The reporting of neuropsychiatric symptoms in electronic health records of individuals with Alzheimer's disease: a natural language processing study

Using AI-Based Technologies to Help Nurses Detect Behavioral Disorders: Narrative Literature Review

BACKGROUND

The behavioral and psychological symptoms of dementia (BPSD) are common among people with dementia and have multiple negative consequences. Artificial intelligence-based technologies (AITs) have the potential to help nurses in the early prodromal detection of BPSD. Despite significant recent interest in the topic and the increasing number of available appropriate devices, little information is available on using AITs to help nurses striving to detect BPSD early.

OBJECTIVE

The aim of this study is to identify the number and characteristics of existing publications on introducing AITs to support nursing interventions to detect and manage BPSD early.

METHODS

A literature review of publications in the PubMed database referring to AITs and dementia was conducted in September 2023. A detailed analysis sought to identify the characteristics of these publications. The results were reported using a narrative approach.

RESULTS

A total of 25 publications from 14 countries were identified, with most describing prospective observational studies. We identified three categories of publications on using AITs and they are (1) predicting behaviors and the stages and progression of dementia, (2) screening and assessing clinical symptoms, and (3) managing dementia and BPSD. Most of the publications referred to managing dementia and BPSD.

CONCLUSIONS

Despite growing interest, most AITs currently in use are designed to support psychosocial approaches to treating and caring for existing clinical signs of BPSD. AITs thus remain undertested and underused for the early and real-time detection of BPSD. They could, nevertheless, provide nurses with accurate, reliable systems for assessing, monitoring, planning, and supporting safe therapeutic interventions.

Optimizing classification of diseases through language model analysis of symptoms

This paper investigated the use of language models and deep learning techniques for automating disease prediction from symptoms. Specifically, we explored the use of two Medical Concept Normalization-Bidirectional Encoder Representations from Transformers (MCN-BERT) models and a Bidirectional Long Short-Term Memory (BiLSTM) model, each optimized with a different hyperparameter optimization method, to predict diseases from symptom descriptions. In this paper, we utilized two distinct dataset called Dataset-1, and Dataset-2. Dataset-1 consists of 1,200 data points, with each point representing a unique combination of disease labels and symptom descriptions. While, Dataset-2 is designed to identify Adverse Drug Reactions (ADRs) from Twitter data, comprising 23,516 rows categorized as ADR (1) or Non-ADR (0) tweets. The results indicate that the MCN-BERT model optimized with AdamP achieved 99.58% accuracy for Dataset-1 and 96.15% accuracy for Dataset-2. The MCN-BERT model optimized with AdamW performed well with 98.33% accuracy for Dataset-1 and 95.15% for Dataset-2, while the BiLSTM model optimized with Hyperopt achieved 97.08% accuracy for Dataset-1 and 94.15% for Dataset-2. Our findings suggest that language models and deep learning techniques have promise for supporting earlier detection and more prompt treatment of diseases, as well as expanding remote diagnostic capabilities. The MCN-BERT and BiLSTM models demonstrated robust performance in accurately predicting diseases from symptoms, indicating the potential for further related research.