Patient preferences for a hospital-based rheumatology Interactive Health Communication Application and factors associated with these preferences.
(33/103)
Does use of the adjuvant! model influence use of adjuvant therapy through better risk communication?
(36/103)
Adjuvant! is a model that provides recurrence and mortality risk predictions for patients with breast cancer considering adjuvant therapies. Although low-risk patients who saw Adjuvant! chose adjuvant therapy less frequently, whether this was because of educational or other aspects of the decision aid is unknown. The authors explored whether Adjuvant! affects choice of therapy through increased patient knowledge. A subset of data were analyzed from a cluster randomized trial in which oncology practices in 2 major United States cities were randomly assigned to use either Adjuvant! or an informational pamphlet to educate patients. Of 405 patients, 48 were low-risk, with 28 assigned to the decision aid and 20 to the pamphlet. Among the low-risk patients, using frequency tables and Fisher exact tests, the authors explored whether Adjuvant! was associated with more accurate patient estimates of survival; whether accuracy was associated with treatment choice; and whether, after controlling for accuracy, any remaining association was seen between Adjuvant! and treatment choice. Adjuvant! was associated with more accurate estimates of baseline prognosis compared with the pamphlet (57% vs. 25%; P = .04). Patients who had more accurate estimates of baseline prognosis were less likely to choose adjuvant therapy (62% vs. 89%; P = .04). After controlling for accuracy, no statistically significant association was found between the use of Adjuvant! and adjuvant therapy (P = .59 and P = .11 for inaccurate and accurate patients, respectively). Adjuvant! seems to influence patient choice through educational rather than other means of persuasion. However, many patients held inaccurate risk perceptions after viewing Adjuvant!. (+info)
Predicting disease risks from highly imbalanced data using random forest.
(37/103)