Skip to main content
ARS Home » Midwest Area » Madison, Wisconsin » U.S. Dairy Forage Research Center » Environmentally Integrated Dairy Management Research » Research » Publications at this Location » Publication #337210

Research Project: Improving Nutrient Use Efficiency and Mitigating Nutrient and Pathogen Losses from Dairy Production Systems

Location: Environmentally Integrated Dairy Management Research

Title: Agreement between quantitative microbial risk assessment and epidemiology at low doses during waterborne outbreaks of protozoan disease

item Burch, Tucker

Submitted to: Meeting Abstract
Publication Type: Abstract Only
Publication Acceptance Date: 3/9/2017
Publication Date: 5/16/2017
Citation: Burch, T.R. 2017. Agreement between quantitative microbial risk assessment and epidemiology at low doses during waterborne outbreaks of protozoan disease. Meeting Abstract. University of North Carolina Water Microbiology Conference, May 15-17, 2017.

Interpretive Summary:

Technical Abstract: Quantitative microbial risk assessment (QMRA) is a valuable complement to epidemiology for understanding the health impacts of waterborne pathogens. The approach works by extrapolating available data in two ways. First, dose-response data are typically extrapolated from feeding studies, which use healthy adults and laboratory-adapted pathogens, to community settings that include general populations and wild-type pathogens. Second, the dose-response data are extrapolated from the range of doses used in feeding studies (> 1 organism) to the range of doses that are relevant for environmental exposures (<< 1 organism) using single-hit models. Dose-response data based on feeding studies are often the best available, and the assumptions underlying single-hit models are simple, biologically plausible, and conservative towards protecting public health. Nonetheless, disease rates determined by QMRA and epidemiology have never been compared comprehensively at environmentally relevant doses. My objective was to compare the two approaches by performing QMRA on data from reported waterborne outbreaks of gastrointestinal disease. I screened more than 2000 papers and identified 9 outbreak reports that supplied the necessary data: responsible pathogen, attack rates from cohort studies or laboratory surveillance, and pathogen concentrations measured in the source water. All 9 outbreaks involved drinking water; 7 were caused by Cryptosporidium and 2 by Giardia. Reported cyst/oocyst concentrations varied between 3×10-5 and 4 per liter, with 4 of 9 values below 1×10-2 per liter. Crude attack rates varied between 1×10-4 and 0.3 cases per person per outbreak. The QMRA accounted for daily drinking water intake, outbreak duration, and used published exponential dose-response models for both organisms. Reported attack rates were adjusted for background rates and reporting bias. Adjustments for reporting bias were made using published factors that varied based on study type; attack rates from retrospective cohort studies and laboratory surveillance were adjusted downwards and upwards, respectively, to account for over- and under-reporting. QMRA predictions correlated well with epidemiological measurements at an order-of-magnitude scale (R2 = 0.61 for log-transformed data). For 8 of 9 data points, QMRA predicted the adjusted attack rate within an average factor of 4.3 (range = 1.6 to 13), while the remaining data point was off by a factor of 280. There was no systematic bias in QMRA predictions compared to epidemiological measurements; QMRA over-predicted attack rates in 5 of 9 cases and under-predicted them in 4. These results demonstrate that QMRA and epidemiology can produce equivalent estimates of disease rates for waterborne Cryptosporidium and Giardia, which should build confidence in drawing relationships between their results. Furthermore, agreement between the two approaches at low doses supports the validity of the assumptions underlying QMRA.