ABSTRACT
Background: The COVID-19 pandemic impacted the delivery of cancer care and outcomes in the United States (US). We examined the association between time-varying state-level weekly COVID19 mortality and progression-free survival (rwPFS), time to progression (rwTTP), and survival (rwOS) among pts with advanced non-small cell lung cancer (advNSCLC). Method(s): This retrospective study used the nationwide Flatiron Health electronic health recordderived de-identified database. The cohort included community oncology pts diagnosed with advNSCLC between March 1, 2020 and December 31, 2021 (follow-up through March 30, 2022). We extracted US data on COVID-19 deaths from the COVID-19 Data Repository by the Center for Systems Science and Engineering at Johns Hopkins University. We calculated state-level weekly COVID-19 death rates as weekly COVID-19 deaths per state population size from the 2019 American Community Survey. We categorized rates into quintiles based on all weekly rates during the observation period. Analyses were restricted to treated pts and indexed to start of first-line therapy. For rwPFS analyses, first occurrence of progression or death was considered an event, and pts were censored at last clinic note date. For rwTTP, only progression (not death) was considered an event, and pts with no event were censored at last clinic note date. For rwOS analyses, pts who did not die were censored at last structured activity. We used Cox proportional hazards models to estimate hazard ratios (HR) and 95% confidence intervals (CI) for associations between weekly time-varying state-level COVID-19 mortality rates and outcomes of rwPFS, rwTTP, and rwOS, adjusted for age at diagnosis, race/ethnicity, and state. Result(s): Among 7,813 advNSCLC pts, the median age at diagnosis was 70 years, the majority of the cohort was non-Hispanic White (59%), had non-squamous cell histology (68%) and a history of smoking (87%). Compared to pts living in states with the lowest quintile of COVID-19 mortality rates (Q1), pts living in states with the highest COVID-19 mortality (Q5) had worse rwOS (Q5 vs. Q1: HR 1.46, 95% CI 1.26-1.69) and rwPFS (Q5 vs. Q1: HR 1.18, 95% CI 1.05-1.33). No association was observed with rwTTP (Q5 vs. Q1: HR 1.05, 95% CI 0.90-1.22). Conclusion(s): In this study of real-world oncology data, we demonstrated the use of publicly-available COVID-19 mortality data to measure the time-varying impact of COVID-19 severity on outcomes in pts with advNSCLC. Higher state-level COVID-19 mortality rates were associated with worse rwOS and rwPFS among advNSCLC pts. The association with increased mortality among pts with advNSCLC may be related to COVID-19-related mortality or other factors such as pre-existing comorbidities which were not explored in this study.
ABSTRACT
Background: The COVID-19 pandemic was associated with declines in in-person clinical visits. While telemedicine visits have increased, uptake has varied. Here we assess demographic and socioeconomic factors associated with telemedicine use among patients initiating treatment for 21 common cancers at community oncology clinics. Methods: This retrospective study uses the nationwide Flatiron Health electronic health record-derived de-identified database of patients with cancer. Patient characteristics were determined using structured and unstructured data curated via technology-enabled ion. We included patients (≥ 18 years) who initiated first-line cancer treatment between March 2020 and September 2021 (follow-up through December 2021). We focused on differences in telemedicine use (≥ 1 telemedicine visit within 90 days after treatment initiation) across race/ethnicity, insurance coverage, rurality (per Rural-Urban Commuting Areas), and socioeconomic status (SES). SES was defined using census block group data from the American Community Survey (2015-2019) (quintiles representing least to most affluent areas) based on patient addresses and measured using the Yost Index (incorporating income, home values, rental costs, poverty, blue-collar employment, unemployment, and education information). We used logistic regression models adjusted for clinical characteristics (i.e., age, sex, performance status, and stage) to examine differences in telemedicine use. Results: This study included 24,164 patients (48.1% women, median age: 69 [interquartile range: 61-77] years), of whom 15.9% used telemedicine services. Black patients were less likely to use telemedicine services than White patients (11.4% vs. 15.6%, odds ratio [OR] 0.69 [95% confidence interval [CI]: 0.59-0.79], p<0.01). Telemedicine use was also lower among patients without documented insurance than well-insured (commercial and Medicare payers) patients (10.7% vs. 15.9%, OR 0.62 [95% CI: 0.54-0.72], p<0.01). Those in rural (9.8%, OR 0.51 [95% CI: 0.45- 0.58], p<0.01) and suburban areas (13.1%, OR 0.71 [95%: 0.64-0.79], p<0.01) were less likely to use telemedicine services than patients in urban areas (17.6%). Finally, patients in the least affluent areas had lower telemedicine use than those in the most affluent areas (10.2% vs. 24.3%, OR 0.35 [95% CI: 0.31-0.40], p<0.01). Conclusions: During the COVID-19 pandemic, nearly one-fifth of patients initiating cancer treatment used telemedicine services. However, there were substantial disparities: Black, uninsured, non-urban, and less affluent patients are less likely to use telemedicine services. While telemedicine may expand access to specialty care, the proliferation of these services may widen cancer care disparities if vulnerable populations do not have equitable access.