A comparison of sociodemographic characteristics of patients undergoing lung cancer screening during the COVID-19 pandemic with pre-COVID patients

ABSTRACT

RATIONALE During the COVID-19 pandemic's spring surge, many lung cancer screening (LCS) programs delayed scans. Upon resuming screening, programs pivoted to telemedicine for shared decision-making (SDM) to minimize COVID transmission risks. This can lead to a cascade of alterations in LCS implementation, exacerbating disparities in screening. We hypothesize these changes have altered the sociodemographic characteristics of our LCS population. METHODS: At our institution in Philadelphia, LCS was deferred beginning March 18, 2020. We analyzed data from patients living in Philadelphia undergoing LCS through our centralized program upon resumption of screening on May 18 and the subsequent 6 months. Patient-level sociodemographic characteristics were compared with a “Pre-COVID” cohort screened May-November 2019. Geospatial analysis was performed for neighborhood-level internet access among patients receiving SDM inperson vs. telemedicine. RESULTS: The 408 patients screened after program reopening had a mean age of 64.2±5.8 years, 246 patients (60.3%) were female, 204 (50.1%) were White and 196 (48.2%) were Black/African-American. Over half were current smokers (n=224, 54.9%), and 172 patients (42.2%) had COPD. Compared with the Pre-COVID cohort, the Post-reopening cohort had no statistically significant differences in age, sex, race, smoking status or intensity, BMI, COPD, personal or family history of lung cancer, or insurance status. Education was significantly different between the two groups, with the Post-reopening cohort comprised of a lower percentage of patients with <HS education (16.9% vs. 11.3%), a greater frequency of patients with a HS diploma (42.6% vs. 46.1%), and fewer patients with education beyond HS (39.2% vs. 35.2%) (p<0.001). There was no significant difference in mean PLCOm2012 lung cancer risk (6.5±5.6% vs. 6.4±5.7%) or frequency of positive screens (≥Lung-RADS 3;8.9% vs. 8.4%). After excluding follow-up patients, education remained the only significant difference between groups. Analysis of neighborhood-level household internet access revealed no difference in the number of patients from census tracts with the lowest availability of household internet. Among the 15 census tracts with the greatest loss of patients from Pre-COVID to Postreopening, only 26.7% were in the highest quartile of internet deficit. CONCLUSIONS: LCS disparities can include disproportionate underscreening of individuals who are Black/African-American, live in rural areas, or have low socioeconomic status. Despite significant changes to our LCS program's workflow, the only significant difference between our Pre-COVID and Post-reopening cohorts was education level. Future research should be directed toward developing strategies to support LCS-eligible patients with low levels of education in the era of telemedicine.