My creative work has recently expanded from sound and perception into another complex system: modern medicine—and, in particular, how institutions decide what counts as “evidence-based” care for women.
Evidence-Selective Medicine: A Six-Level Model of Institutional Bias in Clinical Medicine
Using menopausal hormone therapy (MHT) as a central case study, this analysis asks the question: Are current menopausal hormone therapy restrictions truly evidence-based, or have they been shaped by a systematic, institutional selectivity in what counts as evidence?
For over two decades, the FDA required a black box warning on estrogen-containing MHT products, signaling the highest level of potential risk and harm. That warning, rooted in a narrow, early interpretation of the Women’s Health Initiative (WHI) trial, effectively encoded the message that hormone therapy is inherently dangerous and should be used sparingly, briefly, and with a degree of fear.
On November 10, 2025, the FDA announced that it was removing that boxed warning from estrogen-containing menopausal hormone therapies, acknowledging that the original labeling overstated and over-generalized the risks, especially for women who initiate treatment within 10 years of menopause. The science had been evolving in that direction for many years, and the labeling is only now beginning to catch up.
The article explores a 6-level framework to understand how we got stuck for so long in a narrative that no longer matched the totality of the evidence. And why guidelines remained entrenched as modern MHT formulations replaced the outdated ones (i.e., Premarin, Prempro) that were the basis of the risk and harm signal.
What is “Evidence-Selective Medicine”?
The paper introduces an analytical framework called Evidence-Selective Medicine (ESM). The premise is that there is an institutional tendency to require much higher standards for evidence that challenges the current paradigm. And implicitly accepts lower quality and falsely extrapolated evidence that supports the status quo position.
The article introduces a six-level model to assess for this type of evidence bias:
- Methodological asymmetry – The same flaws (underpowered subgroups, adherence issues, selection biases) are treated as fatal when results challenge the prevailing narrative, and as trivial when they confirm it.
- Reporting asymmetry – Dramatic relative risks are highlighted while minimal absolute risks, non-significant results, and favorable secondary outcomes are buried in the fine print.
- Category error – Data from one very specific treatment regimen and population are generalized to an entire therapeutic class: “hormone therapy”, “antidepressants”, “statins”.
- Semantic obfuscation – generalized, ambiguous language blurs meaningful distinctions in route, dose, timing and formulation, collapsing everything into a single, monolithic risk label.
- Institutional commitment architecture – Guidelines, malpractice norms, quality metrics, and regulatory warnings all align around the early interpretation, making any revision politically and legally expensive.
- Ethical-framework failure – In principle, informed consent and shared decision-making is explicitly endorsed. However, in practice, patients are generally not given the balanced, absolute-risk information that would allow them to make genuinely informed choices when those choices diverge from the status quo position.
How this plays out in menopausal hormone therapy
In the early 2000s, WHI results from one oral CEE + medroxyprogesterone (Premarin) regimen in mostly older women were rapidly interpreted as proof that hormone therapy “causes” heart disease, stroke, blood clots, dementia, and breast cancer. Small absolute increases in certain risks were treated as decisive, while fracture reduction, diabetes reduction, nuanced age effects, and favorable estrogen-alone findings were marginalized as unimportant details.
From there:
- The risk findings were generalized from a single regimen to “hormone therapy” as a whole, including modern transdermal estradiol and micronized progesterone.
- Broad language (“estrogen-containing products”) blurred critical distinctions in formulation, route, dose, and timing.
- The FDA’s black box warning, along with conservative guidelines and malpractice fears, created a commitment architecture that made it far safer to avoid offering MHT than to engage with a more nuanced and accurate balance of risks and benefits.
- Clinically, women often received a very simplified story: use MHT only as a short-term palliative treatment for symptoms like hot flashes, rather than as a preventive therapy for bone, brain, or cardiovascular health.
Over the last decade, the evidence base has shifted: timing matters; formulation and route matter; absolute risks for healthy women starting near menopause are small; benefits for bone and quality of life are substantial. Yet the restrictive narrative lagged behind, anchored by its own institutional weight.
The FDA’s decision to remove the black box warning is not just a regulatory footnote; it is a tacit admission that the original framing was evidence-selective, not simply “evidence-based”.
Although my website is primarily a home for the exploration of polychromatic music, this article expresses a similar underlying preoccupation: the behavior of complex systems, where they may become distorted, and how our language and conceptual structures shape and limit what we perceive.
In music, this is explored through a polychromatic framework: pitch, color, and the limits of multidimensional perception.
In medicine, this is explored through an Evidence-Selective Medicine framework which aims to illuminate institutional blind spots, making them explicit and open to discussion.
The full open-access article, Evidence-Selective Medicine: A Six-Level Model of Institutional Bias in Clinical Medicine, is available at:
SSRN: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5738144
Recent Comments