Abstract
Background The major, common adverse health outcomes for patients with diabetes come from macrovascular, not microvascular, disease, including death, heart attack and stroke. Personalized risk information might help patients make better decisions about prioritizing health behavior change. In this study, we explored diabetic subjects’ beliefs about diabetes-related risks and sources of information along with reactions to presentation of personalized risk information.
Methods We recruited 56 English- and Spanish-speaking adults with diabetes and at least one other major cardiovascular risk factor from a community health center and an academic family medicine practice. In in-depth interviews, we explored sources of risk information, risk-reduction activities, and behavioral intentions. Subjects also ranked 6 diabetes-associated adverse outcomes based on perceived risk, reviewed personalized risk predictions from the UKPDS Outcomes Model, and re-ranked these perceived risk of outcomes. We explored their reasons for changing/not changing risk rankings. Qualitative analysis was used to develop themes and concepts underlying subjects’ risk perceptions and reactions to risk information.
Results Virtually all subjects believed at least some diabetes-related adverse outcomes were modifiable or preventable. A substantial minority expressed a fatalistic view that at least some were not preventable, while others felt that risk factor control meant an outcome was “not going to happen.” A very common theme was that a “warning shot” would occur before many outcomes, providing time and impetus to change bad habits. Provider urging was the other commonly-cited motivator and about 20% of subjects cited each as motivators for past change. While providers were cited as a key source of information, vicarious experience with relatives and acquaintances was cited just as often and ¾ of subjects felt they knew themselves and their risks better than providers or risk models. Subjects reported the personalized risk information was interesting, but <⅓ changed their rankings of mortality (nearly always initially ranked least likely) to match the model predictions (nearly always the highest probability).
Conclusions Our subjects often based their risk perceptions on anecdote and gut feelings, not medical information. Personalized risk information does not appear promising for motivating behavior change or even altering risk perceptions.
