Outbreak investigations are often seen as key examples of field epidemiology.

Definition of an outbreak

What exactly is an outbreak? Different definitions of an outbreak can be found in literature. Here we will define an outbreak simply as 'the occurrence of more cases than expected in a particular population, in a specific geographical area and over a specified period of time'.

Difference between an outbreak and an epidemic

The word 'epidemic' is often used interchangeably with 'outbreak' in professional communication. In the media, the word epidemic usually has a more threatening ring to it, which is why most communication experts tend to avoid using it. Outbreak investigators may decide to use the word 'epidemic' or not depending on whether they want to attract or deflect attention. In September 2014, president Obama made a careful distinction between the two terms to attract attention to the emergency that Ebola constituted. (1)

From the identification of a cluster to the establishment of the existence of an outbreak

To establish the existence of an outbreak, we first of all need to understand what is meant by 'cases'. This needs to be defined. Usually the first signal of an outbreak can come from a telephone call or report from the health care system about a cluster of cases. We may or may not yet know if such a cluster is 'more than expected', so a systematic approach is required. 

At this stage it is important to understand the distinction between a cluster of cases (2 or more cases that are related by sharing similar characteristics in time and/or place and or personal characteristics) and an outbreak (more cases than expected). For example 5 cases of respiratory illness occurring in the same week can be considered 'a cluster in time', yet this may be the usual number of such cases that one would expect in that week of the year.

Similarly cases can be clustered in place (same village / neighbourhood etc) or according to personal characteristics (e.g. cases sharing the same age-category). In each of these examples of clusters, a key question will be: 'is this number of cases more than we would expect?'. If the answer is 'yes' then the cluster can be considered an 'outbreak'.

"More cases than expected", implies that we need to have knowledge of the 'normal' number of cases (or baseline). This knowledge may come from surveillance or surveys. The increase in the number of cases is best documented as a population-based incidence rate. Investigators may want to examine possible artifact in the numerator (e.g., batch reporting of old cases or of prevalent cases) on in the denominator (e.g., population movements, mass gathering).

Steps of an outbreak investigation

The various lists of various steps

Investigating an outbreak requires a systematic approach that is summarized as a number of steps. Unfortunately, various groups have generated various lists where steps differ in sequence and number. One of the most classical lists (2) includes Ten steps summarized as below:

  1. Establish the existence of an outbreak. This maybe achieved as described above by calculating rates, comparing the rate with the baseline and excluding artifacts in the numerator or the denominator.
  2. Confirm the diagnosis. This may be achieved through (a) shortlisting a number of possible diagnoses on the basis of the frequency of signs and symptoms and (b) confirming the diagnosis, most often with support from the laboratory.
  3. Establish a case definition (that needs to have time, place and person elements and that may have different levels of sensitivity and specificity, including possible, probable and confirmed cases) and count cases following a case search strategy that can be passive, stimulated passive or active, but that must be always homogeneous in the area considered.
  4. Orient the data in terms of time, place and person through descriptive data analysis. This will lead to an epidemic curve (time), a map (place) and rates by age and sex (person).
  5. Determine who is at risk of becoming ill (population at risk)
  6. Develop a hypothesis that explains the specific exposures that caused disease and test this hypothesis by appropriate statistical methods (through analytical studies). An article by Werber and Bernard published in Eurosurveillance (3) describes the development of a toolbox consisting to increase the use of analytical studies in the investigation of outbreaks of food borne diseases. In general, analytical epidemiology may use case-control investigations (more adapted if the attack rate is low, under 5-15%) or cohort investigations (more adapted if the attack rate is not too low, above 5-15%)  
  7. Compare the hypothesis with the established facts.
  8. Plan a more systematic study (environmental, microbiological etc)
  9. Prepare a written report (outbreak reporting, to communicate findings to those who need to know)
  10. Execute control and prevention measures (Recommend options for interventions based on the findings)

Some lists have referred to 13 steps (4), adding additional steps such as logistical aspects (e.g., Prepare for field work) while another prepared for foodborne outbreaks was more conceptual with only 7 steps (1. Detecting a possible outbreak, 2. Defining and finding cases, 3. Generating hypotheses about likely sources, 4. Testing the hypotheses, 5. Finding the point of contamination, 6. Controlling the outbreak and 7. Deciding an outbreak is over).

Since 2012, the EPIET and EUPHEM fellowships have used an adaptation of the original 10 steps for teaching and supportive supervision. These 10 steps are very similar from the one above, with minor adjustments derived from an analysis of the common errors in outbreak investigations (5). This adaptation has the advantage of disentangling (a) case definition (step 3) from case search (step 4) and (b) generation of hypothesis (step 5) from hypothesis testing (step 6). Case definition and case search on one side and hypothesis formulating and testing on the other side are quite different processes that can suffer from specific pitfalls and benefit from specific guidance (Hence the benefit in the split). In addition, the 10 steps adapted in such a way places the 'middle' of an investigation between hypothesis generating and testing. This reflects the pivotal thought process that needs to take place at that critical phase of the outbreak investigation when the outbreak investigation team may have to write a mini-protocol.  

How to understand the lists of steps of an outbreak investigation?

The lists of steps for outbreak investigations must not be taken to literally. First, they are ordered in a sort of logical sequence that does not necessarily match the temporal sequence. For example, some outbreak investigations may start with the enforcement of control measures (e.g., implementation of infection control in health care facilities to prevent secondary spread). Second, they summarize a number of steps that should take place for most investigations. However, some this may vary from investigation to investigation. Overall, they can be thought of as a list of 'things one wants to consider' while investigating an outbreak.

The outbreak team

Key to the investigation and control of an outbreak is the constitution of a team.

References:

  1. Remarks by the President on the Ebola Outbreak. https://www.whitehouse.gov/the-press-office/2014/09/16/remarks-president-ebola-outbreak (Accessed 11 May 2015)
  2. Gregg, Michael B. (Editor). Field Epidemiology.Oxford University Press, new York, 1996
  3. Werber D, Bernard H. Reducing the barriers against analytical epidemiological studies in investigations of local foodborne disease outbreaks in Germany – a starter kit for local health authorities . Euro Surveill. 2014;19(8):pii=20714.
  4. US CDC. Principles of Epidemiology in Public Health Practice, Third Edition
    An Introduction to Applied Epidemiology and Biostatistics
    . 2006, updated in 2011. Lesson 6; Section 2: Steps of an Outbreak Investigation.
  5. Murhekar M, Moolenaar R, Hutin Y, Broome C. Investigating outbreaks: practical guidance in the Indian scenario. Natl Med J India. 2009. :252-6.
  6. Heymann, David L. (Editor). Control of Communicable Diseases Manual, 19th Edition.

EPIET Lectures:

Epidemiological Outbreak Investigation

Operational Aspects of Outbreak Investigations