Rhinitis, sinusitis, and ocular allergy
Mining social media data to assess the risk of skin and soft tissue infections from allergen immunotherapy

https://doi.org/10.1016/j.jaci.2019.01.029Get rights and content

Background

Allergen immunotherapy (AIT) treatment for allergic rhinitis and asthma is used by 2.6 million Americans annually. Clinical and sterility testing studies identify no risk of contamination or infection from extracts prepared using recommended aseptic techniques, but regulatory concerns persist. Social media can be used to investigate rare adverse effects not captured by traditional studies.

Objective

We sought to investigate large social media databases for suggestion of AIT skin and soft tissue infection (SSTI) risk and compare this risk to a comparator procedure with a sterile pharmaceutical.

Methods

We analyzed US-restricted data from more than 10 common text-based social media platforms including Facebook, Twitter, and Reddit between 2012 and 2016. We used natural language processing (NLP) to identify posts related to AIT and, separately, influenza vaccination (comparator procedure). NLP was followed by manual review to identify posts suggesting a possible SSTI associated with either AIT or influenza vaccination. SSTI frequencies with 95% CIs were compared.

Results

We identified 25,126 AIT posts, which were matched by social media platform to 25,126 influenza vaccination–related posts. NLP identified 4088 (16.3%) AIT posts that required manual review, with 6 posts (0.02%; 95% CI, 0.005%-0.043%) indicative of possible AIT-related SSTI. NLP identified 2689 (10.7%) influenza posts that required manual review, with 7 posts (0.03%; 95% CI, 0.007%-0.048%) indicative of possible influenza vaccination–related SSTI.

Conclusions

Social media data suggest that SSTI from AIT and influenza vaccination are equally rare events. Given that AIT's SSTI risk appears comparable to the risk using a sterile pharmaceutical based on social media data, current aseptic technique procedures seem safe.

Section snippets

Data source

Public text–based social media data were obtained from a large database compiled by Treato Ltd (Haifa, Israel). Treato uses proprietary methods to extract large volumes of web content from health-related websites (eg, weightwatchers.com), social media (eg, Facebook, Twitter, and Reddit), forums (eg, forums.thebump.com), and blogs. Their database, freely available on their website, includes more than 2 billion users' discussions, reflective of more than 40,000 medications and medical conditions.

Results

We identified 25,126 AIT posts, which were matched by social media platform to 25,126 influenza vaccine posts (Table III).

NLP identified 4088 AIT posts (16.2%) that required manual review. The most common term used to describe AIT in social media was “allergy shot” and its lexical variations (eg, plurals and misspellings). There were 6 posts (0.02%; 95% CI, 0.005%-0.043%) indicative of possible AIT-related SSTI (Table IV). Of the 6 possible SSTI posts, 4 included the terms “infection” or

Discussion

In this study, we analyzed large US-restricted public text–based social media data from sites including Facebook, Twitter, and Reddit, with the aim of discovering rare adverse events less likely to be identified by traditional research methods. We identified mentions of symptoms consistent with SSTI that the poster associated with AIT, a procedure prepared using aseptic technique, or influenza administration, a procedure using a sterile pharmaceutical. In total, only 6 of 25,126 posts related

References (29)

  • M.E. Falagas et al.

    Fluoroquinolones vs beta-lactams for empirical treatment of immunocompetent patients with skin and soft tissue infections: a meta-analysis of randomized controlled trials

    Mayo Clin Proc

    (2006)
  • J. Liu et al.

    Pharmacovigilance from social media: an improved random subspace method for identifying adverse drug events

    Int J Med Inform

    (2018)
  • N. Bhattacharyya

    Incremental healthcare utilization and expenditures for allergic rhinitis in the United States

    Laryngoscope

    (2011)
  • S.R. Durham et al.

    Long-term clinical efficacy of grass-pollen immunotherapy

    N Engl J Med

    (1999)
  • Cited by (11)

    • Natural language processing: A window to understanding skincare trends

      2022, International Journal of Medical Informatics
      Citation Excerpt :

      Another recent article demonstrated that evaluation of social media such as Reddit and Twitter may be an effective modality for examining patient experiences and opinions on sensitive or uncomfortable topics such as treatment of hypogonadism [2]. Likewise, large social media databases with data from Facebook, Twitter, and Reddit have been used for surveillance of rarely reported side effects including skin and soft tissue infection in the setting of allergen immunotherapy [3]. Moreover, data mining in social media may more efficiently capture side effects that are too rare to identify as a pattern in an individual clinic or department.

    • Identifying Urinary Tract Infection-Related Information in Home Care Nursing Notes

      2021, Journal of the American Medical Directors Association
      Citation Excerpt :

      Overall, NLP systems implemented using the baseline or baseline + PubMed terms achieved very good performance when applied on a human expert–annotated gold standard data set of clinical notes. Similar to previous applications of NimbleMiner in other domains,15,20,21 these results prove the feasibility of accurate extraction of UTI-related language from clinical notes. When at least 1 UTI-related category was discovered, labeling using the baseline + PubMed model resulted in 6.1% more episodes than when using just the baseline model.

    View all citing articles on Scopus

    This study was supported by the Division of Rheumatology, Allergy and Immunology, Massachusetts General Hospital. K.G.B. receives career development support from the National Institutes of Health (grant no. K01AI125631), the American Academy of Allergy,Asthma & Immunology Foundation, and the MGH Claflin Distinguished Scholar Award. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

    Disclosure of potential conflict of interest: L. Zhou reports grants from the Agency for Healthcare Research and Quality during the conduct of the study. R. Saadon reports grants from Treato Ltd during the conduct of the study. The rest of the authors declare that they have no relevant conflicts of interest.

    Co-first author.

    View full text