For most of medical history, the menstrual cycle has been treated as a reproductive signal and not much else. Something to monitor when trying to conceive, or to avoid pregnancy. The assumption, largely unexamined, has been that the roughly 450 cycles a woman experiences across her reproductive life are only medically interesting during the small fraction spent trying to get pregnant.
New research suggests that assumption has been leaving an enormous amount of health information on the table.
A new tool built for a neglected dataset
A team of researchers has developed an algorithm called WAVES, short for Women’s Health Assessment through Variability in Endocrine-Related Signals, designed to extract detailed health metrics from physiological signals that follow menstrual patterns. The primary signal it analyzes is basal body temperature, the temperature measured first thing in the morning before any movement or activity.
The tool captures 32 distinct metrics organized across four dimensions, covering temperature levels, timing, within-cycle variations and the overall shape of the temperature wave across a full cycle.
To test WAVES, researchers analyzed more than 5,600 menstrual cycles from 753 participants between the ages of 18 and 42. The data came from a historical collection gathered across seven European centers during the 1990s, with participants recording daily temperature measurements and observations over periods ranging from several months to a few years. No single participant was tracked across her entire reproductive lifespan, but the combined dataset created a picture spanning most of the reproductive years.
The motivation behind the tool was a well-documented gap in how menstrual health research has historically been conducted. Nearly all clinical training, research funding and consumer technology in this space has been built around the roughly one percent of cycles that are conceptive. The other 99 percent have been treated as essentially unremarkable.
What actually changes as women age
When researchers compared participants in two age groups, those between 18 and 35 and those between 35 and 42, they found that 27 of the 32 metrics differed significantly between them.
Older participants showed higher average temperatures across the full cycle, in both the follicular phase before ovulation and the luteal phase after it. Their overall cycle length was shorter, driven specifically by a compression of the follicular phase rather than the luteal phase. The temperature swing between the two phases was also smaller, meaning the rise and fall across the cycle became less pronounced with age. Multiple metrics became more variable overall, including cycle length, luteal phase duration and several measures of temperature fluctuation.
Within individuals tracked over time, 16 of the 32 metrics shifted in measurable ways as participants aged. The follicular phase shortened gradually each year, the lowest temperatures of the cycle crept upward and the amplitude of the cycle’s rhythm decreased incrementally. These findings align with what is already understood about how the ovarian follicle supply declines steadily over time, with only a small fraction of the original reserve remaining by the early forties. The menstrual cycle, it turns out, reflects that biological progression in ways that can be captured and measured.
Your cycle has a personal signature
One of the more striking findings in the research concerns what the authors describe as an individual footprint. Many cycle characteristics are not simply population-level averages that everyone clusters around. They are deeply personal traits that remain remarkably consistent within a single person over time.
Metrics related to temperature levels showed particularly strong individual consistency, meaning that a person’s follicular temperature, luteal temperature and overall temperature range tend to be more stable within them across multiple cycles than they vary between different people. Cycle length and phase durations showed similar patterns of individual stability.
This has meaningful practical implications. Only a small fraction of people actually have a 28-day cycle, and ovulation timing varies considerably even among those who do. Comparing an individual’s cycle to a textbook average is therefore far less informative than tracking that individual’s own patterns over time and watching for meaningful shifts. Research has also found that variability in cycle length carries information about future cardiovascular health outcomes, suggesting the cycle’s relevance extends well beyond the reproductive system.
What this means for how you track your cycle
The researchers are clear that a personalized approach is not simply preferable but may be necessary for menstrual health monitoring to be genuinely useful. The more valuable question is not whether your cycle matches a population standard but whether your own pattern is shifting in ways that warrant attention.
A cycle that has always run longer than average is not inherently problematic. A cycle that was consistent for years and has begun changing noticeably across multiple months is worth discussing with a healthcare provider, particularly if other patterns are shifting at the same time.
Temperature tracking through wearables and dedicated apps can capture the kind of data the WAVES algorithm analyzes, offering a richer picture than period dates alone. The algorithm itself is open-source and available for researchers, which means it could eventually be embedded in the consumer tools many people already use.
Where the research could go next
The study authors identify several conditions where menstrual cycle metrics could eventually serve as accessible and noninvasive health markers, including polycystic ovary syndrome, endometriosis and ovarian cancer. The infrastructure to support this kind of monitoring at a population level is already taking shape through the growing availability of wearable technology that tracks physiological data continuously.
What the research ultimately argues is that the menstrual cycle has been underestimated as a health signal for decades. The data has always been there. The tools to read it meaningfully are only now catching up.
