Escherichia coli Contamination of Water for Human Consumption and Its Associated Factors in Peru: A Cross-Sectional Study

ABSTRACT.

The objective of the study was to determine the factors associated with the presence of Escherichia coli contamination in water supplies for human consumption in Peru. A secondary analysis of the Food and Nutrition Surveillance by Life Stages survey (VIANEV) of 2017–2018 was performed. The presence of E. coli contamination in the water samples for human consumption of the households evaluated was defined as a dependent variable. A supply was considered contaminated when there was at least 1 colony-forming unit of E. coli in 100 mL of water for human consumption. Data from 886 participants were analyzed. It was found that 25.2% of household water supply sources for human consumption had E. coli at the time of sampling. Water reservoirs such as buckets or other containers (adjusted prevalence ratio [aPR]: 1.15; 95% confidence interval [CI]: 1.18–1.93), households belonging to a poor wealth quintile (aPR: 1.82; 95% CI: 1.01–3.25), residing in a rural area (aPR: 1.36; 95% CI: 1.01–1.83), and having a low human development index (aPR: 2.12; 95% CI: 1.15–3.91) were more likely to contain E. coli in water supplies for human consumption. However, households with chlorine concentrations of 0.5 mg/L or more in water (aPR: 0.20; 95% CI: 0.11–0.33) and with household members with a higher education (aPR: 0.67; 95% CI: 0.45–0.99) were less likely to contain E. coli in drinking-water supplies. From 2017 to 2018, one in four Peruvians had contamination by E. coli in the water supply to their homes, which was associated with sociodemographic factors, management, and water treatment.

INTRODUCTION

Water pollution is one of the most common environmental problems in the world.¹ In 2015, the United Nations established 17 Sustainable Development Goals (SDG), which constitute a universal call to improve the lives and prospects of people around the world.² Despite SDG No. 6 seeking to guarantee the availability of safe water and sanitation for all people (through actions aimed at reducing the contamination of this natural resource), approximately two billion people consume water contaminated with feces.² Likewise, in 2019, water contamination generated more than 65 million disability-adjusted life years, more than 7 million years lived with disability and more than one million deaths.³ These figures are related to the quality of water for human consumption, characterized according to physical, chemical, and microbiological parameters⁴ and that must remain within acceptable limits to avoid the development of diseases associated with contaminated water.

Water contamination is related to the transmission and spread of pathogenic microorganisms that generate infectious gastrointestinal diseases due to the ingestion of water with human and animal feces from contaminated water supply sources.^5,6 One of the main pathogens that serves as an indicator of fecal contamination in water supplies is Escherichia coli,⁷ for which a level of zero per 100 mL of water is considered safe for consumption.⁴ This microorganism is a facultative anaerobic Gram-negative bacterium found in the gastrointestinal tract of people (as a component of the colonic microbiota) and animals.⁸ In addition, E. coli presents long-term growth in environmental habitats such as lakes, rivers, beaches, soils, aquatic animals, and plants.⁹ Due to the virulence factors and phenotypic characteristics of this microorganism, it has several strains that are differentiated into diarrheagenic E. coli, nonpathogenic E. coli, and extraintestinal pathogenic E. coli,¹⁰ among which diarrheagenic strains are one of the main causes of diarrhea in people, especially in low- and middle-income countries (LMIC),¹¹ where diarrheal diseases are considered a public health problem, and health systems have a shortage of economic resources for the management of these pathologies.

In LMIC, E. coli is most often found in nonpurified water sources, such as unprotected wells, unprotected springs, tanker trucks, and surface water (rivers, lakes, irrigation canals, ponds, and streams).^12,13 In addition, a higher frequency of this microorganism is observed in households with the poorest compared with the richest quintile and in rural compared with urban areas due to communal sharing of water supplies within the neighborhood exposing people to multiple pathways of fecal–oral contamination.^12–16 In Peru, the water supply of 19.6% of homes is not provided by the public network, and this is especially of note in rural areas, where there are unfavorable water and sanitation conditions compared with urban areas^17–19 and there is a greater detection of E. coli in water for human consumption.^20–22 These problems determine the presence of infectious diseases related to the consumption of contaminated water, an increase in the burden of disease and mortality by these diseases.

Although the main strategies related to the improvement of the quality of drinking water, sanitation, and hygiene (such as the WASH strategy) have generated significant improvements around the world, there are social, economic, institutional, and geographic barriers that make it difficult to improve these indicators, especially in LMIC, where diseases related to the consumption of contaminated water and inadequate hygiene are a persistent public health problem.²³ Most epidemiological studies in Peru have focused on the prevalence and associated factors of E. coli in specific regions located in the highlands of Peru, and a representative sample of the Peruvian population is needed to characterize fully the prevalence and factors contributing to E. coli contamination at a national level.^18–22 Therefore, the objective of this study was to determine the factors associated with the presence of E. coli contamination in water supplies for human consumption in Peru.

MATERIALS AND METHODS

Study design and population.

We performed a cross-sectional analytical study of data obtained from the Food and Nutrition Surveillance by Life Stages survey (VIANEV Spanish acronym) conducted in the period from 2017 to 2018.²⁴

The VIANEV 2017–2018 survey collected information on consumption and anthropometric and biochemical indicators to describe the nutritional status of the Peruvian adult population.²⁴ The VIANEV 2017–2018 survey was made up of three random and independent subsamples (adults from 18 to 59 years old, children under 3 years old, and children from 6 to 13 years old) from the National Household Survey sampling frame of the first quarter of 2017. Information was collected in all households in the subsamples, in which adolescents aged 12 to 17 years and adults aged 60 years and older were found to reside and be members of these households.²⁴ This survey is representative at the national level because it collects data in three strata corresponding to Metropolitan Lima, and the rest of the urban and rural areas. The VIANEV 2017–2018 survey used stratified, multistage, probabilistic, and independent sampling, in which sample selection was carried out in two stages. The first stage involved the selection of conglomerates (composed of a set of 120 households on average), selecting 621 randomly chosen conglomerates (176 in Metropolitan Lima, 260 in the rest of the urban area, and 185 in the rest of the rural area), while in the second stage dwellings with adults aged 18 to 59 years of age were randomly selected.²⁴

The sample size of the VIANEV 2017–2018 survey was based on 1,211 adults of both sexes nationwide, and the sample was divided into three strata: Metropolitan Lima (557 adults), remaining urban (256 adults), and rural (398 adults). To assess the quality of water for human consumption, data from samples of water supplies for human consumption from the homes of 911 adults aged 18 to 59 years were used.²⁴ A total of 886 water samples from the homes of the participating subjects were included in the present study according to the inclusion (subjects aged 18 to 59 years and with data on E. coli in water for human consumption) and exclusion criteria (incomplete data on the independent variables) (Figure 1).

Figure 1.

Flowchart of participants selection.

Peru is divided into three regions based on its geographic features: the coast, on the shores of the Pacific Ocean, which has the most developed urban areas; the highlands, which has the highest altitudes in the nation due to its location within the Andes Mountain range; and the jungle, with its Amazonian biodiversity and a population with limited access to healthcare.²⁵ The natural coastline region has the greatest socioeconomic development and has the largest population density, followed by the highlands and the jungle, in which rural poverty is the norm.²⁵

Variables and measurements.

Determination of the quality of water for human consumption.

The VIANEV 2017–2018 survey considered the determination of the minimum adequate concentration of residual or free chlorine in water for human consumption as an indicator of water quality. Determination of this element in water for human consumption was made with the visual colorimetric method (color comparison) using the Aquamerck^® Chlorine test reagent. The processing of this reagent is based on direct contact with free chlorine in samples of water for human consumption collected from adult homes, which produces a yellow coloration. The intensity of coloration is compared on a color chart, with each color being equivalent to a given chlorine concentration. According to the chlorine determination chart, the chlorine concentrations are as follows: no chlorine, chlorine (0.1 mg/L), chlorine (0.25 mg/L), chlorine (0.5 mg/L), chlorine (1.0 mg/L), and chlorine (2.0 mg/L). A concentration of 0.5 mg/L is considered an adequate minimum concentration of residual chlorine.⁴

Determination of microbiological quality in water for human consumption.

For the determination of the microbiological quality of the water for human consumption used for the preparation of adult food, the microbiological method using the Readycult^® Coliforms culture medium was performed to determine the presence or absence of total coliforms and E. coli. in 24 ± 1 hours in samples of water for human consumption. This ready-to-use, dehydrated, granular culture medium is supplied in a gamma-irradiated pressure pack, in which each unit is added to a 100-mL sample of drinking water. This method requires no confirmation or verification steps, and the detection limit of Readycult Coliforms Culture Medium is 1 colony-forming unit (CFU) of coliform bacteria or E. coli per 100 mL of medium.

Study variables.

The dependent variable of the present study was the presence of E. coli contamination in samples of water for human consumption from the homes of the adults evaluated. The presence of E. coli contamination was defined as containing 1 CFU of E. coli in 100 mL of water and was coded as 1, whereas no E. coli contamination was defined as 0 CFU of E. coli in 100 mL of water, and was coded as 0.

On the other hand, the independent variables were as follows: chlorine concentration (mg/L) in the water samples for human consumption in the home (divided into three categories: 0, > 0 and < 0.5, and ≥ 0.5), the origin of the water (spout [direct]/reservoirs [e.g. bucket or others]), education level (up to primary/secondary/higher education), per capita household spending quintile (richest/rich/medium/poor/poorest), area of residence (urban/rural), natural region (coast/highlands/jungle), and human development index of the district (low/medium/high). These variables have been described and used in the literature as the main independent variables in the study of E. coli in various regions of the world.^{13,21,26–28}

Statistical analysis.

The statistical analyses were performed using Stata 14 software (StataCorp LP, College Station, TX). In all estimations, an svy command was used to include the characteristics of the sample design and the VIANEV weighting factor. In the descriptive analysis, the number of participants is reported with their weighted proportions for the categorical variables. Chi-square tests were used to compare categorical variables. The variables with a P value < 0.20 in the crude analysis obtained by means of a generalized linear model of the Poisson family were included in an adjusted analysis to analyze the associations between the variables of interest, controlling for possible confounding factors and obtaining prevalence ratios (PR) with their confidence intervals (CI). Statistical significance was considered as a two-sided P value < 0.05.

Ethical considerations.

The approval of an ethics committee was not necessary because this was an analysis of secondary data from two surveys that are in the public domain and that did not allow the identification of the participants evaluated. The National Household Survey 2017 microdata were obtained through the National Institute of Statistics and Informatics website (http://iinei.inei.gob.pe/microdatos/), and the VIANEV 2017–2018 database was accessed through a request for access to public information through the National Institutes of Health (https://web.ins.gob.pe/es/transparencia/solicitud-de-acceso-a-la-informacion-publica).

RESULTS

A total of 886 participants were included in the analysis. The majority had a higher education (41.3%), belonged to the poorest quintile of wealth (25.4%), resided in an urban area (79.5%), on the coast (63.7%), and had a low human development index (33.9%). Likewise, in 23.1% of the households the water supply for human consumption was obtained from buckets or other reservoirs, and in 37.1% of household water supply sources for human consumption chlorine concentrations were absent (Table 1).

Table 1.

Characteristics of the participants included in the study in Peru, 2017–2018

Characteristics	Absolute frequency (N = 886)	%*
Amount of chlorine in mg/L
0	413	37.1
> 0 to < 0.5	157	21.6
≥ 0.5	316	41.3
Origin of the water
Spout (direct)	633	76.9
Reservoirs (e.g., bucket or others)	253	23.1
Educational level
Up to primary	240	21.2
Secondary	349	37.5
Higher education	297	41.3
Wealth quintile
Richest	152	21.2
Rich	149	18.5
Medium	143	17.2
Poor	158	17.8
Poorest	284	25.4
Area of residence
Urban	573	79.5
Rural	313	20.5
Region
Jungle	140	12.1
Highlands	218	24.2
Coast	528	63.7
Human development index
High	217	33.0
Medium	300	33.2
Low	369	33.9

^*Estimates include the weights and VIANEV 2017–2018 sample specifications.

In relation to the characteristics of the participants according to the presence of E. coli contamination, it was found that 25.2% of the water supply sources for human consumption in households contained E. coli at the time of sample taking. In addition, chlorine concentrations were absent in 54.2% of the water supply sources for human consumption. On the other hand, 13.7% contained chlorine concentrations > 0 and < 0.5 mg/L, and 5.3% of households had concentrations of ≥ 0.5 mg/L, with these households containing a lower proportion of E. coli in the water supply sources (Table 2). Regarding the origin of water for human consumption, in 57.9% of households that used buckets or other sources E. coli was contained in the water, whereas only 15.4% of households with drinking water by spout showed E. coli in the water. In relation to the characteristics of the people and their homes, the presence of E. coli contamination in the sources of water supply for human consumption was lowest in households of people with a secondary education level (72.0%) and higher (90.0%), in households belonging to a rich (90.1%) and the richest wealth quintile (92.0%) and with a high (93.6%) and medium (79.4%) human development index, as well as in households located in an urban area (84.2%) and in the costal (83.6%) and highlands regions (65.8%) (Table 2).

Table 2.

Characteristics of the participants according to the presence of E. coli contamination in water supplies for human consumption in Peru, 2017–2018

Characteristics	E. coli		P value*
	No	Yes
	% (95% CI)	% (95% CI)
Overall	74.8 (71.0–78.2)	25.2 (21.8–29.0)	–
Amount of chlorine in mg/L
0	45.8 (39.0–52.9)	54.2 (47.1–61.0)	< 0.001
> 0 to < 0.5	86.3 (78.0–91.8)	13.7 (8.2–22.0)
≥ 0.5	94.7 (91.3–96.9)	5.3 (3.1–8.7)
Origin of the water
Spout (direct)	84.6 (80.8–87.7)	15.4 (12.3–19.2)	< 0.001
Reservoirs (e.g., bucket or others)	42.1 (33.3–51.4)	57.9 (48.6–66.7)
Educational level
Up to primary	50.2 (41.5–58.8)	49.8 (41.2–58.5)	< 0.001
Secondary	72 (66.1–77.2)	28 (22.8–33.9)
Higher education	90 (85.4–93.2)	10 (6.8–14.6)
Wealth quintile
Richest	92 (85.2–95.8)	8 (4.2–14.8)	< 0.001
Rich	90.1 (83.7–94.1)	9.9 (5.9–16.3)
Medium	73.1 (64.6–80.2)	26.9 (19.8–35.4)
Poor	67.4 (58.2–75.4)	32.6 (24.6–41.8)
Poorest	55.6 (47.5–63.3)	44.4 (36.7–52.5)
Area of residence
Urban	84.2 (80.0–87.6)	15.8 (12.4–20.0)	< 0.001
Rural	38.3 (30.5–46.7)	61.7 (53.3–69.5)
Region
Jungle	46.3 (35.0–57.9)	53.7 (42.1–65.0)	< 0.001
Highlands	65.8 (56.3–74.1)	34.2 (25.9–43.7)
Coast	83.6 (79.2–87.3)	16.4 (12.7–20.8)
Human development index
High	93.6 (89.0–96.3)	6.4 (3.7–11.0)	< 0.001
Medium	79.4 (72.2–85.2)	20.6 (14.8–27.8)
Low	51.9 (43.9–59.9)	48.1 (40.1–56.1)

Estimates include the weights and VIANEV 2017–2018 sample specifications.

^*The P value was calculated using the Rao–Scott χ² test. CI = confidence interval.

The crude analysis (Table 3) showed a significant association between all independent variables and the presence of E. coli contamination in drinking-water sources. In the adjusted model, E. coli was more likely to be found in the water for human consumption from containers such as buckets or others (adjusted PR [aPR]: 1.15 [95% CI: 1.18–1.93]; P = 0.001), from households of people belonging to a medium wealth quintile (aPR: 2.06 [95% CI: 1.13–3.76]; P = 0.018) and a poor quintile (aPR: 1.82 [95% CI: 1.01–3.25]; P = 0.045), households in a rural area (aPR: 1.36 [95% CI: 1.01–1.83]; P = 0.042), and households of people with a medium (aPR: 2.04 [95% CI: 1.14–3.67]; P = 0.017) and low (aPR: 2.12 [95% CI: 1.15–3.91]; P = 0.016) human development index. On the other hand, the presence of chlorine concentrations > 0 and < 0.5 mg/L (aPR: 0.42 [95% CI: 0.26–0.68]; P < 0.001) and ≥ 0.5 mg/L (aPR: 0.20 [95% CI: 0.11–0.33]; P < 0.001) in the water, and households of people with a higher education (aPR: 0.67 [95% CI: 0.45–0.99]; P = 0.044) were less likely to have E. coli in their water.

Table 3.

Factors associated with the presence of E. coli contamination in water supplies for human consumption in Peru, 2017–2018

Characteristics	Crude model		Adjusted model*
	PR (95% CI)	P value	PR (95% CI)	P value
Amount of chlorine in mg/L
0	1.00 (1.00–1.00)	–	1.00 (1.00–1.00)	–
> 0 to < 0.5	0.25 (0.15–0.42)	< 0.001	0.42 (0.26–0.68)	< 0.001
≥ 0.5	0.10 (0.06–0.16)	< 0.001	0.20 (0.11–0.33)	< 0.001
Origin of water
Spout (direct)	1.00 (1.00–1.00)	–	1.00 (1.00–1.00)	–
Reservoirs (e.g., bucket or others)	3.76 (2.88–4.91)	< 0.001	1.51 (1.18–1.93)	0.001
Educational level
Up to primary	1.00 (1.00–1.00)	–	1.00 (1.00–1.00)	–
Secondary	0.56 (0.44–0.72)	< 0.001	1.06 (0.85–1.33)	0.602
Higher education	0.20 (0.13–0.31)	< 0.001	0.67 (0.45–0.99)	0.044
Wealth quintile
Richest	1.00 (1.00–1.00)	–	1.00 (1.00–1.00)	–
Rich	1.24 (0.59–2.59)	0.571	1.23 (0.66–2.27)	0.517
Medium	3.35 (1.67–6.70)	0.001	2.06 (1.13–3.76)	0.018
Poor	4.06 (2.04–8.08)	< 0.001	1.82 (1.01–3.25)	0.045
Poorest	5.53 (2.86–10.69)	< 0.001	1.50 (0.84–2.66)	0.167
Area of residence
Urban	1.00 (1.00–1.00)	–	1.00 (1.00–1.00)	–
Rural	3.91 (2.97–5.14)	< 0.001	1.36 (1.01–1.83)	0.042
Region
Jungle	1.00 (1.00–1.00)	–	1.00 (1.00–1.00)	–
Highlands	0.64 (0.45–0.90)	0.011	0.80 (0.60–1.05)	0.106
Coast	0.31 (0.22–0.42)	< 0.001	1.03 (0.75–1.42)	0.854
Human development index
High	1.00 (1.00–1.00)	–	1.00 (1.00–1.00)	–
Medium	3.20 (1.70–6.04)	< 0.001	2.04 (1.14–3.67)	0.017
Low	7.49 (4.22–13.29)	< 0.001	2.12 (1.15–3.91)	0.016

CI = confidence interval; PR = prevalence ratio.

^*Adjusted for all variables displayed in the column.

DISCUSSION

The present study sought to evaluate the factors associated with the presence of E. coli contamination in water supplies for human consumption in Peru. It was found that approximately 1 in 4 households had E. coli in their water supplies for human consumption at the time of sampling. Likewise, it was found that water from reservoirs such as buckets or others, households of people belonging to a poor wealth quintile, living in a rural area, and having a medium and low human development index were more likely to contain E. coli in the sources of water for human consumption. Finally, having a chlorine concentration of ≥ 0.5 mg/L in the water and households with members with a higher education were less likely to have E. coli in drinking-water supplies.

This study found that having water from reservoirs such as buckets or others was associated with a higher probability of containing E. coli. This is consistent with what was reported in studies carried out in Thailand and Laos, in which it was found that water storage containers such as jugs and buckets have a higher prevalence of E. coli contamination.²⁶ Likewise, limited access to drinking-water supplies in LMIC leads people in these regions to store water in tanks for later consumption.²⁹ The greater probability of contamination by E. coli in buckets or other containers could be explained by the fact that water containers are not easy to clean and are not cleaned regularly, and thus, the storage of water in these containers could cause the deterioration of water quality over time.²⁶ Moreover, it has been shown that the storage of water for even a short period of time can produce microbial contamination with E. coli.³⁰ In Peru, water contamination (especially in rural and periurban areas) can be attributed to poor storage practices, low water chlorination, inadequate hand hygiene to ensure safe drinking water, unsafe water handling, inadequate cleaning of the water storage container and the geographic barriers that people face in obtaining water.^22,31,32 Although local governments have contributed to the implementation of measures such as the “Global Scaling Up Handwashing Project”,³³ “Directiva Sanitaria para la Promocionar el lavado de manos social como práctica saludable en el Perú”,³⁴ and the “Programa Nacional de Saneamiento Rural”,³⁵ more than 90% of children under 5 years of age have been exposed to drinking water with fecal coliforms.^18,36,37 Therefore, local governments should generate educational strategies of greater impact in rural communities, emphasizing the need to clean water containers and describe the necessary disinfection measures and the frequency with which they should be carried out to avoid microbial contamination.

It was also found that households with people belonging to a poor wealth quintile were more likely to have drinking-water supplies contaminated with E. coli. This is similar to what has been reported in other LMIC in which households belonging to the poorest quintile were more likely to contain E. coli contamination of their water supply compared with the richest quintile.^12,38 This can be explained by the better hygiene practices carried out by people from the richest households. Previous studies have shown that children living with families with a higher socioeconomic level were less exposed to pollutants causing diarrhea because of greater access to sanitation facilities in households of the richest quintiles, allowing better hygiene, sanitation, and more frequent use of health services.^39,40 In Peru, there is a decreasing pattern in the access to safe drinking water among the poorest households, whereas the richest households show an increasing trend.^41,42 Specifically, the poorest households in the Peruvian territory are located in marginal areas or with limited water and sanitation services, which generates high costs for water services.⁴³ These limitations create challenges for government institutions that, through programs such as “Agua Segura para Lima y Callao”,⁴⁴ seek to improve access to quality drinking water and sanitation services to improve the coverage of safe drinking water in all Peruvian households. Therefore, these strategies should improve their coverage and remain on the agenda of government institutions to avoid contamination of water supplies, especially in households belonging to the poorest quintiles.

Living in a rural area has been shown to increase the probability of containing E. coli in drinking-water sources. This is similar to what has been reported in Ethiopia and Thailand where higher levels of E. coli contamination were seen in rural compared with urban areas.^26,45 The availability of sanitary facilities is less likely in households in rural areas, making it necessary to store water or use other water supply sources for handwashing, with the subsequent greater possibility of contamination of the water supply.⁴⁶ In addition, inadequate containment of human excreta resulting from a lack of sanitation facilities can provide other pathways for water contamination and increase the risk of disease transmission related to the consumption of contaminated water.⁴⁷ Bacterial contamination in rural areas in Peru can be attributed to low or no chlorination of water from springs or other natural water resources, with less than 40% of water supply systems being chlorinated.⁴⁸ This lack of compliance with regulations regarding water for human consumption is a public health problem, leading to an increase in cases of diarrhea and other diseases related to the consumption of unsafe water.⁴⁹ Furthermore, rural areas (especially in the highlands of Peru) are close to mining areas, where water contamination may not only be due to microorganisms but also to heavy metals that can cause pathologies in the population that consumes it.⁵⁰ In this sense, governmental institutions have implemented strategies such as the “Plan Nacional de Recursos Hídricos”, which seeks to improve water quality conditions in the most remote areas of Peru and to improve the coverage of water and sanitation services.⁵¹ In this regard, government institutions should continue to implement public health measures and strategies to improve or implement new sanitation systems in rural areas to reduce the likelihood of infections or related diseases and improve access to clean water.

Another relevant problem in the rural areas of Peru is related to the climatic changes experienced by these regions. Specifically, higher altitude and more remote areas experience changes such as melting glaciers, water shortages, and less rainfall to rivers or lakes, which generate a shortage of water in homes that are supplied by these natural resources.⁵² Likewise, these climatic changes generate uncertainty in the populations that reside in the Andes about the lack of water supply due to the melting of the glaciers because this resource is a necessary input for domestic use and agricultural practices, which for the most people, is one of the largest sources of household income.⁵³ This problem of climate change is not only linked to water scarcity but also to the quality of this resource, which cannot be guaranteed due to the lack of water and sanitation infrastructure present in these regions.⁵² As such, government institutions must make climate change a priority social problem and from an ecological point of view they must develop strategies focused on the implementation of the necessary infrastructure for water and sanitation services in the high Andean regions.

We found that E. coli was more likely to be contained in the drinking-water supply of people living in areas with a medium and low human development index. This association has previously been reported in studies carried out in low-income countries with a low or medium human development index, which described a high incidence of E. coli infections caused by the consumption of contaminated water and food.⁵⁴ This finding could be based on the lack of capacity of government organizations to perform periodic controls of water quality and the lack of adequate infrastructure for water maintenance in countries with a low human development index and scarce economic resources, leading to the costs for water quality control outweighing the potential benefits of safe water for the people.^55,56 Countries with a low human development index also have human resources that are poorly trained in the management of water quality monitoring and the benefits of continuous monitoring of water quality are less likely to be perceived by people with a low educational level. This, in turn, generates low standards of water storage and preservation of and a lack of hygienic practices in people.^56,57 Unlike countries with a high human development index, developing countries with a medium or low development index should prioritize the development of low-cost strategies to improve regular monitoring of water quality for human consumption and promote knowledge about the potential benefits of safe water to thereby reduce the burden of disease due to pathologies related to the consumption of contaminated water.

We found that water supplies with a chlorine level ≥ 0.5 mg/L were less likely to have E. coli in drinking water. This correlates with the findings of a systematic review, which described that the presence of chlorine in the water reduces E. coli growth rates, thereby decreasing the risk of contamination by E. coli.^58,59 According to the World Health Organization, the minimum chlorine concentration necessary for water to be free of total coliforms and E. coli must be ≥ 0.5 mg/L. Likewise, the Ministry of Health of Peru has established the same measure as the ideal concentration of chlorine.^4,41 In this sense, supervision of public water sources is essential to guarantee that the companies that provide drinking water provide adequate chlorine concentrations in water.

Respondents with a higher level of education were less likely to have E. coli in drinking-water sources. It has been shown that people with a higher educational level have better practices in the management and care of water.²⁸ Similar to the Essential Health Care Program established in the Philippines to develop good hygiene habits and generate positive health and educational outcomes for children,⁶⁰ in Peru, strategies such as the Global Scaling Up Handwashing Project have had a positive impact on handwashing education in the most remote areas of the country.⁶¹ Therefore, the population must be informed about how to perform proper handwashing, as well as how to collect and store water correctly to reduce the levels of E. coli contamination.

This study has some limitations. One is that because of the cross-sectional nature of the study causal inferences are limited due to the lack of temporality. Additionally, as this was an analysis of secondary data, the factors evaluated are limited to the variables available in the survey, and there may be other influential factors that could better characterize the factors associated with the presence of E. coli contamination, such as the pH of the water, the presence of pesticides, and organic carbon in the reservoirs. Likewise, there could be errors in the interpretation of the results of colors or cultures of E. coli due to the qualitative determination of these variables. Nonetheless, this information is collected by previously trained professionals, thus reducing these possibilities. Likewise, our study measures the contamination E. coli in water reservoirs, but it is not possible to differentiate what type of strain it presents, with some strains being more pathogenic than others. The VIANEV is a nationally representative survey, being the only study that collects information on the level of contamination of water supplies at a national level; therefore, it provides updated information on this subject in the Peruvian population.

In conclusion, from 2017 to 2018, one in four Peruvian households contained contamination by E. coli in their water supply for human consumption at the time of sampling. Sociodemographic, management, and treatment factors were associated with the presence of E. coli contamination. These results indicate a high prevalence of E. coli contamination in water supplies for human consumption at the national level, indicating that Sustainable Development Goals are not met. Therefore, strategies focused on improving the infrastructure of water and sanitation services, increasing the coverage of improved water services in rural areas, promoting educational measures for handwashing, and correct storage and treatment of water should have greater coverage in rural and in poorer areas, as well as greater participation of local governments to improve water quality indicators. Moreover, these measures must consider the factors associated with the prevalence of E. coli contamination of household water supplies described in the present study to reduce the incidence of gastrointestinal diseases related to water contamination.