Estimating the Risks from COVID-19 Infection in Adult Chemotherapy Patients

Matt Williams; Kerlann Le Calvez; Ella Mi; Jiarong Chen; Seema Dadhania; Lillie Pakzad-Shahabi

We used combined data from the China CDC 3 , Italian public health authorities 13 and a COVID outbreak on a cruiseship 14 to estimate case fatality rates (CFR) by age group. We identified studies that estimated risk of harms (including death) in patients with cancer or immunosupression during influenza outbreaks and extracted data on risks of harm. We explored sensitivity by constructing models using central estimates of risk of harm and most optimistic estimates. We constructed clinical scenarios by identifying key clinical trials in the curative and palliative setting in five sentinel tumours sites (breast, lung, colorectal, prostate and brain), and extracted data on absolute benefit from treatment in those trials. For trials in the curative setting, we extracted data on the absolute difference in OS at the timepoint specified by the trial (typically 3 or 5 years). For palliative chemotherapy trials, we extracted data on the difference in absolute survival at 3, 6 and 12 month timepoints. We then compared the improvement in absolute survival from treatment with the potential harm if COVID infected, and visualised these results.