Patient preferences for personalized (N-of-1) trials A conjoint analysis
OBJECTIVE: Despite their promise for increasing treatment precision, Personalized Trials (i.e., N-of-1 trials) have not been widely adopted. We aimed to ascertain patient preferences for Personalized Trials.
STUDY DESIGN AND SETTING: We recruited 501 adults with ≥2 common chronic conditions from Harris Poll Online. We used Sawtooth Software to generate 45 plausible Personalized Trial designs comprised of combinations of 8 key attributes (treatment selection, treatment type, clinician involvement, blinding, time commitment, self-monitoring frequency, duration, cost) at different levels. Conditional logistic regression was used to assess relative importance of different attributes using a random utility maximization model.
RESULTS: Overall, participants preferred Personalized Trials with no costs vs. $100 cost (utility difference 1.52 [standard error 0.07], p<0.001) and with less vs. more time commitment/day (0.16 [0.07], p<0.015), but did not hold preferences for the other 6 attributes. In subgroup analyses, participants ≥65 years, white, and with income ≤$50,000 were more averse to costs than their counterparts (p all <0.05).
DISCUSSION: To optimize dissemination, Personalized Trial designers should seek to minimize out-of-pocket costs and time-burden of self-monitoring. They should also consider adaptive designs that can accommodate subgroup differences in design preferences.
Moise, N., Wood, D., Kuen K Cheung, Y., Duan, N., Onge, T. S., Duer-Hefele, J., ... members of the “Personalized Trial Collaboratory” (2018). Patient preferences for personalized (N-of-1) trials: A conjoint analysis. Journal of Clinical Epidemiology, 102, 12-22. https://doi.org/10.1016/j.jclinepi.2018.05.020