Use of AI with VistA

[Kevin Toppenberg]

With all the recent discussion of AI's in the news lately, I have been wondering how they could be of use to me with my use of VistA.

I haven't come up with an answer to this question yet, but I did find an interesting page that uses AI to help de-identify patient records.

https://huggingface.co/spaces/obi/Medical-Note-Deidentification

It seems to correctly identify elements of a note that contain references to persons and places etc, and then removes them.  It seems pretty cool.  If one scrolls down the page there are examples can be clicked on to see output.  I typed in a sample note of a patient coming in with their daughter and made up an example of them traveling to Florida.  It correctly identified the potentially identifying elements of my fictional patient and cut all that out.

Here is one of the examples:

INPUT:

Consult NotePt: Ulysses Ogrady MC #0937884Date: 07/01/19 Williams Ct M OSCAR, JOHNNY Hyderabad, WI 62297

HISTORY OF PRESENT ILLNESS: The patient is a 77-year-old-woman with long standing hypertension who presented as a Walk-in to me at the Brigham Health Center on Friday. Recently had been started q.o.d. on Clonidine since 01/15/19 to taper off of the drug. Was told to start Zestril 20 mg. q.d. again. The patient was sent to the Unit for direct admission for cardioversion and anticoagulation, with the Cardiologist, Dr. Wilson to follow.
SOCIAL HISTORY: Lives alone, has one daughter living in Nantucket. Is a non-smoker, and does not drink alcohol.
HOSPITAL COURSE AND TREATMENT: During admission, the patient was seen by Cardiology, Dr. Wilson, was started on IV Heparin, Sotalol 40 mg PO b.i.d. increased to 80 mg b.i.d., and had an echocardiogram. By 07-22-19 the patient had better rate control and blood pressure control but remained in atrial fibrillation. On 08.03.19, the patient was felt to be medically stable.

OUTPUT:

Consult NotePt: #Date:
HISTORY OF PRESENT ILLNESS: The patient is a -year-old-woman with long standing hypertension who presented as a Walk-in to me at the on . Recently had been started q.o.d. on Clonidine since to taper off of the drug. Was told to start Zestril 20 mg. q.d. again. The patient was sent to the Unit for direct admission for cardioversion and anticoagulation, with the Cardiologist, Dr. to follow.
SOCIAL HISTORY: Lives alone, has one daughter living in . Is a non-smoker and does not drink alcohol.
HOSPITAL COURSE AND TREATMENT: During admission, the patient was seen by Cardiology, Dr. , was started on IV Heparin, Sotalol 40 mg PO b.i.d. increased to 80 mg b.i.d., and had an echocardiogram. By the patient had better rate control and blood pressure control but remained in atrial fibrillation. On , the patient was felt to be medically stable.

Seems like this could be useful

Kevin

[Kevin Toppenberg]

It's not perfect.  For example my wife and I tried this input:

Mrs Jones 4/5/76 is here today for recheck of HTN.  She recently was one the front page of the Greeneville newspaper and has become quite a local celebrity for her work with the festival. 

And here is the output:

is here today for recheck of HTN. She recently was one the front page of the newspaper and has become quite a local celebrity for her work with the festival.

If someone was familiar with the local newspaper, and if it was known that this was a patient of mine, then it might still be possible to re-identify this fictitious patient.  But if there was no date or location associated with the note, and the data was put into a large database from widely distributed sources, then one would have a somewhat difficult time picking between all women who had appeared on the front page of some newspaper somewhere in the world and where being a "local" celebrities doesn't mean much.

Kevin

[Valerie J H Powell]

Dear Kevin,

From time to time I wonder whether AI experts are aware of the failure to integrate medical and dental data in the private section in the manner done by the VA in VistA, the military EHR, and the Indian Health Service EHR. What do the words "complication," "comorbidity" etc. mean if one omits consideration of the stomatognathic system?

Here is what I normally reference when discussing oral-systemic integration: 

Why Integrate Oral and Systemic Health

Without a comprehensive and integrated approach to medical science that includes dental and oral health, our public policies cannot provide the best answers to health promotion, disease prevention, early detection and treatment for the most common diseases, as well as for many rare ones.

https://www.frontiersin.org/articles/10.3389/fdmed.2020.599214/full

Sommerman M, Mouradian W

Santa Fe Group: Improving Lives Through Oral Health. Available online at: https://santafegroup.org/ (accessed August, 2020)

University College London Institute of Epidemiology and Health Care. Lancet Commission on Oral Health. (2019). Available online at: https://www.ucl.ac.uk/epidemiology-health-care/research/epidemiology-and-public-health/research/dental-public-health/lancet-commission-oral-health (accessed August, 2020)

Atchison KA, Rozier RG, and Weintraub JA. Integrating Oral Health, Primary Care, and Health Literacy: Considerations for Health Professional Practice, Education and Policy. National Academies Press (US) (2018). Available online at: https://www.nationalacademies.org/our-work/integrating-dental-and-general-health-through-health-literacy-practices-a-workshop

National Institute of Dental and Craniofacial Research. NIDCR 2030: Envisioning the Future, Together. (2018). Available online at: https://www.nidcr.nih.gov/research/research-priorities-nidcr-2030 (accessed August, 2020)

--
--
http://groups.google.com/group/Hardhats
To unsubscribe, send email to Hardhats+u...@googlegroups.com

---
You received this message because you are subscribed to the Google Groups "Hardhats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hardhats+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/hardhats/ad976672-51ab-4b46-8edd-d04ddbb1b194n%40googlegroups.com.

--

Valerie J H Powell RT(R) retired, MS, PhD
University Professor Emerita, Computer & Information Systems
Project on Clinical Data Integration (CDI)
Robert Morris University
6001 University Boulevard
Moon Township, PA 15108-1189 USA