National Survey of Lung Cancer Screening Eligibility in United States Veterans.

The study team conducted a cross-sectional mailed, paper survey from January to November 2022. Participants who did not respond to the initial survey received one additional mailing. The survey was pilot-tested in-person at a single Veterans Affairs medical center (VAMC) from August to November 2021 among a randomly selected sample of Veterans in primary care and cardiology outpatient clinics. Minor revisions were made to improve item clarity. Veterans who completed the survey were given the option to receive $10. The study team entered survey responses in VA Research Electronic Data Capture (REDCap). A 5% random sample of surveys was checked to ensure data entry from paper to electronic datasets was accurate. This study was reviewed and approved by the VA Central Institutional Review Board (C-IRB E19–05), and VA Tennessee Valley Healthcare System Research & Development Committee. This study used the VA Corporate Data Warehouse (CDW), which reflects VHA’s electronic health records of all Veterans enrolled in VHA, to identify Veterans alive, over the age of 50 as of December 1, 2021, and who received outpatient or inpatient care at least once at any VHA location from November 2020 to November 2021 within an existing convenience cohort of Veterans and who had at least one chest imaging study (Common Procedural Terminology [CPT] codes 71271, G0297, 71250) between 2011 and 2021( Appendix Table 2 ). 4 , 24 To ensure inclusion of members of smaller demographic groups that are at high-risk of lung cancer and because tobacco use varies by geographic region, 25 – 29 investigators sampled a stratified, random sample of 2,000 Veterans from among 24 strata composed of all combinations of US Census region (Midwest, Northeast, South and West), rural status (rural, non-rural), and race (Black, American Indian or Alaska Native, all others). Investigators oversampled from strata composed of individuals in the American Indian or Alaska Native and Black groups or who live in rural areas and under sampled or sampled proportionally from the remaining strata. Investigators adapted survey items from the Behavioral Risk Factor Surveillance System 30 survey and the PLCOm2012 risk model. 31 The questionnaire asked respondents their current age, height, weight, gender, race, highest level of education, and home zip code. Medical history questions included personal history of cancer other than skin cancer, chronic obstructive pulmonary disease (COPD), and family history of lung cancer. If the participant indicated that he or she had smoked at least 100 cigarettes in his or her lifetime, five items asked a detailed smoking history including current smoking status, average number of cigarettes smoked per day, start age, stop age, and number of quit attempts within the past 12 months ( Appendix Table 3 ). For the USPSTF, CMS, ACS and NCCN guidelines, investigators coded participants as eligible or not eligible based on their survey responses. Definitions are provided in Appendix Table 4 . The PLCOm2012 uses a logistic-regression-based formula to generate a six-year probability of lung cancer for individuals who currently smoke or formerly smoked cigarettes. This study used a threshold of 1.51% or greater to categorize participants as eligible and reported eligibility of those who never smoked cigarettes as “not available and “unsure/don’t know as “no.” 31 Investigators selected 1.51% because this threshold is where lung cancer mortality improved in the LDCT arm compared with the chest x-ray arm in the National Lung Screening Trial, and it has been found to be a reasonable risk level in diverse populations. 32 , 33 Rural status, defined by the Rural-Urban Commuting Areas (RUCA), characterizes Veterans as residing in highly rural, rural, or urban locations. Investigators obtained these data from the CDW and classified Veterans residing in highly rural or rural areas as “rural” and those residing in urban areas as “non-rural” as described previously. 24 For surveys with missing data on patient characteristics, investigators obtained the most recent values in the CDW. If smoking history survey responses were inconsistent within the survey (e.g. a Veteran reported smoking every day but also quit smoking), investigators called Veterans to resolve discrepancies where possible. If Veterans could not be reached, they were classified as “currently smoking.” To account for the original sampling scheme and possible differential response rates across sampling strata, investigators used weights calculated by multiplying the original sampling weights by a stratum-specific non-response weight (which assumed non-response was at random within each stratum). Investigators calculated frequencies and percentages to describe the sociodemographic cohort characteristics. The first analysis reported the proportions of Veterans eligible by USPSTF 2013 and 2021 with Newcombe confidence intervals. A chi square test of independence tested the hypothesis by rural status. Due to weighting, investigators used jackknife standard errors to compute 95% CIs for proportions by rural status and the difference between these proportions. For the secondary analyses, investigators report the proportions eligible, categorized as “yes,” by CMS 2022, PLCOm2012, ACS 2023 and NCCN 2022 overall and stratified by rural status. Proportions eligible by race, region, and smoking status were also calculated. All analyses accounted for study design and survey weighting. Investigators used STATA Version 17 for data analysis. Investigators conducted two sensitivity analyses. First, the team excluded participants with inconsistent self-reported smoking histories that remained unresolved after follow-up phone calls. Second, because the original survey population completed at least one chest imaging study (reflecting engagement in care), it is possible that the smoking prevalence and screening eligibility may not be reflective of the overall population of Veterans age 50 or older. Accordingly, investigators obtained smoking history data available in the CDW for all Veterans age 50 or older who received care at VHA during 2020–2021 and were alive (N= 3,867,488). Smoking history data from the CDW represented what was documented in the VHA’s electronic medical record system as any tobacco use. Among a systematic sample of electronic health records that underwent chart review, this algorithm has a positive predictive value of 88%. The smoking history data was then incorporated into a modified set of sampling, non-response, and analysis weights. The sensitivity analysis was repeated using the primary outcome and the modified weights to obtain estimates representative of the broader United States Veteran population. Investigators calculated the proportion of age-eligible Veterans who met the USPSTF 2013 and 2021 smoking history criteria in the broad United States Veteran population in the second sensitivity analysis.

Of the 717,019 Veterans identified in the study population, there were 716,440 Veterans with sufficient data to be included in the sampling strata. Investigators surveyed a stratified, random sample of 2,000 patients and 754 responded (37.7%) ( Appendix Table 5 ). Of those who responded, 34.0% resided in rural versus 66.0% in non-rural areas ( Figure 1 ). Characteristics of the weighted cohort are shown in Table 1 . In the weighted analysis, respondents’ age ranged from 50 to 93 with 50.9% between ages of 70–79. The majority self-identified as White (74.4%), male (92.6%), and completed post high-school education (56.9%). Over three-quarters reported smoking at least 100 cigarettes in their lifetime (76.4%) while 23.5% reported that they never smoked. Of those with a smoking history, 70.8% formerly smoked. Characteristics of the unweighted cohort are available in Appendix Table 6 . Using the full sampling population, the overall proportions of Veterans eligible for lung cancer screening by USPSTF was 35.5% (95% CI 31.6% to 39.6%) by 2021 criteria and 27.8% (95% CI 24.3% to 31.7%) by 2013 criteria ( Figure 2 ). The proportion of eligible Veterans living in rural areas was significantly higher than the proportion of eligible Veterans living in non-rural areas when applying the USPSTF 2021 guidelines (rural = 41.2%; 95% CI 34.8% to 48.0% vs non-rural = 32.5%; 95% CI 27.7% to 37.7%, difference = 8.7%; 95% CI 0.04% to 17.0%, P = 0.037) ( Figure 3 and Appendix Table 7 ). When applying the CMS 2022, ACS 2023, and NCCN 2022, guidelines to the weighted cohort, 33.3% (95% CI 29.5% to 37.4%), 47.9% (95% CI 43.7% to 52.0%), and 52.1% (95% CI 47.9% to 56.3%) were eligible for lung cancer screening, respectively. Using the PLCOm2012 6-year lung cancer risk threshold of 1.51% yielded 47.4% (95% CI 43.2% to 51.6%) of the cohort as eligible for lung cancer screening ( Figure 2 ). Across all criteria, more rural Veterans were eligible for lung cancer screening compared to non-rural Veterans ( Figure 3 and Appendix Table 7 ). Results of the primary and secondary analyses by other variables accounted for in the sampling scheme (race/ethnicity, geographic region, and smoking status) are provided in Appendix Tables 8 , 9 , and 10 . In the first sensitivity analysis, investigators excluded participants with inconsistent self-reported smoking histories (<5% of the unweighted cohort). Eligibility results did not change by more than 2.4 percentage points and results stratified by rural status were consistent with the full analysis. In the second sensitivity analysis, investigators used modified weights to reflect smoking prevalence in the larger Veteran population ( Appendix Table 10 ). Investigators found that the proportions of Veterans eligible for lung cancer screening were slightly lower (1.7–3.7 percentage points) but represents approximately 1.4 million Veterans nationally. Restricting to age 50 to 80 in this broader population, the proportions of age-eligible Veterans who met USPSTF smoking history criteria was 36.8% by 2021 criteria and 29.0% by 2013. Differences in proportions eligible between rural and non-rural Veterans were consistent with the full analysis ( Appendix Table 12 ). Sensitivity analyses by race/ethnicity, geographic region, and smoking status are provided in Appendix Tables 13 , 14 , and 15 . Proportions meeting the USPSTF 2021 eligibility criteria by Veterans Integrated Services Networks (VISN) are demonstrated in Appendix Table 16 .

This study found that the proportion of Veterans age 50 or older who are eligible for lung cancer screening by USPSTF 2021 is nearly triple the proportion of Americans in the general population. 22 , 23 Of the 4,374,206 Veterans age 50 or older enrolled in VHA in 2022, this translates into approximately 1.4 million Veterans eligible for lung cancer screening. This study also found a higher proportion of USPSTF-eligible Veterans who reside in rural compared with non-rural areas. These results are critical for policy-makers and leaders to plan adequate resource allocation for population-based screening, diagnostic evaluations, and treatment for individuals diagnosed with screen-detected cancer. This study found that nearly 50% of Veterans who ever smoked met the USPSTF 2021 criteria for lung cancer screening, which is higher than 34.7% reported in the general population. 34 The present study also reports that the proportion of age-eligible Veterans who met USPSTF smoking history criteria increased from VHA’s previous national estimate of 32% for USPSTF 2013 criteria 21 to 36.8% for USPSTF 2021, which is an important data point used by VHA and researchers for estimating the eligible population. 4 , 35 The proportion of age-eligible Veterans who meet USPSTF 2013 smoking history criteria in the current study (29.0%) is lower as would be expected due to declining smoking rates 36 and increasing quit time beyond the 15-year threshold, but is generally similar compared with the prior VHA estimate (32%) from the 2013–2015 demonstration project. 21 Investigators found the proportion of Veterans who are eligible for lung cancer screening by USPSTF in the whole population increased by 7.7% or 9.8% for Veterans who ever smoked. This is less than the 13.7% increase reported for the general US population of individuals who ever smoked using BRFSS data. 34 This difference may be due to characteristics of the Veteran population that is predominantly White males. A study examining 2022 BRFSS national data, found that the USPSTF 2021 eligibility criteria were associated with most significant changes in eligibility by race (Asian 88% increase, Black 109% increase and Hispanic 86% increase). Furthermore, women had larger changes in eligibility compared to men using the USPSTF 2021. 22 Lung cancer mortality is higher among individuals who reside in rural areas than metropolitan areas. 19 This study found a higher proportion of rural Veterans were eligible for screening compared with non-rural Veterans regardless of the various eligibility criteria, likely driven by the known higher rates of cigarette smoking among rural Veterans. 37 Furthermore, odds of initial and annual repeat lung cancer screening are lower among rural Veterans compared with non-rural Veterans. 24 Lung cancer screening with LDCT can reduce lung cancer mortality in hig-hrisk individuals. This study’s findings highlight an urgent need for VHA to focus on expanding screening access through mobile CT and telehealth programs to reach rural communities. 22 Approximately 50% of individuals diagnosed with lung cancer would have met the USPSTF 2021 screening criteria at the time of their cancer diagnosis. 38 – 40 A prior study found that among those who formerly smoked cigarettes, 39% of incident lung cancers occurred in individuals who had more than 15 years since quitting, which is the threshold at which USPSTF 2021 guidelines recommend to stop screening. 41 The present study assessing Veteran eligibility for lung cancer screening found that 52.1% were eligible by NCCN 2022, 47.9% by ACS 2023 and 47.4% met the PLCOm2012 1.51% six-year risk threshold. The NCCN eliminated the upper age limit and quit limit among those who formerly smoked. 42 More recently, the American Cancer Society removed the quit year eligibility criteria in their guidelines. 42 Therefore, it is no surprise that this study found the NCCN 2022 and ACS 2023 eligibility criteria captures more Veterans eligible for lung cancer screening. Similarly, this study found that nearly 50% of Veterans would be eligible for lung cancer screening using the PLCOm2012 six-year 1.51% risk threshold. This is markedly greater than the 35.3% reported to be eligible by PLCOm2012 in a study using the 2019 BRFFS data. 34 Although VHA follows USPSTF, each VAMC has local authority to adjust their screening criteria based on local resources and policies and may consider adopting eligibility criteria outside of USPSTF to detect more lung cancers early. There are limitations to this study, including first, social desirability bias and recall bias, which can influence an individual’s response to survey items. The survey was adapted from validated items to lessen the potential impact of social desirability bias. Second, the survey had a relatively low response rate, but is similar to other VHA and non-VHA national surveys. 3 , 43 Third, investigators made an assumption in generating non-response weights, that the probability of response was independent of screening eligibility within each stratum, both of which could limit generalizability to all Veterans aged 50 or older. The sampling approach, which oversampled Veterans from high-risk demographic groups with relatively low numbers in the larger population, helped ensure that the analytical sample would have relatively robust numbers of Veterans from these groups even with a low response rate.

This study found that the proportion of Veterans eligible for lung cancer screening by USPSTF 2021 was nearly three times higher than the general US population (11.2%) and translates into approximately 1.4 million Veterans in 2022 who should be considered for screening. This study also found a higher proportion of USPSTF-eligible Veterans residing in rural areas compared with non-rural areas. These findings are critical for policy-makers and leaders who are planning cancer screening resource allocation and highlight an urgent need for VHA to focus on rural programming for equitable access to lung cancer screening.

Annual low-dose computed tomography (LDCT) detects lung cancer at early stages when it could be curable and improves lung cancer-specific mortality. 1 , 2 However, lung cancer screening is vastly underutilized despite multiple guideline and policy recommendations. 3 , 4 Each recommendation differs in its eligibility criteria. In 2013, the United States Preventive Services Task Force (USPSTF) recommended lung cancer screening with LDCT for individuals age 55 to 80 with at least 30 pack-years of cigarette smoking, and who currently smoke or formerly smoked and quit within the past 15 years in its Grade B recommendation. 5 This was a pivotal policy change because the Affordable Care Act mandates that private insurers cover the cost of any preventive health service receiving a USPSTF Grade A or B recommendation without co-payment. 6 In 2021, the USPSTF broadened its eligibility criteria to include younger individuals starting at age 50 with reduced smoking history of at least 20 pack-years. 7 Since 2015, the Centers for Medicare & Medicaid Services (CMS) have covered screening in a population similar to USPSTF. 8 , 9 The National Comprehensive Cancer Network (NCCN) guidelines is the most inclusive of the professional guidelines and the American Cancer Society (ACS) recently updated its recommendations in 2023 to remove quit time from its eligibility criteria. The Prostate Lung Colorectal and Ovarian (PLCO) m2012 is a risk calculator that is widely used in international settings to determine lung cancer screening eligibility. 10 – 15 Eligibility criteria by type as well as elements included in the PLCOm2012 risk calculator are summarized in Appendix Table 1 . Many Veterans are at-risk for lung cancer due to the high prevalence of cigarette smoking in this population. 16 Furthermore, the prevalence of cigarette smoking is higher among rural compared with non-rural individuals, 17 , 18 and rural residence is associated with a 20% higher lung cancer mortality compared with metropolitan residence. 19 Lung cancer screening with LDCT offers an avenue to decrease the existing lung cancer mortality disparity among rural populations. The Veterans Affairs (VA) National Center for Health Promotion and Disease Prevention issues screening recommendations for the Veterans Health Administration (VHA), and recommends lung cancer screening for individuals meeting the USPSTF 2021 eligibility criteria. 20 VHA conducted a clinical demonstration project from 2013 to 2015 and found that 32% of age-eligible Veterans met the USPSTF 2013 smoking history eligibility criteria, resulting in a national estimate of t approximately 900,000 Veterans eligible for lung cancer screening. 21 An estimated 11.2% of the United States (US) population over age 50 meet the USPSTF 2021 eligibility criteria, 22 , 23 but it is unknown how many Veterans are eligible, and furthermore, how many rural Veterans are eligible for lung cancer screening. Therefore, the primary objective of this study was to evaluate the proportions of USPSTF-eligible Veterans when applying the 2013 and 2021 criteria and to test the hypothesis that more USPSTF-2021-eligible Veterans reside in rural versus non-rural areas. A secondary objective was to calculate proportions of Veterans overall and stratified by rural status meeting other screening thresholds commonly used in the US and internationally: (1) CMS 2022 coverage criteria, (2) the PLCOm2012 model (3) ACS 2023 and (4) NCCN 2022 eligibility criteria.