Sentiment Analysis Reveals Public Perceptions of Coronary Artery Calcium Testing

By Muhammad OsamaApr 15 2024Reviewed by Susha Cheriyedath, M.Sc.

In a recent article published in the journal NPJ Digital Medicine, researchers from the USA comprehensively analyzed contemporary public attitudes and beliefs about coronary artery calcium (CAC) testing using discussions from the social media platform Reddit.

They employed an artificial intelligence (AI) pipeline, including natural language processing and unsupervised machine learning techniques, to systematically understand prevalent sentiments, concerns, and perceptions regarding CAC imaging within the general population. Their research aimed to provide insights into public perceptions that can inform strategies for shared decision-making around atherosclerotic cardiovascular disease (ASCVD) management and public health interventions.

Background

Coronary artery calcium (CAC) imaging is pivotal in cardiovascular risk assessment, detecting calcified plaque in coronary arteries. The presence and extent of CAC serve as indicators of atherosclerosis and aid healthcare practitioners in predicting future cardiovascular events. This technology plays a crucial role in preventive cardiology by assisting in risk stratification and treatment decisions.

By visualizing calcified plaque, CAC imaging provides valuable insights into a patient's cardiovascular health, enabling targeted interventions to mitigate risks and improve outcomes. It serves as a cornerstone in the comprehensive evaluation of cardiovascular disease risk, complementing traditional risk factors with direct visualization of coronary artery pathology.

As a non-invasive imaging modality, CAC imaging offers a safe and efficient means of assessing cardiovascular risk, empowering clinicians to proactively manage at-risk patients. However, public awareness and understanding of CAC testing are crucial for its effective implementation. This study aimed to explore the public's knowledge, attitudes, and beliefs regarding CAC testing through an analysis of social media discussions.

About the Research

In the present paper, the authors conducted a comprehensive study to analyze 5,606 Reddit discussions. They utilized AI techniques to categorize the discussions into different thematic clusters and gain insights into the key themes discussed. They employed topic modeling or clustering analysis and sentiment analysis techniques to analyze social media discussions. These techniques allowed them to gain insights into the prevailing attitudes and sentiments towards CAC testing.

To begin the analysis, the study employed a pre-trained model named all-MiniLM-L6-v2, which was trained on a diverse dataset encompassing Reddit posts and medical journal papers. Leveraging the insights and patterns gleaned from this extensive training data, the authors employed this pre-trained model to examine discussions related to CAC testing.

These discussions were then embedded into a numerical representation, allowing the researchers to transform the text-based discussions into a format that could be analyzed using clustering techniques. Subsequently, clustering analysis was conducted on the embedded discussions to identify overarching themes within the discussion.

Clustering analysis techniques were applied to group discussions related to CAC testing based on their content and similarities. This approach enabled the authors to identify common themes or topics emerging from the discussions, facilitating a comprehensive understanding of various aspects and perspectives surrounding CAC testing. By grouping similar discussions, the researchers gained insight into the different characteristics and features of the topic. This method also enhanced the interpretation of the data, contributing to a more nuanced analysis of the discussions.

In addition to topic modeling, the study also employed sentiment analysis using a separate model known as the robustly optimized bidirectional encoder representations from transformers (RoBERTa) approach. RoBERTa is a variant of the bidirectional encoder representations from transformers (BERT) model that has been trained on social media posts.

Sentiment analysis involves determining the sentiment or emotion expressed in a text. In this study, the researchers utilized RoBERTa to classify the sentiment of the discussions related to CAC testing. This technique provided insights into the emotional tone and attitudes expressed within the discussions, complementing the topic modeling analysis by offering a deeper understanding of the sentiments associated with CAC testing discussions. The sentiment analysis revealed whether the discussions had a positive, neutral, or negative sentiment. Understanding the sentiment of the discussions can provide valuable insights into the prevailing attitudes and perceptions toward CAC testing.

Research Findings

The outcomes revealed 91 topics and 14 groups of discussions covering various aspects of CAC testing, including its benefits, limitations, accuracy, cost-effectiveness, and potential risks. The wide range of topics indicated significant interest and engagement in discussions surrounding CAC testing.

Additionally, the sentiment analysis showed that the majority of the discussions have a neutral or slightly negative sentiment. This suggested that participants may have mixed feelings or concerns about CAC testing, possibly due to its associated risks such as radiation exposure or financial costs.

Furthermore, researchers observed a decline in sentiment over time. They found that discussions became more negative each year from 2013 through 2023, indicating increasing skepticism or dissatisfaction with CAC testing. Additionally, they recommended further investigation to understand the reasons behind this decline in sentiment.

The insights gained from this AI-driven investigation have implications for healthcare providers, practitioners, policymakers, and researchers in cardiovascular health. By understanding public perceptions, misconceptions, or knowledge gaps regarding CAC, stakeholders can customize educational initiatives, enhance shared decision-making practices, and design interventions that promote informed discussions about cardiovascular risk assessment.

Conclusion

In conclusion, the paper effectively utilized AI to analyze and highlight the public perceptions about CAC testing. The authors revealed a diverse range of topics and sentiments surrounding CAC testing discussions on social media platforms. They recommended the need for targeted educational interventions to address misconceptions and improve public understanding of CAC testing. Moreover, they demonstrated the potential of AI-enabled analysis to gather public opinions efficiently and cost-effectively.