Skip to content
Edmon Begoli edited this page Dec 12, 2017 · 17 revisions

A Generator of SOAP-formated Notes for Major Depressive Disorders

Introduction

The objective of this project is to generate synthetic SOAP-formatted notes that follow the structure and expected content of the actual notes for the mental disorders associated with the suicide risk.

Background

Approach

Simple Generation from Template Text, and Randomly Selected Templated Values

Develop the foundation of the note content by using this material as a template.

Generation of Narrative from Examples and with Consistent Random Values

The objective is to have some generic text and make it a little bit more customizable For example:

Subjective narrative is some statement from the patient. This can be, for now, random, English text taken from examples.

Description

This component is in a detailed, narrative format and describes the patients self-report of their current status in terms of their function, disability, symptoms and history.

It may also include information from the family or caregivers and if exact phrasing is used, should be enclosed in quotation marks. It allows the therapist to document the patients perception of their condition as it relates to their progress in rehabilitation, functional performance or quality of life.

In this section, algorithm will mimic the writing and describe the impressions of the client in the subjective section. Including impressions about the client’s/patient’s level of awareness, motivation, mood, willingness to participate. It may also list here anything the patient and/or family may say during a session. For example: The client appeared very alert and cooperative. He stated, “I’m ready to work hard today.”

Template

"Patient is a $age old $gender. It is $possesive_verb $visit_order visit. The patient appears $alertness and $cooperativeness. $possesive_verb has a history of $history. $third_person_report

Template Variables

Randomly select following statements, and insert right pronouns (he/she):

$age=yy - should match other ages/dates of birth used

$gender=(male|female) - should match other gender determinants

$visit_order=(first|repeat)

$gender_verb=(He/She) - has to match other gender determinants

$possesive_verb=(his|her) - should match other gender determinants

$alertness=(alert|very alert|lethargic|drowsy|confused)

$cooperativeness=(cooperative|indifferent|non-cooperative)

$history=(alcohol abuse|drug abuse|sexual abuse|attention problems|problems sleeping)

$third_person_report=(Patient stated "I have been thinking about ending my life"|Sister reports "$gender_verb has been irritable"|Patient stated "I don't want to be here."|Patient stated "this is pointless."|Sister reports "$gender_verb has been sleeping for days."|Mother reports "$gender_verb has been withdrawn."|Patient stated "I've been thinking about killing myself."|Patient stated "I am ready to work hard today."|) Note the last empty value in the $third_person_report. It is possible that patient did not provide any.

An Example of a Populated Template

"Patient is a 44 old male. It is his first visit. The patient appears drowsy and cooperative. He has a history of alcohol abuse. Patient stated "I've been thinking about killing myself."

Direct References

https://www.unodc.org/ddt-training/treatment/VOLUME%20A/Volume%20A%20-%20Module%203/3.Clinical%20Forms/5.Workshop_Handout_SOAP_Note.pdf

Where age, gender and condition are consistent across the note.

For now, we can copy paste some text from here and use it randomly to improve the variability (where multiple examples, search for depression or PTSD):
http://www.unm.edu/~clinic/Procedures/Forms/soap%20notes.pdf
https://www.scribd.com/document/176828204/Sample-Mental-Health-Progress-Note
http://www.soapnoteexample.com/sample-soap-note/
http://www.become-an-effective-psychotherapist.com/Clinical-Case-Notes.html
http://www.nekosmuse.com/soap.pdf
https://images.template.net/wp-content/uploads/2015/12/17103057/Medicine-Soap-Note-PDF-Free-Download.pdf

Modeling of the Disease, Patient, and Physician Characteristics

Generation and Validation with Actual Notes

Results

References

See: References