Start Your Search

N. Mehta



Author of

  • +

    OS30 - Enabling High Value Analytics (ID 44)

    • Event: e-Health 2018 Virtual Meeting
    • Type: Oral Session
    • Track: Clinical Delivery
    • Presentations: 1
    • +

      OS30.05 - Data Mining Twitter to Detect Prescribing Cascades: A New Concept (ID 522)

      N. Mehta, Women's College Research Institute; Mississauga/CA

      • Abstract
      • Slides

      Purpose/Objectives: A prescribing cascade occurs when the adverse event from a drug therapy is misdiagnosed as a new medical condition, resulting in a subsequent drug therapy, medical devices, or diagnostic tests (Figure 1). Traditionally, prescribing cascades have been evaluated using administrative data to assess recorded drug therapies. However, these data do not capture information such as the use of over-the-counter drugs, devices or tests, thus limiting the scope of our understanding of the prescribing cascade. Over 11 million people in the United States have used social media to post information about health and treatment issues. Over 600 million active users are registered on Twitter. Recently, an approach to use Twitter to detect signals of potential drug-to-drug prescribing cascades was described by Hoang et al. We explore the feasibility of using Twitter to identify expanded prescribing cascades, using dementia as an example. Figure 1: The Expanded Prescribing Cascade (adapted from Rochon & Gurwitz, 2017.)picture1.png

      Methodology/Approach: A challenge with searching social media for clinical health information is the wide variety of synonyms, colloquial terms, and informal language used to describe conditions, medications and symptoms. Clinical data dictionaries, including the Consumer Health Vocabulary, were identified, which links lay speech about health to technical terms used by healthcare professionals. Using the Twitter Application Programming Interface, a preliminary search was run to identify the level of tweeting relating to dementia. The search terms used were Dementia, Alzheimer, and Lewy body. The collected tweets were then manually explored for general sentiments and user demographics.

      Finding/Results: Feasibility testing within a five-hour window revealed 872 potentially relevant tweets, suggesting that Twitter users tweet about dementia every 20 seconds on average. Manual exploration of these tweets showed that the majority were posted by caregivers of people with dementia, or healthcare professionals. Of these tweets, 6 pertained to dementia drug therapies. We expect that other medical conditions that are more prevalent in the general population will have more relevant tweets.

      Conclusion/Implications/Recommendations: Twitter is a tool for patients, caregivers or providers to post information relevant to prescribing cascades, and drug therapies in general. This platform is an unexplored resource for identifying potential prescribing cascades, which may allow opportunities to collect previously unavailable data on over-the-counter drugs, devices and tests. These data can also inform future population-level exploratory studies about the consequences of prescribing cascades.

      140 Character Summary: We explored the feasibility of using social media meta data to identify prescribing cascades, using drug therapies for dementia as an example.

      Only Active Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login or select "Add to Cart" and proceed to checkout.