Lesson 6: Investigating an Outbreak

Answers to Self-Assessment Quiz

1. C, D. Most outbreaks come to the attention of health authorities because an alert clinician or a concerned case-patient (or parent of a case-patient) calls. The other methods listed occasionally detect outbreaks, but less frequently.
2. A, B, C, D. Factors influencing a health department's decision to conduct a field investigation include some related to the health problem itself (e.g., severity of illness, number of cases, availability of prevention / control measures), some relate to the health department (e.g., "corporate culture" for conducting field investigations versus handling it by telephone, available staff and resources), and some relate to external concerns (e.g., public or political pressure).
3. B (False). The most important public health reason for investigating an outbreak is disease control and prevention. Protecting and promoting the public's health is our primary mission, even if it interferes with our ability to conduct research.
4. 1. A. As in Answer 3, our primary mission is to protect the public's health, so disease control and prevention measures should take priority whenever possible. Because disease prevention and control measures are often aimed at interrupting transmission, such measures can be implemented if the source and mode of transmission are known.
   2. B. If the agent is known but the source and mode of transmission are not known (example: Salmonella eventually traced to marijuana), then the health department does not know how to target its intervention. Investigation to learn the source and/or mode is necessary.
   3. B. As in Answer 4b, If the agent, source, and mode of transmission are not known (examples: Legionnaires' Disease in Philadelphia in 1976; Kawasaki Syndrome — if it turns out to be an infectious disease), then the health department does not know how to target its intervention. Investigation to learn the source and/or mode is necessary.
5. D. Early steps include confirming that the number of cases exceeds the expected number, verifying the diagnosis, and preparing for field work (which includes talking with laboratorians about specimen collection). Next steps include conducting surveillance to identify additional cases; analyzing the data by time, place, and person; generating hypotheses; and evaluating those hypotheses (for example, by conducting a case-control study).
6. B (False). The order presented in this text is conceptual. In practice, the order can be different. For example, preparing for field work often follows establishing the existence of an outbreak and verifying the diagnosis. When possible, control measures are initiated at the same time the field investigation begins, or even earlier.
7. 1. B. Epidemic, the occurrence of more cases of disease than expected in a given area or among a specific group of people over a particular period of time, tends to refer to more widespread occurrence than outbreak.
   2. C. Outbreak tends to be used for an increase that is localized.
   3. A. Cluster is an aggregation of cases in a given area over a particular period of time that seems unusual or suspicious, but often the usual or expected number of cases is not known.
8. B, C. Even an investigator without a clinical background should, if possible, see and talk to a patient or two to gain a better understanding of the clinical features of the disease (needed for developing a case definition) and to generate hypotheses by asking about possible exposures.
9. A, B, C, D. A case definition for an outbreak should specify clinical criteria as well as appropriate time, place, and person characteristics. The case definition should NOT include the hypothesized exposure of interest. First, the hypothesized exposure may not turn out to be the true exposure, so inclusion of the hypothesized exposure as part of the case definition during the case-finding step may result in missed cases. Second, during the analytic step, disease status and exposure must be determined independently to avoid bias. Including exposure as part of the case definition means that all cases will, by definition, be exposed, while only some of the controls will likely be exposed. As a result, the exposure will appear to be associated with disease, not necessarily because it is the true exposure, but because of the case definition.
10. A (True). A case definition is a decision making tool. It provides criteria for classifying illness as a "case" or "not a case. " However, few case definitions are 100% accurate, because people with mild or atypical or asymptomatic disease are likely to be missed, and people with similar but not the same disease may be included. Even a case definition that requires a laboratory test is not 100% perfect, because laboratory tests themselves are not perfect.
11. B (False). On the one hand, case definitions need to be applied consistently, so that everyone involved in an investigation defines a case in the same way. On the other hand, case definitions can change during the course of an outbreak. For example, for case finding purposes, a case definition might include categories such as confirmed, probable, and possible, to try to include as many cases as possible. Later on, in the analytic phase, the case definition may be restricted to the confirmed cases. As another example, a case definition may initially be restricted to a particular community. If the outbreak spreads beyond that geographic area, the "place" component of the case definition also would need to be expanded.
12. A, B, C, D. To identify additional cases as part of an outbreak investigation, health department staff contact (by telephone, broadcast fax, or e-mail) physicians' offices, clinics, hospitals, and laboratories. Depending on the affected age group, staff might also contact day care centers, schools, employers, or nursing homes. Sometimes a press release is issued to local media outlets that inform the public and suggest that persons with particular symptoms or exposures contact their healthcare providers or health department. In addition, health department staff routinely interview case-patients and ask whether they know any persons with the same exposure, if known, or with the same illness.
13. A, B, C, D, E. A data collection form for an outbreak investigation should include patient identifying information (e.g., name, telephone number), demographic information (e.g., age, sex), clinical information (e.g., date of onset, laboratory confirmation, whether hospitalized), risk factor information (disease-specific, e.g., attended sports banquet (yes/no), previously vaccinated?), and information about who collected the data (e.g., interviewer or abstracter initials, date of collection).
14. B (False). Descriptive epidemiology is essential not only for characterizing the pattern and distribution of the outbreak, but also for generating testable hypotheses about the source, mode of transmission, and risk factors for illness. Two of the suggested ways for generating hypotheses are to review the descriptive epidemiology, particularly (1) the overall pattern of cases and develop hypotheses about what they have in common, and (2) the outliers to determine how they might be linked to the other cases. These hypotheses, in turn, are the ones that are tested using analytic epidemiology.
15. 1. C. This scenario represents an intermittent exposure. The resulting epidemic curve has cases that appear to be occurring sporadically, but in fact occur when the malfunctioning heater is turned on at irregular intervals.
    2. A. This scenario represents a point source exposure. The epidemic curve has a single peak, and all cases occur during a single incubation period.
    3. D. This scenario represents person-to-person transmission. The epidemic curve has a succession of "waves" of cases.
16. D. Subtracting 24 hours (the minimum incubation period) from the time of onset of the first case puts you in the April 20 Dinner interval. Subtracting 33 hours from the median case (which occurred in the 4-8 AM interval) on April 22), puts you in the April 20 4-7 pm interval, near both lunch and dinner that day. While the minimum method points to dinner on April 20, thorough investigators would probably investigate possible exposures at lunch that day, too.
17. A, B, C, D, E. A late case on an epidemic curve has several possible explanations, including a case of a similar but unrelated disease, a secondary case (assuming it occurs one incubation period after another case), a case with an unusually long incubation period, a case that resulted from exposure at a different time (for example, someone who ate leftovers the next day), or an error in recall or in recording the date.
18. A, B. A spot map is useful for pinpointing the geographic location of exposures, residences, employment sites, and the like. The spots represent occurrences, either of exposure or disease. Spot maps are not used to display rates. Rather, area maps (also called shaded or chloropleth maps) are used to display incidence and prevalence rates.
19. A, B, C, D, E, F. Hypotheses can be generated in a variety of ways. One way is based on subject matter knowledge derived by reviewing the literature or talking with experts – what are the usual causes, sources, vehicles, or modes of transmission? Other ways include reviewing the overall pattern and the outliers from the descriptive epidemiology, by asking case-patients if they have any suspicions about the cause of their illness, and by asking the same question of local authorities (if you are from out of town).
20. D. The key feature that characterizes an analytic (epidemiologic) study is presence of a comparison group. Single case reports and case series do not have comparison groups and are not analytic studies. Cohort studies (compares disease experience among exposed and unexposed groups) and case-control studies (compares exposure experience among persons with and without disease) have comparison groups and are analytic studies.
21. A, B, C, D, E. Disease control measures can be directed at the eliminating the agent (e.g., by sterilizing surgical equipment), interrupting transmission (e.g., reducing mosquito population, covering one's mouth when coughing), preventing entry into a host (e.g., wearing a mask, using insect repellant), or improving host defenses (e.g., by immunization).
22. A. A retrospective cohort study is one in which disease has already occurred (hence, retrospective) and the investigator enrolls all (or almost all) of a population (hence, cohort). The investigator then determines exposures and calculates risks (attack rates) for different exposures and risk ratios (relative risks) for those exposed and unexposed. The study described for Questions 22-25 meets this characterization.
23. D. The measure of association recommended for a retrospective cohort study is a risk ratio, calculated as the ratio of the risk of disease among those exposed divided by the risk of disease among those not exposed. The attributable risk percent is a supplemental measure that quantifies how much of the disease could be "explained" or accounted for by a particular exposure. The chi-square is not a measure of association, but a test of statistical significance (which is affected both by the strength of association and number of subjects in the study). The odds ratio is used primarily as a measure of association in case-control studies.
24. B. The wedding cake (risk ratio = 45% ⁄ 5% = 9.0) is the most likely culprit. It has a high attack rate among the exposed group, a low attack rate among the unexposed group, and can account for 45 out of the 50 cases. The five "unaccounted for" cases are within the range that can be "explained away," for example by misreporting (for example, a man takes a bite of his partner's cake but reports "no" for cake because he didn't take a whole piece himself), poor recall, etc. Punch is not associated with illness at all (risk ratio = 25% ⁄ 25% = 1.0). Sushi has an extremely high attack rate among those exposed (91%), but a relatively high attack rate among those unexposed (21%), and most importantly, could only account for 10 of the 50 cases.
25. A, B, C. The results should be communicated to all those who need or want to know, including the concerned family and wedding attendees, local governmental officials, the caterer, the church or facility where the wedding was held, et al. The outbreak is also reportable to the state health department, who in turn is likely to report it to CDC. However, local outbreaks do not need to be reported to the World Health Organization.