%0 Journal Article
%T Reliability of assessing lifestyle and trigger factors in patients with migraine--findings from the PAMINA study.
%A Zebenholzer K
%A Frantal S
%A Pablik E
%A Lieba-Samal D
%A Salhofer-Polanyi S
%A Wöber-Bingöl C
%A Wöber C
%J Eur J Neurol
%V 23
%N 1
%D Jan 2016
%M 26228627
%F 6.288
%R 10.1111/ene.12817
%X <B>OBJECTIVE: </B>Numerous lifestyle factors are blamed for triggering migraine attacks. The reliability of assessing these factors retrospectively is unknown. Therefore, retrospective and prospective assessments of lifestyle in general and of migraine triggers in particular were compared in patients with migraine.<BR><B>METHODS: </B>At baseline, the patients filled in two questionnaires covering the previous 90 days. Thereafter they kept a prospective 90-day diary. Questionnaires and diary included the same set of 45 factors. In the first questionnaire the patients assessed their lifestyle, in the second they rated for each factor the likelihood of triggering a migraine attack, and in the diary they recorded the daily presence of these factors irrespective of headache. Five categories were used for comparing frequencies in questionnaire and diary, defining agreement as identical categories in diary and questionnaire, minor disagreement and major disagreement as overestimation or underestimation by one category and two or more categories, respectively.<BR><B>RESULTS: </B>In all, 327 patients (283 women, age 41.9 ± 12.1 years) who recorded 28,325 patient days were included. Calculating for each factor the percentage of patients with major disagreement the mean proportion was larger for trigger factors than for lifestyle (38.7% ± 6.6% vs. 16.9% ± 6.4%, P < 0.001). The proportion of factors showing major disagreement in more than 20% of the patients was 8.8% for lifestyle but 94.1% for trigger factors (P < 0.001).<BR><B>CONCLUSIONS: </B>Comparing questionnaire and diary assessments of lifestyle and trigger factors in patients with migraine shows that questionnaire assessment of lifestyle is reliable, whereas trigger factors are overestimated and/or underestimated in retrospective questionnaires.