Quantitative microbial risk assessment to estimate health risks attributable to water supply: Can the technique be applied in d&?tpacr=1;INT&?show [nucBreak];RODUCT&?show [/nucBreak];IONDescription of the water supplies in Kampala, UgandaMETHODOLOGYRisk assessment procedureSelection of reference pathogens and indicator organismsEstablishing disease burdens for inclusion in risk modelsCalculating susceptible populationsWater quality dataAnalytical methodsRESULTSRisk assessment findingsDISCUSSIONUsing risk assessment data for investment planningCONCLUSIONSAcknowledgementReferencesQuantitative microbial risk assessment to estimate healthrisks attributable to water supply: Can the technique beapplied in developing countries with limited data?Guy Howard, Steve Pedley and Sarah TibatemwaABSTRACTGuy Howard (corresponding author)Department for International Development,United House,10 Gulshan Avenue,Gulshan 1, Dhaka 1212, BangladeshE-mail: [email protected] PedleyRobens Centre for Public and EnvironmentalHealth,University of Surrey, Guildford,Surrey, GU2 7XH,UKSarah TibatemwaNational Water and Sewerage Corporation,Jinja Road, Kampala,UgandaIn the 3rd edition of its Guidelines for Drinking-Water Quality (2004) (GDWQ) the World HealthOrganization (WHO) promotes the use of risk assessment coupled with risk management for thecontrol of water safety in drinking water supplies. Quantitative microbial risk assessment (QMRA)provides a tool for estimating the disease-burden from pathogenic microorganisms in water usinginformation about the distribution and occurrence of the pathogen or an appropriate surrogate.This information may then be used to inform decisions about appropriate management of thewater supply system. Although QMRA has been used to estimate disease burden from watersupplies in developed countries, the method has not been evaluated in developing countrieswhere relevant data may be scarce. In this paper, we describe a simplified risk assessmentprocedure to calculate the disease burden from three reference pathogens – pathogenicEscherichia coli, Cryptosporidium parvum and rotavirus – in water supplies in Kampala, Uganda.The study shows how QMRA can be used in countries with limited data, and that the outcomecan provide valuable information for the management of water supplies.Key words|drinking water safety, QMRA, risk assessment, WHO GuidelinesINTRODUCTIONFor many years the water sector has relied upon compliancewith end-product standards to ensure water safety. Recently,however, the water sector has begun moving towards the useof risk assessment coupled with risk management as a moreeffective tool for the control of water safety (Deere et al. 2001;Davison et al. 2005). This approach to water safety has beenincorporated into the 3rd edition of the Guidelines forDrinking-Water Quality (GDWQ) (WHO 2004). In particu-lar, the World Health Organisation (WHO) advocates the useof water safety plans (comparable to the Hazard AssessmentCritical Control Point system applied in the food industry),guided by health-based targets, with independent surveil-lance to verify performance. The use of quantitative riskassessment is highlighted by the WHO as a valuable tool forsetting health-based targets and for validation of water safetyplans (WHO 2004).Assessing risks from water supplies is important whentrying to make judgements regarding the level of safetyrequired in the light of alternative and multiple routes ofexposure (Fewtrell & Bartram 2001). These assessments areof particular value in developing countries, where microbialpathogens may be transmitted by a number of routes (Esreyet al. 1991). In many cases, improving the microbial qualityof water may have less impact on disease than investmentsin sanitation, hygiene or level of water supply access (Esreyet al. 1985, 1991; Howard & Bartram 2003).Quantitative microbial risk assessment (QMRA) is atechnique that has been developed for calculating theburden of disease from a particular pathogen. The majortasks of a QMRA have been defined by Haas & Eisenberg(2001) as exposure assessment, dose-response analysis andrisk characterisation. In order to capture and compare thedoi: 10.2166/wh.2005.05849 Q IWA Publishing 2006 Journal of Water and Health|04.1|2006various outcomes from different pathogens, the use ofdisability adjusted life years (DALYs) has been rec-ommended in risk assessment (Havelaar & Melse 2003;WHO 2004). Murray & Lopez (1996) provide data fromwhich to calculate DALYs for health effects derived frominfection by waterborne pathogens.QMRA is typically confined to individual causativeagents and specific disease symptoms rather than undiffer-entiated health effects (Haas et al. 1999). Initial work in theUS developed risk models for Giardia and rotavirus, whichprovided the levels of treatment required to meet targetsrelated to acceptable levels of risk of disease (Regli et al.1991). This analysis was possible because the health datarequired for the analysis were readily available.Completing a QMRA for every pathogen that may betransmitted by water would be time-consuming and thenecessary information is currently not available for manypathogens. To overcome this difficulty, WHO (2004) rec-ommended using a suite of ‘reference pathogens’. A referencepathogen is an organism whose severity of impact andpersistence in water is such that its control would provideconfidence that health risks from pathogens of a similarnature have also been controlled (WHO 2004). The distri-bution and concentration of pathogens in water are highlyvariable, and the use of reference pathogens introducesanother layer of uncertainty into the analysis. Hence, QMRAmodelling often includes Monte Carlo simulation techniquesto help capture the uncertainty and variability withinfrequency distributions. However, the complexity of thesetechniques and the need for proprietary and costly software,largely limits their use to the developed world. To expand thescope of QMRA, a simplified approach has been developedusing point estimates (WHO 2004).The limited data on pathogens in developing countriesrequires that a QMRA be based on the occurrence ofindicator organisms. Despite the weaknesses of usingindicator organisms, Haas et al. (1999) believe that manyinitial QMRAs will have to be performed using data onindicator organisms due to inadequate data for occurrenceof pathogens. The use of indicator organisms does,however, require assumptions to be made about
View Full Document
Unlocking...