Tinnitus Archive

Tinnitus Archive > Introductory Material > Limitations and Cautions

Limitations and Cautions


Data Limitations
  • Patient Self Reports
    Many items in the Tinnitus Archive are based on patient self-reports, and it is well known that such information is limited by several factors: (1) The patient's ability to recall past conditions or circumstances accurately; (2) the extent of the patient's knowledge of appropriate terminology; (3) the patient's willingness to reveal what might be potentially embarassing information. For example, patients may be reluctant to discuss details of their current health status, current employment issues, problems with family relationships resulting from tinnitus, and other types of personal information.
  • Sampling Biases

    • Tinnitus Severity: The OHSU Tinnitus Clinic sees patients who have tinnitus as their primary complaint. As a result, the Clinic population is biased toward the more severe cases of tinnitus.

    • Geographic: Although the Oregon Tinnitus Clinic patient population includes individuals from all 50 states and from many foreign countries, it is predominantly drawn from the Pacific Northwest (Oregon and Washington). Please see Methods: Subject Selection for each Data Set.

    • Racial: Prior to 1993, patients attending the Tinnitus Clinic were not queried reqarding racial or ethnic affiliation. Based on informal observation by Clinic staff, the large majority of Tinnitus Clinic patients have appeared to be "White, not of Hispanic origin." (Please also see Methods: Subject Selection for each Data Set.

    • Gender: Although it varies slightly from year to year, the male/female ratio of the Oregon Tinnitus Clinic is typically around 70% male / 30% female.

    • Specific Groups Excluded: (1) patients found to have medically treatable otologic conditions (e.g. active middle ear disease) are referred to other otologic services without evaluation or treatment at the Oregon Tinnitus Clinic; (2) patients with pulsating tinnitus that is synchronized with cardiovascular sounds are usually referred elsewhere for appropriate cardiovascular diagnosis and applicable surgical or medical treatment; (3) patients with profound hearing loss in both ears are unlikely to benefit from acoustic therapies offered by the Clinic and therefore are usually referred elsewhere for evaluation and potential medical treatment.

  • Selection Criteria for Tinnitus Archive Subjects
    Specific selection criteria were applied to all patients included as subjects in the Tinnitus Data Registry, and it is only those subjects whose data are included in the Archive data summaries. Please see Methods: Subject Selection for each Data Set.

Statistical Analyses: Cautions

  • Exact Format of Questionnaire Items
    The exact format of a question affects patients' responses to that question. In each data summary, please review the format of the questionnaire item carefully, in order to evaluate potential effects of the item format upon the responses summarized here. The exact wording of the relevant item is provided at the top of each data summary.
  • Coding Format
    The exact coding format for each item governs the Registry content for that item, and thus controls the content of data summaries included in the Archive. The coding format for each item can be viewed on the Forms page (under the heading CODING).
  • Sample Number for each Data Summary
    In most data summaries the actual number of patients available for analysis is slightly less than the Overall Sample Number for the Data Set in question, because there are usually a few patients who do not provide an answer for the item in question. By means of careful questionnaire design, and by reviewing patients' questionnaires with them when possible, we have tried to minimize such omissions.

    In a few cases, however, there are specific reasons why data could not be obtained from some subset of the patients. In those cases where the available patient sample is substantially less than the Overall Sample Number for the Data Set, the specific reason for the truncated sample size is stated in each data summary.