Types of Outbreaks

Epidemics (outbreaks) can be classified according to the method of spread or propagation, nature and length of exposure to the infectious agent, and duration (Goodman and Peavy 1996; Kelsey et al. 1996). There are a few main categories.

Common Source Epidemic. Disease occurs as a result of exposure of a group of susceptible persons to a common source of a pathogen, often at the same time or within a brief time period. When the exposure is simultaneous, the resulting cases develop within one incubation period (a point source epidemic) (Last 1995). The epidemic curve in a common source outbreak will commonly show tight temporal clustering (Figure 3-2). The data shown from

*A probable case was defined as onset of diarrhea (two or more loose stools during a 24-hour period) with either fever or bloody stools while at the resort or within 11 days of leaving the resort. A confirmed case additionally required Shigella sonnei isolated from stool. A total of 82 cases were identified, including 67 probable and 15 confirmed.

Figure 3-2. Epidemic curve showing the number of cases of confirmed and probable cases of Shigella sormei,* by date of onset—Idaho, August 6-24, 1995; Source: CDC (1996)

*A probable case was defined as onset of diarrhea (two or more loose stools during a 24-hour period) with either fever or bloody stools while at the resort or within 11 days of leaving the resort. A confirmed case additionally required Shigella sonnei isolated from stool. A total of 82 cases were identified, including 67 probable and 15 confirmed.

Figure 3-2. Epidemic curve showing the number of cases of confirmed and probable cases of Shigella sormei,* by date of onset—Idaho, August 6-24, 1995; Source: CDC (1996)

the outbreak of Shigella sortnei also show the effect of a case definition based on laboratory confirmation versus that based on signs and symptoms. When a common source is present, a sharp upslope and a more gentle downslope are common (Figure 3-3). A waterborne outbreak that is spread through a contaminated community water supply is an example of a common source epidemic.

Propagative Epidemics. The infectious agent is transferred from one host to another. Generally, this involves multiplication and excretion of an infectious agent in the host and sometimes intermediate animal-human or arthropod-human multiplication cycles (Kelsey et al. 1996). Propagative spread usually results in an epidemic curve with a relatively gently upslope and somewhat steeper tail, sometimes including a second but less prominent group of cases later in time (Goodman and Peavy). An example of a propagative epidemic is an outbreak of hepatitis B virus due to intravenous drug use.

Mixed Epidemics. In this category, the epidemic begins with a single, common source of an infectious agent with subsequent propagative spread. Many foodborne pathogens result in mixed epidemics.

August

August

Date of Symptom Onset

Figure 3-3. Epidemic curve showing the number of cases of Shigella dysenteriae type 1 detected at Suan Phung Hospital, Thailand, July-August, 1992; Source: Höge et al. (1995)

Steps in the Outbreak Investigation Process in the Community Setting

Several public health epidemiologists (Goodman et al. 1990a; Tyler and Last 1992; Gregg 1996a) have described a series of steps that should be considered when conducting an outbreak investigation (Table 3-2). While there is considerable overlap between steps and certain aspects may assume more importance depending on the nature of the individual outbreak, the following framework provides a useful guide. A comprehensive discussion of outbreak investigation methods is provided in Gregg et al. (1996).

Determine the Existence of an Epidemic. Ongoing disease surveillance conducted by state and local health agencies can be used to determine whether an excess of disease is occurring (Thacker 1996). Observed numbers can be compared with expected frequencies based on past trends by week, month, or year. This initial determination is often made on the basis of preliminary data, lacking laboratory confirmation of cases.

Confirm the Diagnosis. Clinical diagnosis can be obtained through standard laboratory techniques such as serology and/or isolation and characterization of the agent (Gregg 1996a). Information on the signs and symptoms among cases must be obtained early in an outbreak investigation (Tyler and Last 1992). It is frequently impractical to confirm every case with laboratory evidence; if every reported case has signs and symptoms consistent with laboratory-confirmed cases it is necessary to obtain laboratory verification on approximately 15-20% of cases.

Define a Case and Estimate the Number of Cases. Usually on the basis of preliminary information, the epidemiologist must develop a case definition.

Table 3-2. Ten Key Steps in an Outbreak Investigation

1. Determine the existence of an epidemic

2. Confirm the diagnosis

3. Define a case and estimate the number of cases

4. Orient the data in terms of time, place, and person

5. Determine who is at risk of having the health problem

6. Develop an explanatory hypothesis

7. Compare the hypothesis with the established facts

8. Plan and execute a more systematic study

9. Prepare a written report

10. Execute control and prevention measures

Sources: Goodman et al. (1990a); Tyler and Last (1992); Goodman and Buechler (1996).

In general, the simplest and most objective criteria for case definition are the best (e.g., fever, blood in the stool, elevated white blood cell count) (Gregg 1996a). Often, a preliminary case definition is adequate as an outbreak is unfolding and this definition can later be refined. After a usable case definition has been established, the field epidemiologist should count the cases, collect data on the cases, and determine common features (Tyler and Last 1992). The case definition must be applied without bias to all persons under investigation.

Orient the Data in Terms of Time, Place, and Person. Each case must be defined according to standard epidemiologic parameters: the date of onset of the illness, the place where the person lives or became ill, and the socio-demographic characteristics (e.g., age, sex, education level, occupation). Graphic depiction of cases will aid in showing the relationship between case frequency and their time of occurrence (known as the "epidemic curve") (Figure 3-2). Graphical representation typically shows the number of cases (y axis) over the appropriate time interval of the date of onset (x axis). In addition, a spot map can be useful in determining the spatial relation of cases (Figure 3-4). Although initial description of cases may superficially seem simple, it can lead to complex and difficult issues. For example, if one is investigating an outbreak in a multiracial population, public health surveillance information may be ambiguous in the classification of race and ethnicity due to lack of consensus in measuring these variables (Hahn and Stroup 1994).

Determine Who is at Risk of Having the Health Problem. Preliminary analysis of the data to this point often provides enough information to determine with reasonable certainty how and why the outbreak started (Gregg 1996a). However, in outbreaks that cover large geographic areas, it may be extremely difficult to determine who is at risk. A preliminary survey may be necessary to obtain more specific information about the group of ill persons.

Develop an Explanatory Hypothesis. This step, the first real foray into analytic epidemiology, involves the assessment of the data collected to date and the generation of hypotheses that may explain the outbreak. The goal is to explain the specific exposure(s) that caused the outbreak. This step is often the most difficult to perform. In nearly every outbreak investigation, a priori hypotheses are voiced by public health professionals, affected persons, the media, employers, or others (Palmer 1989). Careful examination of descriptive epidemiologic data can exclude many possible hypotheses and will often stimulate alternative hypotheses.

Syphilis

Gonorrhea

Figure 3-4. Geographical clustering of syphilis and gonorrhea, by zip code group, Dade County, Florida, 1986-1990; Source: Hamers et al. (1995)

Syphilis

Gonorrhea

Figure 3-4. Geographical clustering of syphilis and gonorrhea, by zip code group, Dade County, Florida, 1986-1990; Source: Hamers et al. (1995)

Compare the Hypothesis with the Established Facts. At this point, the goal is to compare the hypothesis with the clinical, laboratory, and other epidemiologic data. The key issue is whether the data collected thus far fit with what is known about the disease (i.e., its method of spread, incubation period, duration of illness, population affected).

Plan and Execute a More Systematic Study. After completion of the initial outbreak investigation, a more systematic study may be necessary, depending on whether there is a need to find more cases, better define the extent of the outbreak, or evaluate a new laboratory or case-finding method. The epidemiologic approaches most commonly used in this phase are case-control and retrospective cohort studies. Both of these designs are discussed in the later section on "Analytic Study Designs" and in Chapter 2. In planning a detailed epidemiologic study of an outbreak it is important to carefully address ethical issues related to informed consent, protections of confidentiality, and reporting of results (see Chapter 1 and Coughlin and Beauchamp 1996).

Prepare a Written Report. The development of a written summary of an outbreak is an essential part of the investigation, just as a hospital discharge summary is key for patient care (Tyler and Last 1992). Of utmost importance, the written report serves as a basis for public health action—namely, implementing prevention and control measures outlined in the next section. The process of writing the report and the report itself can provide (1) information to the public about the outbreak, (2) new insights into the characteristics of the outbreak including the agent and its spread, (3) a record of performance of the outbreak investigation team, (4) documentation of potential legal issues, and (5) a means for teaching epidemiology (Gregg 1996a).

Execute Control and Prevention Measures. The underlying purpose of most outbreak investigations is to develop and implement appropriate measures of prevention and control. Methods to eliminate the current outbreak and to prevent future similar outbreaks may involve one or more of the following: (1) eliminate the source of the pathogen or the exposure of susceptibles to it; (2) interrupt the spread from the source to the susceptibles, and (3) protect the susceptibles from the consequences of exposure even when the source or method of transmission cannot be controlled (Kelsey et al. 1996) (Table 3-3). After a control measure has been implemented, ongoing evaluation is necessary to determine effectiveness.

Several examples illustrate the interplay between etiology, mode of transmission, and disease control measures (Figure 3-5) (Goodman et al. 1990a). A single case of hepatitis A in a daycare setting can lead to prophylactic administration of immune globulin to an entire group of children and staff (CDC 1985). A multistate outbreak of Salmonella muenchen required extensive outbreak investigation, including a case-control study. This investigation found

Table 3-3. Elements of Outbreak Control

Action Example

Control the source of the pathogen Remove the source of contamination

Remove persons from exposure Inactivate or neutralize the pathogen Isolate and/or treat the infected person(s) Sterilize or interrupt environmental sources of spread (e.g., water, milk, air) Control mosquito or insect transmission Improve personal sanitation (e.g., hand washing)

Immunize the susceptibles Use prophylactic chemotherapy

Interrupt the transmission

Control or modify the host response to exposure

Source/Transmission Mode known unknown

Source/Transmission Mode known unknown

Investigation: +

Example: Hepatitis A in day care (CDC 1985)

Example: Salmonellosis and marijuana (Taylor et al. 1982)

Investigation: +++

Example: Parathion poisoning (Etzel et al. 1987)

Example: Legionnaires' disease (Fraser et al. 1977)

Figure 3-5. Relative emphasis of investigative and control efforts (response options) in disease outbreaks as influenced by levels of certainty about etiology and source/ mode of transmission. "Investigation" means extent of the investigation; "control" means the basis for rapid implementation of control measures. Pluses show the level of response indicated: +, low; + +, intermediate; + + +, high. Source: Goodman et al. (1990a)

Figure 3-5. Relative emphasis of investigative and control efforts (response options) in disease outbreaks as influenced by levels of certainty about etiology and source/ mode of transmission. "Investigation" means extent of the investigation; "control" means the basis for rapid implementation of control measures. Pluses show the level of response indicated: +, low; + +, intermediate; + + +, high. Source: Goodman et al. (1990a)

the mode of transmission to be personal use of or household exposure to marijuana (Taylor et al. 1982). In an outbreak of acute organophosphate poisoning, contaminated bread was shown as the source of exposure, allowing public health officials to make preventive recommendations before the etiologic agent (i.e., parathion) was isolated in the laboratory (Etzel et al. 1987). In the Legionnaires' disease outbreak of 1976, extensive field investigation failed to identify either the etiology or source of the outbreak in sufficient time to control the acute problem (Fräser et al. 1977).

Analytic Study Designs

The case-control study and the retrospective cohort study are the two most common designs used in outbreak investigations (Dwyer et al. 1994; Buehler and Dicker 1996). Each of these designs has advantages and disadvantages. Rather than a detailed review of each study design, this chapter will present a brief discussion of the use of each study design in outbreak investigations. Detailed descriptions are available elsewhere (Schlesselman 1982; Hennekens and Buring 1987; Lilienfeld and Stolley 1994; Kelsey et al. 1996).

Case-Control Studies. Over the past few decades, the case-control method has grown in prominence as an important tool for outbreak investigation. A comparison of the use of case-control studies within 6-year periods of the 1960s, 1970s, and 1980s showed a clear increase in frequency of use (Fonseca and Armenian 1991). The proportion of published outbreak investigations that used case-control methods increased from 0.2% in the 1960 period to 5.7% in the 1970 period to 17.0% in the 1980 period. Methodologic improvements were increasingly noted over time including more frequent reporting of control selection, exposures, possible biases, and control of confounders.

Three specific circumstances have been cited where case-control studies are the most appropriate method of formally testing hypotheses in outbreak investigations (Dwyer et al. 1994). These situations take into account the advantages and disadvantages of each study design. First, in a large outbreak where it would be necessary to enumerate (i.e., obtain a complete list of individuals) an entire cohort to conduct a retrospective cohort study, a case-control sample is much faster and less costly because full enumeration is not necessary. Second, the complete cohort at risk cannot be adequately defined or a discrete cohort cannot be fully enumerated. Third, case-control studies can be conducted within cohort studies ("nested case-control studies") to test specific hypotheses that would be too expensive or infeasible to study in an entire cohort.

Retrospective Cohort Studies. The retrospective cohort study is a commonly used design in outbreak investigations. In a retrospective cohort study, the outcomes (i.e., a disease or health condition) have all occurred prior to the start of the investigation and the experience of cohorts is reconstructed through existing records. The general characteristics of a retrospective cohort design are discussed in Chapter 2. They often measure the association between exposure and disease by means of the risk ratio or relative risk (i.e., the attack rate in the exposed divided by the attack rate in the unexposed).

Advantages and Disadvantages. The advantages of the case-control method tend to highlight the disadvantages of the cohort method. The main advantages of the case-control method (and disadvantages of the cohort design) are that results can be obtained relatively quickly, at relatively low costs; multiple exposures can be studied; and enumeration of the whole cohort is not required when a representative sample of cases can be obtained. The major disadvantages of case-control studies (and advantages of the cohort design)

include the inability to directly measure the attack rate, higher potential for bias in retrospectively measuring risk factors, the inability to evaluate multiple disease outcomes, and the potential for bias in control group selection.

0 0

Post a comment