Guidelines for writing outbreak investigation reports
Date: Date of report
To: Supervisor
From: Investigator(s)
Subject:
Location:
Date of departure:
Date of return:
Abstract
Half page or less:
- What was the problem?
- What was done to address the problem?
- What was found?
- What conclusions were drawn?
- What recommendations were made?
- What public health actions were taken?
Background
Nature of the problem and its public health importance:
- Problem description
- Sequence of events leading to the study or investigation
- Why was an investigation undertaken?
Contacts in the field and investigation team
Pertinent background information and situation upon arrival:
- Geographic setting
- Size of community/hospital, etc
- What had been done so far?
- What was known to date?
- Brief statement of the working hypothesis
Objectives of the investigation
Methods
Case definition
Clinical, laboratory, time, place, person
Case finding methods
Source and mode of data gathering (telephone, interviews, record review, etc)
Analytical study-design and rationale
Case-control study
- Control definition
- Control selection
- Definition of exposure(s)
- How was exposure measured and categorised?
- What measure(s) of association were chosen?
- What statistical test(s) were chosen?
- Rationale for stratified and multivariate analysis, if any
Cohort study
- Definition of exposure
- How was exposure measured and categorised?
- What measure(s) of association were chosen?
- What statistical test(s) were chosen?
- Rationale for stratified and multivariate analysis, if any
Cross-sectional, etc
- Idem
Laboratory methods
- Type of samples
- Laboratory examination and methods
- Further typing
Environmental studies
- Type of inspection
- Method for sample collection
Other studies
Results
Descriptive findings
- Response rates
- Number of persons meeting case definition
- Overall attack rate (AR)
- Description by
time (epidemic curve)
place (AR by place)
person (clinical features, AR by demographic characteristics)
Laboratory findings
- Number of samples tested and found positive
- Typing results
Environmental study findings
- Number of samples tested and found positive
- Comparison with human samples
Transition
- What do the descriptive results suggest in terms of risk groups, source, mode of transmission,
exposure?
- Hypotheses generated that will be subsequently tested in analytic studies.
Analytical study results
- Proceed from general to particular
- From univariate to bivariable to multivariable (stratification and then regression) analysis.
Further studies performed, if any
Pending results, including lab
Discussion
Main results
Our investigation suggests that ……
Refutation of findings (Validity)
- Limitations of study design
- Possible biases (information, selection, confounding) that may have lead to the observed results.
Inferences from analytic study results
- Whether the findings fit with what is known about the disease
- Which criteria of causality have been met.
Conclusions
- Present a logical, clear interpretation of the results; explain how the working hypothesis is
confirmed or disproved by the results.
Recommendations, actions
- Feasible recommendations for prevention/control measures based on public health implications of
the findings.
- Rationale for recommendations and actions
- Further or future studies needed
Signatures of investigators and supervisors
Tables
- With a complete legend including time, place, person.
Figures
- With a complete legend including time, place, person.
References
Vancouver style
Annex 5 Example of an outbreak investigation report
Date: 25 September 1996
To: Director of Public Health, Eastern Health Board
From: Thomas Grein, EPIET Fellow, EHB
Subject: Salmonella typhimurium outbreak
Location: Malahide, County Fingal
Date of departure: N/A
Date of return: N/A
Abstract
An outbreak of salmonellosis occurred among 127 persons attending a wedding
reception on 21 August 1996. Of 115 interviewed guests, 57 (50%) met the case
definition (diarrhoea within three days after having eaten at the reception). Thirty-
eight cases visited their GP, seven were admitted to hospital. Forty-six cases
submitted stool samples, of which 39 were culture positive for Salmonella
typhimurium. Turkey was identified as the most likely vehicle for this outbreak
(relative risk ¥). Environmental investigations at the catering facilities showed
deficiencies in food hygiene practices. Eight of 17 asymptomatic kitchen workers
carried S. typhimurium in their stool.
We recommended: to exclude all symptomatic food handlers from work in the hotel
kitchen for 48 hours after their first normal stool; to educate food handlers and
other personnel in the hygienic preparation and serving of food; and to immediately
address the structural and operational deficiencies in the hotel kitchen. Introduction
On 26 August 1996 the Eastern Health Board (EHB) was informed of an outbreak of
gastrointestinal illness among guests of a wedding party that was held in a large
hotel in Malahide on 21 August 1996.
Many guests had fallen ill since the reception and some had required hospitalisation.
Malahide is a popular seaside town approximately twenty kilometres north of Dublin
City.
The same day the EHB started an investigation to assess the extent of the outbreak,
identify the mode and the vehicle of transmission, and initiate appropriate control
measures.
Dr. Darina O’Flanagan, Specialist in Public Health Medicine at the EHB, led the
epidemiological investigations. She was assisted by Dr. Thomas Grein, Fellow of the
European Programme for Intervention Epidemiology Training. Mr. Tom McCarthy,
Principal Environmental Health Officer for food hygiene North Dublin City with
special responsibility for communicable disease, and Mr. Derek Bauer, Principal
Environmental Health Officer for County Fingal, led the environmental investigations
and supervised the implementation of control measures.
Nature of problem
Public health importance
Sequence of events
leading to investigations
Objectives of
investigation
Composition of field
investigation team
Materials and Methods
Case definition
We defined a case as a person who had consumed food at the wedding reception on
21 August 1996 and developed diarrhoea (three or more loose stools in 24 hours)
within the next 72 hours.
Case finding
We obtained the addresses and telephone numbers of all 127 attendees of the
wedding reception. Hotel management provided a copy of the menu and a list of all
food items served during the reception.
Starting 27 August 1996, Environmental Health Officers (EHOs) conducted personal
interviews at the homes of all wedding guests. Hospitalised cases were interviewed
after discharge from hospital. Information was obtained on demographic details,
symptoms of gastrointestinal illness three days prior to and after the wedding
reception, the time of onset and the duration of symptoms, contact with ill persons
not related to the wedding party, secondary spread among family members, foods
consumed during the reception, whether the family doctor was contacted because of
the illness, whether hospitalisation was required, and length of hospital stay if
admitted.
Analytical study design
We conducted a retrospective cohort study to identify the potential vehicle of the
outbreak. The retrospective cohort design was chosen because information could be
obtained on a clearly identifiable risk group.
Definition of exposure.
The outbreak occurred among 127 guests who attended the
wedding reception in the hotel on 21 August 1996. The main meal was served to
108 guests at 1800 hours on 21 August 1996. The meal consisted of honeydew
melon, roast turkey, baked Irish gammon (ham steak), a selection of vegetables and
potatoes, and chocolate eclairs for dessert. At 2200 hours sandwiches (turkey, ham,
chicken, salad, savoury, egg, cheese) were offered to the guests and consumed by
58 individuals. Hotel staff prepared all dishes and sandwiches in a kitchen on the
premises except for a home-made birthday cake and a home-made wedding cake.
Both cakes were brought into the hotel by guests and consumed throughout the
evening. To identify potential risk factors for illness, all guests were asked if they
had consumed any of these food items
The restaurant of the hotel caters for hotel guests and a large number of visitors. No
other functions were held on the day of the wedding reception. The number of
persons who attended the restaurant on 21 August 1996 is unknown.
Analysis of the data was performed with Epi Info software, version 6.041. Food
specific attack rates (AR), relative risks (RR) and 95% confidence intervals (95% CI)
were calculated for the consumption of food items. The c2 test was used to compare
Case definition
Note: Only clinical case
definition was used. If
others would have been
used, describe them here.
Source and mode of data
gathering
Type of analytical study
Rationale
Definition of exposures
Chosen measures of
associations and statistical
tests
Laboratory investigations
All interviewed persons who reported an illness were asked to provide a stool
sample. Stool samples were also collected from some individuals who attended the
wedding reception but did not become ill. Most specimens from non-cases were
obtained from household members of cases. All specimens were submitted to the
Public Health Laboratory for culture. Faecal specimens were also obtained from the
17 kitchen workers who were on duty during the week of the wedding reception,
regardless of their symptoms.
Environmental investigations
Starting 26 August, EHOs inspected the restaurant and the hotel kitchen on several
occasions, investigated food handling practices and interviewed all food handlers for
illness one week prior to and after the wedding. They examined transport, storage
and preparation processes for the foods served at the wedding reception, and
reviewed order and delivery books of the restaurant. The ingredients of incriminated
foods were identified and traced to their sources.
Food specimens from the day of the wedding were no longer available when
investigations commenced. EHOs sampled the same type of food items which were
mentioned on the wedding reception menu and submitted them for culture on 27
August 1996.
Results
Descriptive findings
Of the 127 wedding guests, four individuals had not eaten at the wedding reception
and were excluded from the study. None of them reported an illness. Five guests
refused to participate in the study and three guests could no longer be contacted.
The remaining 115 (93%) individuals were interviewed (table 1). Sixty-two (54%) of
them were female, 100 (87%) between 15 and 64 years of age (table 2).
Sixty-eight guests reported an illness during the interview. The case definition could
be applied to 57 individuals. The overall attack rate among guests was 50%.
Dates and times of onset of illness for the 57 cases are shown in figure 1. There was
a steady increase in the number of cases, starting in the night of 21 August, peaking
during 22 August and declining over the next 48 hours. Two individuals developed
diarrhoea on 25 August 1996 but were not included as cases. The median time
(range) between the main meal and onset of illness in cases was 24 (5-72) hours.
Males were 1.3 times (95% CI 0.9 - 1.9) more likely to be a case than females.
Guests older than 65 years had the highest attack rate (100%) and were 2.3 times
(95% CI 1.7 - 3.2) more likely to become ill than guests 45- 64 years who had the
lowest attack rate with 43%.
The main symptoms of cases were diarrhoea (case definition, 100%), feeling
feverish (89%), general malaise (88%) and nausea (81%). Vomiting was reported
Environmental
investigations
Type of inspection
Methods for sampling
collection
Eligibility
Response rates
Time
Person
Clinical features
less frequently (47%). The duration of illness ranged from two hours to 13 days
with a median of five days (table 4).
Individuals who ate only during the late meal had a 1.7 times (95% CI 1.0 - 2.6)
higher risk of illness than individuals who only ate during the main meal. The attack
rates for guests seated at different tables varied between 25% and 80% (c2 = 11.3,
p = 0.42). The age and sex distribution of guests seated at tables with higher attack
rates (table 5 and 11) was not different from the distribution of guests seated at
tables with lower attack rates (table 3).
Forty-six (81%) cases provided stool samples. Thirty-nine (85%) samples were
culture positive for
Salmonella typhimurium.
All isolates showed the same resistance
pattern to Ampicillin, Amoxycillin, Chloramphenicol and Sulphonamides. One culture
was phage typed at CDSC London (Definitive Type 104). An increase in the number
of
S. typhimurium
isolates unrelated to the outbreak was not observed by hospital
laboratories in the EHB area during this period.
The rapid increase and decline in the number of cases, the single peak, the common
exposure to food consumed at the wedding reception and the absence of an
increase in other laboratory-detected cases of
S typhimurium
suggested a foodborne
point source outbreak among the wedding guests (figure).
Food specific attack rates, relative risks and percentage of cases exposed to the
food items consumed at the wedding reception are given in table 5.
For seven food items, cases had higher attack rates than non-cases: turkey (RR ¥),
savoury sandwich (RR 1.85), birthday cake (RR 1.61), egg sandwich (RR 1.56),
chicken sandwich (RR 1.43), ham (RR 1.22) and turkey sandwich (RR 1.12).
There were no cases among guests who had not eaten turkey during the main meal.
Of the 57 cases, 52 (91%) had consumed turkey during the main meal
Environmental investigations
EHOs noted 23 violations of the food hygiene regulations during the kitchen
inspections. Relevant findings with regard to the wedding outbreak were that frozen
food was thawed in hot water, cooked meats cooled down at room temperature for
indeterminate times and that storage practices in the cold room allowed for possible
cross-contamination of raw meat.
Food items from hotel kitchen and bar buffet were sent to the laboratory on 27
August 1996. The only positive microbiological finding was found for a sample of
cooked turkey (
Salmonella agona).
The examination of the kitchen delivery dockets revealed that ten turkeys were
delivered to the hotel on 19 August. Six of the ten turkeys were used for the
Place
Laboratory results
Summary descriptive
findings:
Identifiable risk groups?
Mode of transmission?
Analytical study results
Univariate analysis
Environmental
investigations
wedding reception. Each of them weighted 20-24 lb. and were cooked on 20 August
at 250oC for thirty minutes and at 180oC for two and a half hours. After cooking
they were put into a non-refrigerated holding cabinet, left at room temperature to
cool down, and later removed to the cold room. We could not determine how long
the turkeys were left in the non-refrigerated holding cabinet. Other turkeys, cooked
at midday on 21 August, were left overnight in the holding cabinet before being
removed to the cold room.
Seventeen kitchen workers were interviewed and stool samples obtained from them.
None reported an illness but eight (47%) stool samples were culture positive for
S.
typhimurium
. Antibiotic resistance was determined for some isolates and matched
that of the cases (resistant to Ampicillin, Amoxycillin, Chloramphenicol,
Sulphonamides).
Discussion
The primary objectives of our study were to identify the mode of transmission, the
vehicle of the outbreak and to initiate appropriate control measures. Our data
suggest that the vehicle of the outbreak was turkey served during the wedding
reception on 21 August, and the infecting agent
S. typhimurium
DT104.
The relative risk for the consumption of turkey was infinite. There were no cases
among guests who had not eaten turkey during the main meal. Of the 57 cases, 52
(91%) had consumed turkey during the main meal. Six other food items showed
statistically significant relative risk estimates greater than. However, all of these food
items were consumed by a small number of cases which makes them implausible
vehicles for this outbreak. Thus epidemiologically turkey appears to be the most
likely vehicle for this outbreak. Isolation of
S. typhimurium
from the stool of cases
supports this finding as the pathogen is frequently found in poultry. Eighty-five
percent of the stool cultures available for the cases were positive for this organism.
As the epidemiological data were obtained from a non-controlled, observational
study some limitations apply to our results. All data were collected by personal
interviews and could not be verified. Some information bias is likely to have existed,
particularly after interviewees learned through the media about legal proceedings
and compensation claims. Although most interviews were conducted within a week
following the outbreak recall bias may have led to wrong exposure status. Selection
bias is unlikely to have influenced our findings as the participation in the study was
high (93%). As most guests ate the same foods stratification for possible
confounding could not be performed for most food items. As we did not enquire
about the amounts of food consumed we were unable to calculate dose response.
The environmental investigations support our epidemiological findings and revealed
severe deficiencies in food handling practices in the hotel kitchen. Stool samples
from eight of the 17 kitchen staff on duty during the week of the outbreak were also
positive for
S. typhimurium
suggesting that the infective food was prepared and
consumed in the hotel kitchen.
Summary of key findings
with regard to objectives
Validity of epidemiological
findings
Limitations of study
design
Do results from
environmental
investigations support
findings?
Six turkeys were identically prepared on the same day and served at 12 tables. We
could not determine if the meat of a whole turkey was served to specific tables or if
the meat of all six birds was cut into pieces and then distributed randomly to all 12
tables. Attack rates for the tables vary between 25% and 80% without statistically
significant differences. As every table had at least two cases it is more likely that
meat of one or more infected birds was served to all tables. The mode of
contamination remains unknown. Poor foodhandling practices may have allowed for
one infective turkey to cross contaminate others, or contamination may have
occurred by an asymptomatic, culture positive food handler.
Our findings are consistent with other foodborne outbreaks related to the
consumption of turkey. It is also a biologically plausible vehicle for the aetiological
agent,
S. typhimurium.
The implicated exposure preceded illness. Consumption of
turkey was positively associated with illness and this association was stronger than
for other food items.
More cases, unrelated to the wedding reception, came to our attention. Of five
golfers lunching in the same hotel on the day of the wedding reception three fell ill
within the next 24 hours. Interviews were conducted with the group. The main
symptoms of the three ill individuals were diarrhoea and general malaise lasting
between four and ten days. All three had consumed turkey salad sandwiches, the
other two unaffected golfers cheese sandwiches. A stool sample was available for
one ill individual which was culture positive for
S. typhimurium
(no definite type
available). These additional cases strongly support the hypothesis that turkey was
the vehicle of the outbreak and
S. typhimurium
the infecting agent.
The Department of Agriculture was informed about the outbreak and subsequently
investigated the poultry farm where the turkeys originated. S
. typhimurium
was
detected in the dust of one of six turkey houses examined. According to a
spokesperson of the Department this is a rare finding on Irish poultry farms. Further
investigations are pending.
Recommendations, actions
We recommended excluding all symptomatic food handlers from work in the hotel
kitchen for 48 hours after their first normal stool. We also advised to educate food
handlers and other personnel in the hygienic preparation and serving of food and to
implement the National Standard Authority of Ireland (NSAI) guideline 340:1994 -
Hygiene in the Catering Sector4. The structural and operational deficiencies in the
hotel kitchen were outlined in a detailed report and hotel management was urged to
correct these deficiencies immediately.
Dr Thomas Grein
EPIET fellow
Department of Public Health, Eastern Health Board
Dr Darina O’Flanagan
Specialist for Public Health
Department of Public Health, Eastern Health Board
Causality criteria
Relevant results from
other studies not part of
this investigation
Recommendations, actions
Acknowledgements
The members of the outbreak control team would like to thank the staff of the EHB, in particular
the Environmental Health Officers involved in the investigation and the laboratory staff of Cherry
Orchard hospital, for their indispensable help. We would also like to thank Dr Alain Moren and Dr
Mike Rowland, EPIET/EUPHEM, for reviewing the manuscript of this report.
Table 1 Study characteristics. Wedding reception, Malahide, 21 August 1996
number (percent)
Wedding cohort
127 (100)
Eligible
123/127 (97)
Refused to participate in study
5/123 (4)
Unable to locate
3/123 (2)
Interviewed (response rate)
115/123 (93)
Table 2 Demographic details of cohort. N = 115. Wedding reception, Malahide, 21 August
1996
number (percent)
Age class (years)
5-14
2 (2)
15-44
46 (40)
45-64
54 (47)
> 65
6 (5)
Unknown
7 (6)
Female
62 (54)
Figure Date and time of onset of diarrhoeal illness among cases. n = 57. Wedding reception,
Malahide, 21 August 1996
10 cases
9
8
7
6
5
4
3
2
1
0
00- 06- 12- 18- 00- 06- 12- 18- 00- 06- 12- 18- 00- 06- 12- 18-
21 August 22 August 23 August 24 August
Date and time of onset
Table 3 Characteristics of cases with attack rates, relative risks (RR) and 95% confidence
intervals (95% CI). n = 57. Wedding reception, Malahide, 21 August 1996.
number attack rate (%) RR (95% CI)
All cases
57
57/115 (50)
Sex
Female
27
27/62 (44)
Male
30
30/53 (54)
1.3 (0.90-1.89)
Age class * (years)
5-14
1
1/2 (50)
1.2 (0.28-4.86)
15-44
25
25/46 (54)
1.3 (0.85-1.92)
45-64
23
23/54 (43)
1.0
65 +
6
6/6 (100)
2.3 (1.72-3.20)
Meals
Main meal only
57
24/57 (42)
Late night meal only
7
5/7 (71)
1.7 (0.97 - 2.57)
Seating arrangements #
Table 1
3
3/10 (30)
1.2 (0.3-5.5)
Table 2
3
3/8 (38)
1.5 (0.3-6.7)
Table 3
5
5/10 (50)
2.0 (0.5-7.7)
Table 4
2
2/5 (40)
1.6 (0.3-8.0)
Table 5
7
7/10 (70)
2.8 (0.8-9.9)
Table 6
4
4/10 (40)
1.6 (0.4-6.6)
Table 7
4
4/8 (50)
2.0 (0.5-8.0)
Table 8
4
4/9 (44)
1.8 (0.4-7.3)
Table 9
2
2/8 (25)
1.0
Table 10
3
3/9 (33)
1.3 (0.3 - 6.1)
Table 11
8
8/10 (80)
3.2 (0.9 - 11.1)
Table 12
5
5/8 (63)
2.5 (0.7 - 9.3)
* 2 = 7.5, p = 0.057; for seven individuals no information about their age
# 2 = 11.3, p = 0.42; seven guests attended only late night meal (no tables assigned),
for three guests table number unknown
Table 4 Clinical and laboratory details of cases. n = 57. Wedding reception, Malahide, 21 August 1996
number (percent) median (range)
Symptoms
Diarrhoea
57 (100)
Feeling feverish
51 (89)
Aches and pains
50 (88)
Nausea
46 (81)
Abdominal cramps
28 (49)
Vomiting
27 (47)
Headaches
16 (28)
Blood seen in / on stool
4 (7)
GP visit
38 (67)
Hospitalisation
7 (12)
Time in hospital (hours)
96 (6 - 312)
Duration of illness (hours)
120 (2 - 312#)
Incubation period (hours)
24 (5 - 72)
Stool samples obtained
46 (81)
Stool sample +ve for Salmonella typhimurium
39/46 (85)
#
Sixteen cases were still symptomatic at time of interview, thus upper range > 312 hours
Table 5 Food specific attack rates (AR), relative risks (RR), 95% confidence intervals (95% CI), and percent of
cases exposed. Wedding reception, Malahide, 21 August 1996.
food eaten
food not eaten
95%
% cases
cases
total
AR %
cases
total
AR %
RR
C.I.
exposed
Main meal
Soup
48
102
47
4
6
67
0.71
0.39-1.29
84
Turkey
52
104
50
0
4
0
91
Ham
48
98
49
4
10
40
1.22
0.56 - 2.70
84
Melon
47
100
47
4
7
57
0.82
0.42-1.61
82
Carrots
46
96
48
4
8
50
0.96
0.46-1.98
81
Potatoes
46
98
47
6
10
60
0.78
0.45-1.35
81
Croquettes
43
84
51
7
19
37
1.39
0.74-2.59
75
éclair
41
90
46
11
17
65
0.70
0.46-1.07
72
Stuffing
40
84
48
11
21
52
0.91
0.57-1.45
70
Cauliflower
40
84
48
12
23
52
0.91
0.58-1.43
70
fresh cream
17
44
39
33
62
53
0.73
0.47-1.13
30
coffee cream
8
14
57
44
93
47
1.21
0.73-1.99
14
Scampi
2
4
50
50
104
48
1.04
0.38-2.83
4
wedding cake
25
53
47
27
54
50
0.94
0.64 - 1.39
44
birthday cake
12
17
71
40
91
44
1.61
1.09 - 2.36
21
Sandwiches
Turkey
3
5
60
23
43
53
1.12
0.52 - 2.42
5
Ham
12
24
50
16
26
62
0.81
0.49 - 1.34
21
Cheese
9
16
56
21
36
58
0.96
0.58 - 1.61
16
Egg
8
10
80
21
41
51
1.56
1.02 - 2.40
14
chicken.
3
4
75
23
44
52
1.43
0.76- 2.70
5
Savoury
3
3
100
26
48
54
1.85
1.42 - 2.39
5
Main meal and/or
sandwiches
Turkey
53
105
50
2
8
25
2.02
0.61 - 6.81
93
Ham
51
104
49
6
10
60
0.82
0.48 - 1.41
89
References
1.
Dean AG, Dean JA, Coulombier D, Burton AH, Brendel KA, Smith DC, Dicker RC, Sullivan K,
Fagan RF, Arner TG. Epi Info, Version 6.04: a wordprocessing, database, and statistics
program for epidemiology on micro-computers. Centers for Disease Control and Prevention,
Atlanta, Georgia, U.S.A., 1996.
2.
Hayes CB, Lyons RA, Warde C. A large outbreak of salmonellosis and its economic cost.
IMJ
1991; 84:65-66.
3.
National advisory committee on microbiological criteria for foods. Hazard analysis and
critical point system.
Int Journal of Food Microbiology
1992; 16:1-23.
4. National standard authority of Ireland. Hygiene in the catering sector; guideline 340,
Dublin, 1994
References
Vancouver style
