Mississippi has 86 nurse practitioners per 100 physicians. Hawaii has 25

We read all 9,798,758 rows of the CMS NPPES monthly full replacement file of September 2026, file date 2026-09-13, and counted every individual whose primary taxonomy code marks them as a physician, a nurse practitioner, a physician assistant, a dentist or a pharmacist. That left 2,595,733 people, each placed in the state their record gives as a practice location and divided by that state's Census 2024 population.

Every count here is people enumerated and not deactivated. An NPI is issued once, never expires, and goes away only when someone files a deactivation, so a person who retired in 2014 and never filed one is still in our numbers. That is why our physician count, 1,247,484, sits 15.3% above the 1,082,187 licensed physicians the FSMB census reports, and why our 362.4 physicians per 100,000 sits 19.2% above the 304 per 100,000 the AAMC publishes for 2024.

We expected to find at least one state where nurse practitioner NPIs outnumber physician NPIs. There is none. The closest is Mississippi at 86.1 nurse practitioners per 100 physicians, then Tennessee at 75.0 and Kentucky at 71.1. Hawaii ends the list at 25.1, a 3.4-fold spread.

What we did not expect is that Mississippi sits at both ends. It has the highest nurse practitioner ratio of any state and the fewest physicians per 100,000, at 255.4. The ratio is high because the denominator is small: at 220.0 nurse practitioners per 100,000 Mississippi is third among the states, behind Tennessee and Kentucky, its neighbours on the ratio list too.

Key findings

1. Mississippi has 86.1 nurse practitioners per 100 physicians and Hawaii has 25.1. Nationally, across the 50 states and DC, the figure is 42.4. Source: ProfessionLens, September 2026.
2. No state has more nurse practitioners than physicians. Not one of the 51 rows reaches 100 nurse practitioners per 100 physicians, and no territory row does either; the highest territory is the Northern Mariana Islands at 28.3.
3. Mississippi has 255.4 physicians per 100,000 residents, the fewest of any state, with Idaho at 255.9 immediately above it.
4. California has 96.9 nurse practitioners per 100,000 residents, the fewest of any state, against 231.3 in Tennessee, the most.
5. The District of Columbia has 960.1 physicians per 100,000 residents, 1.7 times the highest state, Massachusetts at 563.8. DC is a city and we keep it out of every sentence that says state.
6. 33,493 people who took an NPI in 2021 hold a nurse practitioner taxonomy today, against 29,967 who hold a physician taxonomy. 2021 is the first year where the nurse practitioner count is the larger of the two.
7. Physician enumerations ran near 33,000 a year from 2010 to 2020 while nurse practitioner enumerations rose from 11,892 to 30,955 over the same stretch.
8. 55,187 people enumerated so far in 2026 hold the NUCC student code 390200000X as their primary taxonomy, against 1,686 of the 2010 cohort. Many of them will switch to a physician code later, which is why we do not print a nurse practitioner to physician ratio for 2022 or after.
9. The five professions together come to 2,595,733 counted individuals: 1,247,484 physicians, 523,482 nurse practitioners, 327,937 pharmacists, 274,523 dentists and 222,307 physician assistants.

All 51 rows, ranked by nurse practitioners per 100 physicians

Physicians and nurse practitioners are individuals enumerated and not deactivated whose practice address falls in that row. The per 100,000 columns divide by Census POPESTIMATE2024.

Massachusetts has 563.8 physicians per 100,000 residents and Mississippi has 255.4

The physician per capita list runs DC at 960.1 per 100,000, Massachusetts at 563.8, Rhode Island at 497.5, Connecticut at 466.3, New York at 460.9 and Pennsylvania at 455.3. DC's figure is a city counting a regional hospital and federal workforce against its own residents, so we rank it and leave it out of state comparisons. At the other end come Alabama at 271.2, Nevada at 268.7, Wyoming at 262.8, Idaho at 255.9 and Mississippi at 255.4.

Nurse practitioners per 100,000 run in a different order. Tennessee leads the states at 231.3, then Kentucky at 220.5, Mississippi at 220.0, Massachusetts at 213.7 and New Hampshire at 202.4, with DC above all of them at 251.5. California has the fewest of any state at 96.9, below Hawaii at 102.9, Oklahoma at 119.6, Idaho at 126.3 and Oregon at 127.7. It is last on nurse practitioners per capita and 49th on the ratio, and it holds 141,205 physicians, more than any other state.

Idaho has 33.8 physician assistants per 100 physicians and Mississippi has 6.2

The physician assistant ratio does not track the nurse practitioner one. Idaho leads at 33.8 physician assistants per 100 physicians, then Alaska at 30.6, North Carolina at 30.6, Montana at 28.0 and South Dakota at 27.5. At the bottom are Missouri at 9.2, Hawaii at 9.1 and Mississippi at 6.2.

Five professions come to 2,595,733 counted individuals

Those rates divide by the 2024 United States population of 340,110,988 and include people with a territory address or none at all. The state rows sum to a smaller set, 1,232,402 physicians at 362.4 per 100,000 and 522,857 nurse practitioners at 153.7, which are the figures to use alongside a state.

Territories are counted separately and appear in no ranking. Puerto Rico holds 12,253 physicians and 151 nurse practitioners. We reparsed those rows and got the same counts. Another 1,709 counted individuals carry a practice-location value that is not a US state code and sit in the national totals only.

Our 362.4 physicians per 100,000 is 19.2% above the AAMC's 304

The AAMC's 2025 key findings page says, in these words, that "In 2024, the United States had 1,032,365 active physicians of which 866,460 were direct patient care physicians, corresponding to 304 and 255 physicians per 100,000 population respectively." Our 362.4 per 100,000 is 19.2% above that 304 for one reason: an NPI is never surrendered when someone retires or leaves the profession, so we count everyone enumerated and not deactivated while the AAMC counts active physicians.

The gap runs the same way against every other published stock figure: 15.3% above FSMB's 1,082,187, 13.6% above AANP's 461,000, 10.6% above NCCPA's 201,038 board certified physician assistants, and 33.9% above the ADA's 205,088 professionally active dentists, the widest of the five, because the ADA's definition drops retirees who keep an NPI. Pharmacists are the exception: BLS reports about 325,200 jobs in 2025 and we count 327,937, a gap of 0.8%, which is the check that told us the pipeline was doing what we thought.

The AAMC's ordering agrees with ours at both ends. It names Massachusetts and the District of Columbia highest and Idaho lowest; our top two are DC at 960.1 and Massachusetts at 563.8, and our lowest state is Mississippi at 255.4 with Idaho at 255.9 just above it. That page prints no per state number, and we did not open the physician workforce dashboard where its state table now lives.

33,493 people enumerated in 2021 hold a nurse practitioner code today, against 29,967 with a physician code

The table counts people by the year they took their NPI and the primary taxonomy they hold today. The student column is people from that year whose primary taxonomy today is the NUCC student code 390200000X; the last column is people from that year carrying a physician code in any of their 15 slots.

2026 is left off because the file stops at 2026-09-13; its figures are in the CSV. 2005 and 2006 are the registry's own enrolment years, when the existing workforce signed up at once, so the series starts at 2010.

We print no nurse practitioner to physician ratio for 2022 or later, and the CSV leaves those cells blank and flags the rows. Medical students now take an NPI while still students and switch to a physician code years afterwards, so a recent year's physician count is held down by people who have not converted. The student column shows the size of it: 3,606 people from the 2015 cohort hold the student code today, against 52,723 from 2025, while the count holding a physician code in any slot falls from 33,121 to 10,572 over the same two cohorts. Cohorts through about 2019 have largely converted. Those from 2022 on have not, and the crossing year may move from 2021 to 2022 or 2023 as they do.

Forty of the 51 jurisdictions enumerated more nurse practitioner NPIs than physician NPIs from 2020 on

From 2020 to the file date, the 50 states and DC produced 135,043 new physician NPIs and 207,252 new nurse practitioner NPIs, and in 40 of the 51 jurisdictions the nurse practitioner count is the larger. The fact pack ranks this cut by ratio. We ranked it by the gap between the two counts instead, computed from enumeration_by_year_state.csv. Florida has the widest gap, with 22,080 new nurse practitioner NPIs against 8,026 new physician NPIs, then Texas with 17,791 against 9,169 and New York with 12,626 against 8,473. Missouri is exactly level at 3,750 against 3,750, and Illinois is furthest the other way, with 9,339 new physician NPIs against 7,236 new nurse practitioner NPIs.

This cut carries the student effect at full strength. It includes the 2022 and later cohorts, so it counts NPIs by taxonomy code today, not new doctors, and we publish no ratio for it. The per state figures are in enumeration_by_year_state.csv, and the gap is in facts.json.

From 9,798,758 rows to 2,595,733 counted individuals

We test deactivation before entity type, and that order matters. CMS blanks a record when it deactivates the NPI, so 355,329 of the 374,443 rows carrying a deactivation date have no entity type code and no taxonomy left. Filtering on entity type first would file all 355,329 as organisations and report 2,327,387 organisations instead of 1,972,058. The same blanking means exits by profession or state cannot be computed from this file at all. Rows with an unparseable enumeration date: 0. Counted individuals with no taxonomy code: 0.

Method

The source is the CMS NPPES September 2026 monthly full replacement file, file date 2026-09-13, fetched on 2026-09-15 through the Webshare rotating gateway and inflated in flight. We kept 30 of its 330 columns as a 9,798,758-row slim file of 799,860,439 bytes, and took taxonomy names from NUCC code set 26.1.

An NPI carries up to 15 taxonomy codes, each with a primary switch. The primary taxonomy is the code in the first slot whose switch is Y, and slot 1 where no switch is Y; everything except the any-slot column is computed on that one code. Physician is a code starting 207 or 208, the NUCC grouping Allopathic and Osteopathic Physicians, 229 codes across 28 classifications. Nurse practitioner is 363L, 18 codes. Physician assistant is 363A, 3 codes. Dentist is 1223, 13 codes. Pharmacist is 1835 excluding 183700000X, the pharmacy technician code, 18 codes. The five sets cannot overlap, because the rule is applied to a single code.

A record is dropped when it carries a deactivation date and no reactivation date. Population denominators are Census POPESTIMATE2024 from NST-EST2024-ALLDATA.csv, fetched through the same gateway on 2026-09-15; the 2024 United States total of 340,110,988 is what the 50 states and DC sum to. Census publishes that file for the states, DC and Puerto Rico only.

We checked the five totals against six published figures, all refetched through the gateway on 2026-09-15 and quoted as they stood that day:

• FSMB, 1,082,187 licensed physicians: https://www.fsmb.org/advocacy/news-releases/fsmb-physician-census-identifies-1082187-licensed-physicians-in-u.s/
• AANP, more than 461,000 nurse practitioners: https://www.aanp.org/about/about-the-american-association-of-nurse-practitioners-aanp/historical-timeline
• NCCPA, 201,038 board certified physician assistants: https://www.nccpa.net/wp-content/uploads/documents/2025-Statistical-Profile-of-Board-Certified-PAs.pdf
• ADA Health Policy Institute, 205,088 professionally active dentists: https://www.ada.org/resources/research/health-policy-institute/dentist-workforce
• BLS Occupational Outlook Handbook, about 325,200 pharmacist jobs in 2025: https://www.bls.gov/ooh/healthcare/pharmacists.htm
• AAMC 2025 key findings, 1,032,365 active physicians and 304 per 100,000: https://www.aamc.org/data-reports/data/2025-key-findings

Limitations, all of them:

1. An NPI is not a licence and goes only by an affirmative deactivation, so every count is larger than a count of people at work. Do not describe anyone counted here as working in the profession today.
2. The student taxonomy holds down the physician count for cohorts from 2022 on, so we print no nurse practitioner to physician ratio for those years and the CSVs leave the cell blank.
3. Practice-location state is self-reported and is not refreshed when someone moves.
4. Primary taxonomy is whatever the person last set, so someone who changed specialty and never updated is counted under the old code.
5. We use the primary taxonomy only, except in the any-slot column, which runs 1 to 8 per cent higher and bounds the effect.
6. Profession here is a taxonomy prefix. MD against DO is not in the file.
7. The 355,329 deactivated records carry no profession and no state, so exits cannot be attributed from this file.
8. Provider Sex Code is binary M or F, and rows with no value are outside the denominator of sex_by_year.csv.
9. Denominators are the whole resident population, not an adult or insured one.
10. 2026 is partial, to 2026-09-13.
11. 2005 and 2006 are registry enrolment years, not a flow of new entrants.
12. The pharmacist enumeration series jumps in 2010, 2011 and 2020 for reasons this file does not explain. State counts are unaffected, but do not build a time-series sentence on that row.
13. DC is a city. It is ranked and flagged, and out of every sentence that says state.
14. Territory and armed forces rows are separate and unranked. AA, AE and AP are military postal codes, and Census publishes a 2024 population for Puerto Rico only.
15. 1,709 counted individuals carry a practice-location value that is not a US state code, and appear in the national totals only.
16. Counts under 5 print as <5 in every CSV, and a per 100,000 rate is blank wherever its numerator is masked.
17. A Massachusetts figure of 485 physicians per 100,000 attributed to the AAMC is not on that page today. We did not verify it. Ours is 563.8.
18. NPPES has no record of people who never enumerate, so a profession whose members often bill no federal health programme can be undercounted too.

Data

• providers_by_state.csv
• enumeration_by_year.csv
• enumeration_by_year_state.csv
• sex_by_year.csv
• facts.json
• README.md
• METHOD.md

These tables are published under CC BY 4.0. Attribute to ProfessionLens.