With the aid of a smartphone app and a virtual army of citizen-scientists who self-report a range of factors relevant to the coronavirus pandemic, a research team has revealed a "striking" increase in risk for COVID-19 among cancer patients, particularly those treated with chemo or immunotherapy.
During an AACR COVID-19 and Cancer meeting, researchers also announced a greater than two-fold increase in risk among ethnic minorities not fully explained by other known risk factors for COVID-19 or socioeconomic data.
"Through the use of this smartphone app, we have been able to rapidly develop a large, prospectively collected longitudinal data set," said David A. Drew, PhD, Director of the Biobanking, Clinical & Translational Epidemiology Unit at Massachusetts General Hospital.
Beyond the general public, the team also has recruited individuals from ongoing cohort studies and clinical trials, particularly those with significant percentages of minorities and participants living with cancer.
"The mobile app allows us to be nimble and adaptable, allowing us to answer questions in a rapidly evolving pandemic," added Drew. "We are aware that our study may not be a generalized sample. The population may have limited access to smartphone applications, which could be a barrier in study access and individuals likely to volunteer for a survey-based study may impact our results."
The COVID Symptom Tracker was developed by researchers at Harvard T. H. Chan School of Public Health, Massachusetts General Hospital, and King's College London, in collaboration with the health science company called ZOE Global Limited.
The goal is to help physicians and others quickly identify COVID-19 "hot spots" and gauge appropriate health care responses to some of the most vulnerable members of the community. By gathering symptoms from millions of people, researchers say they can build a better picture of the onset and progress of the virus and start to figure out why some people become severely ill while others are only mildly affected.
The app asks contributors to answer a few questions about themselves and their current health-such as age, weight, height, ethnicity, work in health care-and then self-report daily how they feel and any symptoms they may be experiencing. Individuals are also being queried about any testing they've undergone, results, and possible treatments they've sought or received, in addition to preventive measures taken such as social distancing. Through mid-July, the app was being used by nearly 4 million people in the U.S., U.K., and Sweden.
The team recently published results in Science, detailing the project's methodology and ability to identify "hot spots" in real time (2020:368(6497):1362-1367). A second study published in Nature Medicine (2020;26:1037-1040) describes the app's ability to predict COVID-19 infection based on a symptom score, with a strong association revealed for anosmia, or loss of taste and smell.
Research Details
At the meeting, Drew described findings about the association of cancer and race with COVID-19, from March 29 until May 8, 2020; the majority were female with wide age range, but a median age of 44 years.
For both cancer and race cohorts, the researchers created a "fully adjusted" model to present their results, taking into consideration body mass index, sex, smoking status, co-morbidities at baseline such as diabetes and heart disease, and frailty. Additional risk factors included community interaction with known or suspected COVID cases and whether or not an individual is a health care worker, who studies have demonstrated to be a strong risk factor.
Among the findings, the team identified 21,155 users who reported they were living with cancer, with 124 self-reporting cases of COVID-19. Another 8,123 cases were reported among 1,575,259 individuals without cancer. "We see a striking approximate 88 percent risk for testing positive for COVID-19 among those living with cancer compared to those who are cancer-free," Drew said.
Likewise, the results showed the odds of testing positive for COVID-19 were stronger among cancer patients aged 65 and older, and males were more likely to test positive (aOR=1.71, 95% CI, 1.36-2.15) compared to females (aOR=1.43; 95% CI, 1.14-1.79).
Chemotherapy and immunotherapy were associated with a greater than two-fold increased risk for testing positive with COVID-19 (aOR=2.60, 95% CI, 2.023-3.34). And those living with cancer were more likely to seek treatment in a hospital or clinic setting.
Pandemic Disparities
"We are increasingly aware of the disproportionate impact COVID-19 is having on our racial and ethnic minority communities," Drew added. "While that's been part of the story from the beginning of the pandemic, as more and more data is collected the disparities are staggering.
Through the approximately 2 months of the study in the U.S., nearly 9,000 individuals self-reported cases of COVID-19 among 2.3 million non-Hispanic white participants, 93 cases among nearly 20,000 Hispanic participants, 204 among nearly 20,000 Black participants, 608 among almost 65,000 Asian participants, and 352 among more than 65,000 mixed race or other racial minorities.
With a caveat that their cohort consisted heavily of non-Hispanic whites, findings showed a higher risk for testing positive for COVID-19 in the U.S. among Hispanic, Latinx, Asian, and Black users compared to non-Hispanic white participants.
"We observe a greater than two-fold likelihood of reporting of COVID-19 positive among minority groups and the highest risks are observed by Black individuals," Drew said.
To explore the root causes of the racial/ethnic disparity, the researchers examined socioeconomic factors that might help explain the disproportionate risks among minority populations. These include zip codes, median income, and education, in addition to other factors related to risk for exposure, including social distancing, community exposures, and whether the individual was a frontline health care worker.
Drew said the team identified a "partial attenuation" following the adjustment that might explain some of increased risk in the U.S.; similar results were observed in the U.K., including among Middle Eastern and South Asian participants.
"There may be other factors in the United States that remain unmeasured in our analysis, including factors associated with systemic racism or institutionalized barriers," Drew said. "These might include higher proportion of non-health care essential work among vulnerable populations or limited access to health care.
In the future, the team will focus efforts to recruit other existing cohort studies and patient advocate networks that have access to non-academically affiliated subsets of the population in the U.S.
Warren Froelich is a contributing writer.