ABSTRACT
Background: The COVID-19 pandemic and response underscore the urgent need for real-time population-leveldata, especially for vulnerable populations (e.g., cancer patients, racial and ethnic minorities). Smartphoneapplications (apps) facilitate the collection of self-reported data at scale, the results of which can then be rapidlyredeployed to inform the public health response. The COVID Symptom Study is an app that was launched March24, 2020, and is now used by nearly 4 million people in the U.S., U.K., and Sweden. Methods: COVID Symptom Study app users self-report health status (e.g., symptoms, COVID-19 testing, healthcare utilization), comorbidities, demographics, and key risk factors for infection on a daily basis. Multivariableadjusted logistic regression models were used to determine the association of cancer and race with COVID-19prevalence, adjusting for age, sex, comorbidities, and risk factors for infection, from app launch through May 25,2020. Results: Among 23,266 individuals with cancer and 1,784,293 without cancer, we documented 155 and 10,249 self-reports of COVID-19, respectively. Compared to individuals without cancer, those with cancer had an increased riskof COVID-19 (adjusted odds ratio (aOR): 1.60;95% confidence interval (CI): 1.36-1.88). The association wasstronger among older participants >65 compared to younger participants (Pinteraction<0.001) and among males(aOR: 1.71;95%CI: 1.36-2.15) compared to females (aOR: 1.43;95%CI: 1.14-1.79;Pinteraction=0.02).Chemotherapy/immunotherapy was associated with a 2-fold increased risk of COVID-19 (aOR: 2.22;95% CI: 1.68-2.94) and risk of COVID-related hospitalization (aOR:2.47;95% CI: 2.22-2.76). In a separate analysis, wedocumented 8,990 self-reported cases of positive COVID-19 testing among 2,304,472 non-Hispanic whiteparticipants (93.6% of cohort);93 among 19,498 Hispanic participants;204 among 19,498 Black participants;608among 64,429 Asian participants;and 352 among 65,046 mixed race/other racial minorities. Compared with non-Hispanic white participants, the ORs for reporting a positive COVID-19 test for racial minorities ranged from 1.44(mixed race/other races) to 2.59 (Black). After accounting for risk factors for infection, comorbidities, andsociodemographic characteristics, the aORs were 1.37 (95% CI 1.09-1.72) for Hispanic participants, 1.42 (95% CI1.23-1.64) for Black participants, 1.44 (95% CI 1.33-1.57) for Asian participants, and 1.18 (95% CI 1.06-1.32) formixed race/other minorities. Conclusion: Our results demonstrate an increase in COVID-19 risk among ethnic minorities and individuals withcancer, particularly those on treatment with chemotherapy/immunotherapy. The association with minorities was notcompletely explained by other known risk factors for COVID-19 or sociodemographic characteristics. These findingshighlight the utility of app-based syndromic surveillance for quantifying the impact of the COVID-19 pandemic on at-risk populations.