Investigating Individuals' Perceptions Regarding the Context Around the Low Back Pain Experience: Topic Modeling...

Area of Science:

• Computational linguistics
• Social media analytics
• Public health

Background:

• Low back pain (LBP) is a leading global cause of disability.
• Understanding patient beliefs and experiences is crucial for improving LBP outcomes.
• Social media offers a novel platform for large-scale, unsolicited data collection on LBP experiences.

Purpose of the Study:

• To identify contextual variables of the LBP experience from a first-person perspective.
• To gain insights into individuals' beliefs and perceptions regarding LBP.
• To leverage social media data for a deeper understanding of LBP.

Main Methods:

• Analysis of 896,867 tweets related to LBP from 2014-2018.
• Comparison of multiple topic modeling algorithms, including Latent Dirichlet Allocation (LDA).
• Qualitative content analysis of identified topics by domain experts.

Main Results:

• LDA was the best-performing algorithm, identifying 60 topics grouped into 19 categories.
• "Emotion and beliefs" (17.6%) and "physical activity" (13.85%) were the most prevalent themes.
• A significant majority (72%) of tweets within "emotion and beliefs" expressed negative sentiment.

Conclusions:

• Social media content analysis provides a granular perspective on LBP experiences, aligning with traditional qualitative findings.
• This approach can inform the development of more effective and personalized LBP care models.
• Understanding patient-reported themes can lead to improved outcomes for individuals with low back pain.