If Patients Won't Tell Their Docs About Drug Use, Their Search History Will

It’s easy to understand why illegal drug users are reluctant to speak with physicians about their drug use behaviors, but that reluctance hinders the scientific community’s understanding of how illicit substances like cannabis can cause adverse health events.

Israeli scientists have discovered a novel way to bridge this knowledge gap — they’re monitoring search engine queries in an attempt to sidestep the hurdles imposed by cannabis’s status as a schedule I illicit substance in the US.

Shaul Lev-Ran, MD

Shaul Lev-Ran, MD

“One of the challenges in studying adverse events of illicit substances, or those that individuals may not disclose due to several factors, such as social desirability biases and legal concerns, is substantial under-reporting,” said Shaul Lev-Ran, MD, director of addiction medicine at Lev Hasharon Medical Center, in an interview with MD Magazine. “Web searches allow for anonymous collection of real-time information regarding prevalence of use and, particularly, adverse effects.”

To learn more about cannabis-associated adverse drug reactions (ADRs), Lev-Ran partnered with Elad Yom-Tov, PhD, a computer scientist with Microsoft Research in Herzeliya, Israel. The pair collected 6 months of anonymized queries from US-based users of Microsoft’s Bing search engine, and compared it with the prevalence of cannabis use reported in the US National Survey on Drug Use in the Household (NSDUH), and with ADRs reported in the US Food and Drug Administration’s (FDA) 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.

Elad Yom-Tov

Elad Yom-Tov

Results indicated that the predicted prevalence of cannabis use at the US census regional level reached an R² of .71 NSDUH data, implying that 71% of the variance in the regional prevalence is predictable from the fraction of people making these types of queries.

The high accuracy of the model fit may indicate that people who use cannabis (particularly those concerned about adverse effects) ask about it online, perhaps because the internet is an anonymous channel of communication that is more accessible and less stigmatizing than official channels, like talking to a doctor, the authors wrote. Moreover, the findings suggest that it is possible to estimate ADRs via search queries.

Monitoring search queries have an additional benefit, authors said — queries for ADRs made by people who also searched the term “cannabis” revealed many of the known adverse effects of cannabis, like cough and psychotic symptoms, but also plausible unknown reactions, like pyrexia, hyperhidrosis, asthenia and vomiting.

“I think physicians could use our work to understand some of the likely causes of symptoms with which their patients present,” Yom-Tov said. “In areas where cannabis is illegal, patients may not feel comfortable mentioning their use of the drug when discussing a symptom with their doctor. Knowing that symptoms such as the ones we identified could be caused by cannabis can lead the doctor to inquire if the patient is using it, thereby getting to the cause of the symptom. Additionally, our methodology can help identify the adverse events of other illicit drugs.”

The study acknowledged several limitations, including the fact that it is often impossible to ascertain whether a person searching for drugs and ADRs is doing so out of curiosity or conducting research for themselves, a relative, or even for a patient. Additionally, internet users comprise a biased sample of the population, and thus the ADRs discovered may not be fully representative of the entire population.

Nonetheless, the results suggest that the sheer size of the data alleviates these concerns, the authors wrote, and the proposed method is able to identify adverse effects of drugs that are not captured by existing surveillance mechanisms.

“This proof-of-concept study indicates that web search results can be an effective method for studying adverse effects of several substances, in this case cannabis. In addition, our findings regarding specific adverse effects, as well as their temporality, may allow physicians to inquire about cannabis use among individuals presenting with these symptoms, or adversely, inquiring about specific adverse effects among cannabis users,” Lev-Ran said.

The study, Adverse Reactions Associated with Cannabis Consumption as Evident from Search Engine Queries, was published October 26 in The Lancet.