Robert M. Hamm, PhD, University of Oklahoma Health Sciences Center, Oklahoma City, OK, David E. Bard, MA, University of Oklahoma Health Sciences Center, Oklahoma City, OK, Dewey C. Scheid, MD, University of Oklahoma Health Sciences Center, Oklahoma City, OK, and Isaac M. Lipkus, Ph.D., Duke University Medical Center, Durham, NC.
Purpose. Some “50%” risk judgments may belie true beliefs because the respondent means “don't know”. We analyzed an existing data set to determine how often the 50% response is used in patients' probability judgments, and what it correlates with. Methods. 316 men ages 40 to 70 used a labeled probability scale to judge risks related to prostate cancer screening, before and after education on the topic. Available measures included risk judgment accuracy, prostate cancer knowledge, use of “don't know” on the prostate knowledge questions, demographics, and numeracy. Results. After education, prostate cancer knowledge and risk judgment accuracy increased, while 50% was used less often on the risk judgments and “don't know” was used less often on the knowledge questions. The proportion of 50% responses was negatively correlated with age (Pearson r = -0.189, df = 314, p = .001), income (Kendall tau = -0.172, df = 312, p < .001), and education (tau = -0.183, df = 314, p < .001). There were significant correlations between prostate cancer knowledge and the use of 50% on the risk judgment questions (before, r = -0.118, df = 314, p = .035; after, r = -0.254, df = 314, p < .001). Use of “don't know” on the knowledge test was related to use of “50%” on the risk judgments before (r = 0.241, df = 314, p < .001) and after (r = 0.204, df = 314, p < .001) the education. Participants with higher scores on a three-item test of proportions (numeracy measure) used the 50% response less frequently (r = -0.142, df = 266, p = .011). Use of 50% was not related to two other numeracy measures, TOFHLA and subjective numeracy. Conclusion. The data are consistent with the theory that patients use the 50% response on a labeled probability scale to mean “I don't know.” It was used less after education. Respondents who were more educated and more numerate (by 1 of 3 measures) used it less. Those who had more knowledge about the topic (prostate cancer screening) used it less. Those who responded “I don't know” on knowledge questions tended to respond “50%” on risk questions. Respondents should be given the opportunity to indicate they don't know, other than by saying “50%”.
See more of Oral Concurrent Session L - Risk Perception
See more of The 27th Annual Meeting of the Society for Medical Decision Making (October 21-24, 2005)
