Meeting Brochure and registration form      SMDM Homepage

Wednesday, October 24, 2007
P4-7

WHAT I SAID ISN'T WHAT I DID: THE INCONSISTENCY BETWEEN POST-DECISION AID INTENTIONS AND ACTUAL BEHAVIORS

Angela Fagerlin, PhD1, Brian J. Zikmund-Fisher, PhD1, Dylan Smith, PhD1, Holly A. Derry, MPH2, Jennifer B. McClure, PhD3, Sarah M. Greene, MPH3, Sharon Hensley Alford, MA4, and Peter A. Ubel, MD1. (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: Although numerous studies have tested the effectiveness of decision aids using short-term outcome measures, little is known about whether longer term outcomes match initial self-reports. We evaluated the impact of a tamoxifen decision aid on women's behavior and knowledge three months after its viewing.

Methods: 630 HMO patients at elevated risk for breast cancer (Gail Score ≥1.66) viewed a decision aid describing the risks and benefits of taking tamoxifen to prevent a first diagnosis of breast cancer and completed numerous post-test questions. Three months later, they received an e-mail requesting their participation in a follow-up survey. 335 women (53%) completed the follow-up survey. In both surveys, women answered questions about their tamoxifen related behavior. At 3-month follow-up, we asked women to report their risk of breast cancer (in percentage form), their decisional conflict, and their use of the decision aid in the past 3 months.

Results: After initially viewing the decision aid, 6% of women believed they would begin taking tamoxifen, 30% indicated they were likely to talk to their doctor, and 29% were likely to look for more information in the subsequent three months. However, only 3 women (1%) began taking tamoxifen to prevent breast cancer, only 6% of women talked to their doctor and only 5% sought more information about tamoxifen. In addition, only 16% returned to the website to view the decision aid during the 3 month time period, only 30% told anyone about the decision aid, and less than 5% showed anyone the website. At follow-up, women experienced little decisional conflict about whether or not to take tamoxifen (M=2.07, SD=2.07). Finally, only 40% of women could accurately recall (within 5 percentage points) their risk of breast cancer.

Conclusions: Women who viewed a prevention decision aid predicted they would perform more health behaviors related to the decision aid content than they reported actually doing 3 months later. Very few women returned to review the information or shared the information they learned with others, and their memory for personalized risk information explicitly presented in the decision aid was relatively poor. Future research should consider whether the observed inconsistency in behavior reflects weaknesses of current decision aid designs and/or the instability of patient preferences.


See more of Poster Session IV
See more of The 29th Annual Meeting of the Society for Medical Decision Making (October 20-24, 2007)