TESTING A DECISION AID FOR COLORECTAL CANCER SCREENING IN THE ELDERLY

Purpose: We developed and tested a decision aid targeted to older adults to help them decide whether to undergo colon cancer screening.

Methods: To develop the messages about colon cancer screening in the elderly we completed cognitive interviews with 25 adults age 75 and older with 3 or more chronic diseases assessing understanding, acceptability, and importance of the information. Based on this iterative process, we refined the messages and then developed a paper-based decision aid. We then tested the decision aid in a convenience sample of older adults age 75 and older with multiple chronic diseases using pre-post design. The primary outcome was the proportion of participants making an informed choice. An informed choice was prospectively defined as adequate knowledge (10/15 questions correct) and clear values (25 or less on values clarity subscale of decisional conflict scale); secondary outcomes were decisional conflict, and screening intent.

Results: 46 adults participated in our study; mean age was 83; 85% were women; 24% African-American; and 35% had inadequate or marginal health literacy. Mean number of chronic conditions was 6 (range 0-14). The decision aid increased knowledge: 4% responded correctly to >= 10/15 knowledge questions before the decision aid compared to 52% after using the decision aid (p<0.01). Literacy was associated with greater knowledge after decision aid use: 64% with adequate literacy reached the knowledge threshold compared to 31% with lower literacy (p=0.06). The proportion having clear values did not differ significantly after decision aid viewing (61% before and 70% after, p=0.34). However, viewing the decision aid increased the likelihood of making an informed choice: 4% fulfilled the criteria for an informed choice prior to the decision aid vs. 41% afterwards (p<0.01). Adequate literacy was associated with informed choice: 54 with adequate vs. 19% with lower literacy (p=0.03). Decisional conflict was low prior to the decision aid, but decreased further after its use (pre mean score=34; post mean score=28, p<0.01). Prior to viewing the decision aid, most participants indicated that they would have screening (67%); after viewing the decision aid, 61% intended to have screening (p=0.45).

Conclusions: The decision aid increased knowledge, especially for those with adequate literacy, and had smaller effects on decisional conflict, clarity of values, and screening intent.