Indiana Epidemiology Newsletter
E3 is a new feature of the Indiana Epidemiology Newsletter dedicated to exploring the fundamentals of epidemiology. Each month, a different epidemiology concept will be explored to enhance understanding of basic epidemiology
Lynae Granzow, MPH
ISDH Enteric Epidemiologist
Once notified of an outbreak and assignment of the initial case definition, it’s time to decide which study design will be used: cohort or case control. A cohort study is used when there is a defined population, group, party with a known discrete event. Cohort studies can be used in research, epidemiology, or public health. Cohort studies involve the comparison of disease rates between exposed and non-exposed groups. For an outbreak, the only type of cohort study used is retrospective, the event has already occurred. The most common events that would lead to a cohort study include: a group of students, a banquet, a one-time party at a restaurant.
Complaint: The mother-of-the-bride calls to state that several attendees of the wedding/reception held at Banquet Center A on June 5, 2008, are ill with vomiting and diarrhea. Illness began within 12-48 hours of attending.
Known: There is a discrete event and defined group.
Initial Case Definition: Any previously healthy person who attended the reception at Banquet Center A on June 5, 2008 and became ill within 48 hours including diarrhea and/or vomiting. Secondary cases should be determined.
Gather: Menu items, contact info for all attendees and food handlers at the event, and stool and food samples.
Develop a Questionnaire: Should include yes/no questions for symptoms, menu items, and other possible exposures. There are 200 people that attended the event. But not every person has to be interviewed; just enough to get some significant data.
Calculate: Calculations can be done manually, by a statistical program (Epi Info can be downloaded for free), or by a plug-in 2x2 table online.
Risk Ratio (RR) = [A/(A+B)] / [C/(C+D)] = [26/(31)] / [6/(41)] = 5.73
Attendees who ate the cake were 5.73 times more likely to become ill than attendees that did not eat the cake. The RR must be supported by at least a 95% confidence interval (CI) that does not include ≤ 1.0. Please see a future issue for Tests of Validity for further explanation.
1. Friis, Robert. Sellers, Thomas. Epidemiology for Public Health Practice. 3rd Edition. 2004