Show Summary Details
Page of

Outbreak Investigation 

Outbreak Investigation
Chapter:
Outbreak Investigation
Author(s):

Noel S. Weiss

and Thomas D. Koepsell

DOI:
10.1093/med/9780195314465.003.0020
Page of

PRINTED FROM OXFORD MEDICINE ONLINE (www.oxfordmedicine.com). © Oxford University Press, 2016. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a title in Oxford Medicine Online for personal use (for details see Privacy Policy and Legal Notice).

Subscriber: null; date: 23 October 2019

An outbreak or epidemic of disease occurs when the number of new disease cases observed exceeds the number expected in a defined setting over a relatively short period of time. Technically, the terms outbreak and epidemic are defined similarly and are sometimes used interchangeably. However, many epidemiologists use epidemic for larger, more widespread, or longer-term elevations in disease incidence and outbreak for more geographically limited increases in disease incidence. Although most commonly associated with infectious diseases, outbreak investigations also occur for non-infectious causes, including intoxications, injuries, and other adverse health events.

Outbreak investigations can serve several purposes:

  • Limit the scope and severity of an immediate threat to public health. There may be effective disease control interventions, such as treatment for infected persons, vaccine or antibiotic prophylaxis for those susceptible to infection, infection control measures, or withdrawal of a contaminated product from distribution. Meningococcal meningitis, hepatitis A and hepatitis B, pertussis, measles, and varicella are among the communicable diseases with outbreak potential for which effective pharmacologic interventions are available.

    In 2012, over 600 cases of fungal meningitis and other central nervous system infections were traced back to epidural or paraspinal injections of a contaminated steroid medication prepared by a single compounding pharmacy. Once the outbreak source was identified, public health officials worked to ensure that the contaminated lots were discontinued from use, and they contacted nearly 14,000 patients and their physicians nationwide to facilitate prompt recognition and treatment of illness (Kainer et al., 2012; Bell and Khabbaz, 2013).

    Examples of controllable non-infectious disease outbreaks include the sudden appearance of a cluster of a rare condition, eosinophilia-myalgia syndrome, among women in New Mexico in 1990 that was stopped when implicated lots of an L-tryptophan supplement contaminated with an industrial lubricant were recalled (Belongia et al., 1990); and a 2007 cluster of polyradiculoneuropathy among workers at a pork processing plant in Minnesota. The ensuing investigation determined that the use of compressed air to remove brains from pig carcasses aerosolized central nervous system tissue and caused the immune-mediated illness among those working in the vicinity. Once this technique was discontinued, no additional cases were reported (Holzbauer et al., 2010).

  • Prevent future outbreaks. Once the reason(s) for an outbreak is understood, implementing changes in products or processes can help prevent a recurrence. For example, when a sudden outbreak of the rare disease toxic shock syndrome occurred in 1980, a series of investigations were conducted. The disease was found to be associated with the use of a new “super absorbent” brand of tampons, which fostered bacterial growth. This type of tampon was removed from the market, the outbreak ended, and further outbreaks were prevented (Shands et al., 1980).

    The identification of a large outbreak of E. coli O157:H7 due to contaminated undercooked hamburger in 1992 resulted in widespread changes in standard procedures for cooking hamburgers in the fast food industry and a subsequent decline in the frequency of outbreaks from this source (Bell et al., 1994).

    An outbreak of salmonellosis associated with contaminated frozen pot pies highlighted consumer confusion with cooking instructions for microwave ovens of various wattages and led to the recommendation that microwave manufacturers consider labeling units with their output wattage (Centers for Disease Control and Prevention, 2008).

    An investigation of mesothelioma cases in Florence, Italy, led to the discovery (and subsequent cessation) of the reuse of polypropylene bags that had contained asbestos cement as baling material for fabrics (Weiss, 1991).

  • Identify new vehicles of infection. Several outbreaks of disease due to enteric pathogens were associated with consumption of raw sprouts throughout the 1990s (Breuer et al., 2001; Mohle-Boetani et al., 2001; Centers for Disease Control and Prevention, 2002a). Subsequent research determined that seeds were often contaminated with enteric bacteria that thrived under sprouting conditions. These investigations resulted in a recommendation issued by the U.S. Department of Agriculture (USDA) and the CDC that raw sprouts not be consumed by young children, the elderly, and immunocompromised persons, who may be at increased risk for serious complications of enteric infections (U.S. Department of Health and Human Services, 1999).

    In 2012, a cluster of chronic skin infections unresponsive to standard treatment was reported among people who had recently received tattoos. The multi-state investigation identified the same atypical mycobacterial strain in clinical biopsies and in a bottle of the implicated ink, which was intended for printing purposes and not approved for use in tattooing. The outbreak highlighted the importance of establishing and enforcing standards for the regulation of tattoo inks (LeBlanc et al., 2012; Falsey et al., 2013).

  • Monitor the success of intervention programs. The rapid emergence of Salmonella Enteritidis outbreaks associated with intact-shell eggs in the 1980s established that this serotype of Salmonella had become adapted to the hen’s ovary and that even intact eggs could contain S. Enteritidis (St. Louis et al., 1988). The USDA, CDC, and the Food and Drug Administration (FDA) worked with the egg industry to create programs to control exposure of laying hens to S. Enteritidis on the farm. The decreasing frequency with which intact-shell eggs were implicated in subsequent outbreaks of S. Enteritidis suggested that this intervention may have been effective in decreasing the prevalence of this pathogen in eggs (Mishu et al., 1994).

    Trends in the occurrence of outbreaks are also used to gauge the success of national vaccination programs. Following the introduction of widespread childhood hepatitis A vaccination, previously large community-wide outbreaks and overall incidence of hepatitis A cases decreased markedly (Wasley et al., 2005).

    Interventions that successfully halt a health threat may have an impact on commerce and can even lead to litigation. Following a Salmonella outbreak that sickened hundreds of people who had consumed contaminated peanut products, the peanut processor subsequently declared bankruptcy and its executives faced charges for knowingly releasing contaminated products into commerce (Centers for Disease Control and Prevention, 2007; Tavernise, 2013).

  • Identify new pathogens. Legionnaire’s disease was first described after a large outbreak of respiratory disease at an American Legion convention in Philadelphia in 1976. Investigation of the outbreak led to discovery of a new organism, now called Legionella pneumophila, in specimens obtained from outbreak cases (Fraser et al., 1977). It was found to have been transmitted in aerosols from outdoor cooling towers. Subsequently, many other cooling tower-associated outbreaks have been recognized, as well as other routes of transmission (Centers for Disease Control and Prevention, 2000; Den Boer et al., 2002).

    Similarly, although sporadic cases of the Acute Respiratory Distress Syndrome had been seen for years, investigation of an outbreak in the Four Corners area of the southwestern United States in 1993 led to the description of Hantavirus Pulmonary Syndrome and identification of a previously unrecognized etiologic agent (Duchin et al., 1994).

    In February 2003, the World Health Organization was alerted by a physician at the French Hospital in Hanoi, Vietnam, about a severe respiratory disease in a traveler and spread of the infection among hospital healthcare workers. The ensuing multinational investigation led to the recognition of the global outbreak of Severe Acute Respiratory Syndrome (SARS) and the discovery of the causal agent, the SARS coronavirus. Earlier cases reported from China were probably mistaken for influenza and other respiratory pathogens (Reilley et al., 2003).

    Such investigations can also reveal the mode of transmission, incubation period, spectrum of disease, and risk factors for infection. Even if the infectious agent causing illness is undetected at the time of the investigation, outbreaks often provide an opportunity to obtain historical specimens and epidemiologic data from cases that can prove valuable in later years when improved technologies for pathogen detection become available.

Outbreak Detection

The sequence of events leading up to an outbreak investigation typically begins when some kind of unusual health event in the community is detected. Sometimes the unusual event is the occurrence of even a single case of an uncommon disease that poses a clear threat to public health, such as botulism, paralytic shellfish poisoning, or anthrax. Often the unusual event is the recognition of two or more similar cases that appear to have occurred suspiciously close to each other in space or time—a cluster of cases—which may or may not represent an outbreak.

Common ways by which such unusual health events are detected include:

  • An astute health care worker. Clinicians, infection control practitioners, and laboratory staff function as the “eyes and ears” of the public health system. In 1980, a report by an alert physician in California of an increase in the number of patients with Pneumocystis carinii pneumonia led to an investigation of what was originally called Gay-Related Immunodeficiency Syndrome. The disease is now known as AIDS (Centers for Disease Control and Prevention, 1981). Other notable examples of high-profile outbreaks that the public health system was alerted to by astute clinicians include Hantavirus Pulmonary Syndrome (Duchin et al., 1994), the 2001 anthrax bioterror attack (Bush et al., 2001), SARS (Reilley et al., 2003), and the large multi-state outbreak of fungal meningitis described above (Pettit et al., 2012).

  • A citizen. The borreliosis now known as Lyme disease was first recognized when the mother of a child diagnosed with rheumatoid arthritis, a condition uncommon in children, notified the local health department that she knew of at least three other cases of this disease in her neighborhood. The subsequent investigation identified a new pathogen, Borrelia burghdorferi, and the tick vector responsible for the disease outbreak (Steere et al., 1977). Often, outbreaks of foodborne illness are identified when citizens call to report that a number of people became ill after consuming food at a large group event.

  • Reportable disease surveillance. As was described in Chapter 6, each state publishes a list of communicable diseases and conditions that laboratories and health care providers are required to report to local public health authorities. In practice, routine disease reports are often not timely or complete enough to be useful in rapidly detecting outbreaks, especially those due to common conditions. Patients may delay seeking care, appropriate specimens may not be collected or the right tests not ordered, laboratory processing and reporting of results can be delayed, and health care providers and laboratories may not comply with reporting requirements. Nonetheless, frequent detailed analyses of routine surveillance data can be important in detecting smaller or geographically dispersed “hidden outbreaks” that may otherwise escape notice when total case reports remain relatively stable.

    In larger communities, surveillance data can even be reviewed on a daily basis, not only for unusual increases in total numbers of cases, but also for increases among subpopulations defined by age, gender, ethnicity, geography, or other risk factors. For example, although overall rates of hepatitis A remained relatively stable in many areas during the 1990s, analyses of the age, sex, and geographic distribution of cases revealed an increased incidence among young men living in urban areas and led to recognition of increased transmission of hepatitis A among injecting drug users and men who have sex with men (Bell et al., 1998).

    More recently, although the overall rate of infection with invasive meningococcal disease remained similar to that in previous years, a cluster of illness in young men who have sex with men was identified in Berlin in 2012 and 2013. Initially, three cases were reported with smoking and attendance at gay bars as their only identified risk factors; genetic testing of their clinical specimens revealed they were infected with the same bacterial strain. Retrospective data analysis revealed a higher proportion of cases than usual in young men, and two additional outbreak cases were identified. A similar cluster among gay men was also recognized in New York City (Centers for Disease Control and Prevention, 2013b; European Centre for Disease Prevention and Control, 2013).

  • Electronic health data. Electronic health data may provide information from large populations fairly efficiently, both regarding incidence of cases during an outbreak as well as baseline rates for specific conditions. For example, review of electronic emergency department data on patient chief complaints and discharge diagnoses (syndromic surveillance) detected an increase in carbon monoxide poisonings, including clusters in Zip codes with significant power outages, following a large windstorm in Seattle (Baer et al., 2011). Recent changes in the regulation and standardization of computerized electronic health data by healthcare facilities and laboratories will facilitate sharing of data with public health authorities and may improve data completeness and quality, and decrease time to case reporting and outbreak detection (Henricks, 2011; Wurtz, 2013).

Verifying an Outbreak

What Is the Illness?

Once a suspected outbreak is identified, characterizing the specific nature of the illness in question is an important early step, particularly for diseases and syndromes that are new or not yet fully characterized. Usually this task involves reviewing the clinical case history and checking key laboratory results. For new or unusual conditions, verification may involve consulting with clinical or laboratory experts and collecting clinical specimens for specialized testing at national reference laboratories to assist with diagnosis. Even in the absence of a specific diagnosis, systematically summarizing the signs and symptoms of illness can help characterize the disease and to develop a working case definition that can be refined as more information becomes available.

Is There a True Excess?

For serious conditions with outbreak potential, such as meningococcal meningitis, measles, Shiga-toxin producing E. coli, or typhoid fever, even a single case should “raise a red flag,” triggering an epidemiologic investigation. Single cases of these conditions require immediate public health investigation to attempt to identify the probable source, other persons at risk, and to formulate and implement intervention strategies to prevent additional persons from becoming ill. Single cases of certain non-contagious diseases, such as botulism, paralytic shellfish poisoning, and vibriosis, also require immediate investigation to allow prompt identification of exposed persons and recall of contaminated food products before they are consumed by others.

It is always important to determine the degree to which the number of new cases or events observed truly exceeds the number expected in a given geographic area during a defined period of time. The expected number of “sporadic” (or background) cases is usually estimated from historical data. For diseases that are reportable by law to the local health department, baseline surveillance data will usually be available and can often be stratified by age, geographic location, and other variables. For conditions that may vary in incidence with time (e.g., expected seasonal variation), it is helpful to look at incidence data from comparable time periods in recent past years to establish baseline rates and clearly identify the onset of the outbreak.

When standardized surveillance data are not available, medical records, laboratory test results, and other data will need to be reviewed. Data from health care institutions, including hospitals, microbiology laboratories and emergency departments, as well as outpatient clinical facilities, may be useful sources with which to establish baseline incidence. For example, in investigating a cluster of cases of Legionnaire’s disease, hospital records, including hospital pneumonia admissions, intensive care unit admissions, and discharge diagnoses, would be good sources to identify the persons hospitalized with pneumonia during a given time interval. Other potential sources of baseline data include disease registries and causes of death from death certificates available through vital statistics offices. In the absence of existing data, surveys of health care providers and key informant interviews may also help in judging whether an increase in cases is occurring.

An apparent excess in the number of reported cases does not necessarily mean that an outbreak is in progress. Other possibilities to consider are:

  • Change in the population at risk. As noted in Chapter 3, a simple case count can be adequate to compare incidence between time periods if it is safe to assume that the underlying population at risk is relatively constant. In some situations, however, this assumption is untenable—for example, in communities that have had rapid population growth or that experience marked seasonal changes in population demographics, such as tourist centers, university communities, regions hosting mass gatherings, or populations migrating in response to natural disasters. To avoid this source of error, comparisons should be based on rates, not just case counts, whenever possible.

  • Change in case ascertainment. Caution must be used when disease rates obtained from active or enhanced surveillance activities are compared with rates calculated using baseline or passively collected data. For example, conditions that routinely go under-reported will seem to increase when active or enhanced surveillance is used, but the increase in reports can simply reflect better ascertainment and not a true increase in disease incidence (Centers for Disease Control and Prevention, 1995; Glatzel et al., 2002).

    A common phenomenon during outbreaks is that increased awareness of the outbreak among the public and healthcare providers leads to increases in healthcare-seeking behavior and diagnostic testing for the condition in question, uncovering both outbreak-associated and unrelated sporadic cases that would otherwise go undiagnosed and unreported, particularly milder cases. As noted below, laboratory methods for characterizing relatedness of organisms (“molecular epidemiology”) can help to distinguish outbreak-associated from unrelated background cases.

    Even when evaluating clusters of notifiable diseases in the absence of enhanced surveillance, it is important to consider whether some element of the reporting system has changed, producing apparent increases in incidence in the absence of any real change in disease occurrence, including: a new or improved laboratory test; changes in interpretation of test results; heightened awareness of a new disease; changes in clinical practice standards; a new physician in the community with particular expertise in a disease; improved disease reporting by a new hospital infection-control practitioner; or changes in patient or laboratory referral patterns among health care providers (Joce et al., 1995; Centers for Disease Control and Prevention, 1997; Weinbaum et al., 1998; Adderson et al., 2000).

    The expansion of the AIDS surveillance definition in 1993 resulted in an increase in reporting (but not corresponding to an increase in incidence) of over 111%, a dramatic example of the impact a change in case definition can have (Centers for Disease Control and Prevention, 1994). When considering changes in reporting patterns, it may be useful to consult resources that document what has been reportable over time, such as the case definitions established by the CDC and the Council for State and Territorial Epidemiologists (CSTE) (National Notifiable Disease Surveillance System, 2013).

Are the Cases Related?

Besides being more numerous than expected, outbreak cases may cluster in space and time or be related in some other as yet unrecognized way. Sometimes links between them may not become apparent until after an investigation is in progress. However, molecular strain–typing methods are informative laboratory-based epidemiologic tools that can help in understanding disease transmission and determining whether a set of cases of illness due to an infectious agent may be connected or are simply a chance collection of sporadic cases (Centers for Disease Control and Prevention, 2002b).

Molecular strain–typing methods provide genotypic data that are more discriminating than phenotypic methods of characterizing organisms, such as serotyping. Molecular strain typing makes it possible to detect small potentially linked clusters of cases and outbreaks due to related strains of common serotypes that would otherwise go unnoticed. For example, pulsed field gel electrophoresis (PFGE), based on electrophoretic migration patterns of bacterial DNA fragments, is routinely employed by public health laboratories to detect clusters of Salmonella, Shiga-toxin-producing E. coli, and other pathogens, even when the total number of cases remains stable (Bender et al., 1997, 2001).

A well-established CDC-sponsored Internet-based surveillance tool for enteric disease pathogens, called PulseNet, enables laboratories across the country to compare standardized PFGE patterns of local cases against a large library of patterns (Swaminathan et al., 2001). This surveillance network can identify widely geographically dispersed outbreaks that would have otherwise gone undetected, with only a few cases occurring in any single health jurisdiction. Newer molecular typing methods based on complete or partial bacterial gene sequencing provide increasingly rich information, which, when interpreted in the context of epidemiologic data, have the potential to improve detection of outbreaks as well as track the evolution of disease-causing pathogens in populations (Goering et al., 2013).

Investigating an Outbreak

The decision on whether to dedicate resources to the investigation of a cluster of illnesses depends on many factors. Outbreaks of a severe illness or outbreaks involving many cases normally prompt an investigation, as do those involving an unusual or newly recognized illness. Some investigations are conducted simply because public attention to a perceived “outbreak” demands that the situation be evaluated. The epidemiologist must anticipate and evaluate the needs of the community, the resources available, whether an effective public health intervention exists, and the potential social and political consequences of the investigation (or of the failure to conduct one).

Overview of Steps

Although no two outbreaks evolve in exactly the same way, the steps commonly involved in investigating an outbreak are:

  • Establish a case definition

  • Enhance surveillance

  • Describe occurrence of cases according to person, place, and time

  • Develop hypotheses about the nature of the exposure(s) responsible for the outbreak

  • Conduct analytic studies, if appropriate

  • Implement disease-control interventions

  • Communicate status reports and results of the investigation

Each step is described more fully below. In practice, the steps are not necessarily sequential, but rather multiple steps often occur simultaneously; in some investigations not all the steps may be needed. Sometimes the pace of events is rapid, and the outbreak epidemiologist may need to anticipate and prepare for later steps before the results of earlier steps are complete.

A structured approach to managing outbreak investigations is desirable. The team leader(s) and other members of the outbreak investigation and response team should be identified and specific roles and responsibilities assigned. For large or complicated investigations, incident-management systems are often employed to help organize and track response activities and required resources. An “operations” team within an incident-management structure might include designated staff for surveillance, analytic epidemiologic studies, clinical investigation, environmental and/or field investigation, and response activities (Qureshi et al., 2006). Communication and logistics are other components of incident management structures.

Establish a Case Definition

Collecting as much information as feasible on all potential cases early in the investigation can help characterize the full spectrum of disease, save time in the long run, and help maximize the sample size available for later analysis. Once the disease symptomatology and laboratory findings have been established, an explicit working case definition is developed and applied consistently to all potential cases. As noted in Chapter 2, case definitions can include several clinical and/or laboratory criteria. A fairly broad or “loose” case definition is useful early in an outbreak investigation to include as many potential cases as possible, while collecting enough specific information to enable refinement of the definition as the investigation progresses. Early cases are often categorized as confirmed, probable, or suspect, depending on the extent to which a clinically compatible illness, laboratory confirmation, and an epidemiologic link to other confirmed cases are present.

Obtaining appropriate biological specimens for laboratory testing can be important to document or rule out potential cases as confirmed, to provide isolates for molecular epidemiology testing, to identify otherwise obscure etiologic agents, or to identify an outbreak etiology when people would otherwise not seek medical care for testing. Since laboratory testing of each possible case is often neither feasible nor necessary, case definitions can be created using a constellation of symptoms exhibited by the laboratory-confirmed cases.

The goal of the case definition is to include only true cases. Inclusion of non-cases as cases, or of subclinical cases in the comparison group, results in misclassification, which weakens the ability to detect an association with the relevant risk factor(s) (see Chapter 10). For large outbreaks with many cases and high statistical power, misclassification may be less of a concern because it may still be possible to detect an attenuated association. Refinement of case definitions can also be useful during the analysis of data to enhance specificity.

For example, at the start of an investigation of a cluster of suspected influenza cases in a nursing home that occurred over two weeks, one might initially consider anyone in the nursing home with a febrile illness during that two-week period as a suspect case. Collecting additional information about the signs and symptoms of the clinical illness, such as presence of cough or sore throat, duration of illness, and laboratory data including results of respiratory specimen testing, allows further narrowing and refinement of the case definition. Later in the investigation, the case definition might thus become “any resident of the nursing home with laboratory-confirmed influenza infection with onset between June 4 and June 16.”

Enhance Surveillance

When it is suspected that an outbreak is occurring, enhanced surveillance can be useful in identifying additional cases. Enhanced surveillance may involve both heightening awareness to increase passive case reports and implementing active surveillance. For outbreaks requiring widespread notification of the health care community, multiple methods of Internet-based communication, including e-mail Listservs, message boards, and online surveys, as well as “broadcast faxes” to area health care providers, hospitals, emergency departments, laboratories and other relevant groups can be employed. Messages should contain current information about the outbreak, inform clinicians about the syndrome or case definition under surveillance, describe how reporting should be done, and provide contact numbers and resources for questions or additional information.

Key information and resources can also be posted on health department web pages and communicated through social media such as Facebook and Twitter (Howland and Conover, 2011). Such channels may also alert health authorities that an outbreak is occurring. In Minnesota, for example, the Department of Health was notified of multiple Facebook postings suggesting foodborne illness among attendees of a high school banquet. Investigation identified pasta prepared by a team member’s parent as the source of the outbreak of Group A Streptococcus pharyngitis (Kemble et al., 2013).

For large outbreaks or public health emergencies, press releases and the print, radio, online, and television media can also be employed. Outbreak epidemiologists should therefore have a good working relationship with their department’s public information officer, designated spokesperson, or media liaison. Using the news media allows communication directly to the public for identifying cases in persons who may not have sought medical attention and for disseminating disease control recommendations.

Describe Occurrence of Cases According to Person, Place, and Time

Descriptive analysis reveals useful information about the basic features of the cases, the population affected, the geographic scope of the problem, and the pace at which the outbreak is evolving. Early descriptive analyses can also suggest potential etiologic agents, risk factors for acquisition, and mechanisms of transmission of infection. These clues can be a fertile source of hypotheses that can then be tested with analytic study designs, as described below.

To begin, a line list of the cases is prepared, as illustrated in Table 20.1. It shows demographic characteristics and other key descriptors, including components of the case definition and laboratory test results. This format allows quick examination for obvious common features or unusual values. The line list can be created by hand initially for small outbreaks, or using computerized data management programs or spreadsheets to allow easy viewing of multiple variables. For cohort studies of foodborne and other types of outbreaks, line lists should also include rows for the comparison group within the cohort who did not become ill, and columns for individual exposures under investigation.

Table 20.1. Example of a Line List of Data on Hepatitis a Cases

Diagnostic

Lab

Signs and symptoms*

HA

Case #

Initials

Date of report

Date of onset

MD Dx

N

V

A

F

DU

J

IgM

Other

Age

Sex

1

JG

10/12

10/6

Hep A

+

+

+

+

+

+

+

AST↑

37

M

2

BC

10/12

10/5

Hep A

+

+

+

+

+

+

ALT↑

62

F

3

HP

10/13

10/4

Hep A

±

+

+

+

S†

+

AST↑

30

F

4

MC

10/15

10/4

Hep A

+

+

?

+

HbSAg −

17

F

5

NG

10/15

10/9

NA

+

+

+

NA

NA

32

F

6

RD

10/15

10/8

Hep A

+

+

+

+

+

+

+

38

M

7

KR

10/16

10/13

Hep A

±

+

+

+

+

+

AST↑

43

M

8

DM

10/16

10/12

Hep A

+

+

+

+

57

M

9

PA

10/18

10/7

Hep A

±

+

±

+

+

+

52

F

* KEY:

  • S† = scleral icterus

  • F = fever

  • N = nausea

  • DU = dark urine

  • V = vomiting

  • J = jaundice

  • A = anorexia

  • HA IgM = hepatitis A IgM antibody test

(Adapted from Dicker [1998])

Data in the line list can be analyzed using simple descriptive statistics, such as counts, percentages, means, and standard deviations; t-tests and Chi-square tests can be employed to examine differences in key demographics.

Characteristics of the illness can be summarized, and the most frequent signs and symptoms may suggest a likely differential diagnosis if the agent has not yet been identified. Pre-established criteria, such as the Kaplan criteria used for diagnosing suspect outbreaks of norovirus, may help rule a suspect etiology in or out in the absence of confirmatory testing (Karagiannis et al., 2010). The Kaplan criteria are:

  1. 1. A mean (or median) illness duration of 12 to 60 hours,

  2. 2. A mean (or median) incubation period of 24 to 48 hours,

  3. 3. More than 50% of people with vomiting, and

  4. 4. No bacterial agent found.

Given the scenario of multiple people ill with vomiting and diarrhea of 24 hours duration among attendees of a child care center in January, health officials would be likely to hypothesize that an outbreak of norovirus is occurring.

If the disease has been definitively diagnosed, hypotheses regarding source of exposure and transmission can be developed based on known risk factors, incubation period, and known vehicles for that disease. For example, an outbreak of invasive Listeria infections among pregnant Hispanic women suggested a possible foodborne source, later found to be a commercially produced Mexican cheese (Jackson et al., 2011).

Plotting the location where cases reside, work, or engage in recreational activities on spot maps or using geographic information system (GIS) software can assist in identifying potential sources of exposure or routes of transmission. John Snow used spot maps such as the one shown in Fig. 7-9 in his investigations of the 1854 cholera epidemic in central London to illustrate the distribution of cholera cases in Golden Square, showing that cholera deaths were strikingly common around the Broad Street water pump (Snow, 1936). Mapping software programs available today are complex and often require specially trained staff.

Spot maps may be useful when the source of infection is unknown, in order to search for clustering of cases by location of home, work, or recreational activities. They can later be used to show important spatial or geographic relationships once the source or mode of transmission has been identified. Examples include outbreaks caused by common-source exposure to contaminated aerosols, such as legionellosis associated with contaminated cooling towers, decorative fountains, or other sources of aerosol transmission; enteric infections resulting from exposure to contaminated recreational water; and institutional outbreaks in which visualizing spatial relationships among infected persons and potential sources of infection (other patients, environmental reservoirs, or caregivers) may be informative. However, for many outbreaks in a mobile society, cases will have been infected in distant locations or through exposure to widely distributed commercial food products, which will not be reflected in spot maps depicting only their local activities.

The epidemic curve is a standard part of the descriptive epidemiologic analysis. The date (or time) of onset is shown on the X-axis, while the number of new cases with onset in each date or time category is plotted on the Y-axis. Potentially significant case data, such as disease severity, confirmed versus suspect cases, or exposure to some suspected risk factor, can be indicated with different colors or fill patterns.

To produce the most informative epidemic curve, the scale of the X-axis should depend upon the incubation period of the disease (if known) and should include the pre-outbreak period to illustrate the background incidence of the disease. For most outbreaks, a useful scale will indicate time in units approximately one quarter the length of the incubation period. If the disease being investigated has not yet been identified, plotting the data on a variety of scales may reveal a pattern.

When the usual incubation period of the disease is known, and if most or all cases in an initial wave were exposed at about the same time, the epidemic curve can often be used to reveal the likely time of that exposure. For example, consider the epidemic curve from an outbreak of hepatitis A, as shown in Fig. 20.1. The incubation period of hepatitis A averages about 28–30 days, ranging from 15 to 50 days between exposure and onset of symptoms (Heymann, 2008). Suppose that all of the cases shown were suspected of having resulted from a common exposure to a single index case, such as an ill food worker. The analyst would count back 15 days from the earliest case, and 28–30 days from the peak of cases, to focus the investigation on a narrow range of days within which the common exposure may have occurred.

Figure 20.1 Epidemic curve from an outbreak of hepatitis A.

Figure 20.1
Epidemic curve from an outbreak of hepatitis A.

(Based on data from Dicker [1998])

The shape of the epidemic curve often provides information about the likely mode of transmission. A “point source” outbreak, in which all cases were exposed to a single source of infection (a common meal or highly contagious person, for example), classically exhibits a steep upswing and an early peak, followed by a gradual decline in cases (Fig. 20.1). All cases in a point source outbreak should also have onset of illness within one incubation period. Prolonged exposure of a population to a common source (such as a contaminated food product with a long shelf life) will produce a more “smeared-out” epidemic curve in which onset of cases extends well beyond a single incubation period, reflecting a series of “mini-outbreaks.” In contrast, ongoing person-to-person (secondary) transmission classically produces an epidemic curve with a series of small peaks—ideally one incubation period apart, although such a regular pattern is often difficult to demonstrate.

The epidemic curve can also provide clues about the course of the outbreak: a rising curve suggests that the outbreak is in the early stages; a plateau suggests that transmission may be stable or decreasing; and a downward slope suggests that the outbreak is waning. This information is helpful in planning disease control strategy, anticipating resource needs, evaluating interventions, and communicating with outbreak response partners and the public.

As the investigation develops and data on cases are entered into a database, more detailed and sophisticated epidemic curves can be created that convey additional information about primary and secondary cases and the timing of relevant events and interventions.

Develop Hypotheses About the Nature of the Exposure(s) Responsible for the Outbreak

The information needed for the initial line list of cases and for initial descriptive analyses may be obtained through interviews with cases and/or medical record reviews. For most outbreaks of reportable diseases with well-recognized routes of transmission, standardized interview forms are available at local and state health departments and from CDC. Staff of these agencies can often provide technical assistance as well.

For situations in which the clinical illness is not yet characterized, or the route or source of transmission appears to be new, an open-ended, unstructured interview with early cases and other key informants (such as family members or providers) can provide valuable information. These exploratory interviews should usually cover a wide range of potential risk factors and exposures, such as lifestyle characteristics, exposure to the outdoors, homelessness, or other factors. It is helpful if the same person can conduct these initial interviews. If the interviewer hears a similar story several times, patterns of exposure may emerge. Also, it is often worthwhile to ask the patient and family members where or how they think the illness was acquired. They may have already figured it out, and at the very least they will be more cooperative because they have been asked for their opinions. Other good sources of hypotheses can be subject-matter experts (infectious disease specialists, public health veterinarians, laboratorians, toxicologists, industrial hygienists, water system or air handling experts, etc.) and the medical literature.

Outliers are cases that are unusual in some way—they “just don’t fit in” with the rest of the cases—and should be carefully scrutinized. Temporal outliers are usually easy to identify on the epidemic curve and often provide the key to understanding the basis for the outbreak. For instance, a single case occurring one incubation period before the other cases may represent an ill food handler who contaminated a food product, or an index case who exposed a large group of susceptible people. Cases who were only in the geographic area a short time can, by their limited opportunities for exposure, also provide valuable information about means of transmission. Similarly, foodborne outbreaks in which most cases and controls ate the same food items can sometimes be solved if “dietary” outliers who consumed only one of several potentially contaminated items are identified.

Hypothesis generation may lead to a list of additional information needed, such as exposure details, medical history, or laboratory data. This list can then be used to develop more refined case investigation forms for use in collecting additional data.

Conduct Analytic Studies

Once a set of hypotheses has been developed, an analytic study—usually a case-control or cohort study—is often the appropriate next step.

Study Design

Choice of a study design depends both on theoretical considerations (see Chapter 5), such as frequency of the disease and of key exposures, and on logistics and data availability. If the outbreak occurs in a discrete, readily identified group, such as a wedding party or passengers on a cruise ship, a cohort study is the preferred option. As many members of the group as possible should be interviewed. Cohort studies are discussed in Chapter 14. In contrast, if cases are few or occur among a widely scattered group, it may be more efficient to conduct a case-control study, sampling appropriate non-ill controls from the presumed population at risk. This study design is discussed at length in Chapter 15.

In certain instances, case-case studies can be conducted, saving valuable time and resources by using cases with the same illness, but who are not associated with the outbreak, as the comparison group. These cases may occur concurrently with the outbreak-associated cases but have a different strain of the pathogen, or may be temporally matched cases from previous years for whom detailed exposure data already exists. For example, in September 2011, an unusually high number of cases of listeriosis were reported in Colorado residents. Because listeriosis is a relatively rare and severe illness, detailed exposure information is routinely collected on all reported cases. Investigators were able to compare exposures from the first 19 outbreak-associated cases with 85 age- and temporally matched Listeria cases from 2004–2010 and found a strong association with cantaloupe consumption (odds ratio=14.9; 95% CI = 2.4 – ), leading to a rapid product recall (Centers for Disease Control and Prevention, 2011).

A few study design issues of special relevance to outbreak investigation will be discussed below. In an outbreak situation, the working case-definition often limits cases in time and vicinity, and these limits should also pertain to controls. Ideally, controls should be persons who did not develop the disease but who met all other criteria that defined the cases. If the outbreak occurred among patrons of an outdoor festival, for example, controls should be chosen from among attendees of the festival; if at a potluck supper, from among those who attended the supper (or unsuspecting family members who ate the leftovers).

Nonetheless, care should be taken not to over-match or restrict controls in such a way that exposure to a risk factor of interest is essentially predetermined. For example, consider an outbreak of shigellosis among children who visited a recreational area with a swimming pool. A reasonable control group would be children in the same age range without signs or symptoms of Shigella infection who also frequented the recreational area during the exposure period of the cases. If one selected as controls only children who also swam in the pool (i.e., matched on swimming history), the analysis would not be able to determine whether exposure to the possibly contaminated pool water was a risk factor for infection.

Data Collection

In general, the farther in time the investigation is conducted after the relevant exposure has taken place, the more difficult it is to get complete and accurate information from both cases and controls about potential risk factors of interest. In addition, with time it becomes more difficult for controls to recall events with as much certainty as cases, who were more affected and may be more closely scrutinizing their recent exposures. This difference can contribute to recall bias (see Chapter 15). Thus, the time spent to develop the data collection instruments must be balanced by the need for a prompt investigation. Not uncommonly, the data instruments will appear flawed in retrospect. Outbreak investigation forms must often be created “on the fly,” without the luxury of time for the methodical planning that is available in elective studies. For event-based outbreaks, online surveys can often be distributed electronically to ill and non-ill attendees. These surveys have the advantage of rapidly gathering exposure and illness information without requiring staff time to conduct telephone or in-person interviews, though the level of response to such surveys may be variable.

In addition, with the advent of warehouse and supermarket membership or “shopper” cards, investigators can obtain dates of purchase, brand names, and lot numbers to fill in gaps in information gathered from case interviews. For case clusters with no obvious common exposures beyond supermarket memberships, comparing purchase records can also be essential during the hypothesis-generation stages of an investigation. While use of shopper cards may expedite source identification and the product traceback and recall process, care must be taken to respect the privacy of cases and limit record access to the minimum necessary to further the investigation. In 2010, after several months of investigation into an increase in Salmonella Montevideo cases where no source was identified, health officials in Washington State obtained permission from cases to review their purchase records from a large membership warehouse. The record review identified salami products from a single company among the majority of cases, and the outbreak was ultimately traced back to contaminated red and black pepper used to season the deli meats (Centers for Disease Control and Prevention, 2010).

As in elective analytic studies, attention should be paid to minimizing measurement error during collection of data. Helpful techniques include using preexisting questions or instruments; training interviewers to collect data in a standard way, especially if two or more interviewers are needed; and using standardized visual aids such as calendars to help respondents recall the timing of events. Pilot testing of questionnaires and blinding of interviewers to disease status is desirable but often impossible in an outbreak situation.

Special care is needed in interviewing persons under the legal age of consent, which requires permission of a parent or guardian. If interviews are conducted in person, the parent/guardian should be present. If interviews are by telephone, having the parent listen in on another phone line is often reassuring to the parent and can sometimes help the interviewee with recall.

If the outbreak etiology is unknown, it is important to collect necessary biological specimens as soon as possible to establish the diagnosis, isolate the etiologic agent, or further characterize the clinical syndrome. At times, nurses and other investigators with clinical skills collect specimens in the field (e.g., through phlebotomy, or throat or skin cultures). For presumed foodborne outbreaks, cases should be asked to retain and refrigerate any leftover suspect foods, including the original packaging when available. Outbreak investigators may need to issue recommendations to health care providers and laboratories on procedures for diagnostic testing and handling of clinical specimens. Clinical laboratories often discard culture isolates and other diagnostic specimens after a few days unless specifically requested to do otherwise. Local laboratories should be contacted as early as possible to request that specimens be conserved or sent on to the local or state public health department laboratory.

Another reason for prompt specimen collection is that cases are generally most interested in cooperating with the investigation when they are currently or recently symptomatic. Once recovered, they soon tire of requests for additional stool specimens or blood samples, or even one more telephone interview. If specimen collection is crucial to the investigation, it is often worth the effort to send someone directly to the field (e.g., home, restaurant, etc.) to collect appropriate specimens.

The environmental field investigation is often carried out by one or more environmental health specialists working with the communicable disease epidemiologist. The field team should be briefed on the situation; specifically what samples and/or other information are needed. Some field investigations may include assessment and/or sampling of wildlife, water supplies, or potentially toxic substances. In these scenarios, wildlife biologists, public health veterinarians, ecologists, or occupational health specialists may be desirable members of the team. It is important to assure that when necessary, field staff have the appropriate personal protective equipment and are trained in its proper use. Field staff should have documentation of vaccination or immunity against diseases to which they may be exposed during investigations.

The first visit by the field team is the best opportunity to observe pertinent behaviors (e.g. food safety practices, hygiene, infection control practices) and to obtain environmental samples. In foodborne outbreaks, information on suspect products and the methods of preparation may be crucial, including brand name, lot number, size of package used, expiration date, delivery date, and supplier (to assist in traceback efforts when indicated), as well as food preparation, holding, and storage conditions. Personnel (including supervisors) involved in handling and preparing implicated food items should be identified and interviewed about food preparation and handling practices and recent illness. When possible, clinical specimens should be obtained from staff that are ill and/or are suspected of serving as a reservoir for the infectious agent. It may be difficult in these situations to determine whether staff who test positive represent a cause of illness, or were exposed to the same contaminated source as the other cases.

Analysis

Additional analytic studies are used to more completely describe the affected population, risk factors for disease, and mechanisms of transmission. As data come in, they should be examined for completeness and consistency; these data are typically then entered into computerized databases. EpiInfo is a software program developed specifically to support data management and analysis for outbreak investigations; it can be downloaded free of charge from the World Wide Web. Preparing mock-up tables early, even before data-collection instruments are finalized, can help guide the analysis, identify gaps, and anticipate the need to reconcile data from different sources (e.g., different laboratories) onto a common scale.

Outbreaks often evolve over a short time, during which changes in the population at risk may be minor. Hence disease frequency is often expressed in terms of the “attack rate”—technically, not a true “rate” but another term for cumulative incidence. The difference between the attack rates in persons with and without a certain exposure is thus the attributable risk, and the ratio of the two attack rates is the relative risk, as was discussed in Chapter 9. These measures of effect can be used to quantify associations, and hypotheses can be tested quickly using simple cross-tabulations (Bryan et al., 1999). In the example shown in Table 20.2, 120 persons out of 200 attendees became ill. The relative risk and the attributable risk for roast turkey both show a large positive association with illness, suggesting a possible causal relationship. Conversely, eating roast pork was negatively associated with illness, perhaps because those who chose pork avoided the turkey. The other associations are weak and do not suggest that these food items were related to becoming ill.

Table 20.2. Attack Rate Table for a Hypothetical Foodborne Disease Outbreak (N = 200)

Ate

Did not eat

Ill

Not ill

Attack rate

Ill

Not ill

Attack rate

Attributable risk

Relative risk

Roast turkey

104

15

104/119 = 0.87

16

65

16/81 = 0.20

+0.67

4.35

Roast pork

15

45

15/60 = 0.25

105

35

105/140 = 0.75

−0.50

0.33

Mashed potatoes

102

77

102/179= 0.57

13

8

13/21= 0.62

−0.05

0.92

Green beans

63

34

63/97 = 0.65

57

46

57/103 = 0.55

+0.10

1.18

Rolls

87

59

87/146= 0.60

33

21

33/54 = 0.61

−0.01

0.98

Apple pie

76

50

76/126= 0.60

44

30

44/74 = 0.59

+0.01

1.02

A second issue to consider is the fraction of cases that each exposure under consideration could account for—the population attributable risk percent—which depends on both the relative risk and the proportion of cases exposed to each item. In this example, while eating green beans was positively associated with illness (RR = 1.18), only 63 of the 120 ill persons recalled having eaten green beans, so that this exposure (even if it truly were a cause of illness) could account for only (1.18−1)/1.18 × (63/120) = 0.08 = 8% of cases. Roast turkey, besides being more strongly associated with illness, was eaten by 104 of the 120 ill persons and therefore could account for about 4.35−1)/4.35 × (104/120) = 0.67 = 67% of cases. When no single food item or multiple food items (especially those that are ready-to-eat) are significantly associated with illness, it may point to broad contamination during food preparation from contaminated equipment or environmental surfaces or by an ill food handler.

Implement Disease Control Interventions

The results of preliminary and analytical studies often implicate a particular exposure. Intervention approaches to prevent additional cases or future outbreaks depend on that exposure and on what is already known about the disease’s mechanism of spread. Among many examples are recalling a contaminated product from distribution, correcting deficient food-handling practices, and administering medication, immune globulin and/or vaccine to susceptible persons. Such efforts can involve multiple regulatory agencies and or members of the healthcare community and highlight the importance of keeping contact lists current and maintaining good communication, as described below. Subsequently, the results of outbreak investigations can be used to support policy changes to decrease the risk for future outbreaks.

Communication

Effective, clear, and timely communication is a critical component of any outbreak investigation. Important target groups and forms of communication include:

  • The public. Relevant information often includes the signs and symptoms of the disease, and recommendations for evaluation, treatment, and prevention of illness. Information can be disseminated through information hotlines with recorded messages, Web pages, press releases or other news media channels (National Research Council, 1989; Covello et al., 2001). It is important to compose public information in clear language that is understandable by the community. Translation of materials into other languages and outreach to target specific cultural groups or hard-to-reach populations may be necessary.

  • Outbreak response team members. Outbreak investigations often require coordinating the efforts of several people. Weekly or more frequent team meetings are needed to review the status of the outbreak, share information, and update and revise the investigation and response plan. Meeting frequency and participants includes will depend on the pace and specific nature of individual outbreaks. It may be useful to invite the public information officer and representatives from other affected local or state agencies to these sessions. The outbreak team leader needs to manage the overall response, anticipate where the investigation is headed, and ensure that adequate resources are available to sustain the investigation and response activities as long as necessary.

  • Public health and government officials. It is a good idea for the outbreak epidemiologist to keep his or her supervisor aware of the status of an investigation. Health officers and elected officials often do not appreciate learning about outbreaks for the first time through inquiries from local news media.

  • Local health care workers. Local environmental health staff, infection control practitioners, infectious disease experts, and other medical and health care professionals are natural partners in investigations. Having preexisting relationships and contact information readily available is very helpful.

  • Other potentially affected government agencies. At times, what appears to be a localized outbreak or cluster of cases is actually part of a larger regional, national, or even international outbreak that is not initially recognized. If circumstances suggest that the local cases might be part of a larger outbreak (e.g., possibly involving a commercially prepared product with wide distribution, or a travel-associated outbreak), consultation with regional or national health officials is recommended, even before confirmatory laboratory test results are available. These agencies can also provide help in confirming and investigating outbreaks when local resources are not adequate. In addition, for outbreaks involving commercial products or multiple states or countries, federal agricultural or pharmaceutical agencies (e.g., CDC, FDA, U.S. Department of Agriculture) may take the lead on coordinating the outbreak response with support from local partners. Daily scheduled conference calls can be useful in addition to updates as needed to communicate new information. It is wise to know in advance who the relevant contacts are at local, state, and federal agencies and have methods to communicate with them after business hours.

Conclusion

Conducting an outbreak investigation can be an exciting and rewarding, as well as a challenging, experience. Good analytical, social, and political skills can help open doors for the epidemiologist, both by aiding the prompt gathering of accurate information and by communicating results and recommendations to the public.

It is not necessary to “reinvent the wheel” when conducting an outbreak investigation. Readily available reference materials provide an overview of best practices and detailed guidelines for outbreak investigation activities. Examples of these resources, available on the World Wide Web, include CDC’s Manual for the Surveillance of Vaccine-Preventable Diseases (2013a) and the Council to Improve Foodborne Outbreak Response (CIFOR) Guidelines for Foodborne Outbreak Response (2013).

There are lessons to be learned from every outbreak. A formal outbreak review process or debriefing is often worthwhile after a large or complicated outbreak, to evaluate what worked and what did not. Such a review can involve representatives from many agencies and professional areas, and information from it can lead to appropriate changes in the response to future outbreaks.

Outbreak investigations can also provide an opportunity to deliver public health messages to the community. While most public health work takes place quietly behind the scenes, outbreak investigations are often the focus of intense community interest and media scrutiny. Carefully crafted communications can make a lasting impression that may favorably affect risk behavior in the population. A well-conducted outbreak investigation can also increase the public’s understanding of, and appreciation for, the work that public health professionals do.

References

Adderson E, Pavia A, Christenson J, Davis R, Leonard R, Carroll K. A community pseudo-outbreak of invasive Staphylococcus aureus infection. Diagn Microbiol Infect Dis 2000; 37:219–221.Find this resource:

Baer A, Elbert Y, Burkom HS, Holtry R, Lombardo JS, Duchin JS. Usefulness of syndromic data sources for investigating morbidity resulting from a severe weather event. Disaster Med Public Health Prep 2011; 5:37–45.Find this resource:

Becker KM, Moe CL, Southwick KL, MacCormack JN. Transmission of Norwalk virus during football game. N Engl J Med 2000; 343:1223–1227.Find this resource:

Bell BP, Goldoft M, Griffin PM, Davis MA, Gordon DC, Tarr PI, et al. A multistate outbreak of Escherichia coli O157:H7-associated bloody diarrhea and hemolytic uremic syndrome from hamburgers. The Washington experience. JAMA 1994; 272:1349–1353.Find this resource:

Bell BP, Khabbaz RF. Responding to the outbreak of invasive fungal infections: the value of public health to Americans. JAMA 2013; 309:883–884.Find this resource:

Bell BP, Shapiro CN, Alter MJ, Moyer LA, Judson FN, Mottram K, et al. The diverse patterns of hepatitis A epidemiology in the United States—implications for vaccination strategies. J Infect Dis 1998; 178:1579–1584.Find this resource:

Belongia EA, Hedberg CW, Gleich GJ, White KE, Mayeno AN, Loegering DA, et al. An investigation of the cause of the eosinophilia-myalgia syndrome associated with tryptophan use. N Engl J Med 1990; 323:357–365.Find this resource:

Bender JB, Hedberg CW, Besser JM, Boxrud DJ, MacDonald KL, Osterholm MT. Surveillance by molecular subtype for Escherichia coli O157:H7 infections in Minnesota by molecular subtyping. N Engl J Med 1997; 337:388–394.Find this resource:

Bender JB, Hedberg CW, Boxrud DJ, Besser JM, Wicklund JH, Smith KE, et al. Use of molecular subtyping in surveillance for Salmonella enterica serotype typhimurium. N Engl J Med 2001; 344:189–195.Find this resource:

Breuer T, Benkel DH, Shapiro RL, Hall WN, Winnett MM, Linn MJ, et al. A multistate outbreak of Escherichia coli O157:H7 infections linked to alfalfa sprouts grown from contaminated seeds. Emerg Infect Dis 2001; 7:977–982.Find this resource:

Bryan FL, Bartleson CA, Cook OD, et al. Procedures to investigate foodborne illness (5th ed.). Des Moines, IA: International Association of Milk, Food, and Environmental Sanitarians, Inc., 1999.Find this resource:

    Bush LM, Abrams BH, Beall A, Johnson CC. Index case of fatal inhalational anthrax due to bioterrorism in the United States. N Engl J Med 2001; 345:1607–1610.Find this resource:

    Carter AO, Borczyk AA, Carlson JAK, Harvey B, Hockin JC, Karmali MA, et al. A severe outbreak of Escherichia coli O157:H7-associated hemorrhaghic colitis in a nursing home. N Engl J Med 1987; 317:1496–1500.Find this resource:

    Centers for Disease Control and Prevention. Pneumocystis pneumonia—Los Angeles. MMWR Morb Mortal Wkly Rep 1981; 30:250–252.Find this resource:

      Centers for Disease Control and Prevention. Update: impact of the expanded AIDS surveillance case definition for adolescents and adults on case reporting—United States, 1993. MMWR Morb Mortal Wkly Rep 1994; 43:160–1, 167–170.Find this resource:

        Centers for Disease Control and Prevention. Enhanced detection of sporadic Escherichia coli O157:H7 infections—New Jersey, July 1994. MMWR Morb Mortal Wkly Rep 1995; 44:417–418.Find this resource:

          Centers for Disease Control and Prevention. Outbreaks of pseudo-infection with Cyclospora and Cryptosporidium—Florida and New York City, 1995. MMWR Morb Mortal Wkly Rep 1997; 46:354–358.Find this resource:

            Centers for Disease Control and Prevention. Legionnaires’ disease associated with potting soil—California, Oregon, and Washington, May–June 2000. MMWR Morb Mortal Wkly Rep 2000; 49:777–778.Find this resource:

              Centers for Disease Control and Prevention. Outbreak of Salmonella serotype Kottbus infections associated with eating alfalfa sprouts—Arizona, California, Colorado, and New Mexico, February–April 2001. MMWR Morb Mortal Wkly Rep 2002a; 51:7–9.Find this resource:

                Centers for Disease Control and Prevention. Rashes among schoolchildren—14 states, October 4, 2001–February 27, 2002. MMWR Morb Mortal Wkly Rep 2002b; 51:161–164.Find this resource:

                  Centers for Disease Control and Prevention. Multistate outbreak of Salmonella serotype Tennessee infections associated with peanut butter—United States, 2006–2007. MMWR Morb Mortal Wkly Rep 2007; 56:521–524.Find this resource:

                    Centers for Disease Control and Prevention. Multistate outbreak of Salmonella infections associated with frozen pot pies—United States, 2007. MMWR Morb Mortal Wkly Rep 2008; 57:1277–1280.Find this resource:

                      Centers for Disease Control and Prevention. Salmonella Montevideo infections associated with salami products made with contaminated imported black and red pepper—United States, July 2009–April 2010. MMWR Morb Mortal Wkly Rep 2010; 59:1647–1650.Find this resource:

                        Centers for Disease Control and Prevention. Multistate outbreak of listeriosis associated with Jensen Farms cantaloupe—United States, August–September 2011. MMWR Morb Mortal Wkly Rep 2011; 60:1357–1358.Find this resource:

                          Centers for Disease Control and Prevention. Manual for the surveillance of vaccine-preventable diseases (6th ed.). http://www.cdc.gov/vaccines/pubs/surv-manual, 2013a.

                          Centers for Disease Control and Prevention. Notes from the field: serogroup C invasive meningococcal disease among men who have sex with men—New York City, 2010–2012. MMWR Morb Mortal Wkly Rep 2013b; 61:1048.Find this resource:

                            Council to Improve Foodborne Outbreak Response. Guidelines for foodborne outbreak response. http://www.cifor.us/CIFORGuidelinesProjectMore.cfm, 2013.

                            Covello VT, Peters RG, Wojtecki JG, Hyde RC. Risk communication, the West Nile virus epidemic, and bioterrorism: responding to the communication challenges posed by the intentional or unintentional release of a pathogen in an urban setting. J Urban Health 2001; 78:382–391.Find this resource:

                            Den Boer JW, Yzerman EP, Schellekens J, Lettinga KD, Boshuizen HC, Van Steenbergen JE, et al. A large outbreak of Legionnaires’ disease at a flower show, the Netherlands, 1999. Emerg Infect Dis 2002; 8:37–43.Find this resource:

                            Dicker R. Principles of epidemiology. An introduction to applied epidemiology and biostatistics (2nd ed.). Atlanta, GA: Centers for Disease Control and Prevention, 1998.Find this resource:

                              Duchin JS, Koster FT, Peters CJ, Simpson GL, Tempest B, Zaki SR, et al. Hantavirus pulmonary syndrome: a clinical description of 17 patients with a newly recognized disease. The Hantavirus Study Group. N Engl J Med 1994; 330:949–955.Find this resource:

                              European Centre for Disease Prevention and Control. A cluster of invasive meningococcal disease in young men who have sex with men in Berlin, October 2012 to May 2013. Euro Surveill 2013; 18:pii=20523.Find this resource:

                                Falsey RR, Kinzer MH, Hurst S, Kalus A, Pottinger PS, Duchin JS, et al. Cutaneous inoculation of nontuberculous mycobacteria during professional tattooing: a case series and epidemiologic study. Clin Infect Dis 2013; 57:e143–147.Find this resource:

                                Fraser DW, Tsai TR, Orenstein W, Parkin WE, Beecham HJ, Sharrar RG, et al. Legionnaires’ disease: description of an epidemic of pneumonia. N Engl J Med 1977; 297:1189–1197.Find this resource:

                                Glatzel M, Rogivue C, Ghani A, Streffer JR, Amsler L, Aguzzi A. Incidence of Creutzfeldt-Jakob disease in Switzerland. Lancet 2002; 360:139–141.Find this resource:

                                Goering RV, Köck R, Grundmann H, Werner G, Friedrich AW, ESCMID Study Group for Epidemiological Markers (ESGEM). From theory to practice: molecular strain typing for the clinical and public health setting. Euro Surveill 2013; 18:20383.Find this resource:

                                  Henricks WH. “Meaningful use” of electronic health records and its relevance to laboratories and pathologists. J Pathol Inform 2011; 2:7.Find this resource:

                                  Heymann D (ed.). Control of communicable diseases manual (19th ed.). Washington, D.C.: American Public Health Association, 2008.Find this resource:

                                    Holzbauer SM, DeVries AS, Sejvar JJ, Lees CH, Adjemian J, McQuiston JH, et al. Epidemiologic investigation of immune-mediated polyradiculoneuropathy among abattoir workers exposed to porcine brain. PLoS One 2010; 5:e9782.Find this resource:

                                    Howland JF, Conover C. Social network as outbreak investigation tool. Emerg Infect Dis 2011; 17:1765–1766.Find this resource:

                                    Jackson KA, Biggerstaff M, Tobin-D’Angelo M, Sweat D, Klos R, Nosari J, et al. Multistate outbreak of Listeria monocytogenes associated with Mexican-style cheese made from pasteurized milk among pregnant, Hispanic women. J Food Prot 2011; 74:949–953.Find this resource:

                                    Joce RE, Murphy F, Robertson MH. A pseudo-outbreak of salmonellosis. Epidemiol Infect 1995; 115:31–38.Find this resource:

                                    Kainer MA, Reagan DR, Nguyen DB, Wiese AD, Wise ME, Ward J, et al. Fungal infections associated with contaminated methylprednisolone in Tennessee. N Engl J Med 2012; 367:2194–2203.Find this resource:

                                    Karagiannis I, Detsis M, Gkolfinopoulou K, Pervanidou D, Panagiotopoulos T, Bonovas S. An outbreak of gastroenteritis linked to seafood consumption in a remote Northern Aegean island, February–March 2010. Rural Remote Health 2010; 10:1507.Find this resource:

                                    Kemble SK, Westbrook A, Lynfield R, Bogard A, Koktavy N, Gall K, et al. Foodborne outbreak of Group A streptococcus pharyngitis associated with a high school dance team banquet—Minnesota, 2012. Clin Infect Dis 2013; 57:648–654.Find this resource:

                                    LeBlanc PM, Hollinger KA, Klontz KC. Tattoo ink–related infections—awareness, diagnosis, reporting, and prevention. N Engl J Med 2012; 367:985–987.Find this resource:

                                    Mishu B, Koehler J, Lee LA, Rodrigue D, Brenner FH, Blake P, et al. Outbreaks of Salmonella Enteritidis infections in the United States, 1985–1991. J Infect Dis 1994; 169:547–552.Find this resource:

                                    Mohle-Boetani JC, Farrar JA, Werner SB, Minassian D, Bryant R, Abbott S, et al. Escherichia coli O157 and Salmonella infections associated with sprouts in California, 1996–1998. Ann Intern Med 2001; 135:239–247.Find this resource:

                                    National Notifiable Disease Surveillance System. Case definitions. http://wwwn.cdc.gov/nndss/script/casedefDefault.aspx, 2013.

                                    National Research Council. Improving risk communication. Washington, DC: National Academy Press, 1989.Find this resource:

                                      Pettit AC, Kropski JA, Castilho JL, Schmitz JE, Rauch CA, Mobley BC, et al. The index case for the fungal meningitis outbreak in the United States. N Engl J Med 2012; 367:2119–2125.Find this resource:

                                      Qureshi K, Gebbie KM, Gebbie EN. Public health incident command system, Vol. 1. http://www.ualbanyc-php.org/pinata/phics/guide, 2006.

                                      Reilley B, Van Herp M, Sermand D, Dentico N. SARS and Carlo Urbani. N Engl J Med 2003; 348:1951–1952.Find this resource:

                                      Shands KN, Schmid GP, Dan BB, Blum D, Guidotti RJ, Hargrett NT, et al. Toxic-shock syndrome in menstruating women: association with tampon use and Staphylococcus aureus and clinical features in 52 cases. N Engl J Med 1980; 303:1436–1442.Find this resource:

                                      Snow J. Snow on cholera. New York: Commonwealth Fund, 1936.Find this resource:

                                        St Louis ME, Morse DL, Potter ME, DeMelfi TM, Guzewich JJ, Tauxe RV, et al. The emergence of grade A eggs as a major source of Salmonella Enteritidis infections. New implications for the control of salmonellosis. JAMA 1988; 259:2103–2107.Find this resource:

                                        Steere AC, Malawista SE, Snydman DR, Shope RE, Andiman WA, Ross MR, et al. Lyme arthritis: an epidemic of oligoarticular arthritis in children and adults in three Connecticut communities. Arthritis Rheum 1977; 20:7–17.Find this resource:

                                        Swaminathan B, Barrett TJ, Hunter SB, Tauxe RV. PulseNet: the molecular subtyping network for foodborne bacterial disease surveillance, United States. Emerg Infect Dis 2001; 7:382–389.Find this resource:

                                        Tavernise S. Charges filed in peanut Salmonella case (February 22, 2013 p.B5). New York Times, 2013.Find this resource:

                                          US Department of Health and Human Services. Consumers advised of risks associated with raw sprouts. HHS News PS99-13. Washington, D.C.: U.S. Department of Health and Human Services, 1999.Find this resource:

                                            Wasley A, Samandari T, Bell BP. Incidence of hepatitis A in the United States in the era of vaccination. JAMA 2005; 294:194–201.Find this resource:

                                            Weinbaum CM, Bodnar UR, Schulte J, Atkinson B, Morgan MT, Caliper TE, et al. Pseudo-outbreak of tuberculosis infection due to improper skin-test reading. Clin Infect Dis 1998; 26:1235–1236.Find this resource:

                                            Weiss NS. Epidemiologic studies in which a necessary cause is known. Epidemiology 1991; 2:153–154.Find this resource:

                                            Wurtz R. The role of public health in health information exchanges. J Public Health Manag Pract 2013; 19:485–487.Find this resource:

                                            Exercises

                                            1. 1. In September, 1985, an outbreak of E. coli O157:H7 gastroenteritis struck 55 of 169 residents of a nursing home in southwestern Ontario, plus 18 of 137 staff (Carter et al., 1987). This microorganism can be transmitted by ingestion of contaminated food (often inadequately cooked ground beef); by person-to-person spread in such groups as families, child care centers, or custodial institutions; or by swimming in or drinking contaminated water (Heymann, 2008). The incubation period ranges from 3–8 days, with a median of 3–4 days.

                                              The epidemic curves for staff (top panel) and residents (bottom panel) are shown in Fig. 20.2. From the information given, what would you consider to be the most likely way(s) by which the staff and residents became infected during this outbreak and when the key exposure(s) occurred? Briefly explain your answer.

                                            2. 2. An article in the New England Journal of Medicine described an outbreak of norovirus gastroenteritis among members and staff of a North Carolina football team in September 1998 (Becker et al., 2000). This illness is characterized by vomiting and diarrhea. The incubation period is 10–50 hours. Infection occurs by consumption of contaminated food, and person-to-person transmission can also occur from close physical contact. Fig. 20.3 shows the occurrence of cases among North Carolina players and staff in each 12-hour time period over the course of several days.

                                              1. (a) Imagine that you are a field epidemiologist working on this outbreak. What period of time would you investigate most closely for a point exposure that could have initiated the epidemic?

                                              2. (b) Suggest two plausible possibilities for why the cases did not all occur within one incubation period.

                                            Figure 20.2 Epidemic curve from an outbreak of E. coli O157:H7 infection in a nursing home.

                                            Figure 20.2
                                            Epidemic curve from an outbreak of E. coli O157:H7 infection in a nursing home.

                                            (Based on data from Carter et al. [1987])

                                            Figure 20.3 Initial epidemic curve from an outbreak of norovirus among high school football players.

                                            Figure 20.3
                                            Initial epidemic curve from an outbreak of norovirus among high school football players.

                                            (Based on data from Becker et al. [2000])

                                            Answers

                                            1. 1. The most prominent feature is a large wave of cases among residents occurring from September 9–14, within one incubation period. This is most easily explained by a point source exposure on or about September 6. Nursing-home residents rarely go swimming in large groups, so this seems an unlikely form of exposure. Person-to-person transmission is also unlikely to explain the large initial wave, for lack of an apparent index case and because one would have to assume intimate contact between that person and a very large number of residents over a very short period of time. A contaminated meal would be very plausible, however, as it could account for exposure of many residents at essentially the same point in time. (In fact, a lunch on September 5 was strongly implicated.)

                                              Early cases among staff could have represented staff who ate some of the same contaminated food as residents. Later cases in staff and in residents may well have represented person-to-person transmission from earlier cases.

                                            2. 2.

                                              1. (a) The cases did not all occur within one incubation period. However, one would look especially closely at events on September 18, especially between noon and midnight that day. Given the incubation period of 10–50 hours, a point source exposure during this time period could potentially account for the 46 cases that occurred on September 19 and 20, as part of a single “wave.” Further investigation of this outbreak did indeed implicate a box lunch shared by team members on September 18.

                                              2. (b) One possibility is a source of continuing exposure, such as contaminated foods or beverages consumed by team members repeatedly over a period of days. Another possibility is that later cases were infected via close contact with cases in a large initial wave. In this particular outbreak, physical combat on the football field, and contact on the sidelines between healthy teammates and others who were actively sick, were thought to be major routes of exposure for the late-occurring cases. A more detailed epidemic curve is shown in Figure 20.4. Later cases, including some among members of the opposing Florida team, appeared likely to have been due to person-to-person transmission from those in the first wave who had eaten the tainted box lunch.

                                            Figure 20.4 Final epidemic curve from an outbreak of norovirus among high school football players.

                                            Figure 20.4
                                            Final epidemic curve from an outbreak of norovirus among high school football players.

                                            (Based on data from Becker et al. [2000])