Search

Background More than 6 million people in the United States have Alzheimer disease and related dementias, receiving help from more than 11 million family or other informal caregivers. A range of traditional interventions has been developed to support family caregivers; however, most of them have not been implemented in practice and remain largely inaccessible. While recent studies have shown that family caregivers of people with dementia use Twitter to discuss their experiences, methods have not been developed to enable the use of Twitter for interventions. Objective The objective of this study is to develop an annotated data set and benchmark classification models for automatically identifying a cohort of Twitter users who have a family member with dementia. Methods Between May 4 and May 20, 2021, we collected 10,733 tweets, posted by 8846 users, that mention a dementia-related keyword, a linguistic marker that potentially indicates a diagnosis, and a select familial relationship. Three annotators annotated 1 random tweet per user to distinguish those that indicate having a family member with dementia from those that do not. Interannotator agreement was 0.82 (Fleiss kappa). We used the annotated tweets to train and evaluate support vector machine and deep neural network classifiers. To assess the scalability of our approach, we then deployed automatic classification on unlabeled tweets that were continuously collected between May 4, 2021, and March 9, 2022. Results A deep neural network classifier based on a BERT (bidirectional encoder representations from transformers) model pretrained on tweets achieved the highest F1-score of 0.962 (precision=0.946 and recall=0.979) for the class of tweets indicating that the user has a family member with dementia. The classifier detected 128,838 tweets that indicate having a family member with dementia, posted by 74,290 users between May 4, 2021, and March 9, 2022—that is, approximately 7500 users per month. Conclusions Our annotated data set can be used to automatically identify Twitter users who have a family member with dementia, enabling the use of Twitter on a large scale to not only explore family caregivers’ experiences but also directly target interventions at these users.

Background Despite the fact that medication is taken during more than 90% of pregnancies, the fetal risk for most medications is unknown, and the majority of medications have no data regarding safety in pregnancy. Objective Using β-blockers as a proof-of-concept, the primary objective of this study was to assess the utility of Twitter data for a cohort study design—in particular, whether we could identify (1) Twitter users who have posted tweets reporting that they took medication during pregnancy and (2) their associated pregnancy outcomes. Methods We searched for mentions of β-blockers in 2.75 billion tweets posted by 415,690 users who announced their pregnancy on Twitter. We manually reviewed the matching tweets to first determine if the user actually took the β-blocker mentioned in the tweet. Then, to help determine if the β-blocker was taken during pregnancy, we used the time stamp of the tweet reporting intake and drew upon an automated natural language processing (NLP) tool that estimates the date of the user’s prenatal time period. For users who posted tweets indicating that they took or may have taken the β-blocker during pregnancy, we drew upon additional NLP tools to help identify tweets that report their pregnancy outcomes. Adverse pregnancy outcomes included miscarriage, stillbirth, birth defects, preterm birth ( Results We retrieved 5114 tweets, posted by 2339 users, that mention a β-blocker, and manually identified 2332 (45.6%) tweets, posted by 1195 (51.1%) of the users, that self-report taking the β-blocker. We were able to estimate the date of the prenatal time period for 356 pregnancies among 334 (27.9%) of these 1195 users. Among these 356 pregnancies, we identified 257 (72.2%) during which the β-blocker was or may have been taken. We manually verified an adverse pregnancy outcome—preterm birth, NICU admission, low birth weight, birth defects, or miscarriage—for 38 (14.8%) of these 257 pregnancies. We manually verified a gestational age ≥37 weeks for 198 (90.4%) and a birth weight ≥5 pounds and 8 ounces for 50 (22.8%) of the 219 pregnancies for which we did not identify an adverse pregnancy outcome. Conclusions Our ability to detect pregnancy outcomes for Twitter users who posted tweets reporting that they took or may have taken a β-blocker during pregnancy suggests that Twitter can be a complementary resource for cohort studies of drug safety in pregnancy.

Background COVID-19 during pregnancy is associated with an increased risk of maternal death, intensive care unit admission, and preterm birth; however, many people who are pregnant refuse to receive COVID-19 vaccination because of a lack of safety data. Objective The objective of this preliminary study was to assess whether Twitter data could be used to identify a cohort for epidemiologic studies of COVID-19 vaccination in pregnancy. Specifically, we examined whether it is possible to identify users who have reported (1) that they received COVID-19 vaccination during pregnancy or the periconception period, and (2) their pregnancy outcomes. Methods We developed regular expressions to search for reports of COVID-19 vaccination in a large collection of tweets posted through the beginning of July 2021 by users who have announced their pregnancy on Twitter. To help determine if users were vaccinated during pregnancy, we drew upon a natural language processing (NLP) tool that estimates the timeframe of the prenatal period. For users who posted tweets with a timestamp indicating they were vaccinated during pregnancy, we drew upon additional NLP tools to help identify tweets that reported their pregnancy outcomes. Results We manually verified the content of tweets detected automatically, identifying 150 users who reported on Twitter that they received at least one dose of COVID-19 vaccination during pregnancy or the periconception period. We manually verified at least one reported outcome for 45 of the 60 (75%) completed pregnancies. Conclusions Given the limited availability of data on COVID-19 vaccine safety in pregnancy, Twitter can be a complementary resource for potentially increasing the acceptance of COVID-19 vaccination in pregnant populations. The results of this preliminary study justify the development of scalable methods to identify a larger cohort for epidemiologic studies.

Social media has recently been used to identify and study a small cohort of Twitter users whose pregnancies with birth defect outcomes—the leading cause of infant mortality—could be observed via their publicly available tweets. In this study, we exploit social media on a larger scale by developing natural language processing (NLP) methods to automatically detect, among thousands of users, a cohort of mothers reporting that their child has a birth defect. We used 22,999 annotated tweets to train and evaluate supervised machine learning algorithms—feature-engineered and deep learning-based classifiers—that automatically distinguish tweets referring to the user’s pregnancy outcome from tweets that merely mention birth defects. Because 90% of the tweets merely mention birth defects, we experimented with under-sampling and over-sampling approaches to address this class imbalance. An SVM classifier achieved the best performance for the two positive classes: an F1-score of 0.65 for the “defect” class and 0.51 for the “possible defect” class. We deployed the classifier on 20,457 unlabeled tweets that mention birth defects, which helped identify 542 additional users for potential inclusion in our cohort. Contributions of this study include (1) NLP methods for automatically detecting tweets by users reporting their birth defect outcomes, (2) findings that an SVM classifier can outperform a deep neural network-based classifier for highly imbalanced social media data, (3) evidence that automatic classification can be used to identify additional users for potential inclusion in our cohort, and (4) a publicly available corpus for training and evaluating supervised machine learning algorithms.

Background Pre-exposure prophylaxis (PrEP) is highly effective at preventing the acquisition of HIV. There is a substantial gap, however, between the number of people in the United States who have indications for PrEP and the number of them who are prescribed PrEP. Although Twitter content has been analyzed as a source of PrEP-related data (eg, barriers), methods have not been developed to enable the use of Twitter as a platform for implementing PrEP-related interventions. Objective Men who have sex with men (MSM) are the population most affected by HIV in the United States. Therefore, the objectives of this study were to (1) develop an automated natural language processing (NLP) pipeline for identifying men in the United States who have reported on Twitter that they are gay, bisexual, or MSM and (2) assess the extent to which they demographically represent MSM in the United States with new HIV diagnoses. Methods Between September 2020 and January 2021, we used the Twitter Streaming Application Programming Interface (API) to collect more than 3 million tweets containing keywords that men may include in posts reporting that they are gay, bisexual, or MSM. We deployed handwritten, high-precision regular expressions—designed to filter out noise and identify actual self-reports—on the tweets and their user profile metadata. We identified 10,043 unique users geolocated in the United States and drew upon a validated NLP tool to automatically identify their ages. Results By manually distinguishing true- and false-positive self-reports in the tweets or profiles of 1000 (10%) of the 10,043 users identified by our automated pipeline, we established that our pipeline has a precision of 0.85. Among the 8756 users for which a US state–level geolocation was detected, 5096 (58.2%) were in the 10 states with the highest numbers of new HIV diagnoses. Among the 6240 users for which a county-level geolocation was detected, 4252 (68.1%) were in counties or states considered priority jurisdictions by the Ending the HIV Epidemic initiative. Furthermore, the age distribution of the users reflected that of MSM in the United States with new HIV diagnoses. Conclusions Our automated NLP pipeline can be used to identify MSM in the United States who may be at risk of acquiring HIV, laying the groundwork for using Twitter on a large scale to directly target PrEP-related interventions at this population.

Objective We address a first step toward using social media data to supplement current efforts in monitoring population-level medication nonadherence: detecting changes to medication treatment. Medication treatment changes, like changes to dosage or to frequency of intake, that are not overseen by physicians are, by that, nonadherence to medication. Despite the consequences, including worsening health conditions or death, 50% of patients are estimated to not take medications as indicated. Current methods to identify nonadherence have major limitations. Direct observation may be intrusive or expensive, and indirect observation through patient surveys relies heavily on patients’ memory and candor. Using social media data in these studies may address these limitations. Methods We annotated 9830 tweets mentioning medications and trained a convolutional neural network (CNN) to find mentions of medication treatment changes, regardless of whether the change was recommended by a physician. We used active and transfer learning from 12 972 reviews we annotated from WebMD to address the class imbalance of our Twitter corpus. To validate our CNN and explore future directions, we annotated 1956 positive tweets as to whether they reflect nonadherence and categorized the reasons given. Results Our CNN achieved 0.50 F1-score on this new corpus. The manual analysis of positive tweets revealed that nonadherence is evident in a subset with 9 categories of reasons for nonadherence. Conclusion We showed that social media users publicly discuss medication treatment changes and may explain their reasons including when it constitutes nonadherence. This approach may be useful to supplement current efforts in adherence monitoring.

Background Although birth defects are the leading cause of infant mortality in the United States, methods for observing human pregnancies with birth defect outcomes are limited. Objective The primary objectives of this study were (i) to assess whether rare health-related events—in this case, birth defects—are reported on social media, (ii) to design and deploy a natural language processing (NLP) approach for collecting such sparse data from social media, and (iii) to utilize the collected data to discover a cohort of women whose pregnancies with birth defect outcomes could be observed on social media for epidemiological analysis. Methods To assess whether birth defects are mentioned on social media, we mined 432 million tweets posted by 112,647 users who were automatically identified via their public announcements of pregnancy on Twitter. To retrieve tweets that mention birth defects, we developed a rule-based, bootstrapping approach, which relies on a lexicon, lexical variants generated from the lexicon entries, regular expressions, post-processing, and manual analysis guided by distributional properties. To identify users whose pregnancies with birth defect outcomes could be observed for epidemiological analysis, inclusion criteria were (i) tweets indicating that the user’s child has a birth defect, and (ii) accessibility to the user’s tweets during pregnancy. We conducted a semi-automatic evaluation to estimate the recall of the tweet-collection approach, and performed a preliminary assessment of the prevalence of selected birth defects among the pregnancy cohort derived from Twitter. Results We manually annotated 16,822 retrieved tweets, distinguishing tweets indicating that the user’s child has a birth defect (true positives) from tweets that merely mention birth defects (false positives). Inter-annotator agreement was substantial: κ = 0.79 (Cohen’s kappa). Analyzing the timelines of the 646 users whose tweets were true positives resulted in the discovery of 195 users that met the inclusion criteria. Congenital heart defects are the most common type of birth defect reported on Twitter, consistent with findings in the general population. Based on an evaluation of 4169 tweets retrieved using alternative text mining methods, the recall of the tweet-collection approach was 0.95. Conclusions Our contributions include (i) evidence that rare health-related events are indeed reported on Twitter, (ii) a generalizable, systematic NLP approach for collecting sparse tweets, (iii) a semi-automatic method to identify undetected tweets (false negatives), and (iv) a collection of publicly available tweets by pregnant users with birth defect outcomes, which could be used for future epidemiological analysis. In future work, the annotated tweets could be used to train machine learning algorithms to automatically identify users reporting birth defect outcomes, enabling the large-scale use of social media mining as a complementary method for such epidemiological research.

Objective Given the uncertainty about the trends and extent of the rapidly evolving COVID-19 outbreak, and the lack of extensive testing in the United Kingdom, our understanding of COVID-19 transmission is limited. We proposed to use Twitter to identify personal reports of COVID-19 to assess whether this data can help inform as a source of data to help us understand and model the transmission and trajectory of COVID-19. Methods We used natural language processing and machine learning framework. We collected tweets (excluding retweets) from the Twitter Streaming API that indicate that the user or a member of the user's household had been exposed to COVID-19. The tweets were required to be geo-tagged or have profile location metadata in the UK. Results We identified a high level of agreement between personal reports from Twitter and lab-confirmed cases by geographical region in the UK. Temporal analysis indicated that personal reports from Twitter appear up to 2 weeks before UK government lab-confirmed cases are recorded. Conclusions Analysis of tweets may indicate trends in COVID-19 in the UK and provide signals of geographical locations where resources may need to be targeted or where regional policies may need to be put in place to further limit the spread of COVID-19. It may also help inform policy makers of the restrictions in lockdown that are most effective or ineffective.

Objective: After years of research, Twitter posts are now recognized as an important source of patient-generated data, providing unique insights into population health. A fundamental step to incorporating Twitter data in pharmacoepidemiological research is to automatically recognize medication mentions in tweets. Given that lexical searches for medication names may fail due to misspellings or ambiguity with common words, we propose a more advanced method to recognize them. Methods: We present Kusuri, an Ensemble Learning classifier, able to identify tweets mentioning drug products and dietary supplements. Kusuri ("medication" in Japanese) is composed of two modules. First, four different classifiers (lexicon-based, spelling-variant-based, pattern-based and one based on a weakly-trained neural network) are applied in parallel to discover tweets potentially containing medication names. Second, an ensemble of deep neural networks encoding morphological, semantical and long-range dependencies of important words in the tweets discovered is used to make the final decision. Results: On a balanced (50-50) corpus of 15,005 tweets, Kusuri demonstrated performances close to human annotators with 93.7% F1-score, the best score achieved thus far on this corpus. On a corpus made of all tweets posted by 113 Twitter users (98,959 tweets, with only 0.26% mentioning medications), Kusuri obtained 76.3% F1-score. There is not a prior drug extraction system that compares running on such an extremely unbalanced dataset. Conclusion: The system identifies tweets mentioning drug names with performance high enough to ensure its usefulness and ready to be integrated in larger natural language processing systems., This is a pre-copy-editing, author-produced PDF of an article accepted for publication in JAMIA following peer review. The definitive publisher-authenticated version is "D. Weissenbacher, A. Sarker, A. Klein, K. O'Connor, A. Magge, G. Gonzalez-Hernandez, Deep neural networks ensemble for detecting medication mentions in tweets, Journal of the American Medical Informatics Association, ocz156, 2019"

Objective The assessment of written medical examinations is a tedious and expensive process, requiring significant amounts of time from medical experts. Our objective was to develop a natural language processing (NLP) system that can expedite the assessment of unstructured answers in medical examinations by automatically identifying relevant concepts in the examinee responses. Materials and methods Our NLP system, In telligent C l i nical T ext E valuator (INCITE), is semi-supervised in nature. Learning from a limited set of fully annotated examples, it sequentially applies a series of customized text comparison and similarity functions to determine if a text span represents an entry in a given reference standard. Combinations of fuzzy matching and set intersection-based methods capture inexact matches and also fragmented concepts. Customizable, dynamic similarity-based matching thresholds allow the system to be tailored for examinee responses of different lengths. Results INCITE achieved an average F1-score of 0.89 (precision = 0.87, recall = 0.91) against human annotations over held-out evaluation data. Fuzzy text matching, dynamic thresholding and the incorporation of supervision using annotated data resulted in the biggest jumps in performances. Discussion Long and non-standard expressions are difficult for INCITE to detect, but the problem is mitigated by the use of dynamic thresholding (i.e., varying the similarity threshold for a text span to be considered a match). Annotation variations within exams and disagreements between annotators were the primary causes for false positives. Small amounts of annotated data can significantly improve system performance. Conclusions The high performance and interpretability of INCITE will likely significantly aid the assessment process and also help mitigate the impact of manual assessment inconsistencies.

Background In the United States, the rapidly evolving COVID-19 outbreak, the shortage of available testing, and the delay of test results present challenges for actively monitoring its spread based on testing alone. Objective The objective of this study was to develop, evaluate, and deploy an automatic natural language processing pipeline to collect user-generated Twitter data as a complementary resource for identifying potential cases of COVID-19 in the United States that are not based on testing and, thus, may not have been reported to the Centers for Disease Control and Prevention. Methods Beginning January 23, 2020, we collected English tweets from the Twitter Streaming application programming interface that mention keywords related to COVID-19. We applied handwritten regular expressions to identify tweets indicating that the user potentially has been exposed to COVID-19. We automatically filtered out “reported speech” (eg, quotations, news headlines) from the tweets that matched the regular expressions, and two annotators annotated a random sample of 8976 tweets that are geo-tagged or have profile location metadata, distinguishing tweets that self-report potential cases of COVID-19 from those that do not. We used the annotated tweets to train and evaluate deep neural network classifiers based on bidirectional encoder representations from transformers (BERT). Finally, we deployed the automatic pipeline on more than 85 million unlabeled tweets that were continuously collected between March 1 and August 21, 2020. Results Interannotator agreement, based on dual annotations for 3644 (41%) of the 8976 tweets, was 0.77 (Cohen κ). A deep neural network classifier, based on a BERT model that was pretrained on tweets related to COVID-19, achieved an F1-score of 0.76 (precision=0.76, recall=0.76) for detecting tweets that self-report potential cases of COVID-19. Upon deploying our automatic pipeline, we identified 13,714 tweets that self-report potential cases of COVID-19 and have US state–level geolocations. Conclusions We have made the 13,714 tweets identified in this study, along with each tweet’s time stamp and US state–level geolocation, publicly available to download. This data set presents the opportunity for future work to assess the utility of Twitter data as a complementary resource for tracking the spread of COVID-19.

Background: In the United States, 17% of pregnancies end in fetal loss: miscarriage or stillbirth. Preterm birth affects 10% of live births in the United States and is the leading cause of neonatal death globally. Preterm births with low birthweight are the second leading cause of infant mortality in the United States. Despite their prevalence, the causes of miscarriage, stillbirth, and preterm birth are largely unknown. Objective: The primary objectives of this study are to (1) assess whether women report miscarriage, stillbirth, and preterm birth, among others, on Twitter, and (2) develop natural language processing (NLP) methods to automatically identify users from which to select cases for large-scale observational studies. Methods: We handcrafted regular expressions to retrieve tweets that mention an adverse pregnancy outcome, from a database containing more than 400 million publicly available tweets posted by more than 100,000 users who have announced their pregnancy on Twitter. Two annotators independently annotated 8109 (one random tweet per user) of the 22,912 retrieved tweets, distinguishing those reporting that the user has personally experienced the outcome (“outcome” tweets) from those that merely mention the outcome (“non-outcome” tweets). Inter-annotator agreement was κ = 0.90 (Cohen’s kappa). We used the annotated tweets to train and evaluate feature-engineered and deep learning-based classifiers. We further annotated 7512 (of the 8109) tweets to develop a generalizable, rule-based module designed to filter out reported speech—that is, posts containing what was said by others—prior to automatic classification. We performed an extrinsic evaluation assessing whether the reported speech filter could improve the detection of women reporting adverse pregnancy outcomes on Twitter. Results: The tweets annotated as “outcome” include 1632 women reporting miscarriage, 119 stillbirth, 749 preterm birth or premature labor, 217 low birthweight, 558 NICU admission, and 458 fetal/infant loss in general. A deep neural network, BERT-based classifier achieved the highest overall F1-score (0.88) for automatically detecting “outcome” tweets (precision = 0.87, recall = 0.89), with an F1-score of at least 0.82 and a precision of at least 0.84 for each of the adverse pregnancy outcomes. Our reported speech filter significantly (P