Google Searches Could Help Track Cancer Incidence, Mortality

America’s most popular search engine could soon be assisting with cancer research.

Google search volume across the United States could help fill in the gaps on cancer incidence and mortality data, according to a new study by scientists at UC San Francisco and the University of Pennsylvania. This is particularly true for cancers not documented in national registries, like basal and squamous cell carcinomas, the two most common kinds of skin cancer.

Countrywide internet search data offers clues to what interests the public, and can be used to estimate populations searching for specific topics. It’s likely that people Googling cancer-related health topics are patients, people that have yet to be diagnosed, or relatives of cancer patients, the study’s scientists suspected.

“This public and easily obtained data could be a proxy for cancer registry data,” said Eleni Linos, MD, DrPH, associate professor of dermatology at UCSF and senior author of the study published June 28, 2017, in JAMA Dermatology. “It’s important to understand cancer trends in real time in order to identify areas that need more attention and to evaluate the effectiveness of prevention programs,” she added.

Searches Linked to Cancer Incidence and Mortality

The study looked at Google search volume data, cancer incidences and mortalities at the state level, from 2009 to 2013, for the eight most common cancers in the United States. The number of Google searches on the names of the eight cancers – in the case of colorectal, the more commonly used name “colon cancer” was used – was compared to the number of patients diagnosed with these cancers reported to cancer registries.

Linos and her team found that by state, Google search volume correlated with the incidence of five of the country’s eight most common cancer types: lymphoma, melanoma, colon, lung and thyroid cancers. For four of those five – all but thyroid cancer – Google search volume also correlated with mortality rates.

The remaining cancer types analyzed, bladder, breast and prostate, did not correlate significantly with Google searches, which might be explained by robust public awareness campaigns that broadly increase interest. “There’s a clear seasonal pattern for some cancer searches, influenced by public health campaigns like Breast Cancer Awareness month in October,” Linos said.

Going Beyond Cancer

This proof-of-concept study suggests that Google search data could contribute much-needed cancer data, particularly for nonmelanoma skin cancers, which are estimated to affect more than 3 million Americans every year but are not tracked by central registries. That’s more than every other kind of cancer combined.

And, though the study has its limitations – not everyone has access to the internet or uses Google as their primary search engine – Linos and her team believe their methodology can go well beyond cancer research.

“There are many diseases, like autoimmune, cardiovascular or infectious diseases, that we don’t collect registry data for,” Linos said. “Could online search data be used to estimate the burden of these diseases too? I think it’s exciting.”