Deep Learning-Based Natural Language Processing for the Identification and Multi-Label Categorization of Social Factors of Healthcare from Unorganized Electronic Medical Records

Davlatov, Sharipov, Mamatkulova, Boymatova, Oltiboyeva, Shamsutdinova and Kitayeva

Health Leadership and Quality of Life, 2024, vol. 3, .585

Abstract: Social Factors of Healthcare (SFH) are non-medical determinants that may significantly influence patient health outcomes. Nevertheless, SFH is seldom included in Unorganized Electronic Medical Records (UEMR) data, such as diagnostic codes, and is often found in uncontrolled descriptive medical notes. Consequently, discerning social factors from UEMR data has gained paramount significance. Previous research towards using Natural Language Processing (NLP) for the automated extraction of SFH from text often emphasizes a selective approach to SFH. It fails to include the current advancements in Deep Learning (DL). This study proposes Deep Learning-Based Natural Language Processing for the identification and multi-label categorization (DL-NLP-MLC) of SFH from UEMR. Information was obtained from the Medical Information Mart for Intensive Care (MIMIC-III) dataset. The database consisted of 4,124 socially connected phrases derived from 2,785 medical notes. A framework for automatic MLC for multiple SFH types has been established. The database consisted of descriptive medical notes categorized as "SFH" inside the MIMIC-III medical dataset. Four types of categorization models have been trained: Decision Tree (DT), Random Forest (RF), and Long Short-Term Memory (LSTM). The efficacy of DL-NLP-MLC has been evaluated using accuracy, precision, recall, Area Under the Curve (AUC), and F1 score. The findings indicated that, in general, LSTM surpassed the other models of categorization with AUC (98.4%) and Accuracy (94.6%) for drug abuse SFH. The suggested method of training a DL classifier on a dataset rich in structured feature hierarchies may yield a very effective classifier using UEMR. Evidence demonstrates that model performance correlates with the semantic variety used by health practitioners and the automated creation of medical statements for documenting SFH.

Date: 2024
References: Add references at CitEc
Citations:

There are no downloads for this item, see the EconPapers FAQ for hints about obtaining it.

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:dbk:health:v:3:y:2024:i::p:.585:id:.585

DOI: 10.56294/hl2024.585

Access Statistics for this article

More articles in Health Leadership and Quality of Life from AG Editor
Bibliographic data for series maintained by Javier Gonzalez-Argote ().