Search

BACKGROUND Adverse reactions of drugs attract significant concerns in both clinical practice and public health monitoring. Multiple measures have been put into place to increase postmarketing surveillance of drug adverse effects and improve drug safety. These include implementing spontaneous reporting systems and developing automated natural language processing (NLP) systems from electronic health records and social media data to collect evidence of adverse drug events (ADEs) that can be further investigated as possible adverse reactions. OBJECTIVE While using social media for this purpose has potential, it is not clear whether it is a reliable source for this information. Our work aims to 1) develop NLP approaches to identify ADEs on social media and 2) assess the reliability of social media data to identify ADEs. METHODS We propose a collocated Long Short-Term Memory network model with attentive pooling and aggregated, contextual representation generated by a pretrained model. We applied the model on large-scale Twitter data to identify ADE-related tweets. We conducted a qualitative content analysis on those tweets to validate the reliability of social media data as a means to collect such information. RESULTS The model outperforms the variant without the contextual representations during both the validation and evaluation phases. Through the content analysis on ADE tweets, we observe that ADE-related discussions referred to seven ADE themes. Mental health-related, sleep-related, and pain-related ADE discussions were most frequent. We also contrast known adverse drug reactions to f those mentioned in tweets. CONCLUSIONS We observe a distinct improvement in the model using contextual information. However, results reveal weak generalizability of the current systems on unseen data. Additional research is needed to fully utilize social media data, and improve the robustness and reliability of NLP systems. The content analysis, on the other hand, shows that Twitter covers a sufficiently wide range of ADEs as well as known adverse reactions for drugs mentioned in tweets. Our work demonstrates that social media can be a reliable data source for collecting ADE mentions.

Background Extraintestinal manifestations (EIMs) occur commonly in inflammatory bowel disease (IBD), but population-level understanding of EIM behavior is difficult. We present a natural language processing (NLP) system designed to identify both the presence and status of EIMs using clinical notes from patients with IBD. Methods In a single-center retrospective study, clinical outpatient electronic documents were collected in patients with IBD. An NLP EIM detection pipeline was designed to determine general and specific symptomatic EIM activity status descriptions using Python 3.6. Accuracy, sensitivity, and specificity, and agreement using Cohen’s kappa coefficient were used to compare NLP-inferred EIM status to human documentation labels. Results The 1240 individuals identified as having at least 1 EIM consisted of 54.4% arthritis, 17.2% ocular, and 17.0% psoriasiform EIMs. Agreement between reviewers on EIM status was very good across all EIMs (κ = 0.74; 95% confidence interval [CI], 0.70-0.78). The automated NLP pipeline determining general EIM activity status had an accuracy, sensitivity, specificity, and agreement of 94.1%, 0.92, 0.95, and κ = 0.76 (95% CI, 0.74-0.79), respectively. Comparatively, prediction of EIM status using administrative codes had a poor sensitivity, specificity, and agreement with human reviewers of 0.32, 0.83, and κ = 0.26 (95% CI, 0.20-0.32), respectively. Conclusions NLP methods can both detect and infer the activity status of EIMs using the medical document an information source. Though source document variation and ambiguity present challenges, NLP offers exciting possibilities for population-based research and decision support in IBD.

BackgroundAdverse reactions to drugs attract significant concern in both clinical practice and public health monitoring. Multiple measures have been put into place to increase postmarketing surveillance of the adverse effects of drugs and to improve drug safety. These measures include implementing spontaneous reporting systems and developing automated natural language processing systems based on data from electronic health records and social media to collect evidence of adverse drug events that can be further investigated as possible adverse reactions.ObjectiveWhile using social media for collecting evidence of adverse drug events has potential, it is not clear whether social media are a reliable source for this information. Our work aims to (1) develop natural language processing approaches to identify adverse drug events on social media and (2) assess the reliability of social media data to identify adverse drug events.MethodsWe propose a collocated long short-term memory network model with attentive pooling and aggregated, contextual representation generated by a pretrained model. We applied this model on large-scale Twitter data to identify adverse drug event–related tweets. We conducted a qualitative content analysis of these tweets to validate the reliability of social media data as a means to collect such information.ResultsThe model outperformed a variant without contextual representation during both the validation and evaluation phases. Through the content analysis of adverse drug event tweets, we observed that adverse drug event–related discussions had 7 themes. Mental health–related, sleep-related, and pain-related adverse drug event discussions were most frequent. We also contrast known adverse drug reactions to those mentioned in tweets.ConclusionsWe observed a distinct improvement in the model when it used contextual information. However, our results reveal weak generalizability of the current systems to unseen data. Additional research is needed to fully utilize social media data and improve the robustness and reliability of natural language processing systems. The content analysis, on the other hand, showed that Twitter covered a sufficiently wide range of adverse drug events, as well as known adverse reactions, for the drugs mentioned in tweets. Our work demonstrates that social media can be a reliable data source for collecting adverse drug event mentions.

ObjectiveInitiatives to reduce neighborhood-based health disparities require access to meaningful, timely, and local information regarding health behavior and its determinants. We examined the validity of Twitter as a source of information for neighborhood-level analysis of dietary choices and attitudes.Materials and MethodsWe analyzed the “healthiness” quotient and sentiment in food-related tweets at the census tract level, and associated them with neighborhood characteristics and health outcomes. We analyzed keywords driving the differences in food healthiness between the most and least-affluent tracts, and qualitatively analyzed contents of a random sample of tweets.ResultsSignificant, albeit weak, correlations existed between healthiness and sentiment in food-related tweets and tract-level measures of affluence, disadvantage, race, age, U.S. density, and mortality from conditions associated with obesity. Analyses of keywords driving the differences in food healthiness revealed foods high in saturated fat (eg, pizza, bacon, fries) were mentioned more frequently in less-affluent tracts. Food-related discussion referred to activities (eating, drinking, cooking), locations where food was consumed, and positive (affection, cravings, enjoyment) and negative attitudes (dislike, personal struggles, complaints).DiscussionTweet-based healthiness scores largely correlated with offline phenomena in the expected directions. Social media offer less resource-intensive data collection methods than traditional surveys do. Twitter may assist in informing local health programs that focus on drivers of food consumption and could inform interventions focused on attitudes and the food environment.ConclusionsTwitter provided weak but significant signals concerning food-related behavior and attitudes at the neighborhood level, suggesting its potential usefulness for informing local health disparity reduction efforts.

In the past decade, a growing adoption of electronic health record (EHR) systems have made massive amounts of clinical narrative data available in computable form. However, extracting relevant information from clinical narratives remains challenging. Clinical notes often contain abbreviations, medical terms, and other jargon that are easy for health professionals, but challenging for automated approaches to disambiguate. Many EHR systems use non-standard document structures to record critical information about medications, diagnoses, and potential complications. Finally, clinical narratives contain sensitive patient information, which raises privacy and security concerns. Data science and natural language processing (NLP) methods, including the recently popular deep learning-based approaches, can unlock information from narrative text and have received great attention in the medical domain. Many clinical NLP methods based on deep learning models have shown promising results in various information extraction tasks. These methods and tools have also been successfully applied to facilitate clinical research, as well as to support healthcare applications. In this tutorial, we will highlight some methods, tools, and technologies to identify medical concepts and entities in clinical text. Deriving from examples in cohort selection, medication extraction, and de-identification of protected health information, the tutorial presenters will lead a hands-on exercise to develop an NLP pipeline for clinical information extraction. The tutorial will spotlight state-of-the-art approaches with domain examples from multiple clinical domains.

BACKGROUND Informatics tools for consumers and patients are important vehicles for facilitating engagement, and the field of consumer health informatics is an key space for exploring the potential of these tools. To understand research findings in this complex and heterogeneous field, a scoping review can help not only to identify, but to bridge, the array of diverse disciplines and publication venues involved. OBJECTIVE The goal of this systematic scoping review was to characterize the extent; range; and nature of research activity in consumer health informatics, focusing on the contributing disciplines of informatics; information science; and engineering. METHODS Four electronic databases (Compendex, LISTA, Library Literature, and INSPEC) were searched for published studies dating from January 1, 2008, to June 1, 2015. Our inclusion criteria specified that they be English-language articles describing empirical studies focusing on consumers; relate to human health; and feature technologies designed to interact directly with consumers. Clinical applications and technologies regulated by the FDA, as well as digital tools that do not provide individualized information, were excluded. RESULTS We identified 271 studies in 63 unique journals and 22 unique conference proceedings. Sixty-five percent of these studies were found in health informatics journals; 23% in information science and library science; 15% in computer science; 4% in medicine; and 5% in other fields, ranging from engineering to education. A single journal, the Journal of Medical Internet Research, was home to 36% of the studies. Sixty-two percent of these studies relied on quantitative methods, 55% on qualitative methods, and 17% were mixed-method studies. Seventy percent of studies used no specific theoretical framework; of those that did, Social Cognitive Theory appeared the most frequently, in 16 studies. Fifty-two studies identified problems with technology adoption, acceptance, or use, 38% of these barriers being machine-centered (for example, content or computer-based), and 62% user-centered, the most frequently mentioned being attitude and motivation toward technology. One hundred and twenty-six interventional studies investigated disparities or heterogeneity in treatment effects in specific populations. The most frequent disparity investigated was gender (13 studies), followed closely by race/ethnicity (11). Half the studies focused on a specific diagnosis, most commonly diabetes and cancer; 30% focused on a health behavior, usually information-seeking. Gaps were found in reporting of study design, with only 46% of studies reporting on specific methodological details. Missing details were response rates, since 59% of survey studies did not provide them; and participant retention rates, since 53% of interventional studies did not provide this information. Participant demographics were usually not reported beyond gender and age. Only 17% studies informed the reader of their theoretical basis, and only 4 studies focused on theory at the group, network, organizational or ecological levels—the majority being either health behavior or interpersonal theories. Finally, of the 131 studies describing the design of a new technology, 81% did not involve either patients or consumers in their design. In fact, while consumer and patient were necessarily core concepts in this literature, these terms were often used interchangeably. The research literature of consumer health informatics at present is scattered across research fields; only 49% of studies from these disciplines is indexed by MEDLINE and studies in computer science are siloed in a user interface that makes exploration of that literature difficult. CONCLUSIONS Few studies analyzed in this scoping review were based in theory, and very little was presented in this literature about the life context, motives for technology use, and personal characteristics of study participants.

Background Extra-Intestinal Manifestations (EIM) occur in nearly 40% of patients with IBD and impact both disease experience and therapeutic decision-making, but are not well captured by administrative codes. We aimed to pilot computational natural language processing (NLP) methods to characterise EIMs using consultant notes. Methods Subjects with a diagnosis of IBD were identified in a single-centre retrospective review of electronic health records (EHR) between 2014–2017. Gastroenterology (GI) notes were annotated by two reviewers for the presence and activity of EIMs. EIM concepts were identified using NLP methods leveraging UMLS libraries and hand-crafted features. EIM characterisation occurred within a ±25-word window around identified EIMs with classifications including inactive concepts (negated, historical, resolved) and active concepts (improved, worsened, active but unchanged). Decisions on EIM status when repeatedly referenced in a document used section-based weighting for status inference, with greatest to least weight ranking for assessment/plan, subjective, past history, exam, and other, respectively. EIM status was classified as ambiguous when multiple conflicting references were present within the same document of approximately equal weight. Model development and testing used an 80/20 dataset split. Results In 4108 unique IBD patients, 1640 (39.9%) had at least 1 EIM identified. The mean age was 41.9 years, 47.2% were male, and 27.0% had biologic exposure. A total of 1240 manually annotated documents (first GI notes) were comprised of 51.1% arthritis, 16.5% ocular, 16.2% psoriasis, with erythema nodosum (EN), pyoderma gangrenosum (PG), and hidradenitis suppurativa (HS) together comprising 16.2% of the cohort. NLP models performed well for correctly classifying both EIM presence and status in a testing set, with overall accuracy, sensitivity, and specificity of 91.2%, 92.9% and 81.8% across all EIMs in notes automatically classified as non-ambiguous (Table 1). NLP methods identified EIM status classification as ambiguous in 38.9% of cases. Conclusion NLP methods can detect and classify EIMs with reasonable performance and efficiency compared with traditional manual chart review. Though source document variation and ambiguity present challenges, NLP offers exciting possibilities for population-based research and decision support.

Identifying mentions of medical concepts in social media is challenging because of high variability in free text. In this paper, we propose a novel neural network architecture, the Collocated LSTM with Attentive Pooling and Aggregated representation (CLAPA), that integrates a bidirectional LSTM model with attention and pooling strategy and utilizes the collocation information from training data to improve the representation of medical concepts. The collocation and aggregation layers improve the model performance on the task of identifying mentions of adverse drug events (ADE) in tweets. Using the dataset made available as part of the workshop shared task, we show that careful selection of neighborhood contexts can help uncover useful local information and improve the overall medical concept representation.

We participated in Task 1 of the Social Media Mining for Health Applications (SMM4H) 2019 Shared Tasks on detecting mentions of adverse drug events (ADEs) in tweets. Our approach relied on a text processing pipeline for tweets, and training traditional machine learning and deep learning models. Our submitted runs performed above average for the task.