Dataset Information

The Voice of Drug Consumers: Online Textual Review Analysis Using Structural Topic Model.

ABSTRACT: Many web-based pharmaceutical e-commerce platforms allow consumers to post open-ended textual reviews based on their purchase experiences. Understanding the true voice of consumers by analyzing such a large amount of user-generated content is of great significance to pharmaceutical manufacturers and e-commerce websites. The aim of this paper is to automatically extract hidden topics from web-based drug reviews using the structural topic model (STM) to examine consumers' concerns when they buy drugs online. The STM is a probabilistic extension of Latent Dirichlet Allocation (LDA), which allows the consolidation of document-level covariates. This innovation allows us to capture consumer dissatisfaction along with their dynamics over time. We extract 12 topics, and five of them are negative topics representing consumer dissatisfaction, whose appearances in the negative reviews are substantially higher than those in the positive reviews. We also come to the conclusion that the prevalence of these five negative topics has not decreased over time. Furthermore, our results reveal that the prevalence of price-related topics has decreased significantly in positive reviews, which indicates that low-price strategies are becoming less attractive to customers. To the best of our knowledge, our work is the first study using STM to analyze the unstructured textual data of drug reviews, which enhances the understanding of the aspects of drug consumer concerns and contributes to the research of pharmaceutical e-commerce literature.

SUBMITTER: He L

PROVIDER: S-EPMC7277719 | biostudies-literature | 2020 May

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

The Voice of Drug Consumers: Online Textual Review Analysis Using Structural Topic Model.

He Lifeng L Han Dongmei D Zhou Xiaohang X Qu Zheng Z

International journal of environmental research and public health 20200522 10

Many web-based pharmaceutical e-commerce platforms allow consumers to post open-ended textual reviews based on their purchase experiences. Understanding the true voice of consumers by analyzing such a large amount of user-generated content is of great significance to pharmaceutical manufacturers and e-commerce websites. The aim of this paper is to automatically extract hidden topics from web-based drug reviews using the structural topic model (STM) to examine consumers' concerns when they buy dr ...[more]

PMID: 32455918

Similar Datasets

Project description:BACKGROUND: The Food and Drug Administration (FDA) approved drug labels contain a broad array of information, ranging from adverse drug reactions (ADRs) to drug efficacy, risk-benefit consideration, and more. However, the labeling language used to describe these information is free text often containing ambiguous semantic descriptions, which poses a great challenge in retrieving useful information from the labeling text in a consistent and accurate fashion for comparative analysis across drugs. Consequently, this task has largely relied on the manual reading of the full text by experts, which is time consuming and labor intensive. METHOD: In this study, a novel text mining method with unsupervised learning in nature, called topic modeling, was applied to the drug labeling with a goal of discovering "topics" that group drugs with similar safety concerns and/or therapeutic uses together. A total of 794 FDA-approved drug labels were used in this study. First, the three labeling sections (i.e., Boxed Warning, Warnings and Precautions, Adverse Reactions) of each drug label were processed by the Medical Dictionary for Regulatory Activities (MedDRA) to convert the free text of each label to the standard ADR terms. Next, the topic modeling approach with latent Dirichlet allocation (LDA) was applied to generate 100 topics, each associated with a set of drugs grouped together based on the probability analysis. Lastly, the efficacy of the topic modeling was evaluated based on known information about the therapeutic uses and safety data of drugs. RESULTS: The results demonstrate that drugs grouped by topics are associated with the same safety concerns and/or therapeutic uses with statistical significance (P<0.05). The identified topics have distinct context that can be directly linked to specific adverse events (e.g., liver injury or kidney injury) or therapeutic application (e.g., antiinfectives for systemic use). We were also able to identify potential adverse events that might arise from specific medications via topics. CONCLUSIONS: The successful application of topic modeling on the FDA drug labeling demonstrates its potential utility as a hypothesis generation means to infer hidden relationships of concepts such as, in this study, drug safety and therapeutic use in the study of biomedical documents.

Project description:BACKGROUND:In case of a population-wide infectious disease outbreak, such as the novel coronavirus disease (COVID-19), people's online activities could significantly affect public concerns and health behaviors due to difficulty in accessing credible information from reliable sources, which in turn causes people to seek necessary information on the web. Therefore, measuring and analyzing online health communication and public sentiment is essential for establishing effective and efficient disease control policies, especially in the early stage of an outbreak. OBJECTIVE:This study aimed to investigate the trends of online health communication, analyze the focus of people's anxiety in the early stages of COVID-19, and evaluate the appropriateness of online information. METHODS:We collected 13,148 questions and 29,040 answers related to COVID-19 from Naver, the most popular Korean web portal (January 20, 2020, to March 2, 2020). Three main methods were used in this study: (1) the structural topic model was used to examine the topics in the online questions; (2) word network analysis was conducted to analyze the focus of people's anxiety and worry in the questions; and (3) two medical doctors assessed the appropriateness of the answers to the questions, which were primarily related to people's anxiety. RESULTS:A total of 50 topics and 6 cohesive topic communities were identified from the questions. Among them, topic community 4 (suspecting COVID-19 infection after developing a particular symptom) accounted for the largest portion of the questions. As the number of confirmed patients increased, the proportion of topics belonging to topic community 4 also increased. Additionally, the prolonged situation led to a slight increase in the proportion of topics related to job issues. People's anxieties and worries were closely related with physical symptoms and self-protection methods. Although relatively appropriate to suspect physical symptoms, a high proportion of answers related to self-protection methods were assessed as misinformation or advertisements. CONCLUSIONS:Search activity for online information regarding the COVID-19 outbreak has been active. Many of the online questions were related to people's anxieties and worries. A considerable portion of corresponding answers had false information or were advertisements. The study results could contribute reference information to various countries that need to monitor public anxiety and provide appropriate information in the early stage of an infectious disease outbreak, including COVID-19. Our research also contributes to developing methods for measuring public opinion and sentiment in an epidemic situation based on natural language data on the internet.

Project description:IntroductionCorrect interpretation of drug labels instructions (DLIs) is needed for safe use and better adherence to prescribed drugs. DLIs are often too difficult for patients, especially for those with limited health literacy. What is yet unknown, is how specific textual elements in DLIs (e.g., the presentation of numbers, or use of medical jargon) and patients' health literacy skills are related to the comprehension of DLIs. In order to provide concrete directions for health professionals on how to optimize drug prescriptions, we performed a systematic review to summarize the available research findings on which textual elements facilitate or hinder the correct interpretation of DLIs in relation to patients' health literacy.MethodA systematic search was performed in PubMed, EMBASE, PsychINFO, and Smartcat (until April 2019) to identify studies investigating textual elements that facilitate or hinder the correct interpretation of DLIs in relation to patients' health literacy.ResultsA total of 434 studies were identified of which 28 studies met our inclusion criteria. We found that textual elements contributing to the correct interpretation of DLIs were: using explicit time periods in dosage instructions, using plain language, presenting numbers in a numerical format, and providing DLIs in patients' native language. Multistep instructions per instruction line, using abbreviations and medical jargon seem to hinder the correct interpretation of DLIs. Although health literacy was taken into account in a majority of the studies, none of them assessed the effectiveness of specific textual elements on patients' comprehensibility of DLIs.ConclusionBased on our findings, we provide an overview of textual elements that contribute to the correct interpretation of DLIs. Optimizing the textual instruction on drug labels may increase the safety and adherence to prescribed drugs, taking into account that a significant proportion of patients has low health literacy.

Dataset Information

The Voice of Drug Consumers: Online Textual Review Analysis Using Structural Topic Model.

Publications

The Voice of Drug Consumers: Online Textual Review Analysis Using Structural Topic Model.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets