Haodong Yang

Social media mining for drug safety signal detection

Adverse Drug Reactions (ADRs) represent a serious problem all over the world. They may complicate a patients medical conditions and increase the morbidity, even mortality. Drug safety currently depends heavily on post-marketing surveillance, because pre-marketing review process cannot identify all possible adverse drug reactions in that it is limited by scale and time span. However, current post-marketing surveillance is conducted through centralized volunteering reporting systems, and the reporting rate is low. Consequently, it is difficult to detect the adverse drug reactions signals in a timely manner. To solve this problem, many researchers have explored methods to detect ADRs in electronic health records. Nevertheless, we only have access to electronic health records form particular health units. Aggregating and integrating electronic health records from multiple sources is rather challenging. With the advance of Web 2.0 technologies and the popularity of social media, many health consumers are discussing and exchanging health-related information with their peers. Many of this online discussion involve adverse drug reactions. In this work, we propose to use association mining and Proportional Reporting Ratios to mine the associations between drugs and adverse reactions from the user contributed content in social media. We have conducted an experiment using ten drugs and five adverse drug reactions. The FDA alerts are used as the gold standard to test the performance of the proposed techniques. The result shows that the metrics leverage, lift, and PRR are all promising to detect the adverse drug reactions reported by FDA. However, PRR outperformed the other two metrics.

Postmarketing Drug Safety Surveillance Using Publicly Available Health-Consumer-Contributed Content in Social Media

Postmarketing drug safety surveillance is important because many potential adverse drug reactions cannot be identified in the premarketing review process. It is reported that about 5% of hospital admissions are attributed to adverse drug reactions and many deaths are eventually caused, which is a serious concern in public health. Currently, drug safety detection relies heavily on voluntarily reporting system, electronic health records, or relevant databases. There is often a time delay before the reports are filed and only a small portion of adverse drug reactions experienced by health consumers are reported. Given the popularity of social media, many health social media sites are now available for health consumers to discuss any health-related issues, including adverse drug reactions they encounter. There is a large volume of health-consumer-contributed content available, but little effort has been made to harness this information for postmarketing drug safety surveillance to supplement the traditional approach. In this work, we propose the association rule mining approach to identify the association between a drug and an adverse drug reaction. We use the alerts posted by Food and Drug Administration as the gold standard to evaluate the effectiveness of our approach. The result shows that the performance of harnessing health-related social media content to detect adverse drug reaction is good and promising.

Diffusion of Messages from an Electronic Cigarette Brand to Potential Users through Twitter

Objective This study explores the presence and actions of an electronic cigarette (e-cigarette) brand, Blu, on Twitter to observe how marketing messages are sent and diffused through the retweet (i.e., message forwarding) functionality. Retweet networks enable messages to reach additional Twitter users beyond the sender’s local network. We follow messages from their origin through multiple retweets to identify which messages have more reach, and the different users who are exposed. Methods We collected three months of publicly available data from Twitter. A combination of techniques in social network analysis and content analysis were applied to determine the various networks of users who are exposed to e-cigarette messages and how the retweet network can affect which messages spread. Results The Blu retweet network expanded during the study period. Analysis of user profiles combined with network cluster analysis showed that messages of certain topics were only circulated within a community of e-cigarette supporters, while other topics spread further, reaching more general Twitter users who may not support or use e-cigarettes. Conclusions Retweet networks can serve as proxy filters for marketing messages, as Twitter users decide which messages they will continue to diffuse among their followers. As certain e-cigarette messages extend beyond their point of origin, the audience being exposed expands beyond the e-cigarette community. Potential implications for health education campaigns include utilizing Twitter and targeting important gatekeepers or hubs that would maximize message diffusion.

Harnessing Social Media for Drug-Drug Interactions Detection

Adverse drug reactions (ADRs) are causing a substantial amount of hospital admissions and deaths, which cannot be underestimated. Drug-drug interactions (DDIs) are an important patient safety problem and have been reported to cause a large portion of patient adverse events resulting in warning notices or the withdrawal of many drugs from the market. Currently, DDIs detection mainly depends on four kinds of data sources - clinical trial data, spontaneous reporting systems, electronic medical records, and chemical/pharmacologic databases, all of which have some limitations such as cohort biases, low reporting ratio, access issue, etc. In this study, we propose to detect DDIs signals from consumer contributed contents in online healthcare communities using associations mining. We conduct an experiment with thirteen drugs and three DDI associations. Leverage, lift and interaction ratio are used in the experiment. Drug Bank is used as gold standard to test the performance of the approach. The results show that our techniques are promising to detect signals of DDIs and the proposed measure, interaction ratio, performs better than leverage and lift.

Using Health-Consumer-Contributed Data to Detect Adverse Drug Reactions by Association Mining with Temporal Analysis

Since adverse drug reactions (ADRs) represent a significant health problem all over the world, ADR detection has become an important research topic in drug safety surveillance. As many potential ADRs cannot be detected though premarketing review, drug safety currently depends heavily on postmarketing surveillance. Particularly, current postmarketing surveillance in the United States primarily relies on the FDA Adverse Event Reporting System (FAERS). However, the effectiveness of such spontaneous reporting systems for ADR detection is not as good as expected because of the extremely high underreporting ratio of ADRs. Moreover, it often takes the FDA years to complete the whole process of collecting reports, investigating cases, and releasing alerts. Given the prosperity of social media, many online health communities are publicly available for health consumers to share and discuss any healthcare experience such as ADRs they are suffering. Such health-consumer-contributed content is timely and informative, but this data source still remains untapped for postmarketing drug safety surveillance. In this study, we propose to use (1) association mining to identify the relations between a drug and an ADR and (2) temporal analysis to detect drug safety signals at the early stage. We collect data from MedHelp and use the FDAs alerts and information of drug labeling revision as the gold standard to evaluate the effectiveness of our approach. The experiment results show that health-related social media is a promising source for ADR detection, and our proposed techniques are effective to identify early ADR signals.

Identifying Implicit and Explicit Relationships Through User Activities in Social Media

Social commerce has emerged as a new paradigm of commerce due to the advancement and application of Web 2.0 technologies including social media sites. Social media sites provide a valuable opportunity for social interactions between electronic commerce consumers as well as between consumers and businesses. Although the number of users and interactions is large in social media, the social networks extracted from explicit user interactions are usually sparse. Hence, the result obtained through the analysis of the extracted network is not always useful because many potential ties in the social network are not captured by the explicit interactions between users. In this work, we propose a temporal analysis technique to identify implicit relationships that supplement the explicit relationships identified through the social media interaction functions. Our method is based on the homophily theory developed by McPherson, Smith-Lovin, and Cook [31]. We have conducted experiments to evaluate the effectiveness of the identified implicit relationships and the integration of implicit and explicit relationships. The results indicate that our proposed techniques are effective and achieve a higher accuracy. Our results prove the importance of implicit relationships in deriving complete online social networks that are the foundation for understanding online user communities and social network analysis. Our techniques can be applied to improve effectiveness of product and friend recommendation in social commerce.

Identifying implicit relationships between social media users to support social commerce

The Internet is an ideal platform for business-to-consumer (B2C) and business-to-business (B2B) electronic commerce where businesses and consumers conduct commerce activities such as searching for consumer products, promoting business, managing supply chain and making electronic transactions. With the advance of Web 2.0 technologies and the popularity of social media sites, social commerce offers new opportunities of social interaction between electronic commerce consumers as well as social interaction between consumers and e-retailers. The user contributed content provides a tremendous amount of information that may assist in electronic commerce services. Social network analysis and mining has been a powerful tool for electronic commerce vendors and marketing companies to understand the user behavior which is useful for identifying potential customers of their products. However, the capability of social network analysis and mining diminishes when the social network data is incomplete, especially when there are only limited ties available. The social networks extracted from explicit relationships in social media are usually sparse. Many social media users who have similar interest may not have direct interactions with one another or purchase the same products. Therefore, the explicit relationships between electronic commerce users are not sufficient to construct social networks for effective social network analysis and mining. In this work, we propose the temporal analysis techniques to identify implicit relationships for enriching the social network structure. We have conducted an experiment on Digg.com, which is a social media site for users to discover and share content from anywhere of the Web. The experiment shows that the temporal analysis techniques outperform the baseline techniques that only rely on explicit relationships.

Drug-Drug Interactions Detection from Online Heterogeneous Healthcare Networks

Drug-drug interactions (DDIs) are a serious drug safety problem for health consumers and how to detect such interactions effectively and efficiently has been of great medical significance. Currently, methods proposed to detect DDIs are mainly based on data sources such as clinical trial data, spontaneous reporting systems, electronic medical records, and chemical/pharmacological databases. However, those data sources are limited either by cohort biases, low reporting ratio, or access issue. In this study, we propose to use online healthcare social media, an informative and publicly available data source, to detect DDI signals. We construct a heterogeneous healthcare network based on consumer contributed contents, develop heterogeneous topological features, and use logistic regression as prediction model for DDI detection. The experiment results show that the proposed heterogeneous topological features substantially outperform the homogenous ones in the training set but only slightly outperform the homogeneous ones in the testing set, and interesting heterogeneous paths with strong predictive power are discovered.

Mining a Weighted Heterogeneous Network Extracted from Healthcare-Specific Social Media for Identifying Interactions between Drugs

Drug-drug interaction (DDI) detection is an important issue of pharmacovigilance. Currently, approaches proposed to detection DDIs are mainly focused on data sources such as spontaneous reporting systems, electronic health records, chemical/pharmacological databases, and biomedical literatures. However, those data sources are limited either by low reporting ratio, access issue, or long publication time span. In this work, we propose to explore online health communities, a timely, informative and publicly available data source, for DDI detection. We construct a weighted heterogeneous healthcare network that contains drugs, adverse drug reactions (ADRs), diseases, and users extracted from online health consumer-contributed contents, extract topological features, develop weighted path count to quantify the features, and use supervised learning techniques to detect DDI signals. The experiment results show that weighted heterogeneous healthcare network using leverage and lift are more effective in DDI detection than both unweighted homogeneous and heterogeneous network.

Discovering Drug-Drug Interactions and Associated Adverse Drug Reactions with Triad Prediction in Heterogeneous Healthcare Networks

Drug-drug interactions (DDIs) are of great importance in drug safety. Currently, DDI signal detection mainly depends on post-marketing surveillance. Various data sources have been used by researchers for DDI detection such as spontaneous reporting system, electronic health records, Pharmacological Databases, and biomedical literatures. However, these data sources are limited by either high underreporting ratio, access difficulty, or long publication cycle. In this work, we propose to utilize consumer-contributedcontents from online health communities, a publicly available, mountainous, and timely data source, for identifying DDI signals and association adverse drug reactions (ADRs). Specifically, we first construct a heterogeneous healthcare network, extract different topological types of Drug-Drug-ADR triad, explore node-based, link-based, and triad-based features, and then formulate the signal detection as a supervised learning problem. The experiment results show that our proposed techniques are effective in detecting DDI signals and associated ADRs at the same time.