Outbreak Investigation in Bogalusa, Louisiana

1
Outbreak Investigation in Bogalusa, Louisiana

• Person, place, time, outliers?
• Causationcriteria from Hill
• Who?
• What?
• When?
• Where?
• How?
• Why?
• What else?

2
Background

• On 31 OCT 1989, 2 physicians in Bogalusa, LA
  reported more than 50 cases of acute pneumonia in
  a 3 week interval between mid-late October.
• All were adults, 6 cases had been fatal
• The doctors were suspicious of legionellosis

3
Investigation

• You are the doctor assigned to investigate
• What additional information do you need to decide
  whether or not this is a real public health
  problem? Base your answers on the scientific
  method used to investigate an outbreak
• Who, what, when, where, how, why, what else?

4
Additional Information

• Antibody studies from the initial acute patients
  are non-conclusive for Legionella
• However, this is not uncommon within the 1st few
  weeks of most illnesses
• No sputum cultures were available as the small
  hospital did not have the capability of analyzing
  them

5
(No Transcript)
6
What else?

• What elseother than a true outbreak of the
  disease could account for this sudden increase in
  incidence of pneumonia being reported to the
  health department?
• Before you reach the site, what kind of
  preparations do you need to make for your trip to
  Louisiana?

7
Background on Legionella pneumophila

• Legionnaires Disease

8
Background on Bogalusa

• Washington Parish
• Pop 16,000
• Largest employer? Paper Mill
• Near center of town adjacent to main street has 5
  cooling towers to cool the paper machine which
  gets very hot
• Large volumes of aerosol blow across the street
• Locals suspect this mist as a cause for the
  sickness
• There are a few other public buildings that have
  cooling towers

9
Bogalusa Health Care Infrastructure

• 98 Bed Hospital (Private)
• 60 Bed Public Hospital
• 3 more in near-by communities
• Note All patients with the disease were in
  Hospital Athe private hospital

10
Table 1. Number of patients dx with pneumonia
discharged from hospital A by month JAN 86-OCT 89
11
Review of Charts

• Review of charts for pneumonia patients in
  hospital A during October revealed that many
  patients were admitted with a febrile illness
  associated with weakness, lethargy, and mental
  confusion. Some had a dry cough, several had
  watery diarrhea.
• Chest X-rays were consistent with pneumonia

12
Table

• What are possible interpretations for data in the
  table on number of cases?
• If you wanted to intensify this investigation
  what steps would you take next?

13
Defining a case

• Would you want a relatively sensitive or
  relatively specific definition of a case for this
  type of investigation?

14
(No Transcript)
15
Develop the case definition
16
Tips

• Sensitive casebroad enough to identify nearly
  all true cases but may have a few false cases
• Specific casenarrow inclusion criteria such that
  we will exclude the false but could miss some
  true as well if they had irregular
  presentations of the disease

17
Case Continued

• What would you call a case?
• Do you need to identify every case

18
(No Transcript)
19
Hypothesis

• How might you develop a hypothesis to test in
  this investigation?
• What, if any, are your ideas at this point?

20
(No Transcript)
21
At this point in the investigation, the leading
hypothesis was outdoor exposure to cooling
towersprimarily because previous studies had
demonstrated the role of cooling towers as a
source of Legionella pneumophila in other
outbreaks, and there were several such towers in
the town. Rather than jumping to conclusions,
however, investigators began to compile a list of
retail stores and other commercial establishments
that were frequently mentioned by some of the
case-patients who had been interviewed.
Investigators also noted the unusual
preponderance of women who had become ill.
Would you use a case-control study, a cohort
study, or some other method to test the
hypotheses in this outbreak? Why?
22
The investigators conducted a case-control study.
They found that 66 of the patients who were still
living met the possible case definition.
Laboratory results confirmed that 15 of these
patients had Legionnaires' disease and ruled out
the disease for 10. The results for the remaining
41 were pending. What case definition would
you use for your case-control study? What are
some possible sources of controls?
23
Investigators decided to select controls from
office records of physicians who had admitted the
patients with Legionnaires disease to the
hospital. They enrolled two controls per case,
for a total of 28 case-patients and 56 controls.
Case-patients and controls were asked about
exposures to cooling towers and nearby buildings.
Among their findings, they learned that 3 of
the 28 case-patients and 7 of the 56 controls
reported visiting Hospital B, and 7 of the
case-patients (1 was unsure) and 12 of the
controls (6 were unsure) reported visiting the
Post Office. Calculate the odds ratios for
illness with Legionnaires disease among people
who visited Hospital B and those who visited the
Post Office.
24
ORs
2X2 of Hospital B Ca/Co
10
74
2X2 of Post-office Ca/Co
28 56
84
19
58
28-1 unk 56-6 unk
84
25
ORs

• Hospital B
• (349)/(725) 0.8
• Post Office
• (738)/(2012)1.1

26
Why might the numbers of cases and controls in
these two odds ratio calculations differ? How
would you interpret the results?
27
Answers..
The unknowns cause variation. It decreased the
sample sizes due to recall bias However, we
see visits to Hospital B were protective
(OR0.8) Visiting the PO only carried a slight
increase in risk associated with Legionnaires
Disease (OR1.1)
28
Take a Look at Exposures and ORs

• How would you interpret these data?
• In other words, which exposures suggest as
  association with illness?
• Which one accounts for the greatest number of
  cases, and what are the implications?

29
Additional epidemiologic analysis showed a
dose-response relationship between time spent in
Grocery Store A and risk of diseasethat is,
there was a direct relation between the amount of
time people spent in the store and their risk for
disease. The investigators visited Grocery Store
A and looked for sources of water in aerosol
form. An ultrasonic mist machine was operating
over one section of the produce display. No one
at Grocery Store A was familiar with the
maintenance or operation of this machine. With
permission from the store manager, investigators
cultured a specimen of water from the mister's
reservoir. The culture contained Legionella
pneumophila serotype 1 (LP-1). Cultures from
various cooling towers around town also contained
LP-1, but of different subtypes. The
investigators suspected that the misting device
may have been related to the outbreak. At
this point, do you have enough information to
make recommendationsin other words, have the
basic criteria of causation been satisfied? How
would you proceed in this investigation?
30
Before making any recommendations, the disease
detectives took four additional steps They
conducted a serosurvey on all of the grocery
store employees in Bogalusa. (In a serosurvey,
blood samples from a specific group of people are
analyzed for antibodies to a particular bacteria.
In this study, technicians looked for antibodies
against LP-1.) They conducted a second
case-control study to determine whether exposure
to Grocery Store As misting device was
associated with developing Legionnaires disease.
They cultured ten similar misting devices
from other parts of the country. They asked
for permission to perform autopsies on two
patients who had died of pneumonia early in the
outbreak. Here is what the investigators found
The serosurvey showed that employees at
Grocery Store A were three times more likely than
employees at the other grocery stores to have
elevated antibody titers to Legionella. This
finding (3 to 1) is the prevalence ratio, or the
ratio of the prevalence of antibody, in each
group. The p value, or probability that these
findings are significant, is 0.02.
31
Analysis of the second case-control study
revealed a significant association between being
sick and buying produce that was nearest the
mister. Of the 10 mist machines from other parts
of the country, 6 grew Legionella. The
subtype of Legionella found in the grocery store
misting device was also isolated from a small
cooling tower that was far from public access and
not near Grocery Store A.
32
Autopsied Patients

• Autopsies of the 2 patients contained the subtype
  of Legionella found in Grocery Store A
• Up to this point, the news media did was unaware
  of the outbreak, the investigation, or the
  results.
• Who needs to know about these findings?
• How would you go about reporting these findings?

33
Conclusion