PREMENSTRUAL-SYNDROME - COMPARISON BETWEEN DIFFERENT METHODS TO DIAGNOSE CYCLICITY USING DAILY SYMPTOM RATINGS

Background. In the diagnosis of premenstrual syndrome (PMS) the techni que of daily prospective symptom ratings is often used. Several method s of assessing cyclicity, based on the daily prospective symptom ratin gs, have been presented in the literature. In this paper we compare fo ur different methods to assess cyclicity. Methods. Eighty consecutive patients seeking help for PMS at the Department of Gynecology complete d daily symptom ratings using a visual analogue scale (VAS) during one menstrual cycle. They also made an Eysenck Personality Inventory (EPI ) and a thorough case history was taken regarding earlier psychiatric case history. The methods compared were: a) the nonparametric Mann-Whi tney U-test, b) effect size, c) Run-test and d) a 30% change in sympto m degree between the follicular and the luteal phase calculated in two different ways. Results. There was good agreement in the number of cy clic and non-cyclic patients between the different diagnostic methods used with exception of the 30% of change methods as the criteria for c yclicity. Here the number of non-cyclic patients became higher and low er, respectively, than with the other methods. The correlation between the number of symptoms for each patient showing cyclicity was high in all tests. When comparing the median neuroticism score of the EPI-N i nventory the non-parametric, the Run-test and the effect size showed s ignificant differences between the non-cyclic group and the group of p atients with pure PMS, but not the 30%-change methods. Frequency of pa tients with psychiatric history showed similar results. Conclusions. T hree of the methods used seem to identify the same patients as having or not having cyclical changes and probably also then finding the same biological and/or psychological factor being responsible for the cycl icity. There were, however, some differences in outcome of validity te sting and the 30%-change methods seem less Valid than the other three methods. Due to its simplicity and theoretical/statistical advances th e Run-test seems most preferable.