Menopause Diagnosis Statistics 2026: What U.S. Data Actually Show

In a July 2026 PCORnet Population Insights Report—the newest large, publicly detailed U.S. electronic-health-record benchmark identified in this review—2,710,209 of 28,208,203 female patients, or 9.6%, had a menopause-related diagnosis code recorded between 2020 and 2024.Every patient in that denominator met PCORnet’s active-patient criteria: at least one face-to-face encounter and a recorded diagnosis code at a participating site during 2024.

PCORnet publishes the 9.6% headline. Its age table also contains the raw counts needed to calculate rates within each age group, but the percentages printed beside the age rows use the entire menopause-coded cohort as the denominator. We recalculated the within-age rates:

8.30% for ages 35–54. 18.78% for ages 55–64. 25.78% for ages 65–74.

PCORnet also reports that 1,486,534 patients—54.8% of the menopause-coded cohort—fell within its “asymptomatic menopause” category. That does not mean 54.8% had only a status entry: the published condition-category counts demonstrate overlap, and the public report does not publish the overlap between the asymptomatic category and the symptomatic or disorder categories. What it does show is that the answer changes materially depending on whether a study counts menopausal status, symptomatic conditions, postmenopausal conditions, premature menopause, or only two narrow diagnosis codes.

That is why 7%, 8%, 9.6%, and 14.7% can all appear in credible U.S. sources without estimating the same thing. Below, we map each major included source, state exactly what it counts, and show which number fits which claim.

These definitions keep biological menopause, survey-assigned stage, diagnosis codes, encounters, and treatment from being collapsed into one statistic. They are not interchangeable, and most of the confusion in this field begins when one is substituted for another.

Menopause.

The World Health Organization defines natural menopause as occurring after 12 consecutive months without menstruation when no other physiological or pathological cause explains it and no clinical intervention caused the cessation. Menopause is recognized retrospectively.

Perimenopause.

The World Health Organization describes perimenopause as beginning when signs of the menopausal transition are first observed and ending one year after the final menstrual period.

Documented diagnosis.

A qualifying diagnosis code recorded in an electronic health record or insurance claim under a study’s stated rules. It is not the same as biological status, and it is not proof of what was or was not discussed during a clinical visit.

Denominator.

The population used to calculate a percentage or rate. Much of the apparent disagreement between the statistics below is a denominator problem.

ICD-10-CM.

The diagnosis-coding system used in the United States, maintained by the CDC’s National Center for Health Statistics. Codes are one way clinical information becomes structured data.

Encounter rate.

A rate based on diagnosis-bearing encounters or service dates rather than unique people. In HCCI’s methodology, the numerator is unique patient encounter service dates, so the same person can contribute more than one service date over time.

PCORnet is a distributed research network funded by the Patient-Centered Outcomes Research Institute. At the time of the report, its eight clinical research networks included 78 partner sites connected to more than 50 million people. Seventy-five partners contributed data to the menopause query.

The query began with patients who had at least one face-to-face encounter and a recorded diagnosis code at a participating site during calendar year 2024. It then used a five-year lookback—January 1, 2020 through December 31, 2024—to identify menopause-related diagnosis codes in the electronic health record.

The mean age of 64.4 is worth sitting with, but it has to be read correctly. The U.S. Office on Women’s Health reports an average age of natural menopause of 52. PCORnet’s mean is 12.4 years higher than that population milestone, but the report does not reveal when menopause began or when a code was first entered. It describes the age of patients in a 2024 active-patient cohort who had a qualifying code somewhere in a five-year lookback.

PCORnet’s published percentages beside the age rows describe the composition of the menopause-coded cohort: the share of all 2.7 million coded patients falling into each age band. They are not rates within the female population of each age band. Both calculations are valid; they answer different questions.

We calculated the within-age rates.

Two cautions we would rather state than have someone else discover.

The rate falls from 25.78% at ages 65–74 to 19.59% at 75 and older. That is not evidence that menopause becomes less common after age 74. The measure is a five-year record of qualifying codes among patients active in 2024. Code history, record continuity, health-system contact, cohort composition, and documentation practices can all change with age.

And age here is the patient’s age in the report’s population table. It is not age at the final menstrual period, age at symptom onset, or age at first diagnosis.

PCORnet publishes condition-category counts and shares in its tables. Bringing them together exposes a distinction the 9.6% headline cannot carry by itself.

Three things follow from that table.

The largest reported category is asymptomatic menopause. It contains 1,486,534 patients, or 54.8% of the coded cohort, and its mean age is 68.8 years. But categories overlap. The public tables do not show how many of those patients also carried an N95 code, so they do not support a “status code only” percentage.

The five N95 rows cannot be added as unique people. Their published counts total 1,694,442, which is 227,267 higher than the 1,467,175 unique patients in the combined N95 category. That difference proves overlap among the child-code counts.

A narrow code definition selects a much smaller part of the full cohort. N95.1 alone appears for 24.8% of the complete menopause-coded cohort. N95.1 plus N95.9 covers somewhere between 24.8% and 28.3% of the cohort, because the public data do not reveal their overlap.

What the data support.In PCORnet, documented-code rates rise sharply through ages 65–74. The selected code list materially changes the measured cohort. Recorded rates differ in unadjusted comparisons by payer and area-level socioeconomic measures. A large asymptomatic-menopause category sits inside PCORnet’s headline cohort. And several studies show gaps between self-reported symptoms or stage information and structured records.

What the data do not support. A national biological prevalence rate. The share of all U.S. women ever accurately diagnosed. A national misdiagnosis percentage. A causal explanation for demographic differences. The age menopause began. The age a code was first entered. Or a valid average across the source-specific figures. Averaging 7%, 8%, 9.6%, and 14.7% produces a number that measures nothing.

A missing qualifying code is proof that the study did not find that code under its stated method. It is not proof that menopause was absent, that a clinician failed, or that a patient was untreated.

Eleven metrics from 10 source families, each with its native population, denominator, time window, and evidence status preserved. We did not pool, average, or reweight anything across sources.

The evidence map and every original arithmetic result used on this page are available in machine-readable formats. All files are data version 1.1 and were last verified August 1, 2026.

• Evidence map as CSV — 11 metric rows × 28 fields
• Evidence map as JSON — the same source-level records with dataset metadata
• Derived calculations as CSV — raw inputs, formulas, displayed values, interpretations, and limitations
• Derived calculations as JSON — the same calculation record in machine-readable form
• Citation record as BibTeX
• Citation record as RIS

Match the statistic to the claim. Each row below identifies the best-matched source included in this review for that specific statement—and only for that statement.

Because they count different events. A five-year code history, an annual unique-patient percentage, an encounter service-date rate, symptom documentation, survey-classified stage, self-reported uncertainty, treated prevalence, and post-diagnosis treatment answer different questions. Most of the apparent disagreement is methodological.

Komodo Health analyzed nearly 29 million women ages 45–51 and found that 7% had N95.1 or N95.9 recorded during 2019–2023. PCORnet used a broader menopause-related code list and reported 9.6% in a different age distribution, healthcare setting, and active-patient cohort.

Inside PCORnet’s own data, N95.1 appears for 671,009 patients, or 24.8% of the complete menopause-coded cohort. N95.9 appears for 95,751, or 3.5%. Because those patient groups can overlap, the number carrying at least one of the two codes must be at least 671,009 and no more than 766,760—between 24.8% and 28.3% of PCORnet’s full coded cohort.

That calculation isolates one thing cleanly: a two-code definition covers only part of the population captured by PCORnet’s broader query. It does notconvert Komodo’s 7% into PCORnet’s 9.6%, and it does not prove how much of the difference between the two headline rates is caused by coding. The cohorts differ in age, database, care setting, observation rules, and denominator as well as code scope.

Sources: PCORnet Population Insights Report, Table 4; Komodo Health, “Lost in Transition”. Bound calculated by The HRT Index on August 1, 2026.

The 2015 National Health Interview Survey classified menopausal status from menstrual-history questions. It classified a respondent as postmenopausal if she had gone more than one year without a menstrual cycle or had surgical menopause after removal of both ovaries. It classified her as perimenopausal if menstruation had stopped and the last cycle was no more than one year earlier.

Under that operational rule, a still-menstruating 47-year-old with changing cycle length and vasomotor symptoms remains in the premenopausal survey category. That is different from STRAW+10, which begins the menopausal transition with persistent cycle-length change while menstruation is still occurring.

That distinction explains why NCHS could report that among nonpregnant U.S. women ages 40–59, 74.2% were classified as premenopausal, 3.7% as perimenopausal, and 22.1% as postmenopausal. The figures are nationally representative for the survey definition. They are not a clinical diagnosis rate.

Source: NCHS Data Brief No. 286, “Sleep Duration and Quality Among Women Aged 40–59, by Menopausal Status”, September 2017, using 2015 NHIS data.

• Do not compare PCORnet’s 9.6% with HCCI’s 1,969 per 100,000 as if both were percentages of unique women.
• Do not call AARP’s 5% treated prevalence a diagnosis rate.
• Do not call the All of Us survey-to-EHR ratio an underdiagnosis rate.
• Do not call 22.7% clinical-problem-list documentation—or 60% clinic-note documentation—a population prevalence.
• Do not average 7%, 8%, 9.6%, and 14.7%.
• Do not read NHIS’s 3.7% as the share of all midlife women who would meet a clinical menopausal-transition framework.

Before quoting any menopause diagnosis statistic, attach:

1. Age band.
2. Insurance, health-system, survey, or research-cohort scope.
3. Geography.
4. Active-patient or continuous-enrollment requirement.
5. Observation window.
6. Unique people versus encounters or service dates.
7. Diagnosis-code set or survey rule.
8. Whether the outcome is biological, documented, self-reported, classified, or treated.

If you cannot fill in all eight, the number is not ready to publish.

We preserved every source’s native population, denominator, metric class, time window, code definition, and evidence status instead of forcing the figures into one estimate. We calculated a percentage only where a source published both the numerator and denominator, and every calculation is labeled as ours.

Source hierarchy

1. Official government or issuing-body material.
2. Peer-reviewed original research.
3. Original analyses published by healthcare-data producers.
4. Professional-society summaries tied to an identifiable original study.
5. Secondary sources used for discovery only, never as the sole evidence for a consequential claim.

Inclusion criteria

A source entered the evidence map only if it supplied:

• A defined U.S. geography or U.S. cohort.
• An identifiable population or denominator.
• A data period, or a visible disclosure that the public period was unstated.
• A numeric result.
• A defined diagnosis, documentation, classification, uncertainty, treatment, or post-diagnosis measure.
• An identifiable original producer or research team.
• Enough information to state the principal limitation without guessing.

Calculation rules

We preserve source units. We do not pool across sources. We never convert an encounter count into a unique-person count, treatment into diagnosis, or survey capture into clinical underdiagnosis.

Derived values are held to at least four decimal places in the calculation record and displayed to two decimal places. The three principal original calculations on this page are:

• Age-specific PCORnet rates:menopause-coded patients in each age band ÷ female patients in the same age band meeting PCORnet’s 2024 active-patient criteria.
• PCORnet code-set bound: the unique N95.1-or-N95.9 group must lie between the larger single-code count and the sum of both counts because overlap is unpublished.
• Area Deprivation Index rates:menopause-coded patients in each assigned ADI quartile ÷ female patients in the same assigned quartile.

What we excluded, and why

PCORnet’s Hispanic-status rows.Table 1 reports 277,632 menopause-coded patients in the “No” row and 2,166,614 in the “Yes” row, against female denominators of 19,095,658 and 4,595,168. Taken literally, those counts imply rates of approximately 1.45% and 47.15%. We could not reconcile that pattern with the table, and the report provides no clarification. We excluded Hispanic-status calculations rather than infer a labeling error or publish a result we could not defend.

“80% of OB/GYNs are untrained in menopause.” We did not identify an original study supporting that exact statement. We excluded the circulating claim.

Any national underdiagnosis percentage. No harmonized U.S. source in this review supplies one. We did not construct one.

An exact perimenopause-only rate from PCORnet. The public report does not publish the full query specification or a mutually exclusive perimenopause subgroup.

The evidence supports a narrower conclusion: menopause-related symptoms and reproductive-stage information are often missing from structured records, and many surveyed women report uncertainty about their stage. It does not support one national U.S. underdiagnosis percentage.

Komodo’s analysis provides a direct view of claims documentation after symptom-related care. Among women ages 45–51 who sought care for symptoms commonly associated with the transition, the share with N95.1 or N95.9 documented was:

Komodo also reports unadjusted documentation differences within the same cohort: 8% among commercially insured women versus 5% among women on Medicaid, and 11% among White women versus 8% among Black women, 8% among Asian American and Pacific Islander women, and 10% among Hispanic women. These are descriptive claims comparisons. They do not establish why the differences occurred.

A peer-reviewed 2024 study reviewed 229 women in the Mayo Clinic Health System—Northwest Wisconsin Region who reported moderate-or-higher vasomotor symptoms. 22.7% had vasomotor symptoms listed in the EHR clinical problem list, while 60% had the symptoms documented in primary-care clinic notes.

That distinction matters. The structured problem list missed more than the narrative notes did, and neither measure is a formal menopause-diagnosis rate or a national estimate.

Source: Bevry ML, Stogdill ER, Lea CM, et al., “Addressing menopause symptoms in the primary care setting: opportunity to bridge care delivery gaps”, Menopause, 2024;31(12):1044–1048. Survey: March 1–June 30, 2021. EHR review window: May 1, 2019–May 1, 2021.

An All of Us Research Program analysis of approximately 396,000 female participants found that survey instruments captured approximately 193,000 menopause observations, compared with approximately 28,000 structured EHR diagnoses—nearly seven times as many observations in the survey data.

Source: “Menopause in the All of Us Research Program: A Descriptive Summary of Electronic Health Records and Survey Data”, medRxiv, posted April 25, 2026.

A 2026 study of 7,640 U.S. Flo app users ages 35 and older, published in Menopause, found that 34% (95% CI 33%–35%) reported being unsure of their reproductive stage.

That is self-reported uncertainty in an app-user sample collected from December 2024 through May 2025. It is not a diagnosis rate, and app users are not a random sample of U.S. women.

Source: “Exploring Prevalence and Drivers of Perimenopause Uncertainty Among U.S. Women”, Menopause, published July 14, 2026.

Claims data, EHR documentation, survey-to-record comparison, and self-report all show a gap of some kind. That convergence matters.

But the gaps are not interchangeable. One study asks whether two codes appeared after symptom-related care. One asks whether vasomotor symptoms appeared in a structured problem list or in clinic notes. One compares survey fields with structured diagnosis fields. One asks whether a respondent knows her stage.

There is no valid arithmetic that turns those four results into “X% of American women with menopause are undiagnosed.” We looked for one. It does not exist, and we are not going to manufacture it.

The public code definitions used by PCORnet, HCCI, Komodo, and Optum do not create a mutually exclusive category for a still-menstruating person in perimenopause. N95.1 combines menopausal and climacteric states, and N95.9 is unspecified.

PCORnet’s public footnote says its “menopausal and other perimenopausal disorders” category contains N95.0, N95.1, N95.2, N95.8, and N95.9. Two of those are explicitly postmenopausal conditions. N95.8 is “other,” and N95.9 is unspecified. N95.1 carries menopausal and climacteric states together.

HCCI uses the same N95 family alongside asymptomatic and premature-menopause categories. Komodo uses N95.1 and N95.9. The Optum incident-diagnosis study uses ICD-9-CM 627.9 and ICD-10-CM N95.1 and N95.9.

This does not mean the rest of a medical record cannot distinguish a 46-year-old with changing cycles from an older postmenopausal patient. Age, menstrual history, symptoms, notes, labs, and procedures can supply context. It means the diagnosis codes used in these published analyses do not, by themselves, produce a clean perimenopause-only cohort.

The CDC’s National Center for Health Statistics maintains ICD-10-CM and its official browser. As of August 1, 2026, the applicable code set is the FY26 April 1, 2026 release for services through September 30, 2026.

The major frameworks below are not five competing diagnostic rules. They serve different purposes: one stages reproductive aging, two guide clinical care, one addresses premature ovarian insufficiency, and one classifies respondents for a federal survey. Their differences still matter because they determine who enters a study or receives a label.

Two divergences are worth naming.

One FSH result, or two. Current NICE guidance uses two elevated FSH samples taken four to six weeks apart for suspected premature ovarian insufficiency under age 40. The 2024 international POI guideline uses one elevated FSH above 25 IU/L, with repeat testing only when uncertainty remains. The 2025 European Society of Endocrinology guideline follows the one-test default and uses two where the clinical picture is unclear.

Different start lines for the transition.STRAW+10 uses persistent change from a person’s own cycle pattern. The ESE clinical guideline considers menstrual irregularity and/or vasomotor symptoms. NHIS’s operational survey category requires menstruation already to have stopped. The same person can be classified differently because the frameworks are answering different questions.

We are describing what each body published. We are not adjudicating between them, and none of this is individual medical advice.

Documented diagnoses rose modestly in employer-sponsored insurance between 2018 and 2022. OB/GYNs recorded the largest physician share, while the combined physician-assistant and nurse-practitioner share rose by almost half.

The Health Care Cost Institute analyzed employer-sponsored-insurance claims and reported that the annual share of women with any menopause diagnosis rose from 13.7% in 2018 to 14.7% in 2022—an increase of 1.0 percentage point, or 7.3% relative.

Separately, HCCI reports an average monthly diagnosis-bearing encounter service-date rate that rose from 1,793 to 1,969 per 100,000, a 9.8% relative increase. These are different quantities from the same brief, and they are easy to mistake for two versions of the same result. The first is an annual percentage of women. The second is an average monthly rate based on unique patient encounter service dates.

Symptomatic-menopause encounter rates specifically rose from 757 to 840 per 100,000, an 11.0% relative increase.

In 2022, more than 81% of menopause diagnoses were recorded by physicians. The combined physician-assistant and nurse-practitioner share rose from 9.3% in 2018 to 13.8% in 2022, a 48.4% relative increase. Among physician-recorded diagnoses, HCCI reports OB/GYNs at 47%, family medicine at 17%, and internal medicine at 14%.

Source: Health Care Cost Institute, “Menopause Diagnosis Steadily Increased from 2018–2022”, October 18, 2024. Relative changes calculated by The HRT Index from HCCI’s published endpoints.

In an incident-diagnosis claims cohort of 297,928 insured women ages 40–65, 30.7% initiated captured prescription pharmacotherapy during observed follow-up. Treatment measured in claims is not the same as all treatment received, and it is not a judgment about who should have been treated.

The Optum Clinformatics study used claims from 2012 through 2024 and identified first qualifying diagnosis claims from 2013 through 2024. Eligible patients had a first claim carrying ICD-9-CM 627.9 or ICD-10-CM N95.1 or N95.9 after a 12-month baseline period. Mean observed follow-up after diagnosis was 20.7 months, and the median was 20 months.

30.7% initiated captured pharmacotherapy; the remaining 69.3% had no qualifying prescription claim. Among those who initiated treatment, the median time to initiation was three months.

The study’s claims could miss treatment obtained outside the measured pharmacy benefit, including compounded products, over-the-counter products, and therapies not consistently captured. The authors’ disclosures identify six current Bayer employees, one Bayer consultant, and two former interns; the paper reports no funding. We state that because a reader deciding how to use the study should know.

PCORnet offers a different view: medication prescribing recorded during calendar year 2024 among people in its five-year menopause-coded cohort.

Across PCORnet’s entire 28,208,203-female-patient denominator, 11,465 had a 2024 prescription in the report’s “medications for hot flashes” class—0.0406%, or about 4 in 10,000. Within the menopause-coded cohort, the corresponding count was 8,309, or 0.3%.

We are not going to tell you what that means clinically. PCORnet’s category is narrow, and hot flashes can be treated with hormonal therapies and other medication classes listed elsewhere in the table.

AARP’s Public Policy Institute, working with NORC and Medical Expenditure Panel Survey data, estimated that 2.1 million U.S. women ages 45–64—5% of that age group—were treated for menopause.The analysis pooled 2016–2021 MEPS data to estimate an average annual population and presented the result as a 2021 annual estimate.

That is treated prevalence under the report’s definition. It is not a diagnosis rate.

Source: AARP Public Policy Institute, “Women in Menopause Often Go Untreated”, April 17, 2025.

PCORnet, HCCI, and Komodo all report unadjusted differences by area-level socioeconomic measure or payer. The direction is similar across the three sources, but the metrics are not interchangeable and none of the comparisons establishes cause.

PCORnet’s raw counts allow a within-quartile calculation for patients assigned to an Area Deprivation Index quartile.

The Q1 rate is 46.2% higher than the Q4 rate. That is a descriptive relative difference, not a causal estimate.

The table excludes 2,227,472 female patients and 168,046 menopause-coded patients whose ADI value was missing. PCORnet conducted no inferential analysis, and the aggregate table cannot control for age, payer, site composition, geography, healthcare use, record completeness, or other confounders. PCORnet also states that ADI is designed for validity at the nine-digit ZIP or census-block-group level, not the five-digit ZIP level used for this table.

HCCI found a similar unadjusted gradient using a different index and a different metric: 2,193 diagnosis-bearing encounter service dates per 100,000 in the least vulnerable areas versus 1,688 in the most vulnerable areas in 2022. Komodo reported 8% documentation among commercially insured women versus 5% among women on Medicaid in its ages-45–51 cohort.

Three datasets, three measures, one direction. That is worth studying. It is not proof of cause.

The training gap is real. The cleanest way to state it is with the exact surveyed population and response denominator attached.

The underlying findings do not need embellishment.

Christianson et al. reported in 2013 that 20.8% of 510 responding OB/GYN residents said their program had a formal menopause-medicine curriculum; 1,799 people received the survey, for a 28.3% completion rate. Kling et al. reported in 2019 that 12 of 177 residents who answered the preparedness item—6.8%—felt adequately prepared to manage women experiencing menopause. The survey reached 703 residents across 20 programs and received 183 responses, a 26.0% response rate. The 2023 program-director survey received 99 responses from 145 programs, a 68.3% response rate; 31.3% of respondents reported a menopause curriculum, and 71% of those curricula offered two or fewer lectures per year.

Those are serious findings. They are also survey results with specific respondents, not census counts of every practicing clinician or every residency program.

Sources: Christianson et al., Menopause, 2013; Kling et al., Mayo Clinic Proceedings, 2019; Allen et al., Menopause, 2023.

The included data systems are being used to answer a question—how often menopause and perimenopause are recognized—that their published codes and denominators do not answer cleanly. At the same time, large new datasets and updated clinical guidance are making the mismatch easier to see.

Three verified developments frame the current snapshot.

A large public EHR benchmark arrived in July 2026.PCORnet’s report exposes counts across 50.3 million unique patient records and 28.2 million female patients, with enough detail to calculate age-specific and ADI-quartile rates. It remains a descriptive health-system cohort, not a national prevalence survey.

Clinical guidance changed or was updated. The international POI guideline was published in December 2024. The European Society of Endocrinology published its menopause and perimenopause guideline in October 2025. NICE last updated NG23 on April 15, 2026.

Recorded diagnoses rose inside one stable claims source. HCCI’s annual share increased from 13.7% to 14.7% between 2018 and 2022. The data cannot determine how much of that change reflects recognition, coding, healthcare use, cohort composition, or another factor.

That is the point of this page: preserve what each measure can answer, and stop one denominator from being mistaken for another.

No source here is a harmonized national diagnosis census, and every original result on this page is arithmetic performed on published aggregate counts. The limitations below travel with the numbers; removing them changes what the data can support.

• No source here is a national census of menopause diagnosis. Every dataset reflects people who reached a health system, had particular insurance, answered a survey, used an app, or joined a research cohort.
• A missing code proves little about the underlying clinical experience. Coding depends on access, recognition, documentation habits, record continuity, billing rules, and whether symptoms were attributed to menopause or something else.
• PCORnet’s denominator requires a 2024 face-to-face encounter and a recorded diagnosis code. People without a qualifying encounter at a participating site in 2024 are absent by construction.
• PCORnet’s 75+ result is not a biological age curve. It is a five-year code-history rate in an active-patient cohort.
• PCORnet’s published condition-category counts overlap. A patient can appear in more than one category, so category shares cannot be added.
• The public PCORnet report does not list its full diagnosis-code query. It specifies the N95 child codes and says complete query information is available on request.
• NHIS data are from 2015. The operational-definition issue they illustrate is durable; the percentages are a 2015 snapshot.
• Healthcare-data-producer analyses have limited public reproducibility. Komodo and Evernorth publish original results without a full peer-reviewed methods paper or downloadable patient-level dataset. Evernorth’s public article does not state the analysis years or diagnosis-code set.
• One source is a preprint. The All of Us result is labeled as such throughout and may change after peer review.
• Self-reported uncertainty is not clinical adjudication. It measures what respondents report knowing about their own stage.
• Demographic, payer, and area comparisons are unadjusted unless stated otherwise. They cannot establish cause.
• The guideline table compares frameworks with different jobs. A research-staging framework, a clinical guideline, a POI guideline, and a federal survey classification should not be treated as interchangeable authorities.
• This is a structured evidence synthesis, not a registered systematic review or meta-analysis. We identified the strongest current U.S. diagnosis-related measures we could verify, preserved their definitions, and documented our calculations.
• We conducted no original fieldwork. Every input traces to a named source.

These answers preserve the same denominators and limits used in the evidence map. Each one stands alone without changing the underlying metric.

What percentage of women receive a menopause diagnosis?

No single national percentage is supported by the included evidence. PCORnet’s 2026 report found that 9.6% of 28.2 million female patients meeting its 2024 active-patient criteria had a menopause-related code recorded during a five-year lookback. Other source-specific figures include 7% among women ages 45–51 using two codes, 8% in a commercially insured ages-40–64 cohort, and 14.7% in HCCI’s 2022 employer-sponsored-insurance population. They cannot be averaged.

How is menopause diagnosed?

Natural menopause is recognized retrospectively after 12 consecutive months without menstruation when no other physiological or pathological cause explains it and no clinical intervention caused the cessation. Clinical assessment also considers menstrual history, symptoms, age, medications, and other possible explanations. Perimenopause can begin while menstruation is still occurring, and the major frameworks do not all operationalize its onset the same way.

Can a blood test diagnose menopause?

For otherwise healthy people age 45 and older with a typical presentation, NICE and the European Society of Endocrinology state that biochemical testing is not necessary. Hormone levels fluctuate during the transition, so one result can be difficult to interpret. Testing has a different role when premature ovarian insufficiency is suspected under age 40 or when the clinical picture is atypical.

Is perimenopause underdiagnosed?

The evidence supports underdocumentation and uncertainty, but not one national underdiagnosis percentage. Komodo reported that among women ages 45–51 seeking care for selected symptoms, the share with N95.1 or N95.9 documented ranged from 19% for headache or migraine to 28% for sleep disturbances. A separate 229-woman primary-care study found vasomotor symptoms in the structured problem list for 22.7% and in clinic notes for 60% of its selected symptomatic sample.

Is there an ICD-10-CM code that cleanly identifies perimenopause in these studies?

Not in the public code definitions used by the included PCORnet, HCCI, Komodo, and Optum analyses. N95.1 combines menopausal and climacteric states, while N95.9 is unspecified. The rest of a medical record can contain age, menstrual history, symptoms, and other context, but the study code sets do not produce a mutually exclusive perimenopause-only cohort.

Is a treatment rate the same as a diagnosis rate?

No. AARP’s estimate that 5% of U.S. women ages 45–64 were treated for menopause measures treated prevalence under a MEPS-based definition. Optum’s 30.7% measures captured pharmacotherapy after an incident diagnosis. Neither is the percentage of all women diagnosed.

What is the average age of menopause diagnosis?

The average age of natural menopause and the average age of a patient carrying a code are different statistics. The U.S. Office on Women’s Health reports an average age of natural menopause of 52. PCORnet reports a mean age of 64.4 among patients in its 2024 cohort who had a menopause-related code during 2020–2024, but that does not reveal the age of first diagnosis.

Why do published menopause diagnosis rates differ so much?

They use different populations, ages, databases, code sets, and time windows, and some count people while others count encounter service dates, survey classifications, or treatment. One code-scope difference can be measured inside PCORnet: patients with N95.1 or N95.9 make up between 24.8% and 28.3% of its broader menopause-coded cohort. That does not reconcile other databases, but it shows why the code list must travel with the statistic.

The block below gives a complete publication record and version date for neutral attribution.

Derived figures on this page are calculated from published counts in the sources named beneath each table. The formulas are stated so the calculations can be reproduced.

All files are data version 1.1, last verified August 1, 2026.

• Evidence map — CSV (11 metric rows × 28 fields)
• Evidence map — JSON (source-level records with dataset metadata)
• Derived calculations — CSV (inputs, formulas, displayed values, interpretations, limitations)
• Derived calculations — JSON
• Citation record — BibTeX
• Citation record — RIS

These research pages cover adjacent outcomes that should not be folded into diagnosis statistics. Treatment uptake, research funding, and terminology each answer a different follow-up question.

• How Many Women Use HRT? 2026 Statistics — treatment uptake, which is a different outcome from diagnosis documentation
• Menopause Research Funding: NIH 2007–2025 — federal funding totals, definitions, and source-system reconciliation
• Menopause & HRT Glossary — plain-language definitions for menopause, perimenopause, hormone therapy, and related research terms

Every consequential claim above traces to an issuing body, original producer, peer-reviewed article, or clearly labeled preprint. Secondary summaries were not used as the sole evidence for a factual statistic.

Primary datasets and original analyses

• PCORnet. Population Insights Report: Characteristics of the Patient Population with Menopausal Conditions Receiving Health Care Services Across PCORnet Clinical Research Networks in 2024. July 10, 2026.
• Natwick T, Martin K, Ganesan N. Menopause Diagnosis Steadily Increased from 2018–2022. Health Care Cost Institute. October 18, 2024.
• Woodard E, Linkter R, Druet A. Lost in Transition: Despite Growing Awareness, Perimenopause Remains Vastly Underrecognized and Undertreated. Komodo Health. March 6, 2025.
• Evernorth Research Institute. Research Finds Few Women Receive Diagnoses for Perimenopause or Menopause. February 7, 2025.
• Vahratian A. Sleep Duration and Quality Among Women Aged 40–59, by Menopausal Status. NCHS Data Brief No. 286. National Center for Health Statistics. September 2017.
• Carter B. Women in Menopause Often Go Untreated. AARP Public Policy Institute. April 17, 2025.
• Sullivan AP, Heo J, Hansen RN, et al. Characteristics of Insurance-Covered Women Newly Diagnosed with Menopause Symptoms and Their Providers in the United States. Women’s Health Reports. 2026;7.
• Staples JW, et al. Menopause in the All of Us Research Program: A Descriptive Summary of Electronic Health Records and Survey Data. medRxiv preprint. Posted April 25, 2026.
• Exploring Prevalence and Drivers of Perimenopause Uncertainty Among U.S. Women: A Mixed-Methods Study. Menopause. Published July 14, 2026.
• Bevry ML, Stogdill ER, Lea CM, et al. Addressing Menopause Symptoms in the Primary Care Setting: Opportunity to Bridge Care Delivery Gaps. Menopause.2024;31(12):1044–1048.

Clinical definitions, guidelines, and coding standards

• World Health Organization. Menopause fact sheet.
• Office on Women’s Health, U.S. Department of Health and Human Services. Menopause Basics. Updated April 2, 2026.
• CDC National Center for Health Statistics. ICD-10-CM. Updated July 2, 2026.
• CDC National Center for Health Statistics. ICD-10-CM Files.
• National Institute for Health and Care Excellence. Menopause: Identification and Management. NICE guideline NG23. Published November 12, 2015; last updated April 15, 2026.
• Panay N, Anderson RA, Bennie A, et al. Evidence-Based Guideline: Premature Ovarian Insufficiency. Human Reproduction Open. 2024;2024(4):hoae065.
• Lumsden MA, Dekkers OM, Faubion SS, et al. European Society of Endocrinology Clinical Practice Guideline for Evaluation and Management of Menopause and the Perimenopause. European Journal of Endocrinology.2025;193(4):G49–G81.
• Harlow SD, Gass M, Hall JE, et al. Executive Summary of the Stages of Reproductive Aging Workshop +10. Journal of Clinical Endocrinology & Metabolism.2012;97(4):1159–1168.

Menopause education studies

• Christianson MS, Ducie JA, Altman K, et al. Menopause Education: Needs Assessment of American Obstetrics and Gynecology Residents. Menopause.2013;20(11):1120–1125.
• Kling JM, MacLaughlin KL, Schnatz PF, et al. Menopause Management Knowledge in Postgraduate Family Medicine, Internal Medicine, and Obstetrics and Gynecology Residents. Mayo Clinic Proceedings. 2019;94(2):242–253.
• Allen JT, et al. Needs Assessment of Menopause Education in United States Obstetrics and Gynecology Residency Training Programs. Menopause.2023;30(10):1002–1005.

This log records changes that alter interpretation, data, or reproducibility. Cosmetic edits do not receive a new data version.

Version 1.1 — August 1, 2026.Factual and production audit. Preserved the PCORnet headline, age-rate calculations, evidence-map architecture, methodology, and attribution block. Corrected unsupported timing inferences from mean age; made category overlap explicit; removed unpublished PCORnet code mappings; corrected the evidence-map count to 11 metrics from 10 source families; separated HCCI’s unique-woman percentage from its service-date rate; bounded rather than summed N95.1/N95.9 coverage; distinguished primary-care problem-list documentation from clinic-note documentation; corrected the Optum journal and cohort description; added missing ADI records; replaced absolute coding claims with dated study-code-set findings; restructured the guideline comparison by purpose; removed the unsupported Dataset license; updated source and schema dates; and published the evidence map, full calculation record, and citation files in machine-readable formats. No cross-source estimates were pooled.

Version 1.0 — July 31, 2026. Initial release. Original age-specific documented-code rates, evidence-map classification, N95 code-scope comparison, and Area Deprivation Index rates calculated from published source counts. Hispanic-status calculations excluded pending source clarification.

This report is non-commercial even though The HRT Index earns revenue elsewhere. No source paid for, reviewed, or approved the analysis.

The HRT Index earns affiliate revenue elsewhere on the site through referrals to hormone-therapy providers. The Research section carries no affiliate links, advertising, paid placement, or commercial routing. No organization whose data appears on this page reviewed, approved, or paid for this report.

The byline is The HRT Index Editorial Team, the site’s standing collective byline. This page has not been medically reviewed and does not claim otherwise.

This page explains population statistics, coding, and research methods. It cannot determine whether an individual is in perimenopause or menopause. Symptoms, unexpected bleeding, periods stopping at an early age, or other concerns should be evaluated by a qualified healthcare professional.