Castilho Silvia Diez, Nucci Luciana Bertoldi, Assuino Samanta Ramos, Hansen Lucca Ortolan
Departamento de Pediatria, Faculdade de Medicina, Pontifícia Universidade Católica de Campinas.
Faculdade de Medicina, PUC-Campinas.
Arq Bras Endocrinol Metabol. 2014 Jun;58(4):394-7. doi: 10.1590/0004-2730000003080.
To compare the age at menarche obtained by recall method according to the time elapsed since the event, in order to verify the importance of the recall bias.
Were evaluated 1,671 girls (7-18 years) at schools in Campinas-SP regarding the occurrence of menarche by the status quo method (menarche: yes or no) and the recall method (date of menarche, for those who mentioned it). The age at menarche obtained by the status quo method was calculated by logit, which considers the whole group, and the age obtained by the recall method was calculated as the average of the mentioned age at menarche. In this group, the age at menarche was obtained by the difference between the date of the event and the date of birth. Girls who reported menarche (883, 52.8%) were divided into four groups according to the time elapsed since the event. To analyze the results, we used ANOVA and logistic regression for the analysis, with a significance level of 0.05.
The age at menarche calculated by logit was 12.14 y/o (95% CI 12.08 to 12.20). Mean ages obtained by recall were: for those who experienced menarche within the previous year 12.26 y/o (±1.14), between > 1-2 years before, 12.29 y (±1.22); between > 2-3 years before, 12.23 y/o (±1.27); and more than 3 years before, 11.55y/o (±1.24), p < 0.001.
The age at menarche obtained by the recall method was similar for girls who menstruated within the previous 3 years (and approaches the age calculated by logit); when more than 3 years have passed, the recall bias was significant.
根据初潮事件发生后的时间间隔，比较通过回忆法得出的初潮年龄，以验证回忆偏倚的重要性。
对圣保罗州坎皮纳斯市学校的1671名7至18岁女孩进行评估，采用现状法（初潮：是或否）和回忆法（对于提及初潮的女孩，记录初潮日期）来了解初潮情况。通过逻辑回归计算现状法得出的初潮年龄，该方法考虑了整个群体；回忆法得出的初潮年龄则计算为所提及初潮年龄的平均值。在这个群体中，初潮年龄通过初潮事件日期与出生日期之差得出。报告初潮的女孩（883名，占52.8%）根据初潮事件发生后的时间间隔分为四组。为分析结果，我们采用方差分析和逻辑回归进行分析，显著性水平为0.05。
通过逻辑回归计算得出的初潮年龄为12.14岁（95%置信区间为12.08至12.20）。回忆法得出的平均年龄分别为：前一年初潮的女孩为12.26岁（±1.14），1至2年前初潮的女孩为12.29岁（±1.22），2至3年前初潮的女孩为12.23岁（±1.27），3年多前初潮的女孩为11.55岁（±1.24），p<0.001。
对于前3年内月经初潮的女孩，回忆法得出的初潮年龄相似（且接近逻辑回归计算出的年龄）；当超过3年时，回忆偏倚显著。
