TRV-2026-1272Version 2 · Revised

Saved · see the current record

Why this version exists

Editorial pass: sharper headline; filed as context: no real AI gain or problem (gain activity_only, problem absent).

The exact text we saved

This is the version exactly as it was saved, character for character. The SHA-256 hash below is a 64-character code worked out from that text. Run the same text through SHA-256 and you get the same code; change a single letter and the code comes out completely different. Each version also stores the previous version’s hash, which chains them together so no version can be swapped out quietly.

TRUVACE RECORD VERSION
record: TRV-2026-1272
version: 2
kind: revised
reason: Editorial pass: sharper headline; filed as context: no real AI gain or problem (gain activity_only, problem absent).
timestamp: 2026-11-04T09:55:38.439807Z
status: published
lens: general
sector: health
headline: Identifying incident medication-related osteonecrosis of the jaw and antiresorptive drug holidays using natural language processing on clinical notes in men and women prescribed bisphosphonates or denosumab for fracture prevention
dek: Objective Medication-related osteonecrosis of the jaw (MRONJ) is a rare complication of antiresorptive therapy for osteoporosis, with risk potentially influenced by "drug holidays" around dentoalveolar procedures. MRONJ risk is higher in rheumatic diseases. Epidemiologic studies are limited by poor administrative code performance for MRONJ identification and lack of structured pharmacy data capturing drug holidays. We developed and validated a natural language processing (NLP) algorithm to classify MRONJ and det…
gain_title: NLP model classified incident MRONJ and identified antiresorptive drug holidays from free-text EHR notes with high precision and recall in validation sets.
problem_title: (none)
trace_subject: (none)
gain_reading: NLP model classified incident MRONJ and identified antiresorptive drug holidays from free-text EHR notes with high precision and recall in validation sets.
gain_evidence: We developed and validated a natural language processing (NLP) algorithm to classify MRONJ and detect drug holidays from free-text clinical notes in electronic health records (EHR) | For MRONJ classification excluding unclassifiable outputs across all validation notes, precision was 1.00 (95% confidence interval [CI] 0.86-1.00), recall 0.96 (0.80-1.00), and F-Measure 0.95 (0.88-1.00) | For drug holiday detection in the held-out validation set, precision was 0.90 (0.83-0.95), recall 0.98 (0.93-1.00), and F-Measure 0.94
problem_reading: (none)
problem_evidence: (none)
quick_read: Investigators developed and validated a natural language processing algorithm to identify incident medication-related osteonecrosis of the jaw and antiresorptive drug holidays from free-text clinical notes, using Veterans Health Administration records for adults 50 and older prescribed bisphosphonates or denosumab for fracture prevention between 1999 and 2022.

Accurate identification matters because MRONJ is a rare but serious complication where administrative codes perform poorly and pharmacy data miss drug holidays, especially relevant for higher-risk rheumatic disease populations; uncertainty remains about performance outside the VA system and in more diverse, non-veteran cohorts.
limitation: Validation was limited to U.S. Veterans Health Administration EHR for adults >=50 prescribed bisphosphonates or denosumab, and MRONJ is described as rare, constraining generalizability and epidemiologic power.
tag: Context
key_points: Study used U.S. Veterans Health Administration EHR for adults >=50 with >=1 filled prescription for bisphosphonate or denosumab for fracture prevention from 10/1/1999-12/31/2022. | Reference set included 6,421 annotations informing a curated vocabulary of >1,350 terms, with 11 targets and 16 attributes defined in NLP codebook. | Decision tree-based machine-learning model used NLP-extracted features to classify MRONJ, while drug holiday was an NLP target directly.
rundown: Researchers built an NLP codebook with 11 targets and 16 attributes, annotated 870 notes with two independent annotators plus 100 notes for independent validation, producing 6,421 annotations and a vocabulary of over 1,350 terms.

Performance was reported for MRONJ classification excluding unclassifiable outputs and for drug holiday detection in held-out validation, with authors concluding the algorithm is sufficiently accurate to be a promising tool for future studies including in persons with rheumatic diseases.
sources:
- peer_reviewed | Arthritis Care & Research | https://doi.org/10.1002/acr.80173 | 2026-10-03
prev: 30c2420728e8b6ba17de6840c4defc3d223c949314f5f70fcb5bbe9d88b0bc76
SHA-256 hash
f006552d6ae143f13a89b1f0e4ef2d4383cca9aa19800e950725009e1ada0321
Previous version’s hash
30c2420728e8b6ba17de6840c4defc3d223c949314f5f70fcb5bbe9d88b0bc76
Check this record hasn’t been altered
How to check it yourself, without trusting this page

Download any version from /api/record/TRV-2026-1272 and hash its canonical field yourself, for example with shasum -a 256. You should get its content_hash. Each version’s text also ends with prev:and the hash of the version before it (version 1 points to 64 zeros). If anyone had changed a single character of any version after it was saved, those hashes wouldn’t line up.