Hormones, Data, and the AI Breakthrough Traditional Medicine Missed

For decades, women’s health has long been underrepresented in clinical research – understudied, oversimplified, and forced to fit models built around male biological averages. Conditions shaped by hormonal cycles, for instance, have long been reduced to static clinical snapshots that miss longitudinal patterns and misread symptoms, contributing to delayed and wrongful diagnoses.

It was only three decades ago that the FDA allowed women back into clinical research studies as participants, yet, women still account for 30% of trial participants – largely due to the pharmaceutical industry’s overfocus on their childbearing potential and fertility concerns.

The consequences transcend representation. A study by the University of Pittsburgh School of Public Health concluded, in fact, that most drug safety profiles are based on male biology, and fail to account for female hormonal variation. In cardiovascular disease research particularly, the underenrollment of women has directly contributed to delayed diagnoses, dismissed symptoms, and systemic misdiagnoses.

Since 2015, clinical research has made meaningful strides in sex equity, however, as the National Institutes of Health (NIH) released a policy to consider sex as a biological variable in research – entailing the need to consider both female and male variables in further scientific inquiry.

Still, the lag between policy and practice has been substantial. It is only now that a broader surge in clinically rigorous research explicitly accounting for female biology has begun to materialize. And with it, a convergence of institutional investment and artificial intelligence that promises to reshape the field.

The Institutional Wake-Up Call

As clinical data accumulates and new research gaps become increasingly consequential, a new alignment between science and technology is forming, oriented not toward treating women’s biology as a variation but towards systems built from the group up around biological sex differences.

Institutions have formalized that shift. The American Heart Association has strengthened its focus on women’s cardiovascular health anchored by a $75 million USD Go Red for Women Venture Fund and a $15 million research initiative examining the effects of menstrual cycles on cardiovascular-kidney-metabolic (CKM) health.

Simultaneously, the 2024-2028 NIH-Wide Strategic Plan and a White House Executive Order are directing over $100 million USD in ARPA-H funding toward what researchers have deemed “metabolic windows” – specifically menopause and menstrual cycles – in response to evidence that nearly 99% of preclinical aging studies have historically excluded these factors. The initiative, in fact, is backed by a $15.7 billion USD National Academies recommendation.

These structural changes have catalyzed a parallel shift in how women’s health data is collected, interpreted, and acted upon. A new generation of platforms is translating hormonal and cycle data into accessible, clinically relevant insights, enabling women to identify patterns before they escalate into conditions that traditional methodologies may misinterpret.

“My own health crisis was one of the most frightening experiences of my life — and what it revealed quickly was that the burden falls almost entirely on the patient at exactly the moment they are least equipped to carry it,” Adriana Torosian, founder and CEO of Ourself Health, told Unite AI.

Ourself Health is leading a structural shift in how women’s health is perceived, rising from women’s personal experiences with poor health data management, and meant to prevent future issues with misalignment of information or poor access to data interpretation.

The San Francisco-based startup recently unveiled Stella, an AI-powered health companion that combines the world’s leading women’s health research with users’ personal health history.

“Ultimately, the answer for me became my data. I suspected my cycle was directly impacting my condition and brought that hypothesis to leading doctors, who dismissed it entirely. The only way forward was building my own dataset, finding my own answers, and then bringing my doctors along in my process — the complete reverse of how I expected this to go,” Torosian added.

Why AI Changes the Equation

AI is fundamentally reshaping healthcare diagnostics, not by replacing clinical judgement, but by enabling a form of pattern recognition at a scale and continuity that traditional care cannot replicate. Unlike clinical models that depend on episodic encounters, AI systems can continuously analyze medical records, biomarkers, and real-time physiological inputs, detecting correlations that standard care routinely misses.

This result has been measurably earlier and more accurate diagnoses across conditions ranging from cardiovascular disease to cancer – a shift already improving patient outcomes.

In women’s health specifically, this capacity is particularly essential; hormonal systems are dynamic, deeply interconnected, and highly individualized. AI-driven tools are beginning to bridge the diagnostic gap by enabling more precise monitoring, prediction and longitudinal analysis across reproductive health, maternal care, and gynecological conditions.

Emerging applications range from AI-enhanced fetal imaging to noninvasive detection of endometriosis, areas where traditional diagnostics have long struggled.

Ourself Health’s Stella builds on this foundation by operationalizing longitudinal hormonal data, converting patterns into personalized, time-sensitive health recommendations rather than generalized clinical guidance.

“The more data a user brings into the platform, the more precise and personal Stella’s guidance becomes. That data comes from multiple layers: individual symptoms tracked daily within the app, personal notes, documents a user can upload directly, and continuous physiological data from wearables like the Apple Watch,” Torosian explained.

In doing so, the tool moves beyond detection toward decision support – reframing women’s health as a continuous, computable system rather than a series of disconnected clinical verticals.

“The goal is to close the gap between what a woman knows about her own body and what her doctor sees in a brief appointment — and to make sure she arrives at every interaction armed with her own data, the latest relevant research, and a clear plan of action. Stella puts all of that in her hands,” the founder added.

A New Computational Layer for an Old Gap

The distinction between generating insight and producing actionable guidance is subtle, but clinically significant. Stella AI is designed to prioritize the latter, interpreting longitudinal trends to generate individualized and time-sensitive recommendations, calibrated to each user’s hormonal baseline.

At its core, Stella’s design rests on the recognition that no two hormonal systems are identical. The Ourself platform then continuously learns from each user’s inputs, regardless of cycles, symptoms,

“What’s still missing for individual women is the ability to take charge right now — without waiting for the research to catch up. That’s exactly where Ourself comes in; we can’t ask women to pause their lives while institutions slowly close the funding gap. We can give them the tools to understand their own bodies today, build their own health record, make informed decisions, and take actions with what we already know — while the broader research landscape continues to evolve around them,” Torosian stressed.

Such an approach transforms hormonal health management from a reactive discipline into a proactive one, where interventions can be timed and tailored with a precision that conventional care models are structurally ill-prepared to deliver.

But beyond individual cases or visionaries, the emergence of better – and AI-assisted – technologies poses a new computational layer capable of both making that complexity actionable and, more importantly, saving lives.

As institutions like the NIH and the American Heart Association formally redirect resources, AI is translating that momentum into real-world impact. The promise of these tools lies in their ability to personalize, and operationalize what medicine has long observed but struggled to apply: that women’s health is dynamic.

The future of healthcare will not be defined by population-level averages but by precision – where each individual’s longitudinal data forms the foundation of their care. And, in that sense, AI is not replacing medicine, but extending it into territory it was never fully equipped to navigate, until now.