Meeting Brochure and registration form      SMDM Homepage

Wednesday, October 24, 2007 - 10:30 AM
J-1

PICTOGRAPHS CORRECT BIASES THAT REDUCE PATIENT KNOWLEDGE OF SIDE EFFECT RISKS: RESULTS FROM AN RCT OF A TAMOXIFEN DECISION AID

Brian J. Zikmund-Fisher, PhD1, Peter A. Ubel, MD1, Dylan M. Smith, PhD1, Holly A. Derry, MPH2, Jennifer B. McClure, PhD3, Azadeh Stark, PhD4, Rosemarie K. Pitsch, MPH1, and Angela Fagerlin, PhD1. (1) VA Ann Arbor Healthcare System & University of Michigan, Ann Arbor, MI, (2) University of Michigan, Ann Arbor, MI, (3) Group Health, Seattle, WA, (4) Henry Ford Health System, Detroit, MI

Purpose: To experimentally test whether using pictographs (image matrices), incremental risk formats, and varied risk denominators in a decision aid would increase or decrease women's comprehension of statistical information about side effect risks.

Methods: We recruited 631 women at high risk for breast cancer (5 year risk ≥1.66%) from two HMOs to view "Guide to Decide," an Internet-administered decision aid describing the use of tamoxifen to prevent a first case of breast cancer. In the section that discussed the side effects of tamoxifen, participants saw individually-tailored estimates of their own risks of endometrial cancer, blood clotting, cataracts, hormonal symptoms, and sexual problems. We randomly varied the presentation format of the risk statistics using a three factor design: (A) Risk information was displayed either in pictographs or numeric text; (B) the decision aid either reported total risks with and without tamoxifen or highlighted the incremental risk (e.g., "2 more women out of 100 would now get cataracts") and (C) risk estimates used either a 100 or 1000 person denominator. Primary outcome measures included risk perceptions and gist knowledge.

Results: When the decision aid presented risk statistics in numerical text form, participants had reduced knowledge scores if the display also used either (a) incremental risk formats or (b) 100 person denominators. In contrast, pictograph presentation of the same information significantly improved comprehension, erasing knowledge deficits for both incremental risk formats (M = 3.11 vs. 2.49 on a 0-4 scale, t = -3.47, p < 0.001, Cohen's d = 0.40) and 100 person denominators (M = 2.98 vs. 2.61, t = -2.13, p = 0.03, Cohen's d = 0.24). Although numeracy (as measured using the Subjective Numeracy Scale) was highly predictive of participants' knowledge scores, similar patterns occurred with both high-numerate and low-numerate participants.

Conclusions: Pictographs can often eliminate biases caused by other design factors. Because pictographs make risk statistics more concrete and easier to interpret, patients may draw more meaning from the statistics presented to them. We therefore suggest that the use of such graphs is essential for supporting informed treatment decision making in patient decision aids and all other types of education materials that present side effect risks to patients.