Analyzing Dementia Caregivers' Experiences on Twitter: A Term-Weighted Topic Modeling Approach

Area of Science:

• Social Sciences
• Computer Science
• Health Sciences

Background:

• Dementia significantly affects patients and their families, necessitating a deeper understanding of family caregiver needs.
• Family caregivers face unique challenges that require tailored support and resources.

Purpose of the Study:

• To introduce a novel approach for analyzing social media data to understand family caregiver experiences with dementia.
• To identify key topics and challenges faced by family caregivers of individuals with dementia.

Main Methods:

• Collected and preprocessed Twitter (now X) data using advanced natural language processing.
• Enhanced the Gibbs Sampling Dirichlet Multinomial Mixture Model (GSDMM) with term-weighting for improved topic clarity.
• Utilized ChatGPT 4 and human experts to analyze identified topics and caregiver challenges.

Main Results:

• The enhanced GSDMM approach identified semantically rich and contextually coherent topics related to dementia caregiving.
• This method demonstrated superior clarity and consistency compared to the BERTopic model.
• Multifaceted challenges faced by family caregivers were uncovered through qualitative analysis.

Conclusions:

• The developed method offers a valuable tool for analyzing social media data to understand dementia caregiver experiences.
• Findings can inform healthcare professionals, researchers, and support organizations to better address caregiver needs.
• Improved understanding of caregiver challenges can lead to enhanced support and care for families affected by dementia.