Using Natural Language Processing to Improve Fall Documentation in VA Nursing Home Residents.
Laura A Graham, Xiaojuan Liu, Bocheng Jing and 7 others
PMID 41956437WHAT IT FOUND
In VA nursing homes, a text-mining tool found more falls and exact dates than MDS reports, and found chart-documented falls that MDS missed.
Use MDS fall counts cautiously.
Key findings
01The algorithm's fall detection was accurate compared with final chart abstractions, with 96.0% overall accuracy and 94.0% sensitivity.
02NLP identified 109,264 falls, while MDS assessments documented at least 67,977 falls over the same period.
036.3% of residents had falls documented in their charts that were never reported in the MDS, and the median lag from EHR fall documentation to an MDS assessment was 17 days.
STILL TO COME
How it was doneWhat they foundWhat it means for RNs
Read the rest of this summary
You get three full summaries a month, free, and we do not ask for a card. Search, the TL;DRs and your library stay unlimited either way.
What it does not show
The study used the text-mining algorithm itself as the reference standard for MDS misclassification, so errors in the algorithm could affect the comparison. Findings are from Veterans Health Administration nursing homes, a mostly male and white population, and from VA-specific note templates, so they may not apply to other care systems or populations. Falls may still be underreported in the electronic record if staff do not document them, so neither MDS nor the algorithm captures all falls. The algorithm could not determine fall severity or clinical consequences. Manual chart review validation used random samples, and copied note text could affect dates, though event-level validation found no evidence of copied text for fall dates.
Declared interests
No conflicts of interest were reported. Publication types list NIH extramural and non-U.S. government research support.
The easy way to misread this
Do not conclude that MDS fall reporting is complete or that this algorithm can be used in any nursing home without testing. The algorithm was used as the reference standard, and falls may still be missing from electronic notes.