Objective-Hot flashes are valuable indicators of physiologic condition and drug effect; however, subjective and objective measures do not always agree. No study has examined both subjective and objective hot flashes in women prescribed aromatase inhibitors. The study (1) compared subjective and objective hot flash measures, (2) examined changes in subjective and objective hot flashes over time, and (3) evaluated predictors of change in hot flashes in aromatase inhibitor-treated women.Design-Subjects (n=135) were enrolled in a randomized clinical trial comparing exemestane and letrozole for the treatment of breast cancer. Hot flashes were assessed prior to starting drug and 1, 3, and 6 months later. Subjects wore a sternal skin conductance monitor for ≥ 24 hours at each time point. With each perceived hot flash, women pressed an event button and rated intensity and bother in a paper diary.Results-Subjects were a mean age of 60 years and mainly Caucasian (92%). Across time points, monitor hot flashes were (1) significantly more frequent than diary and/or event button flashes (p<. 05) and (2) moderately correlated with subjective measures (.35< r <.56). Monitor hot flashes did not significantly change over time with aromatase inhibitor therapy, whereas both diary and event button frequency significantly varied but in dissimilar patterns (51% non-linear). No consistent predictors of hot flashes across measures or time points were identified.Conclusions-Findings indicated dissimilarities between subjective and objective measures of hot flashes. Despite statistical significance, there was little clinically meaningful change in hot flashes after initiating aromatase inhibitor therapy.Address for correspondence and reprints: Julie L. Otte, PhD, RN, Indiana University, 1111 Middle Drive NU236, Indianapolis, IN 46220, (email) jlelam@iupui.edu. Objective and subjective hot flash measures were analyzed after initiation of two aromatase inhibitors for the treatment of breast cancer. Findings indicated dissimilarities among measures, hot flash frequency and hot flash intensity and bother did not significantly change over time, diary and event button flashes significantly changed but in dissimilar patterns, and few consistent predictors of change in hot flashes were identified.