The Statistical Analysis of Menstrual Bleeding Patterns

Abstract

Conclusion and summary: The proposed method of bleeding pattern analysis avoids the problem of defining the menstrual cycle. The difficulty of maintaining a cyclical approach becomes acute when methods of fertility regulation are used which are known to disrupt seriously previously experienced bleeding patterns.By using eight simple measures relating to the pattern of bleeding over a fixed period of time, it is possible to present sensitive indicators previously available but not used. These indicators will permit a more meaningful comparison of data obtained from different groups of women or for the same woman over time. The eight measures relate to the number, average duration and variability of bleeding episodes taking place within a defined period of time; the variability of the bleeding-free interval; the average duration and variability of the bleeding episode and its succeeding bleeding-free interval; and the number of days on which bleeding has occurred.Agreement was reached by the Workshop participants regarding the terms to be used, the definition of these terms, the statistical approach to be adopted and the analytical procedures to be followed in the analysis of bleeding patterns. It is hoped that this alternative approach to bleeding pattern analysis will assist researchers, clinicians and other investigators in the collection, analysis and interpretation of menstrual bleeding patterns.An example of this approach is given in the following paper by D. Mayes. This example is based on data in the form of diary cards made available to the Family Planning Research Unit, Exeter University, by John Guillebaud of the Churchill Hospital, Oxford.