Data collection

During an outbreak, data will come from a number of sources that might inform hypothesis generation regarding a source. Firstly, data from the cases themselves, who will provide contextual descriptive epidemiological information about where they might have been and when they got ill. The collection of such information should be standardised to enable consistent interpretation, and questionnaires are a simple method of ensuring this. Secondly, any samples taken as a result of environmental investigations should be recorded, whether from a case's home or a potential community source. Thirdly, results of microbiological tests on the cases and environmental samples from potential sources from laboratories will also need to be matched, where available, to each case source as the tests become available. Fourthly, communication messages can engage local media/doctors to warn the public of potential outbreak, inform them of suspect sources (where appropriate) and enable case finding thus leading to potential additional cases. If skills exist in the outbreak control team, then consideration should be given to the creation of a database to store contextual data in a coherent and robust manner. To facilitate data interpretation and management, the data should be collected in a standardised format during the outbreak response. Guidance for potential conventions for formatting key variables from patients and environmental sources can be found here.

The incubation period for Legionnaires' Disease has been reported to be up to 19 days and has been discussed elsewhere. Consideration of a pragmatic upper limit of 14 days for case exposure history, is suggested in this toolbox. However, this may vary operationally from outbreak to outbreak

Case Definitions

It is important to have clear case definitions during an outbreak, to enable comparison and analysis of cases and thus facilitate the public health response, particularly when working internationally. Guidance on formulating case definitions in an outbreak can be found here. A case definition could change during an outbreak from being rather broad at the outset to being more constrained as the outbreak develops and information changes, this is fine provided that changes are documented. It is critical that the case definition is explicitly stated when sharing information on disease occurrence.

Questionnaires

Three types of questionnaires could be used during an investigation into cases of Legionnaires' disease. These service particular needs during a response and are summarised in the table below. It should be noted that a patient may not be available or capable of responding to questions. In this case, a suitably proxy - perhaps a partner or family member - should be identified. Some of the questions, concerning demographic and clinical details, could be answered by medical staff at the hospital, if possible.

An interview logistics and guidance document is available here

A surveillance questionnaire, enables reporting and centralisation of case data, and should facilitate rapid awareness of outbreaks.

Once multiple cases are identified, consideration should be given to a more exhaustive form of questionnaire. Hypothesis generation is a critical step in an outbreak situation. By administering an open-ended hypothesis-generating questionnaire (trawling questionnaire) to some of the first case-patients, the investigator(s) will attempt to learn about potential exposures to known risk factors. General guidance is available on creating a hypothesis-generating questionnaire.

Following a review of the summary information taken from such a process (by for example a study of the line listing of an extract from a database) specific hypotheses about the cause of the outbreak can be formulated. It might be beneficial, in certain circumstances to perform an analytical study, where these hypotheses will become the basis of a more structured hypothesis-testing questionnaire (see for example the analytical study protocol).