Validity of Self-Reported Menstrual Cycle Length

Area of Science:

• Reproductive Health
• Epidemiology
• Biostatistics

Background:

• Menstrual cycle length is a key indicator of reproductive and overall health.
• The accuracy of self-reported menstrual cycle data has been a subject of ongoing debate.
• Previous studies have questioned the reliability of self-reported menstrual cycle length.

Purpose of the Study:

• To assess the agreement between self-reported and calculated menstrual cycle length.
• To evaluate the validity of self-reported menstrual cycle data in premenopausal women.
• To provide empirical evidence for the use of self-reported menstrual cycle length in research.

Main Methods:

• Utilized data from the Cancer Prevention Study-3, including premenopausal participants.
• Self-reported cycle length was categorized, while calculated length was based on prospectively recorded start dates.
• Agreement was analyzed using descriptive statistics, Spearman's correlation, and weighted kappa, incorporating a ±7-day buffer for biological variability.

Main Results:

• Exact agreement was found in 44.9% of participants; this increased to 82.2% with a ±7-day buffer.
• Spearman's correlation improved from weak (ρ=0.17) to moderate (ρ=0.53) with the buffer.
• Weighted kappa indicated poor agreement (κ=0.15) without the buffer, improving to fair (κ=0.33) with it. Agreement varied across demographic subgroups.

Conclusions:

• Self-reported menstrual cycle length demonstrates fair agreement with calculated length when biological variability is accounted for.
• The findings support the use of self-reported menstrual cycle length in epidemiologic studies.
• This study provides justification for incorporating self-reported menstrual data in menstrual-related analyses.