Page 129 - Journal of Special Operations Medicine - Fall 2017

between 0800 and 1000 each day. Emergency services at the Investigators independently pulled food samples from the food
hospitals were open 24 hours a day, 7 days a week. preparation and stock assemblage areas. Samples were not
pulled from the serving line itself, because once the outbreak
A similar GI outbreak occurred on the same compound almost started, these samples could have been cross-contaminated by
a year before, in October 2015. The 2015 outbreak was attrib- already sick patients. Foods considered to be at high risk for
uted to E. coli and Salmonella spp. and had an overall total of contamination were selected for sampling and sent to an inde-
84 cases. Another outbreak in May of 2016 sickened 260 in- pendently certified food laboratory in Amman, Jordan, in ac-
dividuals. There had been two relatively large-scale outbreaks cordance with IOS methods. Specific food items were sampled
per year in May and October for the past 3 years. and shipped on October 3, 4, 5, 10, 11, and 12. Water analyses
and testing continued daily per standard procedure. A surprise
Data Collection and Laboratory Investigations reinspection of the dining facility and private food vendors oc-
Compound personnel who had fallen ill began to present on curred on 8 October. Food operations continued as normal at
30 September 2016. As the frequency of ill patients increased, the dining facility until 10 October, when the line menu items
health-care providers were briefed on 2 October to obtain in- were changed because of a preponderance of evidence.
formation about the situation. A public health investigation
was also initiated on 2 October once it became apparent an Descriptive Epidemiology
outbreak was underway. As part of the investigation, medi- Of the approximate population of 1,000 personnel, a total
cal records at the two large hospitals and two outlying clinics of 123 people had eaten at the dining facility and had been
were reviewed. A case definition was established as any pa- sickened enough to seek medical care (Table 1). Of the over-
tient who had eaten at the compound dining facility or from a all cases, 77% to 88% were men. The median age was ap-
compound vendor; had presented with GI complaints, which proximately 23 years. The age of patients ranged from 19 to
included abdominal pain, diarrhea, or vomiting; and who 40 years. Military personnel were affected more than civilian
was evaluated at one of the Baghdad military or diplomatic personnel at a 4.3-to-1 ratio. The attack rate was approxi-
clinics or hospitals from 30 September to 12 October 2016. mately 12% overall (123 of 1,000 people). The index case was
Other clinical records were later reviewed as the investigation a 21-year-old, female US military member who had reported
progressed. to sick call with vomiting. The type and frequency of symp-
toms are reported in Table 2.
A search for cases was initiated at each of the health-care fa-
cilities, and line lists were updated as new cases were identi- Incubation periods were not precisely calculated: military per-
fied. Case information was collected passively and an active sonnel had either had to wait for sick call or had sought ini-
case search was not pursued because of the magnitude of the tial care from a medic before seeking care at a fixed facility
outbreak and limited public health staff and equipment. Medi- with an established medical recording process. The estimated
cal recordkeeping was not standardized across health-care sys- onset time of symptoms was approximately 12 hours. This
tems, which complicated data collection. For example, several estimated onset time is consistent with the drop off in cases
outlying clinics had not collected patient data aside from chief after the suspect food items had been removed from the din-
complaint, name, and month of visit. When possible, a detailed ing facility. The removal occurred early on the morning of 10
questionnaire was obtained from patients. The questionnaire October. Table 3 is a symptoms-onset chart for known cases.
included queries on travel, age, job description, living quar-
ters, symptoms, date of onset, medications, animal contact, Laboratory Results
food history, and handwashing. Treatments and interventions Table 4 lists the number and percentages of polymerase chain
were recorded for outcome analysis of the 123 patients. A to- reaction (PCR)-confirmed cases by type of pathogen. Not all
tal of 43 personnel filled out questionnaires to varying degrees cases were confirmed among military personnel, due to either
of completion. A subset of three patients were interviewed to the limited availability of resources or operational constraints.
obtain responses; however, most patients were asked to fill out A total of three hospital workers—a 28-year-old asymptomatic
the questionnaire on their own. old man, a 40-year-old symptomatic man, and a 40-year-old
symptomatic woman—decided to run PCRs on their own stool
The dining facility and private food vendors had previously samples. All three workers chose to not report for sick call or see
undergone monthly inspections for sanitation and hygienic a medical provider. The three workers were not counted in the
practices. The dining facility, workers, and private food ven- total cases because they did not meet the case definition criteria.
dors were also inspected on 8 October for food preparation, Two of the workers tested positive for both enteroaggregative
handling, storage, and transportation. The dining facility had E. coli and enteropathogenic E. coli but did not test positive for
also continued to use the licensed medical provider, including other viruses, bacteria, or parasites. The asymptomatic worker
on 8 October, to examine food workers before their shifts for did not test positive for an enteric pathogen or parasite. The
signs of illness or disease. The dining facility had also contin- PCR screening test panel results are listed in Table 5. The most
ued to operate its food laboratory and to collect, hold, and common responses from food histories are listed in Table 6.
test food samples. The authority for the choice of samples had
been delegated to dining facility management personnel, who A response of “salad” on questionnaires was not considered
chose which food samples to test. The dining facility personnel reliable. There had been potato salad, cucumber salad, onion
had collected approximately 3 ounces, by volume, of all food salad, and a self-serve build-your-own salad as options during
types served. Approximately 200 samples were collected for the time of the outbreak. Two respondents specified “cucumber/
each lunch or dinner meal, and approximately 100 samples onion salad,” and the remainder of the responses were undiffer-
were collected for breakfast meals. Of the approximately 500 entiated. All other food items tallied fewer than seven respon-
samples collected daily, only 4 to 10 underwent microbiologi- dents of the 43 total. Certain food items were reportedly used
cal analysis. for meal components, such as the spinach for turkey wraps.

Investigation of Gastrointestinal Illness Outbreak | 125