The identification of disease (or illness) is made by one of several methods (Table 4.1). The difference between disease and illness may be minor in some cases, but from a medical viewpoint and for HAZ ID, these do not necessarily mean the same thing. Disease is defined as the process or mechanism that ultimately results in an illness or a condition that impairs vital functions. A person could have a disease without initially having any symptoms. Symptoms are a state where the effects of the illness (e.g., headache, diarrhea, stomach cramps, vomiting) can be described by the person who is ill. Clinical assessment of the illness is generally defined by a measurable description of the illness (e.g., fever, bloody stool). Infection is colonization of the microorganism in the body and may result in disease and symptoms, as this is the initial step in the microbial disease process. However, this may also result in asymptomatic or subclinical infections. Symptoms and clinical descriptions (e.g., fever, rash, inflammation) can be very specific, such as with measles, which is associated with a specific agent, or they can be generic, such as with diarrhea, which is associated with many different types of microorganisms. An important quantitative piece of information as part of HAZ ID that is needed for QMRA is the incubation time; that is the time from exposure to the appearance of symptoms. The time to excretion of the pathogen by the infected host, duration of excretion, and the concentrations of the pathogen excreted (e.g., in feces) are all important quantitative data that are needed for QMRA. Physicians rarely gather this information specifically and utilize the general medical information on pathogen incubation times and excretion rates to help with the diagnosis. For example, the duration of diarrhea may help to diagnose the difference between Giardia infections and Norovirus infections.

Health Outcomes Associated with Microbial Infections

After exposure to a microorganism, once infection begins (defined by dose–response or attack rates; see the section “Epidemiological Methods for Undertaking HAZ ID” and Chapter 5), there are a number of possible outcomes, including asymptomatic illness, various levels of acute and chronic disease (known as morbidity: mild illness to more severe illness, to chronic problems, to that which requires hospitalization), and potentially death (known as mortality). In particular, there have been inadequate recognition and documentation of the chronic disease and long-term sequelae associated with microorganisms such as degenerative heart disease and insulin-dependent diabetes caused by Coxsackievirus B infection and peptic ulcers and stomach cancer caused by the bacterium Helicobacter pylori.

Figure 4.1 demonstrates the various health outcomes that should be part of the HAZ ID, which can occur after exposure. It has been difficult to quantify what goes in each of these boxes, based on the current health databases. The goal is the development of a quantitative probability of each possible outcome; these are likely to be microorganism specific, even isolate specific, and can depend on the host status. The goal of the HAZ ID, however, is to define these outcomes to the extent possible. Each outcome can be described as a ratio or percentage; the numerator and denominator need to be adequately defined as well as the populations associated with the data. Table 4.2 is an example of the level of symptomatic infections for Salmonella infections. The term used to address this outcome is morbidity. In a total of 12 studies, conducted primarily during outbreak investigations, the percentage of people who exhibited symptoms, divided by the total numbers infected (in this case from a variety of species of Salmonella), was measured by the detection of the bacterium in fecal specimens, indicating infection, and averaged 41%, with a range of 6–80%, or stated conversely, an average of 59% were asymptomatic (ranging from 20% to 94%) [2]. These data do not indicate the type of disease or the severity, and often other databases are needed to address those aspects of the health outcome.

While acute outcomes have been described for most microbial infections, the chronic impacts may be more significant but are rarer (in terms of cases) and generally not as quantifiable. Table 4.3 lists some of the bacteria, protozoa, and viruses for which both acute and chronic effects have been documented. Some of the chronic impacts are more severe and particularly affect the more sensitive populations. For example, Enteropathogenic E. coli (EPEC) causes mild illness in adults (1–2 days of mild diarrhea), while E. coli O157:H7 causes hemorrhagic uremia, with death more likely in children and the elderly [4, 5, 8]. Figure 4.2 quantifies some of the health outcomes documented during the outbreak associated with contaminated hamburgers in the Northwest in 1993 [8].

Salmonella, Shigella, and Yersinia have all been shown to initiate a reactive arthritis, which has been greatly underestimated, with attack rates ranging from 1.5% to as high as 29% in those infected with these bacteria [7]. Toxoplasma gondii is associated with very little disease in adults but causes congenital malformations. It has been estimated that there is a 45% chance (30–60% range) of maternal-to-fetal transmission of the infection if the mother is seropositive. The outcome, while normal for 55% of the newborns, is associated with 2% mortality, 11% retardation, and 6% blindness [6]. Guillain-Barré syndrome (GBS) is a major cause of neuromuscular paralysis in the United States, causing an estimated 2628 to 9575 cases each year, between 20% and 40% caused by infections with Campylobacter [9]. Figure 4.3 shows some of the health outcomes associated with Campylobacter-associated GBS that have been estimated.

In order to examine the severity of the hazard, hospital data can be used. Data were obtained from the Centers for Disease Control and Prevention (CDC) for examining the range of severe outcomes as hospitalization ratios (numerator is the number of cases in the hospital divided by the total number of cases) during waterborne disease outbreaks. It should be kept in mind that the denominator in these outbreaks may be highly uncertain. The ratios have been shown to be highly dependent on the etiologic agent (Table 4.4). The highest percentages of hospitalizations during waterborne outbreaks in the United States were associated with HAV and bacteria such as Shigella, Salmonella, and E. coli. Levels of 1.0% were found for Cryptosporidium; in the outbreak in Texas, the ratio was 0.85%. In the Carrollton, Georgia, outbreak, although no hospitalizations were documented in the literature, the number of visits to the hospital ER increased five- to sixfold, giving approximately an 0.8% ratio (ER visits/total cases) [27]. In Milwaukee, Wisconsin, 4000 hospitalizations were estimated at less than 1% [28], a bit higher than for Giardia and Norwalk-like enteric viruses. The undetermined etiologic agents responsible for AGI also resulted in hospitalizations of less than 1%; however, a large number of cases (about 874 cases) felt that their illness was serious enough to warrant treatment in the ER. Almost four times the number of hospitalized cases of AGI were treated in the ER. The only other illness with significant numbers of cases that was treated in the ER was giardiasis; twice the number of hospitalized cases (60 cases) were seen in the ER (120 cases). While the more severe illnesses appear to be caused by bacteria and HAV infections, Giardia and the etiologic agents of AGI contribute to a significant number of hospitalized cases, representing 5.3% and 22.5% of all hospitalized cases, respectively.

Surprisingly, in the waterborne disease data sets reported in Morbidity and Mortality Weekly Report [29], it can be very difficult to find hospitalization data summarized in association with the outbreak statistics. Thus, other published data sets need to be scrutinized.

A review of outbreaks associated with E. coli O157:H7 from 1982 to 2002 found a 17% hospitalization ratio based on 350 outbreaks (9% of which were waterborne) from a total of 8598 cases [30, 31], which is not so different than the data from the waterborne outbreaks presented in Table 4.4.

Studies on the enteric bacteria are valuable as the severity associated with these infections is often more extreme. Scallan et al. [32] reviewed the hospitalizations and mortality ratios associated with infections in children in the United States. The hospitalization ratio overall was 7.6% (Table 4.5).

Vibrio infections associated with a variety of species but mostly V. vulnificus and V. parahaemolyticus from recreational water exposures resulted in 124 cases, 37 hospitalizations (29.8%), and 4 deaths (3.2%) for 2007 and 2008 [25].

Hospitalizations and mortality (death) are often endpoints that are more likely to be identified, reported, and quantified. The most common matrix for mortality is the case fatality ratio. However, these ratios may be overestimated, as the denominator (total cases) is not well defined and is often underestimated. Often, the severity associated with acute outcomes as well as mortality is much more significant in sensitive populations.

Key pathogens have very specific outcomes. Naegleria fowleri, a free-living amoeba, causes primary amebic meningoencephalitis (PAM). The parasite is acquired mostly through recreational exposure via entrance through the nasal passages and then migration to the brain where it causes neurological damage. There was a 100% case fatality ratio in 6 and 2 known cases in 2007 and 2008 in the United States (AZ, CA, 3 in FL, OK, and 2 in TX), respectively [25]. These cases occur in mostly warm lakes in children or young adults (13–22 years of age for these reports). Overall, from 1985 to 2008, 41 cases of PAM occurred ranging from 1 to 6 cases per year.

In the United States, from 2000 to 2009, there were 22,331 cases of Legionnaires’ disease caused by Legionella infections, and the case fatality ratio was 8% [33]. Transmission occurred through exposure to cooling towers, tap water (showers), building cooling systems, spas, and hot tubs. In these cases, 90% occurred in individuals greater than 40 years of age. In Europe, from 1995 to 2005, more than 32,000 cases of Legionnaires’ disease (600 outbreaks) were reported to the European Working Group for Legionella Infections (EWGLI).

Sensitive Populations

Select populations such as pregnant women, the elderly, infants, and the immunocompromised currently represent almost 40% of the general population (Table 4.6), and their numbers are expected to increase in the years ahead [38]. The elderly account for almost 32% of the vulnerable group, with a great percentage residing in long-term care facilities. Risks of increased severity and mortality are greater for these populations when exposed to a variety of pathogens. In some cases, this may be due to just reduced immunity, but in other cases, there seems to be some specific mechanisms such as with HEV and pregnant women as discussed in the following text.

New microbial hazards are often discovered in sensitive populations that direct attention to further investigation. This includes C. difficile, a problem for those on antibiotics and particularly the elderly, which has now been associated with waterborne disease [39].

Women during Pregnancy, Neonates, and Young Babies

Women during pregnancy may be at increased risk from a number of infectious agents and may also act as a source of infection for neonates. During the past decade, at least 30 outbreaks of hepatitis E have been documented in 17 countries due to contaminated water [40]. Although outbreaks of hepatitis E have not been reported in the United States, cases do occur among tourists returning from developing countries. Waterborne outbreaks have at times involved thousands of people. Overall, case fatality ratios have ranged from 1% to 2% during outbreaks, which is significantly higher than that for HAV. However, for pregnant women, the ratio is generally between 10% and 20% but can be as high as 40% [40]. In contrast, hepatitis A does not appear to manifest itself differently in well-nourished pregnant women than in nonpregnant populations [41]. However, this may not be true for poorly nourished pregnant women. Numerous reports from developing countries recount disease of great severity, often leading to fulminant hepatitis, particularly in the third trimester [41].

Viral infection during pregnancy may also result in the transmission of infection from the mother to the child in utero, during birth, or shortly thereafter. This appears to be a common mode of transmission of Coxsackieviruses and echoviruses [42]. Neonates are uniquely susceptible to enterovirus infections. This group of viruses is capable of causing severe disease and death when infection occurs within the first 10–14 days of life. Acquisition of Coxsackievirus B infections early in life is the most significant risk factor leading to fatal disease. Most fatal cases caused by this virus are probably transmitted transplacentally at term [42]. Among 41 documented cases of fatal infection in infants, 24 of their mothers had symptomatic illness leading to fatal infection in infants, and 24 of their mothers had symptomatic illness consisting of fever, symptoms of upper respiratory tract involvement, pleurodynia, or meningitis.

Symptomatic infection occurred between 10 days antepartum and 5 days postpartum. The observed case fatality ratio for Coxsackievirus B in a New England county was almost 13%, with a morbidity ratio of 50.2 per 100,000 live births. Stillbirth late in pregnancy has been reported for the echoviruses and Coxsackievirus B [43]. Coxsackievirus B has been implicated as a potential important causative agent in spontaneous abortions, and anomalies (urogenital system, heart defects, digestive malformation) in children born to mothers infected with Coxsackievirus B suggested in several studies [43].

Echoviruses can also be transmitted from the mother to the unborn child or shortly after birth with a potentially serious outcome [44]. An average case fatality ratio of 3.4% was observed in 16 documented outbreaks of echovirus in newborn nurseries. In two outbreaks of Coxsackievirus B in nurseries, the infant mortality ratio from myocarditis ranged from 50% to 60%. Coxsackievirus and rotavirus have also been associated as potential cause of sudden infant death syndrome in young children. An outbreak of sudden infant death syndrome associated with rotavirus infection over a 3-week period was observed in the emergency facility of a hospital in which two of five children died [45].

Diabetes

One of the largest-growing segments of the sensitive populations globally is those individuals who have diabetes [46]. While the causes of type 1 diabetes (T1D) are still unknown, viruses (enterovirus including the Coxsackievirus B) have been suggested as an environmental trigger for the disease, but the data are not conclusive. Type 2 diabetes related to diet, lifestyle, and genetics is by far the larger problem. Some of the evidence shows that this group is more susceptible to disease and severity of outcomes. Data on nontyphoid chronic infections found that 15% of 129 studied for over a decade at the Massachusetts General Hospital, Boston, had diabetes as a risk factor [47].

The Elderly

Most epidemiological studies concerning a specific agent in the elderly are focused around nursing homes since the impact can be more easily observed in a confined group of people. Case fatality ratios for specific enteric pathogens are 10–100 times greater in this group than in the general population (Table 4.7). One documented outbreak of rotavirus in a nursing home was characterized by high attack rates (66%), with few, if any, asymptomatic cases [50]. While the number of days of illness was within the range observed for other age groups (1–5 days), the convalescence was prolonged for some people. The case fatality ratio was 1%. Gordon reported a case fatality ratio of 1.3% among a retirement community during a foodborne outbreak of Snow Mountain agent, a calicivirus that is now known as norovirus [49]. They pointed out that several of the residents sustained serious injuries from falling because of near-syncopal episodes due to dehydration from the gastroenteritis. The elderly are expected to be more prone than younger adults to such injuries because of greater illness severity. Outbreaks of noroviruses and enteric adenoviruses have been reported in nursing homes and geriatric wards in hospitals [51, 53]. In the early years investigating the emergence of the norovirus types, no mortality was observed during these outbreaks, and higher attack rates occurred among the residents, as well as a more severe or protracted illness, compared to the staff [53]. In a review of norovirus infections resulting in 158 associated deaths, the reports from 12 countries (from 1988 to 2011) found age a significant factor [54]. Where there were age data, 61% of the deaths were found in those greater than 65 years of age with 22% and 17% occurring in age groups of less than 2 years of age and between ages of 49 and 65, respectively. Data from 11 outbreaks occurring in the United States (eight outbreaks), Finland, England, and Wales reported that 71 deaths occurred in nursing homes and 33 deaths occurred in hospitals.

The most common causes of death for the elderly include pneumonia, influenza, and septicemia responsible for 5.5% (95,640 deaths) for those who are greater than 65 years of age (data obtained in 1997). A 25% increase in mortality between 1980 and 1992 was observed for infectious diseases in the elderly [55].

Cryptosporidium is also a significant cause of severe diarrhea in the elderly as well as mortality [56]. Records were examined from 1991 to 2004 in the United States from Medicare patients who were 65 years of age or older. Hospitalization rates were 0.15 to 0.39/100,000 cases per year (0.00015–0.00039%) for the general elderly population, and these doubled for those greater than 85 years of age. Case fatality ratios were 7.8% and only slightly below HIV patients (10.3%). The rates for hospitalizations for the elderly in a community are dramatically affected and are increased by age group (Fig. 4.4).

HAV usually causes a mild and often asymptomatic infection in children. However, in adults, the illness typically produces clinical illness that can lead to death [58]. Waterborne outbreaks of hepatitis A are often characterized by high attack rates with all or most of the infected persons exhibiting clinical illness [59]. The case fatality ratio of hepatitis A cases from England, Wales, and Ireland for patients less than 55 years of age was 0.02–0.03% (0.9% at 55–64 years of age and 1.5% for older patients). The median age of those dying from hepatitis A is over 60 in the United Kingdom [60].

The elderly also experience higher mortality from enteric bacterial gastroenteritis (Table 4.6). The overall case fatality ratio for foodborne outbreaks in nursing homes in the period 1975–1987 was 1.0%, compared to 0.1% for outbreaks at other locations [51]. For domestically acquired cases of typhoid, the case fatality ratio is higher among those 55 years or older [61]. In a developing country, for E. coli O157:H7, children from birth to 1 year of age and adults greater than 31 years of age were at highest risk of complications and death [52].

The Immunocompromised

New protocols have decreased the incidence of AIDS; thus, there are many more individuals living with HIV infections but are not as susceptible to microbial infectious disease. However, the impact of AIDS can still be seen, particularly in parts of the world where drugs are not readily available or being used to keep the disease under control. An 18-year assessment found that case fatality in patients with HIV dropped dramatically with treatment going from 78.6% (11/14) to 11.4% (5/44) (1992–1999 and 2000–2009 periods, respectively) [62].

AIDS remains an important global disease. It is likely that without treatment/therapy the disease statistics associated with enteric, respiratory, and skin infections will remain relevant. Studies in Brazil found that Cryptosporidium infections were associated with poor observance to the antiretroviral therapy, diarrhea, and CD4(+) T lymphocyte counts less than 200 cell/mm³ [63].

Prior to the use of any treatment, it was found that AIDS patients (50–90%) suffered from chronic diarrheal illnesses that were often fatal [64]. Adenoviruses and rotavirus were the most common enteric viruses isolated in the stools of AIDS-infected persons [64]. A comprehensive study of Australian men showed that 54% of diarrheal illnesses in AIDS patients with clinical disease suffered from adenovirus infections and that 45% of these cases resulted in death within 2 months [65]. The other enteric viruses do not appear to be a significant problem in AIDS-associated gastroenteritis. Enteric bacterial infections are more severe in AIDS patients. For example, patients with Salmonella, Shigella, and Campylobacter often develop bacteremia [66].

Although patients with AIDS may not have more severe illness when infected with Giardia, they do exhibit impaired immune response to the parasite [67]. Cryptosporidium was at one time a serious problem among AIDS patients. A severe and protracted diarrhea results, with fluid losses of several liters per day in some cases. Symptoms persisted for months, resulting in severe weight loss and mortality. Cryptosporidium in some studies caused 7–38% of diarrhea in immunocompromised patients [68]. Case fatality ratios for Giardia are low and do not seem to increase in vulnerable populations. Waterborne outbreaks of Cryptosporidium in the United Kingdom increased the incidence of disease in the AIDS population, with severe consequences [69]. In the 2 years following the waterborne outbreak in Milwaukee, there were 54 cryptosporidiosis-associated deaths, compared to only four deaths in the 2 years prior to the outbreak [70].

Cryptosporidium at one time accounted for 16% of the cases of diarrhea in AIDS patients (and as much as 50% in the developing world), with 87% associated with chronic illness related to CD4 counts (<180 × 10⁶ l⁻¹). Three waterborne outbreaks associated with drinking water have found those with AIDS to be at grave risk. Community-wide exposure did not increase the attack rates in the AIDS patients; however, the outcome of the disease was severe, with 52–68% mortality within 6 months to a year after the outbreaks. During the Milwaukee outbreak, in the cohort of 73 AIDS patients (33 with Cryptosporidium), morbidity was also much more severe, with 400 of 444 hospital days logged in by those with protozoan infection and extra medical costs of $795,699 [71].

One year after Milwaukee, a cluster of cases and deaths in AIDS patients in Las Vegas, Nevada, alerted health officials to another outbreak [72]. Those who drank any unboiled tap water were four times more likely than those drinking only bottled water to develop cryptosporidiosis. It was hypothesized that contamination of the drinking water had been over an extended time period with intermittent low levels of oocysts as opposed to massive contamination event as was the case in the Milwaukee and Oxford/Swindon outbreaks, both associated with rainfall events.

Cancer patients undergo intensive chemotherapy with cytotoxic and immunosuppressive drugs and often radiation treatment in attempts to destroy neoplastic growth. These measures also attack the immune system, leaving the patient with little defense against opportunistic pathogens. For example, in cancer immunosuppressed patients, the mortality ratio for adenovirus infection was 53% [65]. Bone marrow transplantation is an effective therapy in patients with severe aplastic anemia or acute leukemia. However, because of a very weakened immune system, they are very susceptible to infection. The case fatality ratio among bone marrow transplants with enteric viral (rotavirus, Coxsackievirus, adenovirus) infection was 59% in one study [73]. Five of eight patients with rotavirus infections died. The case fatality ratios for adenoviruses for bone marrow patients ranged from 53% to 69%, depending on subgenus (Table 4.8).

Coxsackievirus A1 infection, which seldom causes diarrhea in healthy persons, resulted in the deaths of six of seven bone marrow patients in one outbreak [73]. Hypogammaglobulinemic patients were at increased risk of chronic meningoencephalitis from enteroviruses [74]. Chronic meningoencephalitis was most frequently associated with echovirus infection but has occasionally been reported in association with Coxsackievirus B infection.

Databases for Statistical Assessment of Disease

In the United States, official statistics on diseases are compiled as part of the National Notifiable Diseases Surveillance System (NNDSS). For assessing vital statistics, certain diseases have been reported since 1920; up to 1960, the health statistics were obtained from publications of the National Vital Statistics System. Initially part of the Public Health Service, the National Center for Health Statistics is now a part of the CDC. The number of reported cases is summarized by type of disease, reported month, state, age, and race in some cases. There are some temporal, spatial, and demographic assessments of the health outcomes by disease [75, 76]. At the state level, the cases are generally reported by county. The data represent only clinically identified cases, and then case ratios (cases to total population) or incidence rates are most often reported annually (numbers of cases/total population per year, often referred to as cases per 100,000 population for U.S. data). It is clear that some diseases are easily identified and reported consistently by all states (e.g., human rabies), while other diseases (e.g., salmonellosis) are underreported. Diseases such as giardiasis are reported at the state level but not always at the national level, and some (e.g., infections associated with Helicobacter and enteric viruses) are not reported at all. Generic illnesses or symptoms are not reported, and antibody prevalence is not reported.

Many diseases associated with transmission through the environment are not reportable. The hazards and health outcomes are underappreciated and inadequately quantified, because these rates are reported without regard to who is exposed or who may experience the gravest consequence as a result of the risk. Worldwide, the assessment and reporting in most countries of most diseases are poor. Given the inadequacy of clinical tests and people’s failure to seek medical attention and even get the proper tests done, it is known that these reports woefully underestimate the burden of disease in most populations. Table 4.9 summarizes the diseases reported at the national level in the various reports for the United States (1943, 1970, 1996, and 2011). One can see the changes that have occurred over the last 60 years in the diseases reported; new emerging diseases of concern will be included in some years such as Cyclospora. The use of vaccines have essentially eliminated or reduced dramatically the level of some diseases in the United States.

Often, diseases that rise to national significance get reported in separate documents. In 1996, 2741 cases of E. coli O157:H7 illness were reported, including 102 cases of HUS. Cryptosporidiosis was made reportable in 1995 with 2972 cases reported from 27 states and 2426 cases reported in 1996 from 42 states but was not included in the 1996 table. In 2011, there were 9250 cases of cryptosporidiosis and 16,747 cases of giardiasis.

Hantavirus pulmonary syndrome associated with environmental transmission from rodents was reported in 26 states: 22 cases in 1996 and another 138 cases as of May 1997 (the case fatality was 47.5%; thus, although the incidence was low, the consequence or health outcome was significant). There were 290 cases reported in 2011.

ICD Codes

Within the National Center for Health Statistics are several other databases that are often used to examine health outcomes and relative risks. In 1990, the National Ambulatory Medical Care Survey provided data from office-based physicians through examination of patient records and gave an indication of the number of persons who seek a physician and are diagnosed. Like the FoodNet program developed in 1996, which enlists state health departments and local network systems, these systems have demonstrated that for most mild illnesses, patients do not seek a physician and do not undergo the correct clinical testing for diagnosing the diseases. Therefore, testing and not reporting appears to be the limiting factor for most diseases. Other significant databases that have been used to examine more severe disease outcomes and death include the National Hospital Discharge Survey (begun in 1988 to assess the number of patients treated in hospitals) and National Mortality Followback Survey (representing about 1% of the U.S. resident deaths). These systems use the International Classification of Diseases (ICD), which assigns codes for specific diseases.

Waterborne and Foodborne Outbreaks

The health endpoints measured in epidemiological and surveillance studies of infectious diseases can be divided into several groups: (1) endemic risks, which are the constant low levels of diseases or infections that are present and circulating in a population; (2) epidemic risks, which are disease cases in excess of the number of cases normally found or expected, constituted as an outbreak if limited to a specific population; and (3) outbreaks, where two or more cases are associated with a common exposure in time and place or source. In most cases, these studies rely on routine health surveillance methods, whereby people seek medical attention, submit laboratory samples, and are diagnosed. Often, this is done retrospectively, through the examination of records or through personnel interviews and recall.

There are several estimates of endemic waterborne disease risks that include:

• Approximately 19 million waterborne illnesses/year for community water systems in the United States (5.4 million illnesses from groundwater and 13 million illnesses from surface water systems) [82]
  
• 12 million cases/year [83]
  
• 16 million cases/year [84]

Outbreaks are generally associated with specific population and exposure in time and place. There are community outbreaks that may be associated with the contamination of water or food. There are outbreaks at events (weddings and dinners, generally associated with foods), outbreaks associated with recreational exposure (fecal accidents in swimming pools), and outbreaks associated with place (hospital and day care and cruise shipoutbreaks, associated with contamination of food, water, surfaces, instruments, hands, etc.).

From 1971 to 2008 in the United States, more than 576,853 persons were reported ill during 747 documented waterborne outbreaks caused by bacteria (15%), viruses (9%), protozoa (19%), and chemicals (11%) [10–25, 72, 85–93]. The etiological agents causing acute gastrointestinal illness in a large percentage of the outbreaks were not identified (45%), and in the last record, outbreaks were identified with multiple types of pathogens (<1%). The numbers of outbreaks and cases associated with microbial agents are shown in Table 4.11. During the investigation of drinking water outbreaks, the source of the water (groundwater, spring, river) is also identified as well as the treatment deficiency (e.g., no disinfection).

Recreational outbreaks associated with ambient waters (untreated) and pools and spas, whirl baths, hot tubs, etc. (treated) have also been reported in the waterborne outbreak databases. From 1981 to 2008, 188 outbreaks occurred in natural recreational waters with an associated 7712 cases. Figure 4.5 and Figure 4.6 show microbial hazards (HAZ ID)/etiological agents by outbreaks and cases, respectively [23–25, 72, 85, 88–90, 92–94].