Scholarly article on topic 'Disease burden due to gastrointestinal pathogens in a wastewater system in Kampala, Uganda'

Disease burden due to gastrointestinal pathogens in a wastewater system in Kampala, Uganda Academic research paper on "Animal and dairy science"

CC BY
0
0
Share paper
Academic journal
Microbial Risk Analysis
OECD Field of science
Keywords
{Bacteria / Gastroenteritis / "Intestinal parasites" / "Quantitative microbial risk assessment" / Viruses / Wastewater}

Abstract of research paper on Animal and dairy science, author of scientific article — Samuel Fuhrimann, Mirko S. Winkler, Michelle Stalder, Charles B. Niwagaba, Mohammed Babu, et al.

Abstract In wastewater systems in Kampala, Uganda, microbial contamination has increased over the past two decades. Those people who live or work along the Nakivubo channel and wetland and those who use the recreational areas along the shores of Lake Victoria are at an elevated risk of gastrointestinal infections. A quantitative microbial risk assessment (QMRA) was applied for five population groups, characterised by different levels of exposure to wastewater in the Nakivubo area, namely: (i) slum dwellers at risk of flooding; (ii) children living in these slum settlements; (iii) workers maintaining the drainage system or managing faecal sludge (sanitation workers); (iv) urban farmers; and (v) swimmers in Lake Victoria. The QMRA was based on measured concentrations of Escherichia coli, Salmonella spp. and Ascaris spp. eggs in wastewater samples. Published ratios between measured organism and pathogenic strains of norovirus, rotavirus, Campylobacter spp., pathogenic E. coli, pathogenic Salmonella spp., Cryptosporidium spp. and Ascaris lumbricoides were used to estimate annual incidence of gastrointestinal illness and the resulting disease burden. The QMRA estimated a total of 59,493 disease episodes per year across all 18,204 exposed people and an annual disease burden of 304.3 disability-adjusted life years (DALYs). Incidence estimates of gastrointestinal disease episodes per year were highest for urban farmers (10.9) and children living in slum communities (8.3), whilst other exposed groups showed lower incidence (<4.3). Disease burden per person per year was highest in urban farmers (0.073 DALYs) followed by sanitation workers (0.040 DALYs) and children in slum communities (0.017 DALYs). Our findings suggest that the exposure to wastewater is associated with public health problems, particularly children and adults living and working along the major wastewater and reuse system in Kampala. Our findings call for specific interventions to reduce the disease burden due to exposure to wastewater.

Academic research paper on topic "Disease burden due to gastrointestinal pathogens in a wastewater system in Kampala, Uganda"

Accepted Manuscript

MICROBIAL

RISK ANALYSIS

A/1 Interdisciplinary journal Iw risk analysis work applied to mkrobwl hazards.

Disease burden due to gastrointestinal pathogens in a wastewater system in Kampala, Uganda

Samuel Fuhrimann , Mirko S. Winkler, Michelle Stalder, Charles B. Niwagaba , Mohammed Babu , Narcis B. Kabatereine , Abdullah A. Halage , Jurg Utzinger, Gueladio Cisse , Maarten Nauta

PII: DOI:

Reference:

S2352-3522(16)30010-X 10.1016/j.mran.2016.11.003 MRAN 28

To appear in:

Microbial Risk Analysis

Received date: Revised date: Accepted date:

17 March 2016 11 October 2016 6 November 2016

Please cite this article as: Samuel Fuhrimann , Mirko S. Winkler, Michelle Stalder, Charles B. Niwagaba , Mohammed Babu , Narcis B. Kabatereine , Abdullah A. Halage , JUrg Utzinger, Gueladio Cisse , Maarten Nauta , Disease burden due to gastrointestinal pathogens in a wastewater system in Kampala, Uganda, Microbial Risk Analysis (2016), doi: 10.1016/j.mran.2016.11.003

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Disease burden due to gastrointestinal pathogens in a wastewater system in Kampala, Uganda

Samuel Fuhrimann1,2*, Mirko S. Winkler1'2, Michelle Stalder1,2,3, Charles B. Niwagaba4, Mohammed Babu5, Narcis B. Kabatereine6, Abdullah A. Halage7, Jürg Utzinger1,2, Guéladio Cissé1,2, Maarten Nauta8

1 Swiss Tropical and Public Health Institute, Basel, Switzerland

2 University of Basel, Basel, Switzerland

3 Institute for Biogeochemistry and Pollution Dynamics, ETH Zurich, Zurich

4 Department of Civil and Environmental Engineering, Makerere University, Kampala, Uganda

5 Department of Research and Development, National Water and Sewerage Corporation, Kampala,

Uganda

Schistosomiasis Control Initiative, Imperial College London, London, United Kingdom

7 Makerere University School of Public Health, Kampala, Ugand..

8 National Food Institute, Technical University of Denmark, Soborg, Denmark

d Public Health

Corresponding author Samuel Fuhrimann, !

Switzerland. Tel. +41 61 284-8304; Fax +41 61 284-8105. E-mail: samuel.fuhrimann@unibas .ch

Samuel Fuhrimann, Swiss Tropical and Public Health Institute, P.O. Box, CH-4002 Basel,

61 284-8

Abstract

In wastewater systems in Kampala, Uganda, microbial contamination has increased over the past two decades. Those people who live or work along the Nakivubo channel and wetland and those who use the recreational areas along the shores of Lake Victoria are at an elevated risk of gastrointestinal infections. A quantitative microbial risk assessment (QMRA) was applied for five population groups, characterised by different levels of exposure to wastewater in the Nakivubo area, namely: (i) slum dwellers at risk of flooding; (ii) children living in these slum settlements; (iii) workers maintaining the drainage system or managing faecal sludge (sanitation workers); (iv) urban farmers; and (v) swimmers in Lake Victoria. The QMRA was based on measured concentrations of Escherichia coli, Salmonella spp. and Ascaris spp. eggs in wastewater samples. Published ratios between measured organism and pathogenic strains of norovirus, rotavirus, Campylobacter spp., pathogenic E. coli, pathogenic Salmonella spp., Cryptosporidium spp. and Ascaris lumbricoides were used to estimate annual incidence of gastrointestinal illness and the resulting disease burden. The QMRA estimated a total of 59,493 disease episodes per year across all 18,204 exposed people and an annual disease burden of 304.3 disability-adjusted life years (DALYs). Incidence estimates of gastrointestinal disease episodes per year were highest for urban farmers (10.9) and children living in slum communities (8.3), whilst

other exposed groups showed lower incidence (<4.3). Disease burden per person per year was highest in urban farmers (0.073 DALYs) followed by sanitation workers (0.040 DALYs) and children in slum communities (0.017 DALYs). Our findings suggest that the exposure to wastewater is associated with public health problems, particularly children and adults living and working along the major wastewater and reuse system in Kampala. Our findings call for specific interventions to reduce the disease burden due to exposure to wastewater.

Keywords: Bacteria; Gastroenteritis, Intestinal parasites; Kampala; Quantitative assessment, Viruses; Wastewater 1. Introduction

Urban wastewater is often contaminated with pathogenic organisms, and thus puts people at risk of ill-health (Blumenthal and Peasey, 2002; McBride et al., 2013; Barker, 2014). Untreated wastewater of domestic and industrial origins is of particular concern (Fuhrimann et al., 2015; WHO, 2015). It follows that in urban centres of low- and middle-income countries (LMICs), characterised by the lack of improved sanitation, high prevalence and outbreaks of gastrointestinal diseases caused by bacteria, viruses, intestinal protozoa or helminths are common (Matthys et al., 2007; Pham-Duc et al., 2014). However, there is a paucity of disease burden estimates caused by these pathogenic organisms in LMICs (Labite et al., 2010; Katukiza et al., 2013; Machdar et al., 2013).

Exposure to urban wastewater is multifaceted (Stenstrom et al., 2011; Keraita and Davila, 2015). Direct exposure occurs through accidental ingestion, inhalation or dermal contact in different contexts: (i) during working procedures (e.g. while emptying on-site sanitation facilities, managing wastewater treatment processes or reusing wastewater for irrigation purposes); (ii) while using wastewater for domestic activities (e.g. for cleaning dishes or washing clothes); (iii) during flooding events caused by heavy rains (Cisse, 2013); and (iv) due to recreational activities (e.g. swimming or bathing in lakes or rivers fed by wastewater) (Ferrer et al., 2012; Katukiza et al., 2013; Yapo et al., 2013). Indirect exposure occurs through consumption of contaminated drinking water or wastewater-fed crops and fish (Machdar et al., 2013; Mok and Hamilton, 2014). Reducing exposures to wastewater is therefore a critical inter-sectoral responsibility for protecting public health (Amoah et al., 2011; Keraita and Davila, 2015). Guidelines published by the World Health Organization (WHO) propose control measures to safely manage and reuse wastewater, excreta and greywater, and to protect recreational and drinking water systems (WHO 2003, 2006, 2011a). These guidelines are built around the concept of health-based targets that are grounded on well-defined health metrics (e.g. disability-adjusted life years (DALYs)) and a level of tolerable health burden (Mara et al., 2010). Even though this health-based target was recently revised, allowing for a tolerable additional disease burden of 0.0001 DALYs per person per year (pppy), it is still out of reach in many LMICs (Ensink and van der Hoek, 2009; Mara et al., 2010; WHO, 2015).

Disease burden of pathogenic organisms can be estimated by quantitative microbial risk assessment (QMRA) (Haas et al., 2014; Ichida et al., 2015). QMRA commonly follows four working steps: (i) hazard identification; (ii) exposure assessment; (iii) dose-response assessment; and (vi) risk characterisation (Haas et al. 2014). For water-borne hazards, QMRA is widely used in industrialised countries to estimate health risks for drinking water supply systems (Hunter et al., 2000; WHO, 2011a), flood water events (de Man et al., 2013), wastewater management and reuse (Westrell et al., 2004; Ashbolt et al., 2006), storm water discharge (McBride et al., 2013) or recreational water (Soller et al., 2015). It is also widely used for food-borne hazards, based on principles and guidelines defined by Codex Alimentarius (Anonymous, 1995; Havelaar et al., 2008). In LMIC settings, QMRA is becoming increasingly popular and has been successfully applied for the identification of effective control measures for wastewater reuse in urban agriculture systems in Accra and Bangkok (Seidu et al., 2008; Ferrer et al., 2012; Barker et al., 2014) and for risk profiling along drainage channels in Abidjan and Kampala (Katukiza et al., 2013; Yapo et al., 2013). QMRA has also been used to guide cost-effective interventions for drinking water supply systems in Accra and Kampala (Howard et al., 2006; Machdar et al., 2013). It must be noted, however, that these QMRAs suffer from a lack of data on source contamination with pathogens, setting-specific dose-response relationships and validation of the estimated risks with epidemiological data (WHO, 2006; Barker et al., 2014).

To fill some of the aforementioned gaps, this paper presents a QMRA case study for Kampala, the capital of Uganda. Of note, Kampala has undergone rapid population growth and there are large volumes of wastewater in face of insufficiently equipped sanitary infrastructures (Fuhrimann et al., 2015). Microbial and chemical contamination has increased over the past two decades along the major wastewater system (Kansiime and Nalubega, 1999; Kayima et al. 2008; Fuhrimann et al., 2015). For example, concentrations of Escherichia coli in wastewater samples are exceeding WHO thresholds of 103-104 colony forming units (CFU) E. coli/100 mL by magnitude of at least 10 (Fuhrimann et al., 2015). Hence, these waters should not be reused without considering additional control measures (WHO, 2006). The mean concentration of Ascaris lumbricoides eggs in wastewater and soil samples in slum community areas were above WHO safety standards of <1 egg/L (WHO, 2006; Fuhrimann et al., 2015). Using a QMRA approach, the goal of the present study was to estimate the disease burden resulting from exposure to water-borne pathogens causing gastroenteritis along the major wastewater system in Kampala. By comparing the model estimates with findings from epidemiological surveys, advantages and limitations of the QMRA methodology are discussed.

2. Materials and methods

2.1. Study area

Detailed information and a short video introducing the study area and the sampling scheme have been published elsewhere (Fuhrimann et al., 2014). In brief, Kampala is located at latitude 0° 18' 49.18" N and longitude 32° 36' 43.86" E at an altitude of 1,140 m above the mean sea level. The study area

d to sup d to sup

2.2. Hazard identification

11 vciix/vi IU 01

includes the Nakivubo channel in Kampala city (Figure 1), which is an open storm water channel transporting most of the city's wastewater from the central division (approximately 13,928 m3/day), comprised of wastewater from households (23%) and industries (77%). Further, the channel receives partially treated effluent from the Bugolobi Sewage Treatment Works (BSTW) (up to 12,000 m3/day) (Beller Consult et al., 2004). Downstream of the treatment plant, the wastewater enters into the Nakivubo wetland, where it is reused for urban agriculture (main crops: sugar cane, yams and maize). Alongside the Nakivubo wetland, there are informal slum communities prone to flooding events (Fuhrimann et al., 2016). The water is finally discharged into the Inner Murchison Bay in Lake Victoria, a popular recreational area for Kampala's inhabitants, especially along its shores. In addition, only 4 km away from the discharge point, the lake water is pumped and treated to supply Kampala with drinking water (Howard et al., 2006).

The hazards considered for the QMRA are seven pathogenic organisms, for which the incidence and disease burden of gastroenteritis due to exposure to wastewater are estimated: two viruses (norovirus and rotavirus), three bacteria (Campylobacter spp., pathogenic Salmonella spp. and E. coli), one intestinal protozoon (Cryptosporidium spp.) and one soil-transmitted helminth species (A. lumbricoides). All of these pathogens are characterised by the faecal-oral transmission route, can persist for weeks or months in the environment and are difficult to inactivate with conventional wastewater treatment processes (WHO 2006; Machdar et al. 2013; Fuhrimann et al. 2015).

Hazard selection was motivated by findings from an environmental assessment and a cross-sectional survey conducted in the study area between October and December 2013 during the short rainy season (Fuhrimann et al., 2015, 2016), along with previous studies about major pathogens giving rise to gastroenteritis in Uganda and elsewhere in the world (Becker et al., 2013; Barker et al., 2014; Gibney et al., 2014; Katukiza et al., 2013). Our selection is further justified on the following grounds:

• Rotavirus is one of the leading causes of childhood diarrhoea, responsible for about 7.3% of deaths among children below the age of 5 years in Uganda and is considered to account for most of the disease burden in slums in Kampala (Katukiza et al. 2013; Sigei et al. 2015).

• Norovirus is the major cause of diarrhoeal disease in adults and its secondary attack rate is known to be high causing epidemic situations, especially in densely populated slum areas (Teunis et al., 2008; Katukiza et al., 2013).

• Campylobacter spp. are zoonotic bacteria that cause campylobacteriosis, with Campylobacter jejuni being a common cause of diarrhoea in LMICs (Kaakoush et al., 2015).

• E. coli bacteria are part of the normal gastrointestinal microflora of warm-blooded animals and humans, whilst enterohemorrhagic E. coli (EHEC) is considered pathogenic, with the serotype E. coli 0157:H7 responsible for the largest public health impact (Okeke, 2009; Hynds et al., 2014).

• Salmonella spp. have more than 2,000 sero-groups, with only a few being of concern for human health (S. typhi and S. para-typhi A, B, and C, and the enteric salmonella strains) (Kariuki et al.,

• Cryptosporidium spp. is a zoonotic intestinal protozoon that can result in severe health implications in children and immunocompromised individuals e.g. for HIV-positive people. As the HIV prevalence in Uganda is estimated to be 7.4% - the 10th highest in the world -, Cryptosporidium spp. are likely to be of higher public health relevance than, for example, other intestinal protozoa such as Giardia intestinalis and Entamoeba histolytica (Kaijuka et al., 2011).

• Helminth infections are endemic in the Great Lake region (Karagiannis-Voules et al., 2015; Lai et al., 2015). Prevalence rates for hookworm, Trichuris trichiura, Schistosoma mansoni and A. lumbricoides in urban farmers in the Nakivubo area were found at 28%, 26%, 23% and 18%, respectively (Fuhrimann et al. 2015). For the current QMRA, the commonly used reference organism A. lumbricoides is considered as its eggs are known to persist in the environment longer than any of the other helminth species (Stott et al., 2003)

2.3. Exposure assessment

The exposure scenarios for the QMRA are based on information derived from a survey of 915 people in the Nakivubo area. The findings of this survey have been reported elsewhere (Fuhrimann et al. 2015). Overall, the QMRA only included accidental ingestion of contaminated water and the following exposure pathways were excluded based on the given context: (i) ingestion of contaminated soil, dermal contact, inhalation and drinking of potentially contaminated water (due to lack of data); (ii) consumption of contaminated food crops (not in direct contact with wastewater (sugar cane and maize) or sufficiently cooked (yams)); and (iii) exposure to contaminated water used for bathing or washing clothes (uncommon local practice).

Five exposure scenarios in four study areas (Nakivubo channel, Nakivubo wetland, community areas and shores of Lake Victoria) were developed and assumptions about exposure groups, number of people exposed, exposure frequency and volume of ingested water are made (Figure 2).

• Scenario 1 (Sjiooding): Slum dwellers (all age groups) living in close proximity to the Nakivubo wetland are located on low altitude and, hence, were prone to flooding events. Almost half (47%) of the people living in these communities reported having been exposed to flooding events in the previous year (i.e. 5,640 out of 12,000 people) (Fuhrimann et al., 2016). According to Lwasa and colleagues (2010), six flooding events may occur during the two rainy seasons (March to May and September to November) in one year. During a flooding event, ingestion of water due to unintentional immersion is assumed to be between 10 and 30 mL (Katukiza et al. 2014).

• Scenario 2 (Sworking): There are 231 registered sanitation workers. 90 workers are employed by the National Water and Sewerage Corporation (NWSC) and responsible for the maintenance of the drainage system and the operation of the BSTW. Those working for the Pit Emptier

2015).

Association (PEA) (n=141), are responsible for emptying of on-site toilet facilities and transfer of the faecal sludge to BSTW (Fuhrimann et al., 2016). Their working practices expose workers to wastewater and faecal sludge (Stenstrom et al., 2011). On average, workers report 312 days on duty per year, with 70% of the workers regularly wearing boots and gloves. An accidental ingestion between 1 and 5 mL per working day is assumed (10-times less compared to workers without protective equipment) (WHO, 2006; Labite et al., 2010; Mara and Bos, 2010).

• Scenario 3 (Sfarming): The farming areas in the Nakivubo wetland are frequently flooded with polluted water from Nakivubo channel, combined with effluent from BSTW, and the area bordering Lake Victoria is defined as floating wetland (Kansiime and Nalubega, 1999). Thus, the likelihood of accidental ingestion of wastewater is considerable. Overall, approximately 650 urban farmers (>15 years of age) reported working within the Nakivubo wetland. On average, farmers reported to work 297 days per year and only 3% of them regularly wear boots and gloves (Fuhrimann et al., 2016). An accidental ingestion between 10 and 50 mL per working day is assumed (WHO, 2006; Labite et al., 2010; Mara and Bos, 2010).

• Scenario 4 (Splaying): Children living in slum communities are at an elevated risk of daily accidental ingestion of water (Katukiza et al. 2010). Due to flooding events and poor sanitation infrastructures, slum environments are constantly contaminated (Fuhrimann et al., 2016) and children aged <15 years are considered at risk when playing (49.4%, 5,880 children out of 12,000 people) (UBOS, 2013). Exposure is assumed to be daily (365 days per year) with an accidental ingestion rate between 1 and 5 mL per day (Labite et al. 2010; Katukiza et al. 2013).

• Scenario 5 (Sswimming): Several studies found adverse health outcomes associated with exposure to contaminated recreational water (WHO, 2003; Schets et al., 2011; Yapo et al., 2013). In Kampala, swimming at the local beaches along the Inner Murchison Bay (e.g. Miami Beach, Ggaba Beach, KK beach) is popular (Beller Consult et al., 2004). However, in our previous study, only three out of 915 individuals interviewed (0.32%) reported to have swum in Lake Victoria in the year preceding the survey (Fuhrimann et al., 2016). This is likely to be an underestimate as only people from lower socioeconomic strata were included in the survey and may not be able to access the private beaches. Still, when extrapolating this to the 1.8 million inhabitants of Kampala, 5,760 people can be estimated to be swimming in the Inner Murchison Bay of Lake Victoria. It is assumed these people swim six times per year in the Lake, ingesting 20 to 50 mL per swimming event (WHO, 2003; Schets et al., 2011, Fuhrimann et al., 2016).

2.4. Measurements of pathogenic organisms along the wastewater system

Between October and December 2013, wastewater samples were collected at 23 sentinel sites along the Nakivubo channel (five points) and wetland (12 points), community areas bordering the Nakivubo wetland (two points) and within the Inner Murchison Bay in Lake Victoria (four points). The samples were tested for E. coli, thermotolerant coliforms (TTC), Salmonella spp. and helminth eggs. Details of

the methodology and sampling strategy, including measurements of heavy metals and physicochemical parameters, have been published elsewhere (Fuhrimann et al., 2015). According to guidance documents put forth by WHO, the ratio between measured E. coli and the pathogens (ppath) can be simplified and therefore assumed to vary between 10-6-10-5 (rotavirus, norovirus and Campylobacter spp.) and 10-7-10-6 (Cryptosporidium spp.) (Haas et al., 1999). The ratio between pathogenic and non-pathogenic strains of E. coli (ppath) was set to vary between 7.6 x 10-4 and 1 x 10" (Shere et al., 2002; Soller et al., 2010; Hynds et al., 2014). In the absence of data, the same ratio was assumed for Salmonella. For Ascaris spp. it was assumed that each egg detected represents A. lumbricoides (ppath = 1, not considering the occurrence of other species such as A. suum) (Mara and Sleigh, 2010).

2.5. QMRA structure, implementation and analysis 2.5.1 QMRA structure

In most points, our QMRA approach follows the descriptions of the WHO 2006 guidelines and Karavarsamis and Hamilton (WHO 2006; Karavarsamis and Hamilton 2010; Mara et al. 2010). However, we purposely do not adopt the entire approach, because we do not aim to study the overall infection risk but the disease burden, which needs estimates of the number of cases for each of the hazards separately. Hence, in addition, three adjustments were made. First, we used quantitative data on E. coli, Salmonella spp. and Ascaris spp. eggs obtained in the research area and fitted them to lognormal distribution, while prevalence estimates were used for Ascaris eggs (Fuhrimann et al. 2015). Second, our QMRA allows people to get ill from more than one hazard at the same time and the mean risk of illness was calculated over the duration of one year for each person, while assuming that each individual can become infected with each exposure event (without considering immunity) (Haas et al. 2014). Third, the corresponding disease burden for each of the seven selected pathogens was calculated according to published probability estimates for mild, moderate, severe and fatal gastroenteritis (Havelaar et al., 2000; Brooker, 2010; Katukiza et al., 2013; Gibney et al., 2014). Note that probability estimates for the severity grade were taken from other countries than Uganda, as no local estimates exist. Burden estimates for mild, moderate and severe diarrhoea episodes were taken from the Global Burden of Disease Study 2010 (Salomon et al., 2012). Mortality was calculated according to the average life expectancy at birth in Uganda of 54.20 years from 2008 (World Bank, 2016). Finally, sequelae such as Guillain-Barre syndrome, reactive arthritis or irritable bowel syndrome are not considered in the model.

2.5.2 QMRA implementation and analysis

As summarised in Table 1, spatial and temporal variability of the number of CFU of E. coli, Salmonella spp. and number of eggs of Ascaris spp. were measured from October to December 2013 during the short rainy season at 23 sampling points over the four study areas. For E. coli and

Salmonella spp., we fitted normal distributions to the log-transformed enumeration data on concentration in the water (Cwater), using a maximum likelihood estimation (MLE) method, allowing inclusion of censored data and accounting for the abundance, while considering the measured prevalence of the indicator bacteria in the water along the four systems (Lorimer and Kiermeier, 2007), in Excel 2013 (Microsoft Corporation, Redmond; WA, USA). As a result, the data fitting provided estimates for the true prevalence of contaminated water samples, and the distribution of concentrations in these contaminated samples. For Ascaris spp. eggs, this approach was not possible as only four out of 168 (excluding Lake Victoria) samples were positive. Hence, with a value of 0.024, a positive count is expected, between 1 and 100 eggs/L, which is included with uniform distribution on a

rm distri T) distri

log-scale (Fuhrimann et al., 2015). Project evaluation and review techniques (PERT) distributions are fitted to minimum, most likely and maximum ratio of pathogen concentration per E. coli (ppath) for rotavirus, norovirus, Campylobacter spp. and Cryptosporidium spp. Uniform distribution were fitted to E. coli and Salmonella spp. ratio of pathogen concentration per measured E. coli and Salmonella spp. (WHO, 2006; Katukiza et al., 2013). This is implemented in the model by assuming that a fraction ppath of the ingested volumes of water consists of a pathogenic strain of the bacterial species. PERT distributions are also fitted to assumed minimum, most likely and maximum ingestion rates (volume (V) in mL water) per exposure event. In a Monte Carlo simulation, values are sampled for these three variables and the ingested amount of pathogens (dose; d) is calculated as:

unt of pathoge unt of pathoge

d = Cwater x pPath x V(equation 1).

The variation in Cwater is implemented as variability per exposure event, the variation in ppath and V is implemented as variability per person (i.e. for practical reasons it had the same value for all exposure events for one person in one iteration of the Monte Carlo simulation). As ingested bacteria are discrete units, assumed to be homogeneously distributed in the water, ingested doses are assumed to be Poisson distributed (d ~Poisson (d) as e.g. in (Nauta et al., 2012).

Doses (d) are used as input in the dose-response relations to obtain the probability of illness PI(d) (equations 1, 2 and 3). Monte Carlo simulations are performed for 100,000 iterations using @Risk, version 6 (Palisade Corporation; Newfield, NY, USA), where one iteration simulates all the n exposure events and associated Pin(d) of one person in a year. Based on this, the expected frequency of illness for this person per year (which, in our approach, can be more than one) can be calculated as the sum of the n values of Pin(d) obtained. Model outputs are presented as number of cases per year, DALYs pppy and total DALYs per year (see equations 6, 9 and 10).

2.6. Dose-response models

Well-established dose-response models for the various pathogens were used to determine the relationship between quantity of exposure (i.e. number of organism ingested) and the effective health outcome (i.e. infection and illness) (Haas et al. 2014). For the QMRA, the simplified Beta-Poisson dose-response models for rotavirus, Campylobacter spp., E. coli O157:H7, pathogenic Salmonella spp. and A. lumbricoides were employed (Teunis and Havelaar, 2000; McBride et al., 2013; Haas et al., 2014), as defined as

P,(d) = 1 - 11 + (d)| / (equation 2)

with a median infectious dose defined as with a median infectious dose defined as

1 ) (equation 3)

For norovirus, a hypergeometric function is presented by Teunis et al. (2008), which is fit to run with the @Risk software and to include the uncertainty about the dose-response. We used an approximation for the mean probability of infection (Haas, 2002) of the Beta-Poisson dose-response model:

n -r r( a+ P)r(d + P) / A\

P(d ) = 1 - r(g+P+d)rw (equation 4)

where Г(.) represents Eulers gamma function. For Cryptosporidium spp., an exponential model was used (Westrell et al., 2004; de Man et al., 2013; McBride et al., 2013):

P i(d) = 1 - ( 1 - r)d (equation 5).

In brief, ( ) represents the probability of infection, is a single dose of the pathogen, whereas the pathogen infectivity constants , and characterise the dose-response relationship. To account for the proportion of infections that turn into symptomatic gastroenteritis cases ( ( )) we used for each pathogen a constant value (2) (i.e. illness to infection ratio):

thogen a c thogen a c

Pi11(d) = Pi(d) x 2 (equation 6).

Table 1 provides the parameter values used in the QMRA for each pathogen. Moreover, for Pm(d) we use the probability of a symptomatic gastroenteritis for each of the seven pathogens (or hazards) h, Piii,h(dt), which is a function of the ingested dose di at exposure event i.(Haas et al., 2014).

2.7. Risk characterisation

2.7.1. Incidence: the number of cases per year

Our model assumed that each exposure event i is independent and that there is no immunity after a previous infection (no dose-response available for the context of LMICs and, hence, it is not possible to include immunity status of exposed population groups in this model). For an average of nh exposures to the hazard per person per year, with population size PopE, the expected number of cases in the population is:

The incidence estimate for pathogen h is:

Casesh = P°VEPii1,h(.di) (equation 7).

6 Cast = —

xoenteritis per ye

DALY metri

h (equation 8).

The combined incidence estimate, Inccomb (episodes of gastroenteritis per year), for all seven pathogens h is defined as:

1ncC0mb = Zaii h^Sh (equation 9).

2.7.2. Estimation of disease burden The disease burden was expressed by the DALY metric. This metric combines morbidity (years lived with disability) and premature death (years of life lost) (Murray et al., 2012). For each pathogen h, DALY per case of gastrointestinal illness (DALYh) was estimated as the sum of the product of the probability of developing disease symptom j (i.e. j = mild, moderate and severe diarrhoea or death) given a case of gastroenteritis, relative frequency of the symptom f), duration of the developed symptom in years (D;) and the respective severity factor (Sj) (Table 2) (Salomon et al., 2012):

DALYh = Zj(fjXDj X Sj) (equation 10).

The total ( sease burden (TotalDALYs,h) per hazard was the product of cases (Casesh) and DALYs per pathogen:

TotalDALYs,h = Casesh x D ALYh (equation 11).

2.8. Sensitivity analysis to detect uncertainty and effect of potential interventions

To explore the uncertainty around the model outputs in DALYs, a nominal range sensitivity analysis

(NRSA) was done. The NRSA was employed for scenario Spicying only to exemplify the impact of

small changes in some of the model parameters used. Selected individual inputs were varied over a certain range, while holding all other inputs at their nominal values (Table 3). The following five parameter groups were investigated to obtain information about their uncertainty and effect of a potential intervention: (i) (V) volume ingested per exposure event were based on assumptions and, hence, values were multiplied by 0.1 and 10; (ii) (ppath) ratio between indicator and pathogenic organisms vary considerably between different contexts and, hence, values were multiplied by 0.1 and 10; (iii) (Cwater) water contamination in community areas and the potential effect of treatment while varying the concentration with 1 log; (iv) (n) number of exposure events per year vary according to implemented health education programmes and, hence, we simulated a reduction by half (153 events) and to one event per year; and (v) (PopE) population at risk per exposure scenario may also vary according to the exposure event, we multiplied the population by 0.1

The study protocol was approved by the institutional research commission of the Swiss Tropical and Public Health Institute (Swiss TPH; Basel, Switzerland; reference no. FK 106) and the Uganda National Council for Science and Technology (UNCST; Kampala, Uganda; reference no. HS 1487). Ethical approval was obtained from the ethics committee in Basel (EKBB; reference no. 137/13) and the Higher Degrees Research and Ethics Committee of Makerere University, School of Public Health (Kampala, Uganda; reference no. IRBOOO11353). This study is registered with the clinical trial registry ISRCTN (identifier: ISRCTN13601686).

3.1. Incidence of gastroenteritis per year

The combined estimated incidence (Inccomb) was highest for urban farmers (Sfarming), children living in slum communities (Splaying) and sanitation workers (Sworking) who suffer from 10.9, 8.3 and 4.3 gastroenteritis episodes per person per year (mean values; Figure 3 and Table 4). The lowest risk was estimated for people swimming in Lake Victoria (Sswimming) who suffer from 0.18 gastroenteritis episodes per year. With regard to the individual pathogens, the risk of gastrointestinal infection was highest for children (Splaying) and urban farmers (Splaying) for rotavirus (4.4 and 3.9 episodes per year, respectively) and E. coli (2.7 and 1.7 episode per year, respectively). Considerably lower incidences were estimated for the same scenarios for A. lumbricoides (0.062 and 0.001 episode per year, respectively) and norovirus (between 0.46 and 0.34 episode per year, respectively).

3.2. Number of gastroenteritis cases per year

Among the overall 18,204 exposed people in the Nakivubo area, 59,493 cases of gastroenteritis were estimated due to any of the seven pathogens, five scenarios and over the duration of one year (Figure 3 and Table 4). Gastrointestinal infection due to rotavirus, E. coli and Cryptosporidium contributed most

2.9. Ethical considerations

3. Results

to the total cases (46%, 21% and 17%, respectively), followed by Campylobacter spp. (12%) and norovirus (4%). Together, 82% of all cases were concentrated in the 5,928 children living in slum communities (Splaying: 48,882 cases) and 650 urban farmers (Sfarming: 7,111 cases).

3.3. Total disease burden per year

Across all five scenarios, the model estimated a burden of 304.3 DALYs per year due to exposure to wastewater in the Nakivubo area among 18,204 exposed people, for all seven pathogens together (Figure 3 and Table 4). The main responsible pathogens were E. coli, rotavirus and Campylobacter, accounting for 46%, 29% and 12% of DALYs, respectively. Children living in slum communities (Splaying; n=5928) and urban farmers (Sfarming; n=645) were most vulnerable with burdens of 236.8 and 47.8 DALYs, respectively.

3.4. Disease burden per person per year Combined DALYs pppy for all scenarios (summed-up for the seven pathogens and all exposed individuals) were far above the revised WHO reference level of 0.0001 DALYs pppy (Figure 3 and Table 4). The highest impact was estimated for urban farmers in the Nakivubo wetland (Sfarming), children living in slum communities (Splaying) and workers maintaining sanitation infrastructures (SWorking), with DALYs pppy of 0.074, 0.040 and 0.017, respectively. In terms of different pathogens in Sfarming, E. coli, Salmonella spp. and rotavirus had the largest share with 0.031, 0.018 and 0.014 DALYs pppy, respectively.

nalysis, st

3.5 Sensitivity analysis

The effect of the sensitivity analysis, stratified for uncertainty and intervention scenarios, is shown in Figure 3 for total cases of gastroenteritis. Uncertainty analysis revealed the highest change when adapting the volume of water accidentally ingested (-0.69 and 0.58). The pathogen ratio showed highest variation for rotavirus, Cryptosporidium and E. coli. Intervention strategies to reduce the water contamination with E. coli would have an impact of 0.70 and 0.58. Reducing the number of exposure events by half or to one event per year would reduce the number of gastroenteritis episodes by 0.29 and 2.56, respectively. The change of population at risk was found to be proportional to the indicated people exposed.

4. Discussion

4.1. Estimated burden due to gastroenteritis

Our estimated disease burden of 304.3 DALYs across all 18,204 exposed people per year corresponds to the estimates by the global burden of diseases study for entire Uganda of 0.017 DALYs pppy (612,202 DALYs considering a total population of Uganda 35.4 million people) (GBD 2010, UBOS 2013). Broken down to the individual exposure groups in our model, urban farmers, children in slum

communities and sanitation workers experience a 7, 3 and 2 times higher disease burden due to gastroenteritis than the general population in Uganda, respectively. These estimates are still lower than the disease burden estimates made for a typical slum area in Kampala of 10,172 DALYs (15,015 people) (Katukiza et al., 2013) and the 31,979 DALYs for 286,833 people being exposed to the urban wastewater systems in Accra (Labite et al., 2010). This large discrepancy between the QMRA estimates can partly be explained by different pathogens used for the QMRA, applied DALY estimates per pathogens (e.g. we excluded sequelae) and methodological differences (e.g. in the calculation of risk of illness). A common finding of QMRA studies in Africa is that E. coli and rotavirus together cause more than half of the disease burden (Katukiza et al., 2013; Machdar et al., 2013). The relatively high model-based estimate for pathogenic E. coli is supported by a recent case-control study in Cote d'lvoire, which found that enterotoxigenic E. coli is indeed one of the most prevalent pathogens, being the causative agent in 32% of all participants with persistent diarrhoea (>2 weeks) (Becker et al., 2015). The importance of rotavirus infection was demonstrated in an investigation at the Mulago Hospital in Kampala, where the rotavirus was detected in 177 out of 390 children aged 3 to 59 months (45.4%) presenting with acute diarrhoea (Nakawesi et al., 2010). Accordingly, the WHO reports that rotavirus is the main diarrhoea-causing agent in this age group, causing 7.3% of all deaths in children aged under 5 years in Uganda (Sigei et al., 2015).

4.2. Estimated incidence of gastroenteritis Few papers have compared or even validated QMRA estimates with diarrhoea episodes assessed by epidemiological studies (Bouwknegt et al., 2014; Haas et al., 2014). In conjunction with the environmental assessment carried out for the QMRA, we have implemented a cross-sectional epidemiological survey in slum dwellers at risk of flooding events, urban farmers and sanitation workers (Fuhrimann et al., 2016). In this survey the self-reported 14-day incidence of diarrhoea episodes ranged from 0.25 in slum dwellers and urban farmers to 0.29 in workers along the sanitation system. When extrapolating this 14-day incidence to 1 year (52 weeks, without considering seasonality), the annual incidence would range between 6.6 and 7.6 episodes pppy. The comparison of the incidence estimate from the cross-sectional survey with the combined incidence of all seven pathogens used for the QMRA reveals that estimates are similar (i.e. 10.9 versus 6.6 in urban farmers and 4.3 versus 7.6 in sanitation workers, 0.3 versus 6.6 in community members). Hence, the estimates might match even more had the QMRA included the level of immunity (reduction in estimates) and additional exposures adding to diarrhoea incidence such as human to human transmission or consumption of contaminated water and food (Machdar et al. 2013; Barker 2014). Further, when focusing on helminth infections, the QMRA estimated 92 people to be infected with A. lumbricoides over the course of a year, with 52 cases occurring in urban farmers (8% of the total number). The incidence estimate is in contrast to findings from our cross-sectional survey, which showed high prevalence in adult urban farmers: A. lumbricoides: 18.4%; T. trichiura: 26.1%; hookworm: 27.8%;

and S. mansoni 22.9%, respectively (Fuhrimann et al., 2016). The comparatively low model-based estimates can be explained by the low prevalence and concentration of helminth eggs measured in wastewater samples. Indeed, only four samples were positive for Ascaris spp. eggs out of 168 (Fuhrimann et al., 2015). On the other hand, the high prevalence in urban farmers may result from accumulation of worms over time as deworming was reported not to be done on a regular basis (Fuhrimann et al., 2016). Similar tendencies, i.e. higher risk estimates by the QMRA than suggested by epidemiological surveys while predicting a lower number of infection, have also been shown by other studies (Havelaar et al., 2008). This points at the need for further cross-comparison between epidemiological surveys and QMRA with the ultimate goal to develop a standardised procedures to assess incidence and burden of diarrhoeal episodes and intestinal parasitic infections to make use of both tools.

4.3. QMRA limitations

Our model framework entails several limitations. The model relies on a range of assumptions pertaining to the volume of ingested water, the indicator to pathogen ratio and the number of exposure events, which were addressed in the uncertainty analysis and may lead to overestimation of the results. Especially the use of pathogen ratios is a key limitation, as E. coli counts may not reflect the densities of enteric virus in water bodies accurately (O'Toole et al., 2014). The dose-response models applied in this model are based on feeding studies (e.g. norovirus) or rely on epidemiological evidence (e.g. Ascaris spp.) conducted with healthy individuals in high income countries. Thus, dose-response may be considerably different due to acquired immunity related to exposure history or vaccination (Haas et al., 2014; Havelaar and Swart, 2014). The simplification of allowing no immunity after an exposure event in the QMRA may result in an overestimation of the disease burden (Haas et al., 2014). Certain pathogens, such Rotavirus, may have considerable different health impact in different age classes, which have not be taken it to account in this QMRA (Sigei et al., 2015). The helminth eggs in water can vary greatly as seasonal and clustered transmission of helminths is common (Cairncross et al., 1996). The model excluded exposure pathways such as dermal contact and inhalation and people living in a slum-like environment may also be exposed to contaminated drinking water and greywater, which should be taken into account in future QMRAs (Howard et al., 2006; Machdar et al., 2013). Further, it is known that in Kampala other pathogens such as adenoviruses, hepatitis A virus, Vibrio cholerae, hookworm and S. mansoni are present in the environment (Bwire et al., 2013; Katukiza et al., 2013); these were not included in the model although they can cause diarrhoea and other adverse health effects (Karagiannis-Voules et al., 2015; Fuhrimann et al., 2016).

4.4. Sensitivity analysis

In the sensitivity analysis we showed a considerable effect of different volume of water accidentally ingested. These values might, however, not be very accurate as exposure to wastewater very much

depends on individual behaviours in the given environment and is also influenced by age, sex, level of education and socio-economic status (WHO, 2006; Haas et al., 2014). Clearly, there is a need to generate specific estimates for accidental ingestion of water during different exposures in low-income countries in the global south. In absence of valid information on the pathogenic strains, assumptions were drawn based on published ratios between E. coli and the pathogenic strain. We showed that especially values for rotavirus, E. coli and Campylobacter have a considerable effect on the total number of gastroenteritis cases. Studies have reported that temporal and spatial variation of environmental pollution is common in urban wastewater systems (Ensink, 2006; Katukiza et al., 2013; Fuhrimann et al., 2015). Furthermore, the E. coli to pathogen ratio might change over time as E. coli is secreted by humans and animals continuously, whereas pathogens are secreted only by a proportion of infected people over a short period of a few days (Mara, 2004).

kiza et al., 2 over time as E. c

nly by a pr< &

4.5. Proposed mitigation strategies for exposure scenarios

In view of our findings, and acknowledging inherent limitations, a set of options for reducing disease

burden for each of the five exposure groups are proposed (from highest burden to lowest burden).

Importantly, the choice of the appropriate mitigation strategy needs to take into account cost-

efficiency as well as acceptability in concerned population groups (Machdar et al., 2013).

• Splaying : in order to protect children (<15 years) living in slum communities, oral rotavirus vaccination could be added to the Ugandan immunisation schedule and given to children at the age of >6 weeks (The Republic of Uganda, 2012; UNAS, 2014). This may be supplemented with bi-annual hygienic and deworming campaigns at schools, as well as at the level of households for targeting women of childbearing age (WHO, 2011b; Sigei et al., 2015).

• Sfarming: the health of farmers working in the Nakivubo wetland can be promoted by means of farmer field schools. For example, in workshops on occupational health risks the value of personal protective equipment and the importance of sanitation and personal hygiene can be introduced (Van Den Berg and Takken, 2007).

• Sworking : sanitation workers employed by the NWSC and those working for the PEA could be trained on the recognition of health risks and effective use of personal protective equipment. Moreover, bi-annual hygienic and deworming campaigns could be implemented (Van Den Berg and Takken, 2007; Strande et al., 2014).

• Sfl00&ng: slum dwellers living in close proximity to the Nakivubo wetland could be protected against flooding events through the construction of small dams and drainage systems. In addition, access to frequently flooded areas along the Nakivubo channel might be restricted via perimeter fences (Katukiza et al., 2010).

• Sswimming : swimmers at the local beaches along the Inner Murchison Bay (e.g., Miami Beach, Ggaba Beach, KK beach) could be informed about the risks involved with swimming in Lake Victoria (e.g. warning signs at unsafe places). Regulations to restrict swimming at certain

commercial beaches close to the discharge point of the Nakivubo Channel could be introduced and the beach water quality could be monitored on an ongoing basis (Soller et al., 2015). Some general recommendations to reduce the exposure of the population to pathogens in contaminated water, proposed by others, are (i) re-establish the Nakivubo wetland flora to reclaim its function as a natural maturation pond and retention pool to protect the Murchison Bay in Lake Victoria (Mbabazi et al., 2010; Fuhrimann et al., 2015;); (ii) fight faecal contamination of slum areas by promoting sanitation coverage in combination with safe collection, treatment or disposal of faecal sludge (Fuhrimann et al., 2016); and (iii), in the longer term, increase wastewater treatment and reuse capacity of Kampala city (Strande et al., 2014; Fuhrimann et al., 2015;).

ent and ent and

5. Conclusions

By using a QMRA approach, we estimated high risk, and considerable burden, due to water-borne pathogens among different population groups being exposed to wastewater in Kampala. Disease burden was estimated to be highest in children living in slum communities and in urban farmers with 0.199 and 0.06 DALYs pppy, respectively. Indeed, the DALYs pppy for these two population groups were several thousand fold above the revised WHO tolerable level of 0.0001 DALYs pppy. Hence, exposure to wastewater is anticipated to have considerable public health implications, calling for action to reduce E. coli and rotavirus infection, which were found to be of major concern. The presented QMRA provides a case study on how the risk assessment framework of the WHO guidelines for the safe use of wastewater, greywater and excreta can be applied to fit an urban low-income setting. The QMRA framework was built on context specific data of indicator microorganism and recent burden estimates for the most common water-borne pathogens resulting in gastroenteritis in Uganda. We showed the need to further link QMRA and epidemiological studies and to elaborate on how to assess the number of exposure events, the indicator organism to pathogen ratio and the dose-response relationship. Such risk assessment frameworks can make an important contribution to our understanding of health impacts in different population groups related to specific exposures, and thus promote the development of targeted mitigation strategies in resource-constrained settings.

Acknowledgements

We thank Prof. Peter Teunis for his help in using the dose-response relation for norovirus. We are grateful to Dr. Jan Hattendorf and Prof. Anders Dalsgaard for the advice on the model parameters and Prof. Christian Schindler for his valuable inputs to the overall QMRA concept. We also thank Dr. Katherine Gibney for sharing her experience with DALY calculation.

References

Amoah, P., Keraita, B., Akple, M., Drechsel, P., Abaidoo, R.C., Konradsen, F., 2011. Low-cost options for reducing consumer health risks from farm to fork where crops are irrigated with

polluted water in West Africa. Colombo.

Anonymous, 1995. Application of risk analysis to food standards issues. World Health Organization and Food and Agriculture Organisation, Geneva.

Ashbolt, N.J., Petterson, S.R., Roser, D.J., Westrell, T., Ottoson, J., Schönning, C., Stenström, T.A., 2006. Microbial risk assessment tool to aid in the selection of sustainable urban water systems, in: Strategic Planning of Sustainable Urban Water Management. IWA Publishing, London, pp. 43-55.

Barker, S.F., 2014. Risk of norovirus gastroenteritis from consumption of vegetables irrigated with highly treated municipal wastewater-evaluation of methods to estimate sewage quality. Risk Anal. 34, 803-817.

Barker, S.F., Amoah, P., Drechsel, P., 2014. A probabilistic model of gastroenteritis risks associated with consumption of street food salads in Kumasi, Ghana: evaluation of methods to estimate pathogen dose from water, produce or food quality. Sci. Total Environ. 487, 130-142.

Becker, S.L., Chatigre, J.K., Gohou, J.-P., Coulibaly, J.T., Leuppi, R., Polman, K., Chappuis, F., Mertens, P., Herrmann, M., N'Goran, E.K., Utzinger, J., von Müller, L., 2015. Combined stool-based multiplex PCR and microscopy for enhanced pathogen detection in patients with persistent diarrhoea and asymptomatic controls from Côte d'Ivoire. Clin. Microbiol. Infect. 21, 591.e1-591.e10.

Becker, S.L., Vogt, J., Knopp, S., Panning, M., Warhurst, D.C., Polman, K., Marti, H., von Müller, L., Yansouni, C.P., Jacobs, J., Bottieau, E., Sacko, M., Rijal, S., Meyanti, F., Miles, M.A., Boelaert, M., Lutumba, P., van Lieshout, L., N'Goran, E.K., Chappuis, F., Utzinger, J., 2013. Persistent digestive disorders in the tropics: causative infectious pathogens and reference diagnostic tests. BMC Infect. Dis. 13, 1-21.

Beller Consult, MacDonald Mott, M&E Associates, 2004. Sanitation strategy & master plan for Kampala city. Kampala.

Blumenthal, U., Peasey, A., 2002. Critical review of epidemiological evidence of the health effects of wastewater and excreta use in agriculture. London School of Hygiene and Tropical Medicine, London.

Bouwknegt, M., Knol, A.B., van der Sluijs, J.P., Evers, E.G., 2014. Uncertainty of population risk estimates for pathogens based on qmra or epidemiology: a case study of campylobacter in the Netherlands. Risk Anal. 34, 847-864.

Brooker, S., 2010. Estimating the global distribution and disease burden of intestinal nematode infections: adding up the numbers--a review. Int. J. Parasitol. 40, 1137-1144.

Bwire, G., Malimbo, M., Maskery, B., Kim, Y.E., Mogasale, V., Levin, A., 2013. The burden of cholera in Uganda. PLoS Negl. Trop. Dis. 7, e2545.

Cairncross, S., Blumenthal, U., Kolsky, P., Moraes, L., Tayeh, A., 1996. The public and domestic domains in the transmission of disease. Trop. Med. Int. Heal. 1, 27-34.

Cisse, G., 2013. Water-related disaster management and adaptation to climate change: bridges and challenges? Water Int. 38, 11-16.

de Man, H., van den Berg, H.H.J.L., Leenen, E.J.T.M., Schijven, J.F., Schets, F.M., van der Vliet, J.C., van Knapen, F., de Roda Husman, A.M., 2013. Quantitative assessment of infection risk from exposure to waterborne pathogens in urban floodwater. Water Res.

Ensink, J.H.J., 2006. Wastewater quality and the risk of hookworm infection in Pakistani and Indian sewage farmers. London School of Hygiene and Tropical Medicine, London, England.

Ensink, J.H.J., van der Hoek, W., 2009. Implementation of the WHO guidelines for the safe use of wastewater in Pakistan: balancing risks and benefits. J. Water Health 7, 464-468.

Ferrer, A., Nguyen-Viet, H., Zinsstag, J., 2012. Quantification of diarrhea risk related to wastewater contact in Thailand. Ecohealth 9, 49-59.

Fuhrimann, S., Stalder, M., Winkler, M.S., Niwagaba, C.B., Babu, M., Masaba, G., Kabatereine, N.B., Mage, A.A., Schneeberger; p.RR, monger, J., СМ. а, 2Ш. ^ЫЫ and ^ contamination of water, sediment and soil in the Nakivubo wetland area in Kampala, Uganda. Environ. Monit. Assess. 187, 475.

Fuhrimann, S., Winkler, M.S., Kabatereine, N.B., Tukahebwa, E.M., Halage, A.A., Rutebemberwa, E., Medlicott, K., Schindler, C., Utzinger, J., Cisse, G., 2016. Risk of intestinal parasitic infections in people with different exposures to wastewater and fecal sludge in Kampala, Uganda: a cross-sectional study. PLoS Negl. Trop. Dis. 10, e0004469. doi: 10.1371/journal.pntd.0004469

Fuhrimann, S., Winkler, M.S., Schneeberger, P.H.H., Niwagaba, C.B., Buwule, J., Babu, M., Medlicott, K., Utzinger, J., Cisse, G., 2014. Health risk assessment along the wastewater and faecal sludge management and reuse chain of Kampala, Uganda: a visualization. Geospat. Health 9,251-255.

Gibney, K.B., O'Toole, J., Sinclair, M., Leder, K., 2014. Disease burden of selected gastrointestinal pathogens in Australia, 2010. Int. J. Infect. Dis. 28, 176-185.

Haas, C., Rose, R., Gerba, C., 1999. Quantitative microbial risk assessment. John Wiley & Sons, New York.

Haas, C.N., 2002. Conditional dose response relationships for microorganisms: development and application. Risk Anal. 22, 455-463.

Haas, N.C., Rose, B.J., Gerba, P.C., 2014. Quantitative Microbial Risk Assessment, 2nd ed. John Wiley & Sons, Inc., New Jersey.

Havelaar, A.H., Evers, E.G., Nauta, M.., 2008. Challenges of quantitative microbial risk assessment at EU level. Trends Food Sci. Technol. 19, 22-29.

Havelaar, A.H., Swart, A.N., 2014. Impact of acquired immunity and dose-dependent probability of illness on quantitative microbial risk assessment. Risk Anal. 34, 1807-1819.

Havelaar, a H., de Wit, M. a, van Koningsveld, R., van Kempen, E., 2000. Health burden in the

Netherlands due to infection with thermophilic Campylobacter spp. Epidemiol. Infect. 125, 505522.

Howard, G., Pedley, S., Tibatemwa, S., 2006. Quantitative microbial risk assessment to estimate health risks attributable to water supply: can the technique be applied in developing countries with limited data? J. Water Health 4, 49-65.

Hunter, P.R., Payment, P., Ashbolt, N., Bartram, J., 2000. WHO risk assessement along the dringking wate chain: Chapter 3. World Health Organization, Geneva.

Hynds, P.D., Gill, L.W., Misstear, B.D., 2014. A Quantitative Risk Assessment of Verotoxigenic E. coli (VTEC) in Private Groundwater Sources in the Republic of Ireland. Hum. Ecol. Risk Assess. 20,1446-1468.

Ichida, A., Schaub, S., Soller, J., Nappier, S., Ravenscroft, J., 2015. Microbial risk assessment tools, methods, and approaches for water media. Microb. Risk Anal. 1, 12.

Kaakoush, N.O., Castaño-Rodríguez, N., Mitchell, H.M., Man, S.M., 2015. Global epidemiology of campylobacter infection. Clin. Microbiol. Rev. 28, 687-720.

Kaijuka, E., Kaija, E., Cross, A., Loaiza, E., 2011. Uganda demographic and health survey 2011, Uganda Ministry of Health. Kampala.

Kansiime, F., Nalubega, M., 1999. Wastewater treatment by a natural wetland: the Nakivubo Swamp, Uganda. Wageningen Agricultural University, Delft.

Karagiannis-Voules, D.-A., Biedermann, P., Ekpo, U.F., Garba, A., Langer, E., Mathieu, E., Midzi, N., Mwinzi, P., Polderman, A.M., Raso, G., Sacko, M., Talla, I., Tchuenté, L.-A.T., Touré, S., Winkler, M.S., Utzinger, J., Vounatsou, P., 2015. Spatial and temporal distribution of soil-transmitted helminth infection in sub-Saharan Africa: a systematic review and geostatistical meta-analysis. Lancet Infect. Dis. 15, 74-84.

Karavarsamis, N., Hamilton, A.J., 2010. Estimators of annual probability of infection for quantitative microbial risk assessment. J. Water Health 8, 365-373.

Kariuki, S., Gordon, M.A., Feasey, N., Parry, C.M., 2015. Antimicrobial resistance and management of invasive Salmonella disease. Vaccine 33 Suppl 3, C21-C29.

Katukiza, A., Ronteltap, M., Oleja, A., Niwagaba, C., Kansiime, F., Lens, P., 2010. Selection of sustainable sanitation technologies for urban slums: a case of Bwaise III in Kampala, Uganda. Sci. Total Environ. 409, 52-62.

Katukiza, A.Y., Ronteltap, M., van der Steen, P., Foppen, J.W.A., Lens, P.N.L., 2013. Quantification of microbial risks to human health caused by waterborne viruses and bacteria in an urban slum. J. Appl. Microbiol. 116, 447-463.

Kayima, J., Kyakula, M., Komakech, W., Echimu, S.P., 2008. A study of the degree of pollution in Nakivubo channel, Kampala, Uganda. J. Appl. Sci. Environ. Manag. 12, 93-98.

Keraita, B., Dávila, J.M., 2015. Risk mitigation for wastewater irrigation systems in low-income countries: opportunities and limitations of the WHO guidelines, in: Alternative Water Supply

Systems. IWA Publishing, London, pp. 367-369.

Labite, H., Lunani, I., Van Der Steen, P., Vairavamoorthy, K., Drechsel, P., Lens, P., 2010. Quantitative microbial risk analysis to evaluate health effects of interventions in the urban water system of Accra, Ghana. J. Water Health 8, 417-430.

Lai, Y.-S., Biedermann, P., Ekpo, U.F., Garba, A., Mathieu, E., Midzi, N., Mwinzi, P., N'Goran, E.K., Raso, G., Assare, R.K., Sacko, M., Schur, N., Talla, I., Tchuente, L.-A.T., Toure, S., Winkler, M.S., Utzinger, J., Vounatsou, P., 2015. Spatial distribution of schistosomiasis and treatment needs in sub-Saharan Africa: a systematic review and geostatistical analysis. Lancet Infect. Dis. 15,927-940.

Lorimer, M.F., Kiermeier, A., 2007. Analysing microbiological data: Tobit or not Tobit? Int. J. Food Microbiol. 116, 313-318.

Machdar, E., van der Steen, N.P., Raschid-Sally, L., Lens, P.N.L., 2013. Application of Quantitative Microbial Risk Assessment to analyze the public health risk from poor drinking water quality in a low income area in Accra, Ghana. Sci. Total Environ. 449, 134-142.

Mara, D., 2004. Domestic wastewater treatment in developping countries. Earthscan, Sterling, VA.

Mara, D., Bos, R., 2010. Risk analysis and epidemiology: the 2006 WHO guidelines for the safe use of wastewater in agriculture, Wastewater Irrigation. International Water Management Institute, Colombo.

Mara, D., Hamilton, A., Sleigh, A., Karavarsamis, N., 2010. Discussion paper: options for updating the 2006 WHO guidelines. World Health Organization, Geneva.

Mara, D., Sleigh, A., 2010. Estimation of norovirus and Ascaris infection risks to urban farmers in developing countries using wastewater for crop irrigation. J. Water Health 8, 572-576.

Matthys, B., Tschannen, A.B., Tian-Bi, N.T., Comoé, H., Diabaté, S., Traoré, M., Vounatsou, P., Raso, G., Gosoniu, L., Tanner, M., Cissé, G., N'Goran, E.K., Utzinger, J., 2007. Risk factors for

Schistosoma mansoni and hookworm in urban farming communities in western Côte d'Ivoire. Trop. Med. Int. Heal. 12, 709-723.

Mbabazi, J., Kwetegyeka, J., Ntale, M., Wasswa, J., 2010. Ineffectiveness of Nakivubo wetland in filtering out heavy metals from untreated Kampala urban effluent prior to discharge into Lake Victoria , Uganda. African J. Agric. Res. 5, 3431-3439.

McBride, G.B., Stott, R., Miller, W., Bambic, D., Wuertz, S., 2013. Discharge-based QMRA for estimation of public health risks from exposure to stormwater-borne pathogens in recreational waters in the United States. Water Res. 1-16.

Medema, G.J., Teunis, P.F., Havelaar, A.H., Haas, C.N., 1996. Assessment of the dose-response relationship of Campylobacter jejuni. Int. J. Food Microbiol. 30, 101-11.

Mok, H.-F., Hamilton, A.J., 2014. Exposure factors for wastewater-irrigated asian vegetables and a probabilistic rotavirus disease burden model for their consumption. Risk Anal. 34, 602-613.

Murray, C.J.L., Vos, T., Lozano, R., Naghavi, M., Flaxman, A.D., Michaud, C., 2012. Disability-

adjusted life years (DALYs) for 291 diseases and injuries in 21 regions, 1990-2010: a systematic analysis for the Global Burden of Disease Study 2010. Lancet 380, 2197-2223.

Nakawesi, J.S., Wobudeya, E., Ndeezi, G., Mworozi, E.A., Tumwine, J.K., 2010. Prevalence and factors associated with rotavirus infection among children admitted with acute diarrhea in Uganda. BMC Pediatr. 10, 69.

Nauta, M.J., Sanaa, M., Havelaar, A.H., 2012. Risk based microbiological criteria for Campylobacter in broiler meat in the European Union. Int. J. Food Microbiol. 158, 209-217.

O'Toole, J., Sinclair, M., Fiona Barker, S., Leder, K., 2014. Advice to risk assessors modeling viral health risk associated with household graywater. Risk Anal. 34, 797-802.

Okeke, I.N., 2009. Diarrheagenic Escherichia coli in sub-Saharan Africa: status, uncertainties and necessities. J. Infect. Dev. Ctries. 3, 817-842.

Pham-Duc, P., Nguyen-Viet, H., Hattendorf, J., Cam, P.D., Zurbrügg, C., Z insstag, J., Odermatt, P., 2014. Diarrhoeal diseases among adult population in an agricultural community Hanam province, Vietnam, with high wastewater and excreta re-use. BMC Public Health 14, 978.

Salomon, J.A., Vos, T., Hogan, D.R., Gagnon, M., et al., 2012. Common values in assessing health outcomes from disease and injury: disability weights measurement study for the Global Burden of Disease Study 2010. Lancet 380, 2129-2143.

Schets, F.M., Schijven, J.F., de Roda Husman, A.M., 2011. Exposure assessment for swimmers in bathing waters and swimming pools. Water Res. 45, 2392-23400.

Seidu, R., Heistad, A., Amoah, P., Drechsel, P., Jenssen, P.D., Stenström, T.-A., 2008. Quantification of the health risk associated with wastewater reuse in Accra, Ghana: a contribution toward local guidelines. J. Water Health 6, 461-471.

Shere, J.A., Kaspar, C.W., Bartlett, K.J., Linden, S.E., Norell, B., Francey, S., Schaefer, D.M., 2002. Shedding of Escherichia coli O157:H7 in dairy cattle housed in a confined environment following waterborne inoculation. Appl. Environ. Microbiol. 68, 1947-1954.

Sigei, C., Odaga, J., Mvundura, M., Madrid, Y., Clark, A.D., 2015. Cost-effectiveness of rotavirus vaccination in Kenya and Uganda. Vaccine 33 Suppl 1, A109-18. doi:10.1016/j.vaccine.2014.12.079

Soller, J.A., Eftim, S., Wade, T.J., Ichida, A.M., Clancy, J.L., Johnson, T.B., Schwab, K., Ramirez-Toro, G., Nappier, S., Ravenscroft, J.E., 2015. Use of quantitative microbial risk assessment to improve interpretation of a recreational water epidemiological study. Microb. Risk Anal. 1, 2-11.

Soller, J., Embrey, M., Tuhela, L., Ichida, A., Rosen, J., 2010. Risk-based evaluation of Escherichia coli monitoring data from undisinfected drinking water. J. Environ. Manage. 91, 2329-35.

Stenström, T.A., Seidu, R., Nelson, E., Christian, Z., 2011. Microbial exposure and health assessments in sanitation technologies and systems. Stockholm Environment Institute, Stockholm.

Stott, R., Mayr, E., Mara, D.D., 2003. Parasite removal by natural wastewater treatment systems: performance of waste stabilisation ponds and constructed wetlands. Water Sci. Technol. 48, 97-

Strande, L., Ronteltap, M., Brdjanovic, D., 2014. Faecal sludge management: systems approach for

implementation and operation, 1st ed. IWA Publishing, London. Teunis, P.F.M., Havelaar, A.H., 2000. The Beta Poisson dose-response model is not a single-hit

model. Risk Anal. 20, 513-520. Teunis, P.F.M., Moe, C.L., Liu, P., Miller, S.E., Lindesmith, L., Baric, R.S., Le Pendu, J., Calderon,

R.L., 2008. Norwalk virus: how infectious is it? J. Med. Virol. 80, 1468-1476. The Republic of Uganda, 2012. Uganda national expanded programme on immunization multi year

Van Den Berg, H., Takken, W., 2007. Viewpoint: a framework for decision-making in integrated vector management to prevent disease. Trop. Med. Int. Heal. 12, 1230-1238.

Westrell, T., Schönning, C., Stenström, T.A., Ashbolt, N.J., 2004. QMRA (quantitative microbial risk assessment) and HACCP (hazard analysis and critical control points) for management of pathogens in wastewater and sewage sludge treatment and reuse. Water Sci. Technol. 50, 23-30.

WHO, 2015. Sanitation safety planning: manual for safe use and disposal of wastewater, greywater and excreta. World Health Organization, Geneva.

WHO, 2011a. Guidelines for drinking-water quality, 4th ed. World Health Organization, Geneva, Switzerland.

WHO, 2011b. Helminth control in school-age children. World Health Organization, Geneva.

WHO, 2006a. WHO guidelines for the safe use of wastewater, excreta and greywater. Volume I-IV. World Health Organization, Geneva.

WHO, 2006b. WHO guidelines for the safe use of wastewater, excreta and greywater. Volume I: policy and regulatory aspects. World Health Organization, Geneva.

WHO, 2003. Guidelines for safe recreational water environments. World Health Organization, Geneva.

World Bank, 2016. World Bank Open Data: Uganda [WWW Document]. URL http://data.worldbank.org/country/uganda (accessed 10.7.16).

Yapo, R.I., Kone, B., Bonfoh, B., Cisse, G., Zinsstag, J., Nguyen-Viet, H., 2013. Quantitative microbial risk assessment related to urban wastewater and lagoon water reuse in Abidjan, Cote

plan 2012-2016. The Republic of Uganda, Kampala. UBOS, 2013. 2013 statistical abstract. Uganda Bureau of Statistics, Kampa] UNAS, 2014. Improving Vaccine and Immunisation Coverage in Uganda. of Sciences.

d'Ivoire. J. Water Health 1-11. doi:10.2166/wh.2013.051

Figures caption

Figure 1 Schematic flow of Kampala's main wastewater system along the Nakivubo channel, Nakivubo wetland and Lake Victoria with indication of the five exposure scenarios (Sf00ding,

Sworking, Sfarming, Splaying and Sswimming) E-" - — ■ - - Li-n = ri lt! Kj rr.pl 11 =in

C: 1 1 4. S tA»

Figure 2 Exposure scenarios (Siding, Sworking, Sfarming, Splaying and Sswimming) considered in the quantitative microbial risk assessment (QMRA) to estimate the burden of Campylobacter spp., Escherichia coli 0157:H7, Salmonella spp., norovirus, rotavirus, Cryptosporidium spp. and Ascaris lumbricoides along the major wastewater system in Kampala.

1. Hazards identification

Measured indicator organisms in water: total faecal coliform, Escherichia coli, Salmonella spp.. Ascaris spp.

Considered pathogens in QMRA: norovirus, rotavirus, Campylobacter spp., E. coli 0152:H12, Salmonella spp,, Cryptosporidium spp,, Ascaris lumbricoides

Source of contamination

Exposure scenarios

Exposure groups

Nakivubo channci

Slum dwellers n = 5,640

Nakivubo wetland

Community areas

Sanitation workers n = 231

Urban farmers (> 15 years) n = 645

Children (<15 years) n = 5,928

Lake Victoria

1 « ■ ^flooding f «r n "' ^working T tic W Jfarming d "Ç " pittyutg C iiÇ » ^swimming

Slums at risk of Work along the Playing in slum

flooding channel Farming community Swimming

Ml ♦

Swimmers

n = 5,760

Exposure everts per year

n = 312

n = 297

n = 365

Accidental ingestion of water per exposure event

10 - 30 raL

I - 5 mL

10 - 50 mL 1-5 mL

20 - 50 mL

3. Dose-rcsponsc

Probability of infection per person per exposure event ^ probability of illness per person per exposure event

4. Risk

characterisation

Incidence of illness per person ■ ► number of cases I DALYs per person per year ^^ total DALYs

Figure 3 Estimated gastroenteritis incidence per year (Inc. py), number of cases, disability-adjusted life years (DALYs) per year (py) and per person per year (pppy). (a) and (b) showing estimates of the respective outcomes per Siding, Sworking, Sfarming, Splaying and Summing- (c) and (d) is indicating the contribution of individual pathogens and scenarios, respectively, to the total estimated numbers per outcome along the major wastewater system in Kampala.

No. of cases ♦ Incidence per year

50'000 40'000 30'000 20'000 lO'OOO 0 J

ç/ Ç,/ c^ c,f\f

12.00 10.00 8.00 6.00 4.00 2.00 0.00

DALYs per year ■ DALYs per person per year

150 ^ h 0.010

100 50

0-1— — ™ ™ — L 0.001

/" / / <?f ^

Pathogens

80% 70% 60% 50% 40% 30% 20% 10% 0%

23.01 59.472 304.3 0.0167 Inc. py Cases DALY DALY V Ascaris lubricoides pppy

□ Cryptosporidium spp. H Salmonella spp.

□ Escherichia coli

■ Campylobacter spp.

■ Rotavirus

□ Notovirus

playing ^fanning working Hooding

Figure 4 Sensitivity analysis showing change in log(proportion) of gastroenteritis cases according to the baseline scenario (Splaying, children in slum communities) and stratified in uncertainty and intervention scenarios. Green line = upper estimates; red line = lower estimates. *indicating change in parameter values (upper; lower estimates).

Baseline scenario: 48,877 cases of gastroenteritis

(V) Volume ingested per exposure event (x 0.1 ; 10)

Norovirus

(Ppnth)

Ratio between indicator and pathogenic organisms (x 0.1;10)

Rotavirus

Campylobacter E. coli

Salmonella spp.

I 2.1 2.2

Cryptosporidium 2.6

(Cm!tr) Water E. coli

quality in community SaimoneUa spp. areas (-/+ 1 log)

Ascaris spp. (n) Exposure events per year (n 1;153> (PopE) Population at risk (x 0.I;10)

.....5

-1.5 -] -0.5 0 0.5 I 1.5

Log proportional change of gastroenteritis cases according to baseline scenario

Table 1 QMRA model parameter, distributions and assumption

Description

(Cwter) Concentrations:

Escherichia coli Salmonella spp. Ascaris spp. (Cwaer) Concentrations:

E. coli

Salmonella spp. Ascaris spp. (Cwaer) Concentrations:

E. coli

Salmonella spp. Ascaris spp. (Cwaer) Concentrations:

E. coli

Salmonella spp. Ascaris spp._

water Nakivubo channel logio (CFU/100mL) logio (CFU/iOOmL) logio Eggs/iL water Nakivubo wetland logio (CFU/iOOmL) logio (CFU/iOOmL) logio Eggs/iL water Community areas logio (CFU/iOOmL) logio (CFU/iOOmL) logio Eggs/iL water Lake Victoria

logio (CFU/iOOmL) logio (CFU/iOOmL) logio Eggs/iL

Distribution and/ or values

References

Normal(6.0;1.1) ; prevalence = 1 Normal(2.7;0.8)*; prevalence = 1 Uniform(0;2)**; prevalence = 0.024

Normal(5.0;1.5)*; prevalence = 1 Normal(2.1;1.3)*; prevalence = 0.95 Uniform(0;2)**; prevalence = 0.024

Fuhrimann et al., 2015

Normal(5.9;1.4) ; prevalence = 1 Normal(2.1;1.3)*; prevalence = 1 Uniform(0;2)**; prevalence = 0.024

Normal(1.8;3.0)*; prevalence = 0.69 Normal(0.6;2.0)*; prevalence = 0.50 Uniform(0;2)**; prevalence = 0.024

(ppath) Ratio between indicator and pathogenic organisms

A. lumbricoides to Ascaris spp Campylobacter spp. to E. co Cryptosporidium spp. to E. coli Pathogenic E. coli: O157:H7 to E. coli

Point estimate: ppath = 1 PERT(0.1;0.55;1)"* per 105 E. Coli PERT(0.01;0.055;0.1)*" per 105 E. coli Uniform (7.6 x 10"4;1x10"2)**

Norovirus to E. coli Rotavirus to E. coli

Pathogenic Salmonella to Salmonella spp.

PERT(0.1;0.55;1)* PERT(0.1;0.55;1)*

per 105 E. Coli per 105 E. Coli

Uniform (7.6 x 10-4;1 x 10"2)*'

Mara et al., 2010 WHO, 2006 WHO, 2006

Shere et al., 2002; Soller et al., 2010;

Hynds et al., 2014

WHO, 2006

Katukiaza et al., 2013

Shere et al., 2002; Soller et al., 2010;

Hynds et al., 2014_

(V) Volume ingested per exposure event for each scenario

Sfoodmg Sworking Sfa

)playing

mL mL mL mL mL

PERT(10;20;30) '

PERT(1;3;5)***

PERT(10;35;50)*'

PERT(1;3;5)***

PERT(20;35;50)*

Katukiaza et al., 2013 WHO, 2006; Labite et al., 2010 WHO, 2006; Labite et al., 2010 Katukiaza et al., 2013 Schets et al., 2011; Yapo et al., 2013

Dose-response models

A. lumbricoides Point estimate: a = 0.0104; N50 = 859 Mara et al., 2010

Campylobacter spp. Point estimate: a = 0.145; N50 = 896 Medema et al., 1996

Cryptosporidium spp. Point estimate: r = 0.0042 Haas et al., 1999

E. coli O157:H7 Point estimate: a = 0.49; N50 = 596,000 Teunis et al., 2008

Norovirus Point estimate: a = 0.04; ß = 0.055; Teunis et al., 2008

Rotavirus Point estimate: a = 0.253; N50 = 6 Teunis and Havelaar, 2000

Pathogenic Salmonella spp. Point estimate: a = 0.3126; N50 = 23,600 Haas et al., 1999

(A) Illness to infection ratio

A. lumbricoides Point estimate: 0.39 Mara et al., 2010

Campylobacter jejuni Point estimate: 0.3 Machdar et al., 2013

Cryptosporidium Point estimate: 0.79 Machdar et al., 2013

Pathogenic E. coli Point estimate: 0.35 Machdar et al., 2013

Norovirus Rotavirus

Pathogenic Salmonella spp.

Point estimate: eta = Point estimate: 0.5 Point estimate: 1

0.00255; r = 0.086

Teunis et al., 2008 Barker et al., 2014 McBride et al., 2013

(n) number of exposure events per year

Sfooding Sworking Sfa

Fuhrimann et al., 2016, UBOS, 2013

Splaying

Point estimate: 6 Point estimate: 312 Point estimate: 297 Point estimate: 365 Point estimate: 6

(Pops) population at risk per exposure scenario Sji„„ding people at risk of

flooding

Sworking workers

Sjarming urban farmers

Spiaying children in slum

communities

Fuhrimann et al., 2016

people swimming in Lake Victoria

Point estimate: 5,640

Point estimate: 231 Point estimate: 645 Point estimate: 5,928

Point estimate: 5,760

(DALYh) disease burden per A. lumbricoides Campylobacter spp. Cryptosporidium spp. Pathogenic E. coli Norovirus Rotavirus

Pathogenic Salmonella J££

pathogenic organisms DALYs/case DALYs/case DALYs/case DALYs/case DALYs/case DALYs/case DALYs/case

(Disability-adjusted life years (DALYs) calculation is indicated in Table 2) Point estimate: 0.0029 Point estimate: 0.0053 Point estimate: 0.0022 Point estimate: 0.0013 Point estimate: 0.0008 Point estimate: 0.0032 Point estimate: 0.0719

Uniform distribution (min; max):

Normal distribution (mean; standard deviation); (min; most likely; max)

Table 2 Disease burden due to gastroenteritis expresse calculated by means of severity (mild, moderate, severe of the respective severity grade per pathogen.

uation and review techniques (PERT)

disability-adjusted life years (DALYs) id fatality), probability, and duration

Severity weights

Gastroenteritis Mild Moderate Severe Fatal Total DALYs References

DALYs per severity grade of gastroenteritis 0.06 0.20 0.28 1.00 Salomon et al., 2012

Probability 0.92 0.07 0.01 7.80 x 10"6 Gibney et al., 2014

Norovirus Duration (days) 2.10 2.40 7.20

Duration (years) 0.01 0.01 0.02 54.20*

DALYs 3.24 x 10"4 9.56 x 10"5 3.33 x 10"5 4.23 x 10"4 8.75 x 10"4

Probability 0.85 0.10 0.05 3.37 x 10-5 Gibney et al., 2014

Rotavirus Duration (days) 4.90 7.10 7.70

Duration (years) 0.01 0.02 0.02 54.20*

DALYs 6.94 x 10-4 4.01 x 10-4 2.96 x 10-4 1.83 x 10-3 3.22 x 10-3

Cryptosporidium spp. Probability Duration (days) Duration (years) 0.86 5.00 0.01 0.12 15.00 0.04 0.02 33.00 0.09 Gibney et al., 2014

V. DALYs 7.19 x 10-4 1.02 x 10-3 4.32 x 10-4 2.17 x 10-3

Probability Campylobacter spp. Duration (days) 0.80 3.50 0.18 9.70 0.02 14.40 6.72 x 10"5 Havelaar et al., 2000 Gibney et al., 2014

Duration (years) 0.01 0.03 0.04 54.20*

DALYs 4.70 x 10"4 9.72 x 10"4 1.77 x 10"4 3.64 x 10"3 5.26 x 10"3

Pathogenic Salmonella spp. Probability Duration (days) Duration (years) 0.21 2.50 0.01 0.66 6.00 0.02 0.14 12.00 0.03 1.26 x 10"3 54.20* Gibney et al., 2014

DALYs 8.61 x 10"5 2.18 x 10"3 1.27 x 10"3 6.85 x 10"2 7.20 x 10"2

Probability 0.94 0.06 0.09 2.00 x 10-4 Katukiza et al., 2013

Pathogenic Escherichia coli Duration (days) Duration (years) 5.60 0.02 10.70 0.03 16.20 0.04 1.00 54.20*

DALYs 8.80 x 10"4 3.55 x 10"4 1.12 x 10"3 1.08 x 10"2 1.13 x 10"2

Ascaris lumbricoides

Probability Duration (days) Duration (years) DALYs

0.95 35.0 0.05 2.90 x 10"3

0.05 28.0 0.01 1.83 x 10"5

Brooker, 2010

2.92 x 10"3

*Average life expectancy at birth in Uganda from 2008 (World Bank, 2016)

Table 3 Sensitivity analysis of input parameters of the model indicated for scenario Sp

1playing.

Scenarios Description Units Distribution Estimates Upper Lower

Uncertainty (V) Volume ingested per exposure event for each scenario 1 Splaying mL PERT(1;3;5)*** 10;30;50 0.1;0.3;0.5

Ratio between indicator and pathogenic organisms

Norovirus to E. coli Rotavirus to E. coli

Campylobacter spp. to E. coli

Pathogenic E. coli to E. coli Pathogenic Salmonella to Salmonella spp.

2.5 Cryptosporidium spp. to E. coli

2.1 2.2

per 105 E. coli PERT(0.1;0.55;1)* per 105 E. coli PERT(0.1;0.55;1): per 105 E. coli Uniform(7.6 x 10-4;1 x 10"2)**

Uniform(7.6 x 10-

PERT(0.01;0.0:

;1 x 10-2)** 55;0.1)***

0.01;0.055;0.1

0.01;0.055;0.1

0.01;0.055;0.1 1x10-3-7.6 x10-5 1-7.6 x10-2 1x10-3-7.6 x10-5

0.1;0.5.5;1 0.001;0.0055;0.01

Interventions

(Cwater) Water contamination in 3.1 Escherichia coli log10 (CFU/100mL) Normal(5.9;1.4)* 6.9 4.9

community areas 3.2 3.3 Salmonella spp. Ascaris spp. log10 (CFU/100mL) log10 eggs/1L Normal(2.1;1.3)* ¿Uniform(0;2)" 3.1 0-3 2.1 0-1

(n) Children in vVV

Number of exposure 4 slum Events Point estimate: 365 1 153

events per year communities \

(POPE) Population at risk per 5 Number of children Children > Point estimate: 5,928 592 59,280

exposure scenario /V.l..

Project evaluation and review techniques (PERT)

(min; most likely; max)

Table 4 QMRA estimates for annual incidence of illness, number of cases per year, total DALYs per year, DALYs per person per year across all five exposure scenarios (Sfiooding, Sworking, S^arming,

playing

and Ssw

Exposure scenario (n = exposed population) Sfooding (n= 5,640) Sworking (n = 231) Sfarming (n = 645) Splayitig (n = 5,928) Sswimmmg (n = 5,760) (n = Total 18,204)

Incidence of gastroenteritis per year (Inch)

Norovirus 0.010 0.106 0.459 0.342 0.012 0.93

Rotavirus 0.127 1.812 4.402 3.880 0.055 10.28

Campylobacter 0.032 0.425 1.158 0.985 0.018 2.62

E. coli 0.053 0.675 2.697 1.679 0.032 5.14

Salmonella spp. 0.001 0.010 0.251 0.001 0.009 0.27

Cryptosporidium 0.039 0.371 1.893 1.355 0.050 3.71

A. lumbricoides 0.001 0.003 0.062 0.003 0.001 0.07

Total number 0.26 3.31 10.92 8.25 0.18 23.01

No. cases per year (Casesh)

Norovirus 57 25 299 2'030 70 2'480

Rotavirus 719 419 2'861 23'000 319 27'317

Campylobacter 178 98 753 5'837 101 6'966

E. coli 300 156 1'753 9'955 182 12'345

Salmonella spp. 7 2 163 4 51 227

Cryptosporidium 218 86 1'230 8'035 287 9'856

A. lumbricoides 5 1 40 17 8 71

Total number 1,483 996 7,099 48,877 1,017 59,472

DALYs per year (TotalDALYs,h) Norovirus Rotavirus Campylobacter E. coli Salmonella spp. Cryptosporidium A. lumbricoides 0.0 2.3 0.9 3.4 0.5 0.5 0.0 0.0 1.3 0.5 1.8 0.2 0.2 0.0 0.3 9.2 4.0 19.8 11.7 2.7 0.1 1.8 74.1 30.7 112.5 0.3 17.4 0.1 0.1 1.0 0.5 2.1 3.6 0.6 0.0 2 88 37 139 16 21 0

Combined DALYs 7.7 4.0 47.8 236.8 8.0 304.3

DALYs per person per year (DALYpppy,h) Norovirus 0.0000 Rotavirus 0.0004 Campylobacter 0.0002 E. coli 0.0006 Salmonella spp. 0.0001 Cryptosporidium 0.0001 A. lumbricoides 0.0000 0.0001 0.0058 0.0022 0.0076 0.0007 0.0008 0.0000 0.0004 0.0142 0.0061 0.0305 0.0181 0.0041 0.0002 0.0003 0.0125 0.0052 0.0190 0.0000 0.0029 0.0000 0.0000 0.0002 0.0001 0.0004 ^ 0.0006 0.0001 0.0000 0.0001 0.0048 0.0020 0.0076 0.0009 0.0012 0.0000

Combined DALYs 0.0014 0.0172 0.0741 0.0400 0.0014 0.0167

Graphical

abstract

- ■ T. I fyj ■HHH^HBB^SSl