Internet Symptom Searches Reflect CHD Patterns

Internet searches for common heart disease symptoms, such as chest pain, correspond with regional, seasonal, and diurnal trends for coronary heart disease (CHD) hospitalizations, new research suggests.

Investigators used Google Trends to determine the frequency of Internet searches relative to chest pain symptoms between 2010 and 2017 in the United States, the United Kingdom, and Australia, correlating these trends with data on prevalence of CHD hospitalizations for the same period.

The researchers looked at geographic and seasonal variations, as well as the time of day when the most searches were conducted.

They found that Internet searches were correlated with the regions with the highest rates of CHD. Moreover, there was a threefold increase in searches in all three countries during winter months compared with summer months and a 100% increase during peak searching hours, which were between 6 AM and 8 AM local time.

“One of the things we can conclude is that for individuals with chest pain, one of the first resources they go to is Google, and search for ‘chest pain’ and use Google to understand their symptoms,” senior author Amir Lerman, MD, professor of medicine and director, Chest Pain and Coronary Physiology Clinic, Division of Ischemic Heart Disease and Critical Care, Department of Cardiovascular Medicine, Mayo Clinic, Rochester, Minnesota, told | Medscape Cardiology.

“On the one side, this can be good and an opportunity for them to learn more; but on the other side, it can potentially delay the time in which they actually seek medical care, so the medical community should try to collaborate with search engines to direct people toward getting help,” he said.

The study was published online October 4 in JAMA Cardiology.

Relating Habits to Symptoms

“Increasingly, patients are turning to online search engines such as Google to evaluate their symptoms prior to turning to a physician,” the authors write.

Google Trends, a free, unbiased, publically availably tool, analyzes search data, with each data point representing the proportion of searches for the queried item by the total searches on the geography of the area selected during a given time range selected.

The researchers used Google Trends to investigate the trends patients show in searching for information about chest pain — specifically, whether there is a relationship between region, season, and time of day of the searches, since CHD has been shown to exhibit geographic, seasonal, and diurnal variations.

“We used Google Trends to look at the patterns by which individuals are searching for words like ‘chest pain’ and whether those coincided with areas of high prevalence of heart disease, times of day when there is a higher prevalence of MI, and seasons when there is a higher prevalence of MI,” Lerman said.

To track search queries they analyzed the frequency of people’s use of Google to search for chest pain, chest discomfort chest pressure, or angina.

To investigate geographical variation, the researchers obtained data on annual CHD hospitalization rate per 1000 Medicare beneficiaries from the Centers for Disease Control and Prevention (CDC) Atlas of Heart and Stroke Statistics from January 2010 to December 2014.

They compared the query trends to the hospitalization data during the same time frame.

A similar query was investigated during the same time frame for knee pain, a control symptom.

The researchers also divided the geographic analysis of 2010-2014 into three 2-year categories: 2012-2014, 2011-2013, and 2010-2012.

Additionally, they examined data from three large, predominantly English-speaking countries from July 2012 to July 2017.

They used a multivariate linear regression model, using month and year as predictor variables and relative search frequency as the outcome variable.

To investigate diurnal variation, the researchers also analyzed the volume of searches in the United States, United Kingdom, and Australia for four 1-week intervals, adjusting hourly data to the most populous time zone in each country.

“Our initial motivation for this study was talking to patients and asking them what is the first thing they actually do when they have chest symptoms, and many told us that they went to search for their symptoms online,” Lerman recounted.

“We wanted to understand how the habits of people using search engines related to their symptoms and how prevalent that was,” he said.

Key Peak Times

“Chest pain” was the most frequently used search word of four keywords used to describe chest pain symptoms and was used 4.5-fold more than “angina” and 13-fold more than “chest pressure” or “chest tightness.”

These search terms were collectively queried 1.6-fold more than “knee pain.”

For CDC data for each state from January 2012 to December 2014 and the search frequency in each state over the same period, the Pearson correlation coefficient was 0.81 (P < .001).

Residents of Kentucky searched most often, and roughly twice as frequently as residents of Oregon, the state with the fewest relative searches.

The Pearson correlation coefficient from January 2011 to December 2013 was 0.81 (P < .001) and from January 2010 to December 2012 was 0.78 (P < .001).

By comparison, the correlation coefficient for knee pain searches from January 2012 to December 2014 was 0.31 (P < .001).

“State-by-state geographical variation of CHD hospitalizations correlated well with searches for related symptoms in all 3 of the time intervals studies,” the authors comment.

“This identifies a linear relationship between search frequency and disease burden at a population level,” they add.

In all three countries evaluated, searches for chest pain symptoms demonstrated significant monthly differences.

In the United States and United Kingdom, the searches peaked during the winter months, with a trough during the summer.

In the United States, the peak month was January, showing an 11% increase over the June, which was the trough month, while in the United Kingdom, the peak month was December, which showed a 39% increase in searches, compared with the trough month of July.

In Australia, July was the peak month, showing a 19% increase over the trough month of December.

In all three countries, “chest pain” searches consistently peaked between 6 AM and 8 AM local time, with a nadir in the late afternoon between 2 PM and 6 PM.

In the United States, searches peaked at 7 AM — a 110% increase from the nadir (between 2 PM and 4 PM).

Likewise, in Australia, searches peaked at 7 AM, representing a 157% increase over the nadir (between 3 PM and 5 PM). The difference was greatest in the United Kingdom, where searches peaked at 6 AM, a 189% increase over the nadir.

“The key peak times are the time frame of highest peak in people having heart attacks,” Lerman commented.

“If we know that this [an online search] is the first thing a patient does when having symptoms, this is a way we can reach the patient and direct them to the right resource,” he said.

Powerful Engines

Commenting on the study for | Medscape Cardiology, Roxana Mehran, MD, professor of medicine (cardiology) and professor, population health science and policy, Icahn School of Medicine, Mount Sinai, New York City, who was not involved with the research, said it is “incredibly important to actually understand what the public is doing with the use of the current incredible amount of information that’s available out there and the power that current systems like Google have to impact the prevalence of disease.”

She said she hopes that “Google and other important venues listen up and know that they are very much at the forefront of being able to make important differences in the burden of cardiovascular disease in man.”

Additionally, “these powerful engines, with their ability to reach, gather information, and make connections, can hopefully help us disseminate not only information but improve and enhance education, getting patients into a healthier lifestyle, enhancing their diet, and avoiding risk factors,” she said.

Also commenting on the study for | Medscape Cardiology, Timothy Henry, MD, an interventional cardiologist at Cedars Sinai Medical Center, Los Angeles, California, who was not involved with the study, called it “really fascinating because it raises awareness of how important social media is right now and that a significant segment of the population get their information that way.”

The study is “novel because it provides insight into changing patterns and how people get their information.”

Previous literature suggests reasons that people might turn to the Internet first when they are experiencing medical symptoms, he said.

“They don’t want to bother you [the medical professional] or waste their money and time,” he suggested.

He noted that it is unknown whether these searches “delay or expedite therapy.”

But knowing that these are patients’ patterns “has the potential to alter the way we treat and alert people,” he said.

The authors add, “Surveillance of online search engine activity may grow to become an important, emerging data source for cardiovascular research and care.”

No study funding was listed. Lerman and coauthors, Mehran, and Henry have disclosed no relevant financial relationships.

JAMA Cardiology. Published online October 4, 2018. Abstract

 For more Medscape Cardiology news, join us on Facebook and Twitter

AddSearch Custom Site Search

Be the first to comment

Leave a Reply

Your email address will not be published.