with the 60% prevalence of use reported by Moser et al. and Heri et al. However, these authors did not report the interval from the introduction of SODIS until assessment of later prevalence of use.^25,26 The results of the case-control analysis, however, did not suggest that any of the putative factors that may facilitate or impede SODIS sustainability had any effect (Objective 2). This could be because of true lack of effect, or because we did not adequately capture the factors that might influence SODIS use. The case-control analysis failed to confirm our second hypothesis that the presence of small children or recent episodes of diarrheal illness would be associated with increased SODIS use.

Our study provided other information concerning potable water. There was no use of purchased bottled water, and few respondents indicated chlorine use. As in the past, when information was lacking on boiling as a water treatment option,³² there is scant literature concerning the current sources of water or the treatment of water in economically distressed communities in the developing world. There is also limited information on the determinants of alternative methods for water treatment. One case study revealed that there could be underlying objections to chlorination as a water-treatment option based on its other uses. For example, in an indigenous Kuna community in Panama, chlorine serves two functions: fishermen often use bleach to dislodge octopus from nearby reef systems and women use it to wash clothing. As a result, discussion of chlorination as a water treatment option was met with strong resistance because of its perceived potent and perhaps harmful effects.³³

A sobering reality is that 77% of families said that they have taken a child to a local health clinic for treatment of diarrheal disease, 54% of whom indicated they have done so within the last six year. Combined with the reported universal application of the boiling method (100%) and significant use of SODIS (42%), there appears to be a large and persistent problem despite families’ reported attempts to make their water potable.

Limitations

This study had several limitations. First, respondents might have tried to answer in a manner that they thought would be pleasing to the interviewer. To minimize this potential bias, interviewers tried to set the interviewees at ease with small talk prior to administering the questionnaire. In addition, interviewers tried not to reveal specific study interests. Interviewers introduced themselves saying they would like to ask a few questions about drinking water usage and water-purification methods in general. SODIS-specific questions were only raised after asking about seeking care for diarrheal disease, knowledge and attitudes toward potable water, and family demographics.

A second limitation was the sample size. None of the variables in the case-control analysis were found to be statistically significant, which could be attributable to the small sample, or the fact that there is no association between these variables and the sustained use of SODIS. The large confidence intervals suggest that our study had inadequate sample size.

A third limitation was the inconsistency between the reported use of SODIS and the lack of visible SODIS use at the respondents’ homes. The respondents may have overestimated their current use of SODIS. There was also a lack of validation of other responses such as prevalence of water boiling and treatment at a local health clinic for childhood diarrhea.

Conclusion

Given the limited information about the prevalence of SODIS use years after targeting, and the lack of information about factors associated with the use of SODIS, the modest power of this study is adequate for the purposes of generating preliminary estimates. Further and more definitive studies are needed to assess the sustainability of SODIS.

However, results were favorable. After seven years, 42% of families that were using SODIS immediately after the intervention reported still using it, indicating that SODIS could be a sustainable, important, and practical tool for reducing child mortality associated with diarrheal disease. The prevention of waterborne illness necessitates a clear assessment of whether SODIS is effective in the short term and sustainable.

SODIS is a promising new intervention. For 12 years, SODIS projects have been implemented worldwide.³⁴ Now is an appropriate time to revisit communities that have introduced SODIS to estimate whether it is sustainable and effective in preventing diarrheal disease and to identify factors that facilitate or impede its adoption and sustained use. This study will help inform larger efforts in the future.