Interview logistics and guide

Interviewing is an important, though sometimes difficult task. There are several things to be covered in interviewer training:

• Interviewers should be provided with an overview of the outbreak situation and review the purpose of the questionnaire.
• Interviewers should also be aware of the respondent selection process.
• Respondents will often ask the interviewer how they got their contact information.
• A majority of the training should focus on the questionnaire - how to use it, the intent and meaning of each question, and how to record or code responses.
• Discuss how the interviewer should respond to questions from the respondents.
• Questions should always be asked in the same way for each participant. The interviewer should not prompt or lead the respondent to answers, which could introduce bias into the study.

The logistics of conducting the interviews should be agreed at the start of the outbreak investigation. This includes the hours during which it is acceptable to call, how to track the calls, how many times should the interviewers call a prospective respondent, whether they should leave a message if they get an answering machine, and what to do with completed questionnaires. Finally, and most importantly, discuss confidentiality of the interviews and questionnaires. At the interviewer training, it is good practice to provide materials to the interviewers in a manual (what exactly is included in a manual will depend on the outbreak situation). You might include a calendar to help track dates, a map of a facility, or guidelines related to but not directly associated with the outbreak that might be useful if there are questions. Also, you might create a list of frequently asked questions and answers that can be used by interviewers as a quick reference tool. Interviewers might also find it useful to have some background information available on the outbreak - the organism, an epidemic curve, etc. Depending on each outbreak scenario, decide which interview method would be most appropriate (face-to-face, postal or telephone) and why. The relative strengths and weaknesses of these approaches are summarised in the table below:

Method	Advantage	Disadvantage
Face to face interview	• Highest response rate • Allows collection of complex data	• Time consuming • Resource intensive
Postal questionnaire	• Least resource intensive	• Lowest response rate • Slow process
Telephone interview	• Relatively high response rate • Relatively quick	• Respondent must have a phone
Online self administered	• Quickest method of data collection	• Respondents must have the technology • Data security issues

If data are to be collected by a postal questionnaire, particular attention will need to be given to the design of the questionnaire. Clear instructions and formatting are important to aid the navigation and completion of postal questionnaires. The questionnaire will need to be short and avoid open or complex questions. Unlike in research studies, there may not be time for a thorough pretesting or piloting of the questionnaire before it is used to collect data. However, where possible the questionnaire should be at least completed by a convenient sample of colleagues or others before use to identify issues with presentation, wording, navigability or content. Whichever method is chosen, data collection should be preceded by an interviewer script (or cover letter for postal questionnaires) explaining who is conducting the investigation and why the investigation is taking place, giving assurances on data confidentiality and security, and thanking respondents for their participation. The questionnaire should be checked after completion, as this is the best time to clarify anything with the respondent. In phone interviews, a courteous and knowledgeable interviewer can be the difference between a hang-up and a completed questionnaire. By maintaining a professional, but friendly, approach throughout an interview, an interviewer can obtain important information that will help investigators identify the cause of an outbreak.

At the draft analytical study protocol stage itself, it is useful to decide how to manage non-responders. They can be either withdrawn from the study or treated as missing values.