Social media monitoring for adverse events
Adverse events are less reported in research studies, particularly randomized clinical trials and pharmacovigilance trials. A method that researchers can use to detect more complete safety profiles for treatments is to use social media analytics. However, the patient's point of view on the ethical problems associated with using trial subject reports of adverse drug events on social media is not clear.
Keyword
Adverse effects on Social Media; Social media; Digital health; Surveillance; Ethics for Social Media Use.
Introduction
Adverse events caused by drugs are one of the most important public health problems. Social media has empowered more patients to share their drug use experiences and has become an important source for the detection of practically unreported adverse events and adverse drug reactions (ADRs). Since many user posts do not mention any ADR, accurate detection of the presence of ADRs in each user post is mandatory before further research can be conducted. Previous methods focus on extracting more shallow linguistic features that are not able to capture deep information in the context, ultimately failing to provide satisfactory accuracy.
Objective
The objective of the study is to explore the ethics of using social media for identifying and monitoring adverse drug reactions and adverse events for research purposes using a multi-methods approach.
ADR detection over social media
User posts describing ADRs must be identified before further detection can occur because approximately 90% of drug-related posts of users are not associated with ADRs. User posts mentioning at least one adverse drug reaction can be categorized as ADRs, while other user posts are classified as non-ADRs. Users sometimes mention multiple symptoms throughout a
post, for example, “Took for insomnia, didn’t make me groggy.” Two symptom words, “insomnia” and “groggy”, are written in this post. When people judge whether an ADR is present in a post, they generally read the post comprehensively, analyze the meaning of every ADR word and its corresponding predicate, and then pass judgment.
At present, the detection of AEs and ADRs primarily depends on two methods: active surveillance of electronic health records (EHRs) based on electronic medical data mining, such as the mining, and passive monitoring based on the spontaneous reporting system (SRS).
Social Media as a Tool for Monitoring of Isotretinoin Adverse Effects
The objective of this study was to investigate the nature of content posted on the social media platform Instagram concerning the systemic acne medication isotretinoin.
The search term “#accutane” was written on Instagram to generate all public posts using the hashtag between February 1 and May 31, 2018. Four independent investigators then criticized the post for eligibility.
The inclusion criteria were posts mentioned in English, accessible by URL, having a prior focus on isotretinoin, and posted by users of the medication.
Data regarding multiple variables from each post was then entered into a Microsoft Excel template. Of 7,661 posts, 3,082 were eligible.
Among posts that contained negative tone (n=1312), this element was more commonly due to the presence of side effects (65%) than lack of improvement in skin appearance (33%).
Overall, 1,263 posters (41%) mentioned adverse effects of oral isotretinoin, most commonly dry facial skin (17%), dry/cracked lips (16%), or arthralgias/myalgias (8%).
Neuropsychiatric side effects were also documented, with users reporting fatigue (4%), mood changes (3%), and headache (2%). This concluded with reported side effects of oral isotretinoin on Instagram closely tracked its known side effects in frequency.
Social media might be a valuable tool to survey the general pattern and burden of adverse effects for patients undergoing treatment for dermatologic conditions.
Common challenges, concerns, and limitations
To understand the risks involved in utilizing digital media for adverse events purposes, it is mandatory to reflect on the challenges, concerns, and limitations that were voiced by the target populations. The most prompted concerns among the respondents were ethicality and confidentiality. This stresses the need for a transparent system of trust between the various populations, particularly between trial subjects and HCPs. According to a report, only 34% of the public respondents say that they would be comfortable with pharmaceutical companies using health-related information from their posts online. A significantly higher percentage of respondents would give consent to HCPs to utilize such information (72%). While there might be a high degree of trust from the public towards HCPs, almost 90% of the respondents agreed on the need for initial consent.
From a pharmaceutical industry perspective, the responses from the public indicate that it would be difficult for MAHs to obtain the consent of patients and consumers for the use of their health-related information. Some of the public respondents alluded to the “fear of retaliation” from pharmaceutical companies. As reflected in the results, patients would be more inclined to report to HCPs or directly to health regulatory authorities. Despite the major ethical challenges that are present, the need for requiring consent before the use of an individual’s publically posted information is still a topic of debate. As stated by one of the pharmaceutical company respondents, “Once the person posts something within the digital
media, it means he/she consents to the use of his/her data”. Studies have estimated that the quantity of fully public Facebook accounts is approximately 25% while for Twitter, around 90% of the feeds are believed to be fully public. With a few of the companies indicating that they are already screening external social media sites for ADR-related information, the data privacy regulations of certain social media sites do not hinder individuals or organizations from screening and mining for health-related information.
Conclusions
Social media users were usually positive about their data being able to be used for research purposes, especially for research on adverse events. However, the decision to approve was dependent on the potential benefit of the research and that trial subjects are protected from harm. More studies are required to establish when consent is required for an individual's social media data to be used.
References
1. Golder S, Scantlebury A, Christmas H. Understanding Public Attitudes Toward Researchers Using Social Media for Detecting and Monitoring Adverse Events Data: Multi Methods Study. J Med Internet Res. 2019 Aug 29;21(8): e7081. doi: 10.2196/jmir.7081. PMID: 31469079; PMCID: PMC6740159.
2.X. Wang, G. Hripcsak, M. Markatou, C. Friedman
3. Active Computerized Pharmacovigilance Using Natural Language Processing, Statistics, and Electronic Health Records: A Feasibility Study J. Am. Med. Informatics Assoc., 16 (2009), pp. 328-337.
4. C. Friedman, discovering novel adverse drug events using natural language processing and mining of the electronic health record, in Proc. 12th Conf. Artif. Intell. Med., 2009.
5. E. Aramaki, Y. Miura, M. Tonoike, T. Ohkuma, H. Masuichi, K. Waki, K. Ohe
Extraction of adverse drug effects from clinical records
Stud. Health Technol. Inform., 160 (2010), pp. 739-743,
6. G. Stanovsky, J. Eckle-Kohler, Y. Puzikov, I. Dagan, I. Gurevych, Integrating Deep Linguistic Features in Factuality Prediction over Unified Datasets, in Proc. Of the 55th Annu. Meet. Of the Assoc. Comput. Linguist., 2017: pp. 352–357. Doi: 10.18653/v1/P17-2056.
7. Omar I, Harris E (2016) The Use of social media in ADR Monitoring and Reporting. J Pharmacovigilant 4: 223. doi:10.4172/2329-6887.1000223
Student Name: Khan Abdul Azim
Student ID: CLS_190/092023
Qualification: B. Pharmacy
E-mail ID: azeemkhan2904@gmail.com
Comments