Search

Due to the high economic and public health burden of chronic pain, and the risk of public health consequences of opioid-based treatments, there is a need to identify effective alternative therapies. The evidence basis for many alternative therapies is weak or nonexistent. Social media presents a unique opportunity to gather large-scale knowledge about such therapies self-reported by sufferers themselves. We attempted to (i) verify the presence of largescale chronic pain-related chatter on Twitter, (ii) develop natural language processing (NLP) and machine learning for automatically detecting chronic pain sufferers, and (iii) identify the types of chronic pain-related information reported by them. We collected data from Twitter using chronic pain-related hashtags and keywords. We manually performed binary annotation of a sample of 4998 posts to indicate if they were self-reports of chronic pain experiences or not, and obtained inter-annotator agreement of 0.82 (Cohen’s kappa). We trained and evaluated several state-of-the-art transformer-based text classification models using the annotated data. The RoBERTa model outperformed all others (F1 score = 0.84; 95% CI: 0.80-0.89), and we used this model to classify a large number of unlabeled posts. We identified 22,795 self-reported chronic pain sufferers and collected their past posted data. Via manual and NLP-driven analyses, we found information about but not limited to alternative treatments, sufferers’ sentiments about treatments, side effects, and self-management strategies. Our social media-based approach will result in an automatically growing massive cohort over time, and the data can be leveraged to identify self-reported effective alternative therapies for diverse chronic pain types.

BackgroundSocial media have served as lucrative platforms for spreading misinformation and for promoting fraudulent products for the treatment, testing, and prevention of COVID-19. This has resulted in the issuance of many warning letters by the United States Food and Drug Administration (FDA). While social media continue to serve as the primary platform for the promotion of such fraudulent products, they also present the opportunity to identify these products early by employing effective social media mining methods. In this study, we employ natural language processing and time series anomaly detection methods for automatically detecting fraudulent COVID-19 products early from Twitter. Our approach is based on the intuition that increases in the popularity of fraudulent products lead to corresponding anomalous increases in the volume of chatter regarding them. ResultsWe utilized an anomaly detection method on streaming COVID-19-related Twitter data to detect potentially anomalous increases in mentions of fraudulent products. We compared the anomaly signal generation date for each product with the corresponding FDA letter issuance date. Issue dates ranged from March 6, 2020 to June 22, 2021, and 44 key phrases representing fraudulent products were included. From 577,872,350 posts made between February 19, 2020 to December 31, 2020, our unsupervised approach detected 34/44 (77.3%) signals about fraudulent products earlier than the FDA letter issuance dates, and an additional 6/44 (13.6%) within a week following the corresponding FDA letters. ConclusionsOur proposed method is simple, effective, and easy to deploy, and does not require high-performance computing machinery, unlike deep neural network-based methods. The method can be easily extended to other types of signal detection from social media data.

Traditional substance use (SU) surveillance methods, such as surveys, incur substantial lags. Due to the continuously evolving trends in SU, insights obtained via such methods are often outdated. Social media-based sources have been proposed for obtaining timely insights, but methods leveraging such data cannot typically provide fine-grained statistics about subpopulations, unlike traditional approaches. We address this gap by developing methods for automatically characterizing a large Twitter nonmedical prescription medication use (NPMU) cohort (n = 288,562) in terms of age-group, race, and gender. Our natural language processing and machine learning methods for automated cohort characterization achieved 0.88 precision (95% CI:0.84 to 0.92) for age-group, 0.90 (95% CI: 0.85 to 0.95) for race, and 94% accuracy (95% CI: 92 to 97) for gender, when evaluated against manually annotated gold-standard data. We compared automatically derived statistics for NPMU of tranquilizers, stimulants, and opioids from Twitter with statistics reported in the National Survey on Drug Use and Health (NSDUH) and the National Emergency Department Sample (NEDS). Distributions automatically estimated from Twitter were mostly consistent with the NSDUH [Spearman r : race: 0.98 ( P < 0.005); age-group: 0.67 ( P < 0.005); gender: 0.66 ( P = 0.27)] and NEDS, with 34/65 (52.3%) of the Twitter-based estimates lying within 95% CIs of estimates from the traditional sources. Explainable differences (e.g., overrepresentation of younger people) were found for age-group-related statistics. Our study demonstrates that accurate subpopulation-specific estimates about SU, particularly NPMU, may be automatically derived from Twitter to obtain earlier insights about targeted subpopulations compared to traditional surveillance approaches.

Nonmedical use of prescription drugs (NMPDU) is a global health concern. The extent of, behaviors and emotions associated with, and reasons for NMPDU are not well-captured through traditional instruments such as surveys, prescribing databases and insurance claims. Therefore, this study analyses ∼130 million public posts from 87,718 Twitter users in terms of expressed emotions, sentiments, concerns, and potential reasons for NMPDU via natural language processing. Our results show that users in the NMPDU group express more negative emotions and less positive emotions, more concerns about family, the past and body, and less concerns related to work, leisure, home, money, religion, health and achievement, compared to a control group (i.e., users who never reported NMPDU). NMPDU posts tend to be highly polarized, indicating potential emotional triggers. Gender-specific analysis shows that female users in the NMPDU group express more content related to positive emotions, anticipation, sadness, joy, concerns about family, friends, home, health and the past, and less about anger, compared to males. The findings of the study can enrich our understanding of NMPDU.

ObjectiveSocial media can be an effective but challenging resource for conducting close-to-real-time assessments of consumers’ perceptions about health services. Our objective was to develop and evaluate an automatic pipeline, involving natural language processing and machine learning, for automatically characterizing user-posted Twitter data about Medicaid.Material and MethodsWe collected Twitter data via the public API using Medicaid-related keywords (Corpus-1), and the website’s search option using agency-specific handles (Corpus-2). We manually labeled a sample of tweets into five pre-determined categories or other, and artificially increased the number of training posts from specific low-frequency categories. We trained and evaluated several supervised learning algorithms using manually-labeled data, and applied the best-performing classifier to collected tweets for post-classification analyses assessing the utility of our methods.ResultsWe collected 628,411 and 27,377 tweets for Corpus-1 and -2, respectively. We manually annotated 9,571 (Corpus-1: 8,180; Corpus-2: 1,391) tweets, using 7,923 (82.8%) for training and 1,648 (17.2%) for evaluation. A BERT-based (bidirectional encoder representations from transformers) classifier obtained the highest accuracies (83.9%, Corpus-1; 86.4%, Corpus-2), outperforming the second-best classifier (SVMs: 79.6%; 76.4%). Post-classification analyses revealed differing inter-corpora distributions of tweet categories, with political (63%) and consumer-feedback (43%) tweets being most frequent for Corpus-1 and -2, respectively.Discussion and ConclusionThe broad and variable content of Medicaid-related tweets necessitates automatic categorization to identify topic-relevant posts. Our proposed pipeline presents a feasible solution for automatic categorization, and can be deployed/generalized for health service programs other than Medicaid. Annotated data and methods are available for future studies (LINK_TO_BE_AVAILABLE).

The performances of current medical text summarization systems rely on resource-heavy domain-specific knowledge sources, and preprocessing methods (e.g., classification or deep learning) for deriving semantic information. Consequently, these systems are often difficult to customize, extend or deploy in low-resource settings, and are operationally slow. We propose a fast summarization system that can aid practitioners at point-of-care, and, thus, improve evidence-based healthcare. At runtime, our system utilizes similarity measurements derived from pre-trained domain-specific word embeddings in addition to simple features, rather than clunky knowledge bases and resource-heavy preprocessing. Automatic evaluation on a public dataset for evidence-based medicine shows that our system’s performance, despite the simple implementation, is statistically comparable with the state-of-the-art.

Breast cancer patients often discontinue their long-term treatments, such as hormone therapy, increasing the risk of cancer recurrence. These discontinuations are often caused by adverse patient-centered outcomes (PCOs) due to hormonal drug side effects or other factors. PCOs are not detectable through laboratory tests and are sparsely documented in electronic health records. Thus, there is a need to explore other sources of information for PCOs associated with breast cancer treatments. Social media is a promising resource, but extracting true PCOs from it first requires the accurate detection of breast cancer patients. We describe a natural language processing (NLP) architecture for automatically detecting breast cancer patients from Twitter based on their self-reports. The architecture employs breast cancer-related keywords to collect streaming data from Twitter, applies NLP patterns to pre-filter noisy posts, and then employs a machine learning classifier trained using manually-annotated data (n=5019) for distinguishing firsthand self-reports of breast cancer from other tweets. A classifier based on bidirectional encoder representations from transformers (BERT) showed human-like performance and achieved F1-score of 0.857 (inter-annotator agreement: 0.845; Cohen's kappa) for the positive class, considerably outperforming the next best classifier--a deep neural network (F1-score: 0.665). Qualitative analyses of posts from automatically-detected users revealed discussions about side effects, non-adherence, and mental health conditions, illustrating the feasibility of our social media-based approach for studying breast cancer-related PCOs from a large population.

ObjectiveTo mine Twitter and quantitatively analyze COVID-19 symptoms self-reported by users, compare symptom distributions across studies, and create a symptom lexicon for future research.Materials and MethodsWe retrieved tweets using COVID-19-related keywords, and performed semiautomatic filtering to curate self-reports of positive-tested users. We extracted COVID-19-related symptoms mentioned by the users, mapped them to standard concept IDs in the Unified Medical Language System, and compared the distributions to those reported in early studies from clinical settings.ResultsWe identified 203 positive-tested users who reported 1002 symptoms using 668 unique expressions. The most frequently-reported symptoms were fever/pyrexia (66.1%), cough (57.9%), body ache/pain (42.7%), fatigue (42.1%), headache (37.4%), and dyspnea (36.3%) amongst users who reported at least 1 symptom. Mild symptoms, such as anosmia (28.7%) and ageusia (28.1%), were frequently reported on Twitter, but not in clinical studies.ConclusionThe spectrum of COVID-19 symptoms identified from Twitter may complement those identified in clinical settings.

Opioid use disorder (OUD) is a public health emergency in the United States. Over 47,000 overdose-related deaths in 2017 involved opioids. Medication-assisted treatment (MAT), in particular, buprenorphine and buprenorphine combination products such as Suboxone®, is the most effective, evidence-based treatment for OUD. However, there are a limited number of conclusive scientific studies that provide guidance to medical professionals about strategies for using buprenorphine to achieve stable recovery. In this study, we used data-driven natural language processing methods to mine a total of 16,146 posts about buprenorphine from 1933 unique users on the anonymous social network Reddit. Analysis of a sample of these posts showed that 74% of the posts described users’ personal experiences and that the top three topics included advice on using Suboxone® (55.0%), Suboxone® dosage information (35.5%) and information about Suboxone® tapering (32.0%). Based on two models, one that incorporated ‘upvoting’ by other members and one that did not, we found that Reddit users reported more successful recovery with longer tapering schedules, particularly from 2.0 mg to 0.0 mg (median: 93 days; mean: 95 days), as compared to shorter tapering schedules investigated in past clinical trials. Diarrhea, insomnia, restlessness, and fatigue were commonly reported adverse events. Physical exercise, clonidine, and Imodium® were frequently reported to help during the recovery process. Due to the difficulties of conducting longer-term clinical trials involving patients with OUD, clinicians should consider other information sources including peer discussions from the abundant, real-time information available on Reddit.Significance StatementOpioid use disorder (OUD) is a national crisis in the United States and buprenorphine is one of the most effective evidence-based treatments. However, few studies have explored successful strategies for using and tapering buprenorphine to achieve stable recovery, particularly due to the difficulties of conducting long-term studies involving patients with OUD. In this study, we show that discussions on the anonymous social network Reddit may be leveraged, via automatic text mining methods, to discover successful buprenorphine use and tapering strategies. We discovered that longer tapering schedules, compared to those investigated in past clinical trials, may lead to (self-reported) sustained recovery. Furthermore, Reddit posts also provide key information regarding buprenorphine withdrawal, cravings, adjunct medications for withdrawal symptoms and relapse prevention strategies.