Epiville

SARS Outbreak Study 1

Data Analysis

Back at the DOH:

You suspect that what you have on your hands is an outbreak of SARS. Before going any further with your investigation you need to get guidance from your supervisor. Dr. Zapp compliments you on the job well done and outlines the following steps of the outbreak investigation to aid your work:

  1. Define the epidemic
    1. Define the "numerator" (cases)
    2. Define the "denominator" (population at risk for developing disease)
    3. Calculate attack rates
  2. Examine the distribution of cases by person, place, and time (PPT)
  3. Look for combinations (interactions) of relevant variables
  4. Develop hypotheses based on the following:
    1. Existing knowledge (if any) of the disease
    2. Analogy to diseases of known etiology
  5. Test hypotheses
    1. Further analyze existing data utilizing an appropriate study design
    2. Collect additional data
  6. Recommend control measures

You return to your modest cubicle at the DOH and eagerly embark upon the first steps of an outbreak investigation: descriptive analysis.

The epidemic should be characterized using the criteria in the case definition.

NOTE: Despite your hard work, no case definitions are perfect because of possible case misclassification. For instance, a diagnostic test may show that a person has the disease when in fact they do not (false positives), or conversely, that a person is free of disease when in fact he or she is truly infected (false negatives). It is very important to have reliable and valid diagnostic tools to decrease the number of false positives and false negatives. The costs of misdiagnosing a case and delaying the appropriate treatment, especially in the case of such a deadly infection as SARS, can be significant. (For more information on this topic, please see Aschengrau & Seage pp. 418-422)

1. Based on what you just have learned, how would you go about preparing a case definition for this outbreak in Epivillle?

  1. Compile a table of symptoms from all suspected SARS cases and from patients with other diagnosed respiratory illnesses, then provide information in terms of PPT.
  2. Compile a table of symptoms of all patients who were admitted to the hospital during the same time that the first patients were admitted to the Epiville General Hospital.
  3. Compile a table of symptoms of all patients who were admitted to the Epiville General Hospital over the last year and who lived in the Amoy Apartment Complex.
Answer (a) — correct: This answer meets both the clinical criteria of suspected SARS and PPT restriction of the case definition (i.e., this answer fulfills the first step of an outbreak investigation, defining a case, as well as the second step of an outbreak investigation, defining the epidemic in terms of PPT).
Answer (b) — incorrect: This answer does not meet the clinical criteria of SARS. It does, however, meet the time requirement of PPT (i.e., this answer fulfills the first step of an outbreak investigation, defining a case, but not the second step of an outbreak investigation, defining the epidemic in terms of PPT.
Answer (c) — incorrect: Here you are ignoring the clinical criteria of the case definition in an attempt to capture the place element of PPT (i.e., everyone who lived in the Amoy Apartment Complex). Thus, this answer fulfills the second step of an outbreak investigation, investigating PPT, but not the first step of an outbreak investigation, defining a case.

With your case definition at hand, you go back to the Epiville General Hospital to review the charts of the 70 persons who reside at the Amoy Apartment Complex. Based on your review of the charts, you confirm that 53 out of the 70 suspected cases in the apartment complex are most likely to be SARS cases. Out of the 17 patients who did not meet the case definition, 5 had similar symptoms as SARS cases but their chest X-rays were clear, others (6) recovered within a week (this was an indication that they just had flu), and the rest (6) had pneumonia of known etiology. All of these 53 individuals are alive and battling the disease. The case definition you came up with is remarkably similar to the case definition of SARS developed by the World Health Organization (WHO) in 2003.

You remember from your reading at the beginning of the day that surveillance is an essential feature of epidemiologic practice and is often divided into two "types": passive and active. Most of the surveillance done on a routine basis is called passive surveillance. In passive surveillance, physicians, laboratories, and hospitals are required to report diseases from the list of reportable diseases (usually a list of 70-80 diseases) and are given the appropriate mailing forms and instructions, with the expectation that they will report all of the cases of reportable disease that come to their attention. Active surveillance requires periodic telephone calls or personal visits to the reporting individuals/hospitals/laboratories to obtain required data.

2. Would you categorize each of the following activities as passive or active surveillance?

  1. Hospitals reporting SARS cases to the Department of Health
  2. Your visit to the Amoy Apartment Complex
  3. Your review of case records
Answer (a) — none: Passive surveillance
Answer (b) — none: Active surveillance
Answer (c) — none: Active surveillance

3. What type of surveillance do you think is more difficult to carry out and why?

  1. Passive surveillance
  2. Active surveillance
Answer (a) — incorrect: Although it may seem that passive surveillance is more difficult to carry out because you need to get physicians, hospitals, and laboratories to comply with reporting of disease on a routine basis, it is relatively simple to set up a system where health care workers fill out the necessary reports and submit them to the DOH.
Answer (b) — correct: Active surveillance is more difficult to carry out. Active surveillance is more labor-intensive and costly because the DOH has to hire additional staff who will have to make telephone calls and/or visit the medical facilities.

The next morning, you are eager to show your case definition and report on your visit to the Amoy Apartment Complex to your supervisor. Dr. Morissa Zapp compliments you on your good work and recommends that you contact two other hospitals in the city to see if they have similar cases that were not reported through passive surveillance. Your advisor's suspicions are confirmed. There have been 3 suspicious deaths at the Star hospital. You set out to do your detective work immediately.

The hospital administrator tells you that they had a 70 year old man admitted two weeks ago with respiratory symptoms, and that a doctor and nurse who were taking care of him got sick within a week of his admission. When the administration realized how contagious the disease was, the patient, doctor and nurse were put into isolation. Unfortunately, by the time extraordinary precaution measures were instituted, 20 additional staff members had become infected, including 5 medical residents. All of them had visited the patient's room while attending the weekly rounds and may have had close physical contact with the patient. The hospital administration notified the DOH by a phone call about the situation, but failed to follow up with proper documents.

Employing your understanding of PPT, you decide to review the chart of the elderly man who was admitted to the hospital. His chart reveals that he lives in the Amoy Apartment Complex. Several days before he got ill, on August 1, 2003, he had attended an annual luau party held in the garden of the Amoy Apartment Complex. Approximately 300 other tenants also attended.

Now that you have gathered information on all the suspected SARS cases, you are ready to create a spot map (For more information on mapping, please see Giesecke pp. 23-24).

4. Given the place of residence of each "SARS affected-case" map the distribution of cases.
[ Open map ]
Do you see any clustering?

  1. Clustering around the hospital
  2. Clustering around the Amoy Apartment Complex
  3. No clustering
Answer (a) — incorrect: Although the second outbreak occurred in the Star Hospital, the affected personnel live all around the city (aside from the five medical residents who live in the housing near the hospital). If you created a spot map by place of work, you would have observed clustering around the hospital.
Answer (b) — correct: You observe clustering around the Amoy Apartment Complex because most of the cases have the same place of residence. Your map confirms a geographic clustering of cases around the apartment complex. You add place of residence as a necessary component of your case definition.
Answer (c) — incorrect: You do observe a clustering of cases around the Amoy Apartment Complex.

Now that you have mapped out all the cases according to the place of residence, you are ready to do some data analysis. The best measure to use in describing an epidemic is cumulative incidence. Cumulative incidence is calculated as the number of new cases over a defined study period, divided by population at risk over that period. Cumulative incidence is usually expressed per 100, per 1,000, per 10,000, or per 100,000 (For more information, please see Aschengrau & Seage pp. 42-44).

Let's calculate the cumulative incidence for 5 residential areas of Epiville

AreaaPopulation SizeSARS Cases From the Outbreak at the Amoy Apartment ComplexSARS Cases From the Outbreak at the Star HospitalcTotal SARS Cases Reported from 08/03 to 08/24
A14,000033
B53,00066b470
C13,000055
D12,000088
E8,000022

a The Amoy Apartment Complex is located in Area B, the Star Hospital is located in Area C, the staff of 110 persons who works at the Star Hospital resides all over Epiville and Epiville suburbs.

b 65 cases were hospitalized at the Epiville General Hospital and 1 case was hospitalized at the Star Hospital.

c 22 cases from the Star hospital include only staff members. Hospital staff reside in different areas of Epiville.

5. In the table above, Epiville is divided into five residential areas: A, B, C, D, and E. Calculate the cumulative incidence of SARS (per 1,000) in each area. Cumulative incidence is calculated as follows: [# of new cases of disease / # in candidate population] over a specified period of time (Please see Aschengrau pp. 42-44).

  1. What is the cumulative incidence of SARS (per 1,000) in Area A?
  2. What is the cumulative incidence of SARS (per 1,000) in Area B?
  3. What is the cumulative incidence of SARS (per 1,000) in Area C?
  4. What is the cumulative incidence of SARS (per 1,000) in Area D?
  5. What is the cumulative incidence of SARS (per 1,000) in Area E?
Answer (a) — none: (3/14,000)*1,000=0.21 per 1,000 over a three-week period.+

[+Please note that a specified period of time is a critical element in calculating incidence.]
Answer (b) — none: (70/53,000)*1,000=1.32 per 1,000 over a three-week period.
Answer (c) — none: (5/13,000)*1,000=0.38 per 1,000 over a three-week period.
Answer (d) — none: (8/12,000)*1,000=0.67 per 1,000 over a three-week period.
Answer (e) — none: (2/8,000)*1,000=0.25 per 1,000 over a three-week period.

6. In which area do you observe the highest cumulative incidence of SARS? Why?

  1. Area D because the largest number of Star hospital employees resides there
  2. Area B because it has the largest population
  3. Area B because this is where the Amoy Apartment Complex is located and where the largest number of new cases live.
Answer (a) — incorrect: While it is true that the greatest number of Star personnel reside in Area D, Area D does not have the greatest overall number of SARS cases. Proportionally, Area D has a lower cumulative incidence of SARS than one other area in Epiville.
Answer (b) — incorrect: Area B has the largest population and therefore, the largest population at risk for developing SARS. Cumulative incidence is, however, defined as the number of new cases divided by the number of persons at risk during a specified time period. Therefore, knowing the number of persons at risk is not enough to conclude that the cumulative incidence is highest in area B.
Answer (c) — correct: Cumulative incidence is a function of both the number of new cases and the number of people at risk. You will get the greatest cumulative incidence when your numerator (number of new cases) is large and/or when the denominator (number of people at risk) is small. Since the greatest number of cases happens to have occurred in the Amoy Apartment complex that is located in Area B, the cumulative incidence is highest in this area as well (1.32 per 1,000 per three-week period).

Disease transmission refers to any mechanism by which an infectious agent is spread from a source or reservoir to another person. The mechanisms, or modes of transmission, are divided into two categories: direct and indirect. For instance, a disease can be transmitted from person to person by means of direct contact (e.g., influenza). Indirect transmission usually occurs through a common vehicle (e.g., contaminated air or a contaminated water supply) or through a vector such as a mosquito]. (For more information, please see Giesecke pp. 16-17 and the Epiville glossary).

7. Based on the cases from the Amoy Apartment Complex, which mode of transmission is most probable?

  1. Indirect: food or water.
  2. Direct: Person-to-Person
  3. Indirect: vector-borne (cockroaches)
Answer (a) — incorrect: Symptoms of the disease are not indicative of food or water poisoning, or of intestinal disease. Additionally, if the disease were transmitted via food or water, we would anticipate that the majority of the Amoy Apartment Complex residents would have fallen ill subsequent to the luau.
Answer (b) — correct: All cases occurred close in time-- shortly after each of the cases attended the luau where they were in close contact with one another for several hours.
Answer (c) — incorrect: Only 66 of the 600 residents at the Amoy Apartment Complex fell ill with SARS. It is unlikely that these 66 persons all came in contact with cockroaches within such a short time period unless the apartment complex was really infested, which was not the case.

8. Based on the cases from the Star Hospital, which mode of transmission is most probable?

  1. Direct: Only through bodily fluids
  2. Indirect: Air-borne through the hospital ventalation system
  3. Direct: Person-to-Person
Answer (a) — incorrect: While it is possible that those who took care of the index patient at the Star Hospital got infected by touching the patient, it fails to explain why 20 additional medical personnel got sick.
Answer (b) — incorrect: The index case was hospitalized in the regular inpatient ward. If the disease were air-borne, it would have quickly spread around the hospital through the building's common airshafts.
Answer (c) — correct: All cases occurred close in time--shortly after admission of the index case.This suggests that the source of infection was the same for all individuals at Star hospital. Further, the statements given by the Star Hospital cases suggests that they were in close physical contact with the initial case while visiting the index case. This suggests that the transmission occurred via direct contact with contaminated water droplets emanating from the sick case's sneezing or coughing.