Outbreak Surveillance Data Research Articles

A Prediction Tool to Identify the Causative Agent of Enteric Disease Outbreaks Using Outbreak Surveillance Data.

Information on the causative agent in an enteric disease outbreak can be used to generate hypotheses about the route of transmission and possible vehicles, to guide environmental assessments, and to target outbreak control measures. However, only about 40% of outbreaks reported in the United States include a confirmed etiology. The goal of this project was to identify clinical and demographic characteristics that can be used to predict the causative agent in an enteric disease outbreak and to use these data to develop an online tool for investigators to use during an outbreak when hypothesizing about the causative agent. Using data on enteric disease outbreaks from all transmission routes (animal contact, environmental contamination, foodborne, person-to-person, waterborne, unknown) reported to the U.S. Centers for Disease Control and Prevention, we developed random forest models to predict the etiology of an outbreak based on aggregated clinical and demographic characteristics at both the etiology category (i.e., bacteria, parasites, toxins, viruses) and individual etiology (Clostridium perfringens, Campylobacter, Cryptosporidium, norovirus, Salmonella, Shiga toxin-producing Escherichia coli, and Shigella) levels. The etiology category model had a kappa of 0.85 and an accuracy of 0.92, whereas the etiology-specific model had a kappa of 0.75 and an accuracy of 0.86. The highest sensitivities in the etiology category model were for bacteria and viruses; all categories had high specificities (>0.90). For the etiology-specific model, norovirus and Salmonella had the highest sensitivity and all etiologies had high specificities. When laboratory confirmation is unavailable, information on the clinical signs and symptoms reported by people associated with the outbreak, with other characteristics including case demographics and illness severity, can be used to predict the etiology or etiology category. An online publicly available tool was developed to assist investigators in their enteric disease outbreak investigations.

Foodborne Outbreak Rates Associated with Restaurant Inspection Grading and Posting at the Point of Service: Evaluation Using National Foodborne Outbreak Surveillance Data

A previously conducted national survey of restaurant inspection programs associated the practice of disclosing inspection results to consumers at the restaurant point of service (POS) with fewer foodborne outbreaks. We used data from the national Foodborne Disease Outbreak Surveillance System (FDOSS) to assess the reproducibility of the survey results. Programs that participated in the survey accounted for approximately 23% of the single-state foodborne illness outbreaks in restaurant settings reported to FDOSS during 2016 to 2018. Agencies that disclosed inspection results at the POS reported fewer outbreaks (mean = 0.29 outbreaks per 1,000 establishments) than those that disclosed results online (0.7) or not at all (1.0). Having any grading method for inspections was associated with fewer reported outbreaks than having no grading method. Agencies that used letter grades had the lowest numbers of outbreaks per 1,000 establishments. There was a positive association (correlation coefficient, R2 = 0.29) between the mean number of foodborne illness complaints per 1,000 establishments, per the survey, and the mean number of restaurant outbreaks reported to FDOSS (R2 = 0.29). This association was stronger for bacterial toxin-mediated outbreaks (R2 = 0.35) than for norovirus (R2 = 0.10) or Salmonella (R2 = 0.01) outbreaks. Our cross-sectional study findings are consistent with previous observations that linked the practice of posting graded inspection results at the POS with reduced occurrence of foodborne illnesses and outbreaks associated with restaurants. Support for foodborne illness surveillance programs and food regulatory activities at local health agencies is foundational for food safety systems coordinated at state and federal levels.

Cost-effectiveness of pediatric norovirus vaccination in daycare settings

ObjectiveNoroviruses are the leading cause of acute gastroenteritis in the United States and outbreaks frequently occur in daycare settings. Results of norovirus vaccine trials have been promising, however there are open questions as to whether vaccination of daycare children would be cost-effective. We investigated the incremental cost-effectiveness of a hypothetical norovirus vaccination for children in daycare settings compared to no vaccination. MethodsWe conducted a model-based cost-effectiveness analysis using a disease transmission model of children attending daycare. Vaccination with a 90% coverage rate in addition to the observed standard of care (exclusion of symptomatic children from daycare) was compared to the observed standard of care. The main outcomes measures were infections and deaths averted, quality-adjusted life years (QALYs), costs, and incremental cost-effectiveness ratio (ICER). Cost-effectiveness was analyzed from a societal perspective, including medical costs to children as well as productivity losses of parents, over a two-year time horizon. Data sources included outbreak surveillance data and published literature. ResultsA 50% efficacious norovirus vaccine averts 571.83 norovirus cases and 0.003 norovirus-related deaths per 10,000 children compared to the observed standard of care. A $200 norovirus vaccine that is 50% efficacious has a net cost increase of $178.10 per child and 0.025 more QALYs, resulting in an ICER of $7,028/QALY. Based on the probabilistic sensitivity analysis, we estimated that a $200 vaccination with 50% efficacy was 94.0% likely to be cost-effective at a willingness-to-pay of $100,000/QALY threshold and 95.3% likely at a $150,000/QALY threshold. ConclusionDue to the large disease burden associated with norovirus, it is likely that vaccinating children in daycares could be cost-effective, even with modest vaccine efficacy and a high per-child cost of vaccination. Norovirus vaccination of children in daycare has a cost-effectiveness ratio similar to other commonly recommended childhood vaccines.

Attributing Human Foodborne Diseases to Food Sources and Water in Japan Using Analysis of Outbreak Surveillance Data

In Japan, strategies for ensuring food safety have been developed without reliable scientific evidence on the relationship between foodborne diseases and food sources. This study aimed to provide information on the proportions of foodborne diseases caused by seven major causative pathogens (Campylobacter spp., Salmonella, enterohemorrhagic Escherichia coli [EHEC], Vibrio parahaemolyticus, Clostridium perfringens, Staphylococcus aureus, and norovirus) attributed to foods and to explore factors affecting changes in these source attribution proportions over time using analysis of outbreak surveillance data. For the calculation of the number of outbreaks attributed to each source, simple-food outbreaks were assigned to the single-food category in question, and complex-food outbreaks were classified under each category proportional to the estimated probability. During 2007 to 2018, 8,730 outbreaks of foodborne diseases caused by seven pathogens were reported, of which 6,690 (76.6%) were of unknown source. We estimated the following source attribution proportions of foodborne diseases: chicken products (80.3%, 95% uncertainty interval [UI] 80.1 to 80.4) for Campylobacter spp.; beef products (50.1%, UI 47.0 to 51.5) and vegetables (42.3%, UI 40.9 to 45.5) for EHEC; eggs (34.6%, UI 27.8 to 41.4) and vegetables (34.4%, UI 27.8 to 40.8) for Salmonella; finfish (50.3%, UI 33.3 to 66.7) and shellfish (49.7%, UI 33.3 to 66.7) for V. parahaemolyticus; grains and beans (57.8%, UI 49.7 to 64.9) for S. aureus; vegetables (63.6%, UI 48.5 to 74.6), chicken products (12.7%, UI 4.6 to 21.5), and beef products (11.1%, UI 8.5 to 13.1) for C. perfringens; and shellfish (75.5%, UI 74.7 to 76.2) for norovirus. In this study, we provide the best available evidence-based information to evaluate the link between foodborne diseases and foods. Our results on source attribution for Campylobacter spp. and EHEC suggest that the strict health regulations for raw beef were reflected in the proportions of these diseases attributed to this food.

Factors associated with influenza-like illness in care homes in Cheshire and Merseyside during the 2017–2018 influenza season

ObjectivesThe aim of the study was to identify care home characteristics associated with reported care home influenza outbreaks and factors associated with increased transmission of influenza-like illness (ILI) in care homes in Cheshire and Merseyside during the 2017–2018 influenza season. Study designThis is a matched case-control study comparing characteristics between care homes with and without a declared influenza outbreak and a retrospective risk factor analysis of care home residents with ILI. MethodsRoutinely collected outbreak surveillance data on symptomatic residents and staff, antiviral prophylaxis and influenza vaccination history, which were reported to Public Health England, were extracted from health protection electronic records. Further care home characteristics were extracted from the Care Quality Commission (CQC) website. Care homes with declared influenza outbreaks were matched with care homes without outbreaks. Chi-squared tests and logistic regression were used to examine associations between care home factors and ILI. ResultsThere were no significant differences in characteristics between 77 care homes with declared influenza outbreaks and 77 matched care homes without outbreaks. Of 2,744 residents from the homes with a declared outbreak, 644 (24%) developed an ILI. The care home risk factors were having a low CQC score and activation of antiviral prophylaxis and the protective factors were having higher numbers of residents, specializing in dementia care and having the highest CQC score. Significantly more cases occurred in residential homes than in nursing homes, in homes with lower CQC scores and in homes where eligible residents were given antiviral prophylaxis. ConclusionsIn homes with declared outbreaks, certain characteristics including activation of antiviral prophylaxis were associated with an increased risk of ILI. Further research is needed, particularly focussing on temporality between provision of prophylactic antivirals and the onset of ILI.

Investigation of early epidemiological signals of COVID-19 in India using outbreak surveillance data

Background: The epidemic of novel coronavirus disease (COVID-19) started in Wuhan, China, in December 2019, has been progressed to a pandemic level with high morbidity and mortality affecting many countries globally. Due to varied clinical characteristics of the disease, diagnosis can be missed easily among the affected population. The epidemiological data and early signals of COVID-19 infection are crucial for the disease investigation. Aim: To assess the presence of early signals of COVID-19 in India before official reporting of cases in the country and to compare epidemiological characters using different surveillance sources in the study. Methods: This study uses Epiwatch outbreaks surveillance data from November 2019 to March 2020 was used to determine the trend of pneumonia of unknown causes in India in order to assess if the COVID-19 could have been detected before the first case was reported.The search was carried out using the search terms defined in the study to identify the early signals of COVID-19.The COVID-19 line list extracted from the crowdsourced database (https://www.covid19india.org/) was additionally used to determine the demographic characteristics of cases. A descriptive analysis was conducted. Results: A total of 379 reports was extracted from Epiwatch. In 2019, there were only six reports of pneumonia of unknown causes in India, and the majority of the reports were focused on swine flu or undetected influenza cases. In 2020, there were 312 news reports of unknown pneumonia and COVID-19. The early signals of COVID-19 in India were observed from 24th January 2020 as the trend of pneumonia news had moderately increased. The trend of pneumonia reporting was low before the first confirmed case was reported in India. The study findings also suggested that there was under-reporting of COVID-19 cases in India by the World Health Organization (WHO) and the Ministry of Health, India when compared to data reported by Epiwatch and the line list. Conclusion: The study findings suggested that there were early signals of COVID-19 in India around 24th January when there was a slight increase in the trend of pneumonia reports in the country. The number of COVID-19 cases in India is under-reported by the Ministry of Health and the WHO. The use of online media surveillance reports such as Epiwatch and the line list could provide real-time assessment and timely investigation of emerging diseases globally.

Measles outbreak in Romania: understanding factors related to suboptimal vaccination uptake.

BackgroundA large measles outbreak started in Romania in 2016. Current study aimed to (i) clarify who was affected by the outbreak, (ii) identify their barriers and drivers to vaccination and (iii) explore variation by population group.MethodsThis was a two-component study. Outbreak surveillance data for 6743 measles cases were reviewed to identify key characteristics. A survey was administered via telephone to 704 caregivers of measles cases (520 respondents) to explore capability, opportunity and motivation barriers to vaccination. Data were summarized descriptively for respondent characteristics and statements. Differences by population group (education, household income, ethnicity, setting and mobility) were explored using χ2 tests, Fisher’s exact tests or regression models.ResultsMost cases were unvaccinated and lived in low coverage areas. Ethnic minorities were disproportionally affected. Most caregivers felt welcome at health facilities. Some were less satisfied with the waiting time and had found the vaccine out of stock. Not everybody knew that vaccines were free of charge. Less than half knew the child’s next vaccination date, some had not been informed and did not know where to seek this information. Some said their peers did not vaccinate. Beliefs were generally supportive of vaccination; but many were concerned about vaccine safety and found they had not received good information about this. Conclusions varied greatly between minorities and less educated groups, compared with people with higher education levels.ConclusionsIdentifying characteristics of the population affected and underlying factors can inform a strategy to avoid future outbreaks and further research to obtain deeper insights.

Transmission Dynamics and Control Strategies During the 2017 Diphtheria Outbreak in Jakarta, Indonesia: A Modelling Study

Background: At the end of 2017, Indonesia, including Jakarta, suffered a large diphtheria outbreak. The disease has been well controlled by vaccination in recent decades, thus there are many unknowns about the contemporary drivers of diphtheria transmission and the best strategies for its control. Methods: Using 2017 diphtheria outbreak surveillance data and mathematical modelling techniques, we estimated i) age-specific susceptibility to diphtheria infection in Jakarta before the outbreak and ii) transmission parameters of diphtheria, including the basic (R0) and effective (Rt) reproduction numbers. We then used simulations to assess the impact of interventions undertaken during the outbreak, and the comparative impact of alternative interventions. Findings: 250 diphtheria cases were reported in Jakarta from November 2017 to February 2018. We estimated population susceptibility immediately prior to the outbreak to be 17-25%, with the highest susceptibility in children, particularly 0-4 year olds. We estimated to be 3·66 and that asymptomatic carriers contributed around 67% to the overall during the outbreak. We estimated that the contribution of asymptomatic carriers to transmission is higher in populations with higher immunity, compared to those with lower immunity. We estimated that tracing and treatment of contacts was vital for bringing the outbreak under control. Interpretation: Prompt tracing and treatment of cases and contacts should be the priority for control measures especially when vaccination coverage is high. Higher primary vaccination coverage than the currently estimated protective threshold of 80-85% and booster vaccinations for preschool and school-age children are vital in Jakarta to prevent future outbreaks of diphtheria. Funding Statement: Wellcome Trust, UK. Declaration of Interests: The authors declare no conflicts of interest. Ethics Approval Statement: The data used in the study are secondary data routinely collected by the Jakarta Provincial Health Office thus did not require ethics review board review.

Prospective cohort study to investigate the burden and transmission of acute gastroenteritis in care homes: epidemiological results

ObjectivesTo estimate the incidence of gastroenteritis in individuals in care homes.DesignProspective cohort study.SettingFive participating care homes in North West England, UK.ParticipantsResidents and staff present at the five study care homes...

Use of prophylactic antivirals and care home characteristics associated with influenza in care homes with confirmed outbreaks

ObjectivesTo assess the association of antiviral prophylaxis and care home characteristics with the spread and severity of influenza-like illness in care homes with influenza outbreaks in North West England in the 2017/2018 influenza season. Study designThis is a retrospective observational study. MethodsRoutinely collected outbreak surveillance data reported to Public Health England were extracted from health protection electronic records. Data included use of antiviral prophylaxis, influenza-like illness or confirmed influenza, hospital admissions and deaths. Care home characteristics were obtained from the Care Quality Commission website. Single variable analysis and multivariable logistic regression were used to examine associations between care home characteristics, antiviral prophylaxis and influenza-related outcomes. ResultsIn the 109 homes, there were 3498 residents; of whom, 855 (24%) developed an influenza-like illness. Antiviral prophylaxis was given to residents of 67 of the 109 care homes with outbreaks (61%). A significantly higher attack rate was observed among residents of homes given antiviral prophylaxis (27%) than among residents of homes not given antivirals (20%) (P < 0.001). Significantly more deaths occurred in homes for people with learning disabilities and homes that received antiviral prophylaxis (P < 0.001). ConclusionsIn homes given antiviral prophylaxis, there were a higher number of residents with influenza-like illness and deaths. To improve our understanding of the impact of antiviral prophylaxis use in real life, enhanced and timely data collection is needed for identification of temporal associations between exposure and administration of antiviral prophylaxis. Consideration needs also to be given to ensure people with learning disabilities are protected through the seasonal influenza vaccine and timely antiviral prophylaxis when appropriate.

Global and regional source attribution of Shiga toxin-producing Escherichia coli infections using analysis of outbreak surveillance data.

Shiga toxin-producing Escherichia coli (STEC) infections pose a substantial health and economic burden worldwide. To target interventions to prevent foodborne infections, it is important to determine the types of foods leading to illness. Our objective was to determine the food sources of STEC globally and for the six World Health Organization regions. We used data from STEC outbreaks that have occurred globally to estimate source attribution fractions. We categorised foods according to their ingredients and applied a probabilistic model that used information on implicated foods for source attribution. Data were received from 27 countries covering the period between 1998 and 2017 and three regions: the Americas (AMR), Europe (EUR) and Western-Pacific (WPR). Results showed that the top foods varied across regions. The most important sources in AMR were beef (40%; 95% Uncertainty Interval 39-41%) and produce (35%; 95% UI 34-36%). In EUR, the ranking was similar though with less marked differences between sources (beef 31%; 95% UI 28-34% and produce 30%; 95% UI 27-33%). In contrast, the most common source of STEC in WPR was produce (43%; 95% UI 36-46%), followed by dairy (27%; 95% UI 27-27%). Possible explanations for regional variability include differences in food consumption and preparation, frequency of STEC contamination, the potential of regionally predominant STEC strains to cause severe illness and differences in outbreak investigation and reporting. Despite data gaps, these results provide important information to inform the development of strategies for lowering the global burden of STEC infections.

Outbreaks of influenza-like illness in care homes inthe East of England: impact of variations in neuraminidase inhibitor provision.

To examine the differences in outcomes related to variable provision of antivirals in care home respiratory outbreaks. This is a retrospective observational study. Routinely collected outbreak surveillance data reported from care home staff was recorded using a standard template and extracted from the Public Health England health protection electronic records. Data included numbers of people affected, provision of oseltamivir, hospitalization, and deaths during the outbreak in the care home. Oseltamivir provision was categorized by proportion of eligible residents prescribed it on advice. Additional data on microbiological diagnoses were obtained directly from the hospital laboratories. Logistic regression was used to examine associations between oseltamivir provision and hospitalizations and deaths in care homes. One hundred and sixty-eight outbreaks were reported from 28th July 2016 until 27th March 2017, affecting 1459 residents and 347 staff. There were 76 hospital admissions and 37 deaths overall. Although deaths and hospital admissions also occurred in outbreaks caused by other respiratory viruses, outbreaks caused by influenza had the highest median number of people affected and a greater proportion of hospital admissions and deaths. Of the 56 outbreaks caused by influenza, there was a significant increase in the odds of the care home reporting deaths when oseltamivir was not used (odds ratio=8.15, 95% confidence interval=1.38-48.20, P=0.02). There were also non-significant reductions in duration of outbreak and hospital admissions in care home outbreaks with oseltamivir treatment. Partial or no provision of oseltamivir was associated with poorer outcomes in laboratory-confirmed influenza outbreaks.

Characterizing and Discovering Spatiotemporal Social Contact Patterns for Healthcare.

During an epidemic, the spatial, temporal and demographic patterns of disease transmission are determined by multiple factors. In addition to the physiological properties of the pathogens and hosts, the social contact of the host population, which characterizes the reciprocal exposures of individuals to infection according to their demographic structure and various social activities, are also pivotal to understanding and predicting the prevalence of infectious diseases. How social contact is measured will affect the extent to which we can forecast the dynamics of infections in the real world. Most current work focuses on modeling the spatial patterns of static social contact. In this work, we use a novel perspective to address the problem of how to characterize and measure dynamic social contact during an epidemic. We propose an epidemic-model-based tensor deconvolution framework in which the spatiotemporal patterns of social contact are represented by the factors of the tensors. These factors can be discovered using a tensor deconvolution procedure with the integration of epidemic models based on rich types of data, mainly heterogeneous outbreak surveillance data, socio-demographic census data and physiological data from medical reports. Using reproduction models that include SIR/SIS/SEIR/SEIS models as case studies, the efficacy and applications of the proposed framework are theoretically analyzed, empirically validated and demonstrated through a set of rigorous experiments using both synthetic and real-world data.

Norovirus genotype profiles associated with foodborne transmission, 1999-2012.

Worldwide, noroviruses are a leading cause of gastroenteritis. They can be transmitted from person to person directly or indirectly through contaminated food, water, or environments. To estimate the proportion of foodborne infections caused by noroviruses on a global scale, we used norovirus transmission and genotyping information from multiple international outbreak surveillance systems (Noronet, CaliciNet, EpiSurv) and from a systematic review of peer-reviewed literature. The proportion of outbreaks caused by food was determined by genotype and/or genogroup. Analysis resulted in the following final global profiles: foodborne transmission is attributed to 10% (range 9%%-11%) of all genotype GII.4 outbreaks, 27% (25%-30%) of outbreaks caused by all other single genotypes, and 37% (24%%-52%) of outbreaks caused by mixtures of GII.4 and other noroviruses. When these profiles are applied to global outbreak surveillance data, results indicate that ≈14% of all norovirus outbreaks are attributed to food.

Stability of the Age Distribution of Measles Cases Over Time During Outbreaks in Bangladesh, 2004–2006

Despite recommendations from WHO to conduct measles outbreak response vaccination campaigns based on the age distribution of cases at the beginning of an outbreak, few data exist to specifically examine whether the age distribution of cases remains constant over time in a measles outbreak. This analysis explores this question with use of measles outbreak surveillance data from Bangladesh from the period 2004-2006. Pearson χ(2) tests were conducted of age distributions over 2 periods during 41 large laboratory-confirmed measles outbreaks. Statistically significant changes in age distribution over time were observed in 24% of the outbreaks. No single pattern was detected in the shifts in age distribution; however, an increase in the proportion of cases occurring among infants <9 months of age was evident in 6 outbreaks. These findings suggest a need to consider the possibility of a shift in the age distribution over time when planning an outbreak response vaccination campaign.

Source attribution of nontyphoid salmonellosis in new zealand using outbreak surveillance data.

In this study, 204 New Zealand outbreaks of non-typhoid salmonellosis reported from 2000 to 2009 were analyzed for information on the sources of human infection. Data were extracted from the outbreak module of EpiSurv, New Zealand's notifiable diseases database, and augmented with information from individual case reports and separate investigation reports. The outbreaks involved 1,426 cases, representing an estimated 9% of the total salmonellosis cases reported for the study period. Salmonella Typhimurium was the causative serotype in 78% of 172 outbreaks for which a serotype was available, involving 71% of outbreak cases. The most commonly reported outbreak setting was the home (47% of outbreaks), followed by commercial food operations (31%). Foodborne transmission was reported for 63% of the 123 outbreaks for which only one mode of transmission was reported, followed by person-to-person transmission (32%), waterborne transmission (3%), and zoonotic transmission (2%). However, evidence for the mode of transmission was weak or absent for 107 (63%) of the 169 outbreaks for which a mode of transmission was reported. For only 22 outbreaks was laboratory evidence successfully used to identify a potential source of infection. Of these 22 outbreaks, 7 were foodborne, 11 involved an infected food handler, 2 were attributed to contact with animals, 1 was attributed to consumption of drinking water, and 1 was attributed to multiple sources. The laboratory-confirmed contaminated foods were diverse and included imported and domestically produced foods. The results of this analysis support the hypothesis that non-typhoid salmonellosis is primarily a foodborne disease in New Zealand, but there is insufficient evidence to confirm important food vehicles.

Dairy-borne Disease Outbreak and Milk Demand: A Study Using Outbreak Surveillance Data

We utilize the outbreak surveillance data from the Centers for Disease Control and Prevention (CDC) to examine whether consumer demand is impacted by the outbreak of food-borne disease. An additional person sickened due to the ingestion of tainted cheese products at home is found to decrease per capita milk demand in New York State by 0.13 percent (or 0.07 pound), while milk-and ice cream-borne disease outbreaks, occurring at home or in public places, are found to have no impact on fluid milk demand. Our results imply the existence of word-of-mouth effects, which cannot be tested by the popular information/media index approach used to measure food-borne disease outbreaks. We also find that a 7 percent increase in generic advertising expenditures or a 10 percent increase in expenditures on non-advertising marketing activities can offset the negative influence of one ill person in cheese-borne disease outbreaks at home, while holding other demand factors constant.

Epidemiology of human cryptosporidiosis in Ireland, 2004-2006: analysis of national notification data

Since cryptosporidiosis became a notifiable human disease in 2004 in Ireland, evidence has been growing as to the national burden of illness caused by this pathogen. Nationally, crude incidence rates of between 8.7 and 13.4 per 100,000 were reported annually in the period 2004-2006.

Risk factors for waterborne enteric infections

Risk factors for waterborne enteric infections are deduced primarily from outbreak surveillance data; however, in the United States, only a fraction of the estimated water-related outbreaks are reported through passive surveillance. In the past several years, advances in molecular detection techniques have furthered our knowledge about foodborne and waterborne causes of gastroenteritis, allowing the association of certain pathogens with biologic and exposure-related susceptibilities in their hosts. This article summarizes some of the recent data characterizing susceptibility to three common waterborne pathogens:Cryptosporidium, a protozoan; Norwalk-like virus; and the bacterium Escherichia coli O157:H7. The infectious dose of Cryptosporidium varies by several orders of magnitude by strain, and repeated low-level exposure in drinking water may be protective. Some people may be innately immune to Norwalk-like virus, despite multiple exposures. A major risk factor for E. coli O157:H7 infection is exposure to shallow groundwater sources contaminated with animal waste.