Objective: To quantify precision in patient reports of different attributes of dizziness.
Patients and methods: In a cross-sectional study, we interviewed consecutive adult patients with dizziness at 2 urban academic emergency departments (EDs) from July 2, 2005, to August 26, 2005. We excluded patients who were too sick for an interview or who posed a risk to the interviewer. We included those who were "dizzy, light-headed, or off-balance" for 7 days or less or previously "bothered" by the same conditions. We assessed descriptions of dizziness quality elicited by 4 questions in different formats (open-ended, multiresponse, single-choice, and directed). Clarity was assessed qualitatively (vague, circular) and quantitatively (overlap of types of dizziness). Consistency was measured by frequency of mismatched responses across question formats. Reliability was determined by test-retest.
Results: Of 1,342 patients screened, 872 (65%) were dizzy, light-headed, or off-balance in the past 7 days (n=677) or previously bothered by dizziness (n=195). Among these 872 patients with dizziness, 44% considered dizziness "the main reason" or "part of the reason" for the ED visit. Open-ended descriptions were frequently vague or circular. A total of 62% selected more than 1 dizziness type on the multiresponse question. On the same question, 54% did not pick 1 or more types endorsed previously in open description. Of 218 patients not identifying vertigo, spinning, or motion on the first 3 questions, 70% confirmed "spinning or motion" on directed questioning. Asked to choose the single best descriptor, 52% picked a different response on retest approximately 6 minutes later. By comparison, reports of dizziness duration and triggers were clear, consistent, and reliable.
Conclusion: Descriptions of the quality of dizziness are unclear, inconsistent, and unreliable, casting doubt on the validity of the traditional approach to the patient with dizziness. Alternative approaches, emphasizing timing and triggers over type, should be investigated.