H3IT: Home Healthcare, Hospice, and Information Technology Conference Nashville, TN, 2015

Using Natural Language Processing to

Automatically Identify Wound Information in

Narrative Clinical Notes: Application Development

and Testing.

Maxim Topaz RN, PhD

1, 2

, Dawn Dowding RN, PhD

3, 4

, Victor J. Lei BA

, Anna Zisberg RN, PhD

, Kathryn H. Bowles

RN, PhD

2, 6

, Li Zhou MD, PhD

1, 1

his study developed and validated one of the ﬁrst automated natural language processing (NLP) applications

to extract wound information (wound type, pressure ulcer stage, wound size, anatomic location, and wound

treatment) from free text clinical notes.

Methods: First, two human annotators manually reviewed a purposeful training sample (n=360) and random

test sample (n=1,100) of clinical notes (including 50% discharge summaries and 50% outpatient notes, including homecare

notes), identiﬁed wound cases, and created a gold standard dataset. We then trained and tested our NLP system (known as

MTERMS) to process the wound information. Finally, we assessed our automated approach by comparing system-generated

ﬁndings against the gold standard. We also compared the prevalence of wound cases identiﬁed from free-text data with coded

diagnoses in the structured data.

Results: The testing dataset included 101 notes (9.2%) with wound information. The overall system performance was

good (F-measure =92.7%), with best results for wound treatment (F-measure =95.7%) and poorest results for wound size

(F-measure =81.9%). Only 46.5% of wound notes had a structured code for a wound diagnosis.

Conclusion: The NLP system achieved good performance on a subset of randomly selected discharge summaries and

outpatient notes. In more than half of the wound notes, there were no coded wound diagnoses, a fact that highlights the

signiﬁcance of using NLP to enrich clinical decision making. Our future steps will include expansion of the application’s

information coverage to other relevant wound factors and validation of the model with external data. We are also conducting

a validation and further system development with homecare notes from the Visiting Nurse Services of New-York and will

share the results at the H3IT conference.

Brigham and Women’s Hospital, Boston, MA, USA

Harvard Medical School, Boston, MA, USA

Visiting Nurse Service of New York, NY, USA

School of Nursing, Columbia University, NY, USA

The Cheryl Spencer Department of Nursing, Haifa University, Israel

School of Nursing, University of Pennsylvania, PA, USA