The NPI Staleness Report: 46% of Provider Records Have Never Been Updated
We analyzed every record in the CMS NPPES file (9,551,447 NPIs, August 2026 vintage). 46.1% of active records were never touched after the day they were registered, and 54.6% have gone five or more years without an update.
Published September 2026 · Recomputed with each NPPES release
What We Measured
Every US healthcare provider has an NPI, and every NPI has a record in NPPES, the government registry that the entire healthcare data industry builds on. Each record carries two dates that make this study possible: the enumeration date (when the NPI was created) and the last update date (the most recent time anyone changed anything on the record).
Providers are required to report changes to NPPES within 30 days. There is no penalty for not doing it. So the gap between those two dates measures something the industry rarely talks about: whether anyone is maintaining the record at all.
We streamed the full August 2026 dissemination file (9,551,447 records), excluded deactivated NPIs, and computed the age of every active record's last update as of September 1, 2026. Specialty figures use each provider's primary classification from the NUCC taxonomy reference, reported as plain-language specialty names.
Nearly Half of All NPI Records Were Registered and Never Touched Again
4,247,169 active records (46.1%) carry a last update date identical to their enumeration date. The provider or their staff filled out the application once, got the NPI, and never came back.
To be precise about what that means: an untouched record is not automatically a wrong record. A solo practice that never moves has nothing to report. But provider mobility is the norm, not the exception, in most of these professions. Clinicians change employers, practices relocate and consolidate, and phone systems get replaced. When a record has sat unmaintained for five or ten years in a profession where people move every few, the address and phone on file describe a past practice, not a current one.
The distribution across all 9.2 million active records: 10.1% updated within the last year, 20.6% between one and three years ago, 14.8% between three and five, 22.5% between five and ten, and 32.1% more than ten years ago. The registry the industry treats as ground truth is, by its own timestamps, mostly a historical document.
Staleness by Specialty
Record maintenance varies sharply by profession. The table covers the fifteen largest clinical classifications by active record count.
| Specialty | Active records | Share not updated in 5+ years | Median record age since last update |
|---|---|---|---|
| Counselor | 630,683 | 54% | 5.8 years |
| Nurse Practitioner | 510,521 | 41% | 4.0 years |
| Social Worker | 447,570 | 53% | 5.5 years |
| Pharmacist | 321,395 | 75% | 8.9 years |
| Internal Medicine | 302,985 | 55% | 5.6 years |
| Physical Therapist | 282,686 | 63% | 7.5 years |
| Dentist | 269,563 | 75% | 13.5 years |
| Registered Nurse | 235,870 | 57% | 6.5 years |
| Physician Assistant | 217,511 | 43% | 4.1 years |
| Speech-Language Pathologist | 194,919 | 63% | 7.7 years |
| Family Medicine | 163,823 | 55% | 5.7 years |
| Occupational Therapist | 155,844 | 64% | 7.6 years |
| Psychologist | 152,756 | 64% | 8.4 years |
| Pediatrics | 110,371 | 59% | 6.8 years |
| Chiropractor | 103,821 | 75% | 12.6 years |
The pattern is not random. The stalest specialties (dentists, chiropractors, pharmacists) skew toward established, owner-operated practices where the NPI was registered long ago and no one has a reason to revisit it. The freshest (nurse practitioners, physician assistants) are the professions growing fastest, where a large share of records are simply new. Freshness here reflects recency of registration more than diligence of maintenance.
For anyone building outreach lists, the practical reading: the specialties that look most reachable in raw NPPES data (small owner-operated practices with a single decision-maker) are exactly the ones whose registry records are least likely to be current.
New Provider Registrations Keep Accelerating
The other side of the file is enumeration: new NPIs entering the registry. Registrations grew every single year from 2020 through 2025.
| Year | New NPI registrations |
|---|---|
| 2020 | 427,476 |
| 2021 | 463,900 |
| 2022 | 468,402 |
| 2023 | 509,906 |
| 2024 | 548,012 |
| 2025 | 631,277 |
That is 48% growth in five years, driven heavily by behavioral health roles: behavior technicians, counselors, and social workers. (2026 shows 243,697 registrations through the August file cut; partial-year figures lag the file's processing window, so we exclude 2026 from the trend claim.)
Recent enumeration dates are one of the most useful signals in the file: a provider enumerated in the last 6 to 12 months is usually a new practitioner or a new practice, which is the highest-intent moment for most healthcare vendors. Our provider enumeration glossary entry covers how to use that signal.
What This Means If You Buy or Build Provider Lists
Every healthcare data vendor, including us, starts from NPPES. The difference between vendors is what happens after that. This study quantifies why the after matters: a list exported straight from the registry inherits records that are, in the majority, five or more years unmaintained.
Three practical rules follow from the data. First, treat registry age as a risk score: a dentist record untouched since 2013 needs verification before a rep dials it, and a nurse practitioner enumerated in March probably does not. Second, verify against live sources (state boards, practice websites, phone and email validation) rather than against another compiled database that shares the same NPPES ancestry. Our guide to bulk NPI verification walks through the free CMS tooling. Third, when you evaluate vendors, ask when their records were last verified, not how many they have. Database size is a claim about enumeration; accuracy is a claim about maintenance. The vendor comparison covers how the major providers approach this.
Methodology
Source: CMS NPPES full data dissemination file, August 2026 vintage (9,551,447 records), joined to the NUCC Health Care Provider Taxonomy reference for specialty names. Analysis date: September 1, 2026.
Definitions and exclusions: records with a deactivation date and no reactivation date are excluded from all staleness figures (343,321 records). "Never updated" means the last update date equals the enumeration date. Specialty is each provider's primary taxonomy classification (individual providers, Type 1, only); classifications that are not practicing clinical roles (for example, students in training programs) are omitted from the specialty table. Median record ages are computed from a systematic 1-in-7 sample within each specialty; shares are computed from full counts. All figures recompute when we ingest a new NPPES release, so numbers on this page track the latest file rather than the publication date.
Cite This Study
This analysis is published under CC BY 4.0. Use any figure with attribution:
Provyx, "The NPI Staleness Report" (September 2026), https://getprovyx.com/research/npi-data-staleness/ : 46.1% of active NPI records have never been updated since registration.
Journalists and researchers: we can cut these aggregates by state, entity type, or specialty on request. Write [email protected].
Frequently Asked Questions
How often are NPI records updated?
Far less often than most buyers assume. In the August 2026 NPPES file, 46.1% of active NPI records were never updated after initial registration, 54.6% had not been touched in five or more years, and only 10.1% were updated within the last year. Providers are required to report changes within 30 days, but there is no enforcement mechanism.
Does a stale NPI record mean the data is wrong?
Not automatically, and the distinction matters. Last Update Date measures record maintenance, not confirmed accuracy. A record can sit untouched because nothing changed. But providers relocate, change employers, and retire constantly, so in high-mobility professions a record that nobody has touched in a decade is unlikely to reflect the current practice. Treat staleness as risk, then verify.
Which specialties have the most outdated NPI records?
Dentists, chiropractors, and pharmacists top the list: 75% of records in each have gone five or more years without an update. The median dentist record is 13.5 years old. Nurse practitioners and physician assistants have the freshest records, mostly because so many of them are newly enumerated.
Can I use NPPES data directly for outreach?
You can start there, and you should: it is free, public domain, and the authoritative identity source. But it carries no email addresses, and this analysis shows most records have not been maintained. Teams that mail or call raw NPPES contacts inherit its staleness. Verify against live sources before spending on outreach.
Where does this data come from?
Every number is computed from the CMS NPPES full data dissemination file (August 2026 vintage), covering all NPI records, joined to the NUCC taxonomy reference for specialty names. Deactivated NPIs are excluded from staleness figures. The methodology section describes classifications, exclusions, and sampling.
Related Resources
Get Records That Were Verified This Week, Not This Decade
Provyx builds provider lists at order time and verifies against live sources. Test us with a sample.
Trusted by healthcare sales teams, medical device companies, and health IT vendors across the US.