Text Fingerprinting and Topic Mining in the Prescription Opioid Use Literature

Background: Prescription opioids are powerful pain-reducing medications, but they may cause a variety of adverse effects. Long-term prescription opioid use (POU) is contributing to an opioid-related epidemic of addiction and death, and the scope of the opioid crisis continues to expand. As such, there is a need to identify the adverse effects associated with prescription opioid use (POU). Thousands of articles that focus on POU and its associated medical disorders have been published. However, it is time-consuming and labor-intensive to extract and understand the information of all POU-related published articles. Methods: In this study, we applied the well-adapted topic modeling method, Latent Dirichlet Allocation (LDA), to perform text mining on POU-related literature. We compiled six large academic abstract datasets by searching PubMed using the Medical Subject Headings (MeSH): prescription opioid, codeine, morphine, hydrocodone, oxycodone, and methadone. We then applied topic modeling to identify topics and analyze topic similarities/differences in these six datasets. Word clouds and histograms were used to depict the distribution of vocabularies over each topic in which the most prevalent words conveyed a topic’s substance. Results: The LDA topics recaptured the search keywords in PubMed, and further revealed relevant themes, such as patients, drugs, side effects, and association links between different POU and risk factors, such as gender and age. Moreover, based on the topic modeling results, TreeMap was used to fingerprint abstracts, which revealed the possibility of constructing a visualized literature index by combining topic modeling and visualization tools such as TreeMap. Meanwhile, while performing trend analysis to explore the prevalent topic dynamics in the POU-related literature, we found that an increasing trend in opioid prescription and its associated health risks are assessed as the most central issues. Conclusion: The topic modeling results presented in this study not only convey an understandable and thematic structure of the POU literature, but also provide a means to discover which documents contain information about medical disorders associated with POU, thus, reducing the time and effort needed to review the literature for relevant articles. These results can be used as a preliminary study to systematically understand the risk factors related to increased POU-associated medical disorders.