Adverse Reactions Associated With Cannabis Consumption as Evident From Search Engine Queries

Elad Yom-Tov, Shaul Lev-RanJMIR Public Health and Surveillance 2017 · OpenReview

Abstract

Cannabis remains one of the most widely used psychoactive substances globally. However, its legal status in many jurisdictions makes studying its adverse drug reactions (ADRs) challenging. This proof-of-concept study investigates whether anonymized internet search engine queries can be used to detect ADRs associated with cannabis use. By analyzing Bing search data from millions of US users over a six-month period, researchers compared predicted usage and symptoms against traditional survey data (NSDUH) and official reporting systems (FAERS). The results demonstrate that search queries can accurately estimate the prevalence of cannabis use and identify both known and potentially novel adverse reactions.

Introduction

Traditional pharmacovigilance relies on reporting systems like the FDA’s Adverse Drug Reporting System (FAERS). While effective for regulated pharmaceuticals, these systems often fail to capture data on illicit or semi-legal substances due to social stigma and legal concerns. Users may be more likely to search for symptoms anonymously online than to report them to a healthcare provider. This study explores the potential of "digital breadcrumbs" to fill this critical data gap in public health surveillance.

Methodology

The research team analyzed anonymized US-based Bing search queries from November 2016 to April 2017. The methodology involved several key steps:

  1. User Identification: Grouping queries by anonymized IDs to track longitudinal behavior.
  2. Lexicon Development: Curating 123 cannabis-related terms (including slang) and 195 ICD-10 symptoms.
  3. Statistical Scoring: Utilizing metrics like the Query Ratio (QR) and Query Log Reaction Score (QLRS) to measure the association between cannabis searches and subsequent symptom queries.
  4. Temporal Mapping: Tracking symptom queries relative to "Day 0"—the first time a user searched for a cannabis-related term.

Geographic Prevalence of Use

A key finding of the study was the high correlation between search behavior and real-world drug use. The model's predicted prevalence reached an $R^2$ of 0.71 at the US census regional level and an impressive 0.93 at the state level when compared to National Survey on Drug Use and Health (NSDUH) data.

Predicted county-level prevalence of cannabis use based on Web-search queries.

This geographic alignment suggests that search data is a robust proxy for population-level substance use, providing a near real-time alternative to annual surveys.

Identifying Adverse Reactions

The study successfully identified several known acute and chronic adverse reactions. Most psychological symptoms, such as anxiety, paranoia, and hallucinations, spiked on "Day 0," suggesting immediate reactions to consumption. Conversely, "cough" queries peaked approximately 40 days after the initial cannabis search, reflecting the cumulative respiratory irritation associated with chronic smoking.

The symptoms most strongly associated with cannabis searches included:

  • Anxiety
  • Pain
  • Overdose
  • Paranoia
  • Depression

Interestingly, the data also highlighted plausible but less-documented reactions, such as pyrexia (fever) and seizures. While cannabis is often used therapeutically for seizures, these findings suggest it may act as a trigger in certain contexts, warranting further clinical investigation.

Conclusion

This research demonstrates that search engine queries are a powerful, low-cost tool for monitoring the health impacts of substances that are difficult to study through traditional means. By validating search data against official FDA reports ($r = 0.42, P < .001$), the study proves that anonymized digital logs can provide early warning signals for adverse drug reactions, potentially improving public health responses to emerging drug trends.

Summary updated

Abstract

Background: Cannabis is one of the most widely used psychoactive substances worldwide, but adverse drug reactions (ADRs) associated with its use are difficult to study because of its prohibited status in many countries.Objective: Internet search engine queries have been used to investigate ADRs in pharmaceutical drugs. In this proof-of-concept study, we tested whether these queries can be used to detect the adverse reactions of cannabis use.Methods: We analyzed anonymized queries from US-based users of Bing, a widely used search engine, made over a period of 6 months and compared the results with the prevalence of cannabis use as reported in the US National Survey on Drug Use in the Household (NSDUH) and with ADRs reported in the Food and Drug Administration’s Adverse Drug Reporting System. Predicted prevalence of cannabis use was estimated from the fraction of people making queries about cannabis, marijuana, and 121 additional synonyms. Predicted ADRs were estimated from queries containing layperson descriptions to 195 ICD-10 symptoms list.Results: Our results indicated that the predicted prevalence of cannabis use at the US census regional level reaches an R2 of .71 NSDUH data. Queries for ADRs made by people who also searched for cannabis reveal many of the known adverse effects of cannabis (eg, cough and psychotic symptoms), as well as plausible unknown reactions (eg, pyrexia).Conclusions: These results indicate that search engine queries can serve as an important tool for the study of adverse reactions of illicit drugs, which are difficult to study in other settings.JMIR Public Health Surveill 2017;3(4):e77doi:10.2196/publichealth.8391
BibTeX
@inproceedings{yomtov2017adverse,
  author = {Elad Yom-Tov and Shaul Lev-Ran},
  title = {{Adverse Reactions Associated With Cannabis Consumption as Evident From Search Engine Queries}},
  booktitle = {JMIR Public Health and Surveillance},
  year = {2017},
  url = {https://openreview.net/forum?id=Rv05VhfxBC},
}
Figure