JMIR Publications

ABSTRACT

Background:

Opioids are often prescribed by doctors to relieve pain following surgery, accidents, or diseases such as cancer. However, prescribed opioids are risky and subject to misuse and addiction. Since 1964, the World Health Organization (WHO) has used the term "opioid dependency" instead of "opioid addiction." Understanding opioid dependency’s risk factors among analgesic patients may potentially inform clinical care and opioid prescribing policies aimed at reducing opioid addiction. One of the effective methods for treating addiction is Medication-assisted treatment (MAT) which is a combination of three FDA-approved medications for opioid dependence: methadone, buprenorphine, and naloxone, with behavioral therapies. These medications help reduce opioid use gradually and reduce symptoms and the desire to seek out unprescribed opioids.

Objective:

While MAT has been shown to be effective initially, there is a need for more information from the patients' perspective, especially in real-world settings. This information gap is more about the patient satisfaction with various opioid dependence treatment modalities across the spectrum of opioid use disorder severity, the ideal duration of treatment, the impact of individual factors, and the long-term transition off these treatments. It would be challenging and time-consuming for health providers to assess the day-to-day experience of all these patients. However, a broad survey of patients' viewpoints can be obtained through social media and drug review forums and assessed using automated methods to discover experiences expressed at the moment over long periods. Several prior studies utilized natural language processing (NLP) to predict patients’ outcomes in a number of healthcare settings. Determinedly, the primary aim of this study is to use NLP methods to detect the patients' perceived satisfaction with two well-studied opioid dependence medications, methadone, and buprenorphine/naloxone, from text posted on health-related social media.

Methods:

To build the predictive model for detecting the patients’ satisfaction with the two targeted opioid dependence medications, we first employed different analyses to select features: 1) applying an unsupervised clustering technique to identify topic modeling and 2) identifying biomedical concepts by applying MetaMap, a widely used tool created by researchers at the National Library of Medicine. Using two healthcare forum websites, we performed the topic modeling and MetaMap on 4,353 patient reviews related to methadone and buprenorphine/naloxone drugs from 2008 to 2021. We then built a dataset using the vectorized text, topic models, MetaMap categories, and duration of treatment as features and using the numerical patients’ satisfaction rating as the class. We compared prediction models developed using Logistic Regression, Elastic Net, LASSO, Random Forest Classifier, Ridge Classifier, and XG Boost.

Results:

We developed a predictive model to recognize patient satisfaction with opioid dependence treatments from health forums. The F-score of the predictive models across all methods ranged from 83.4% to 90.6%. The Elastic Net model, a regularized regression method, outperformed the other models.

Conclusions:

The patient's assessment of satisfaction with opioid dependency treatment can be predicted by using automated text analysis. We found that adding biomedical concepts such as symptoms, drug name, and illness, along with the duration of treatment and topic models, could improve predicting patient satisfaction with opioid treatments.