In what may be another encouraging sign that the surge of the Omicron variant is subsiding in Massachusetts and around the country, the volume of Google search trends for COVID-19 symptoms is declining.
Data provided by Google on trends in searches for COVID-19 symptoms showed that after rising through most of December, the number of searches for symptoms like fever, chills, and cough began to drop in the last days of 2021 in the United States and Massachusetts.
The decline in search volume for certain COVID symptoms appears to align with data from the state’s Department of Public Health that show COVID-19 cases are declining in the state. According to state data, the seven-day average of new cases is 30 percent lower than when it peaked last week. And in the United States, the seven-day average of daily cases is beginning to tick downwards after appearing to reach a peak a few days ago, according to data from the Centers for Disease Control and Prevention.
While studies have indicated Google search data can be a useful tool, one expert said it’s better to consider that data in addition to other, more reliable metrics before making conclusions about the direction of the pandemic.
Jagpreet Chhatwal, an assistant professor at Harvard Medical School and the associate director of Massachusetts General Hospital’s Institute for Technology Assessment, which has created a simulator for COVID trends, said multiple measures can be considered together to show that COVID-19 levels are declining in the state. The MGH model is projecting that Massachusetts is now reaching the peak of the Omicron surge, while other models have said the state has reached the peak, or soon will. The level of COVID-19 in Boston-area waste water is also declining, and the Google search data points to “a very similar trend.”
“Look at Google Analytics, the prediction models, and waste water all pointing to the same trends,” Chhatwal said. “That tells us something about what’s happening in the state, which is trends should be going downwards from this point onwards.”
A study published in the scientific journal Nature in July 2021 looked into whether Internet searches for COVID symptoms can predict spread by analyzing Google Trends in Turkey, Italy, Spain, France, and the United Kingdom from January 2020 through August 2020.
That study found that Internet search interest for COVID symptoms “is a reliable predictor of later reported cases for the first wave of the COVID-19 pandemic,” and that search interest through Google can be “used to alert the healthcare system to prepare and allocate resources needed ahead of time.” The study identified that there was a lag time between searching for COVID symptoms and new cases that varied by country.
The study also noted that experts disagree over the validity of using Google Trends data as an epidemiological tool, but their research concluded that when studying the relationship between the trends data and real world data like COVID-19 cases, “the resulting correlation can be more reliable.”
In September 2020, Google made data on search trends available in order for researchers to “study the link between symptom-related searches and the spread of COVID-19″ in the hope that the data “could lead to a better understanding of the pandemic’s impact.”
The Google Trends data includes trends for more than 400 symptoms and health conditions in US counties and also contains trends from the past three years to account for changes in search due to seasonality. The data represent the volume of searches per symptom relative to a baseline of what’s typical for a given region.
Chhatwal said the Google Trends data is best used in tandem with other measurements because it depends on human behavior.
For example, if a new COVID-19 symptom were to arise, searches may spike for that particular symptom. Once the symptom becomes familiar among the population, Google searches will go down even as cases may continue to rise, Chhatwal said.
“It’s not always that the Google Trends is going to exactly follow what’s happening with the the actual epidemiology,” Chhatwal said.
In another example, if a family member is infected with COVID, a person might search a particular symptom at that time. At a later point, if they themselves or another family member is infected, they may not conduct the search again because they’ve already searched for it, Chhatwal said.
“The search is not as strongly correlated as the other things,” Chhatwal said. “Once the knowledge has been gained, then what is the purpose of doing more searches on Google? That’s what I’m wondering. That’s where things may get a little discordant.”
Waste water, by contrast, is directly linked to the amount of virus in the community, so if COVID levels are increasing or decreasing, “we should expect there’s a direct correlation,” Chhatwal said.
The levels of COVID-19 detected most recently in the Greater Boston area continued its downward trend, falling to less than a quarter of their Omicron-fueled peaks in early January. Biobot Analytics, which conducts the testing, said they have found that the amount of virus detected is correlated with newly diagnosed coronavirus cases several days later.
“Ballpark, from this week, we should start seeing things wind down from next week onwards. That’s what most of the models are projecting,” Chhatwal said.
Previous Globe material was used in this report.
Amanda Kaufman can be reached at email@example.com. Follow her on Twitter @amandakauf1.