Algorithmic fairness in pandemic forecasting: lessons from COVID-19.

Racial and ethnic minorities have borne a particularly acute burden of the COVID-19 pandemic in the United States. There is a growing awareness from both researchers and public health leaders of the critical need to ensure fairness in forecast results. Without careful and deliberate bias mitigation, inequities embedded in data can be transferred to model predictions, perpetuating disparities, and exacerbating the disproportionate harms of the COVID-19 pandemic. These biases in data and forecasts can be viewed through both statistical and sociological lenses, and the challenges of both building hierarchical models with limited data availability and drawing on data that reflects structural inequities must be confronted. We present an outline of key modeling domains in which unfairness may be introduced and draw on our experience building and testing the Google-Harvard COVID-19 Public Forecasting model to illustrate these challenges and offer strategies to address them. While targeted toward pandemic forecasting, these domains of potentially biased modeling and concurrent approaches to pursuing fairness present important considerations for equitable machine-learning innovation.

Understanding myocardial infarction trends during the early COVID-19 pandemic: an infodemiology study.

During the first months of the coronavirus disease 2019 (COVID-19) pandemic in early 2020, Google Trends data in the United States showed a strong increase in search query frequency for chest pain symptoms despite a concurrent decrease in search interest for myocardial infarction. This suggests a reduced attention to acute coronary syndrome (ACS) and chest pain as its main symptom during this time period. These observations could help explain why cardiovascular mortality rose dramatically despite a strong decrease in hospitalisation rates for ACS.

Addressing the Disproportionate Impact of COVID-19 on Communities of Color.

With Black and Hispanic communities across the USA experiencing more detrimental negative effects from the COVID-19 pandemic as compared with other demographic groups, the virus has exposed the racial and ethnic disparities in treatment and care that public health experts have been grappling with for years. This paper explains how the systematic collection of racial and ethnic data gleaned from COVID-19 testing in underserved communities can be used to better understand this pandemic and inform measures within our control to prevent the spread of disease in the future.

Declining interest in clinical imaging during the COVID-19 pandemic: An analysis of Google Trends data.

OBJECTIVE: Current evidence suggests a decrease in elective diagnostic imaging procedures during the COVID-19 pandemic with potentially severe long-term consequences. The aim of this study was to quantify recent trends in public interest and related online search behavior for a range of imaging modalities, and "nowcast" future scenarios with respect to imaging use. METHODS: We used Google Trends, a publicly available database to access search query data in systematic and quantitative fashion, to search for key terms related to clinical imaging. We queried the search volume for multiple imaging modalities, identified the most common terms, extracted data for the United States over the time range from August 1, 2016 to August 1, 2020. Results were given in relative terms, using the Google metric 'search volume index'. RESULTS: We report a decrease in public interest across all imaging modalities since March 2020 with a subsequent slow increase starting in May 2020. Mean relative search volume (RSV) has changed by -19.4%, -38.3%, and -51.0% for the search terms "Computed tomography", "Magnetic resonance imaging", and "Mammography", respectively, and comparing the two months prior to and following March 1, 2020. RSV has since steadily recuperated reaching all-year highs. CONCLUSION: Decrease in public interest coupled with delays and deferrals of diagnostic imaging will likely result in a high demand for healthcare in the coming months. To respond to this challenge, measures such as risk-stratification algorithms must be developed to allocate resources and avoid the risk of overstraining the healthcare system.