Tinnitus Archive

Tinnitus Archive > Data Sets > 2 > Methods > Questionnaire development and administration

Questionnaire development and administration


Questionnaire development

One of the major purposes of the Tinnitus Data Registry was to learn more about previously-undescribed aspects or characteristics of tinnitus. The existing literature prior to 1980 provided relatively little quantitative information concerning the actual range of subjective variations to be expected within the patient population. Therefore the patient questionnaires were designed to provide maximum scope and flexibility in regard to the range of answers allowed. To capture the individual nature of patients' symptoms and tinnitus histories, the questionnaires allowed the patients as much freedom as possible in describing the subjective nature of their tinnitus sensations.

An initial set of three questionnaires, dealing with Tinnitus Description and History, Hearing History and Occupational Exposure, and Medical and Health Information, was designed in 1980-81. These questionnaires were based on 5 years' experience evaluating a total of 1806 OHSU Tinnitus Clinic patients who attended the Clinic between 1976 and 1980. In addition the tinnitus literature was reviewed for relevant information, particularly from previous studies of tinnitus involving patient questionnaires (Nodar, 1978; Hazell, 1979a). Further guidance for the development process was obtained from a diverse sample of audiologists and otologists in the U.S. and the U.K. The three questionnaires are still in use at the Oregon Tinnitus Clinic, although minor revisions and some additions have occurred over the years.

Thanks are due to a large number of audiologists, otologists, and other health professionals who generously sent us their comments. Particular mention should be made of R.R.A. Coles, Edwin C. Everts, Jonathon W.P.Hazell, John W. House, Alexander Schleuning, James D. Smith, and W. Dixon Ward, for their numerous helpful suggestions.

Questionnaire structure and administration
  1. Multiple-choice questions (consisting of a fixed set of response alternatives) were used when feasible to facilitate coding of answers for computer data entry. Patients were allowed to select the answer "Other" if appropriate and were asked to explain that answer whenever they selected it.

  2. Open-ended questions (i.e. no specific response alternatives provided) were used for a few questions where it was important to allow the patient maximal flexibility in answering. Open questions also have the advantage that they do not suggest any particular answer(s) as being typical or appropriate, thus avoiding potential response bias.

  3. The three questionnaires presented to subjects were:

  4. The questionnaires were sent to prospective OHSU Tinnitus Clinic patients to fill out about 2 weeks before their scheduled Tinnitus Clinic visit. Many patients found it helpful to have an extended period in which to think about their answers, particularly in regard to details of their medical history. Advance mailing of the questionnaires allowed time for discussion of medical details with the patient's spouse or other relatives.

Format of Questionnaire Items

Single response items - Most questionnaire items in the OHSU Tinnitus Clinic questionnaires permitted only one answer per patient, such as the following example:

	Did you become aware of your tinnitus suddenly or more gradually?

		Suddenly (1 week or less)....................................1
		More gradually...............................................2
		Do not know..................................................3

The respondent was permitted to check or circle only one of the three responses provided.

Multiple response items - There were, however, several questions in the Oregon Tinnitus Clinic questionnaires that permitted the respondent to cite more than one answer. (A common example of such multiple-response questions occurs in standard medical history forms where patients are asked to list any surgeries they have experienced previously.)

Multiple-response items are problematic when computing frequency distributions. Because the frequency distribution for each question shows the number or percentage of patients who checked each of the various response options for that question, it is clear that patients who cite more than one response option will be counted more than once (i.e. once for each answer they checked). The total number of answers for such a multiple-answer question is therefore greater than the total number of patients.

In general we have handled that difficulty by inserting the warning symbol: intersecting groups.gif: Certain questionnaire items permit patients to provide multiple responses (e.g. "IN THE LIST BELOW, PLEASE CHECK SOUND(S) THAT MOST CLOSELY RESEMBLE YOUR TINNITUS:"). When such items are tallied, patients with multiple tinnitus sounds will be counted multiple times, thus raising the response total to a value greater than 100% whenever the frequency distribution includes patients who selected more than one response option. Following is a list of all multiple-response items currently included in the Archive:

Onset Factors Reported
Predominant Tinnitus Sound(s)
Tinnitus Localization: All Locations Reported

Confirmatory interviews

A fundamental aspect of quality-control for patient self-reports was to make sure that each patient (1) understood the question, and (2) wrote the answer that he/she actually intended. Questionnaires filled out by patients often contain errors, particularly if the questionnaires are lengthy and detailed. In the case of tinnitus, where the relevant terminology is unfamiliar to most patients, it is important to corroborate the written information by a structured interview. Following are the major aspects of the confirmatory interview conducted for the Tinnitus Data Registry:

  1. Interviews were conducted by staff members of the Registry who received special training in how to conduct the confirmatory interview.
  2. Efforts were made to ensure consistency between interviewers, both in regard to the use of terminology and in regard to the scope and duration of the interviewer-patient interaction.
  3. Interviewers recorded the results of the interview in the form of written notes, and discussed these notes subsequently with other staff members in order to achieve consensus on potential areas of disagreement.

We considered the confirmatory interview so important to the quality of the data in the Registry that we did not include patients for whom the interview could not be conducted (see Methods: subject selection: criteria for inclusion of subjects in the Tinnitus Data Registry).