Racial and Socioeconomic Disparities in Lung Cancer Screening in the US: A Systematic Review

Abstract

Background:

Non-small cell lung cancer (NSCLC) is the leading cause of cancer deaths. Lung cancer screening (LCS) reduces NSCLC mortality; however, a lack of diversity in LCS studies may limit the generalizability of their results to marginalized groups who face higher risk for and worse outcomes from NSCLC. Identifying sources of inequity in the LCS pipeline is essential to reduce disparities in NSCLC outcomes.

Methods:

We searched three major databases for studies published from January 1, 2010, to February 27, 2020, that met the following criteria: 1) included humans subjects between the ages of 45 and 80 years who were current or former smokers, 2) written in English, 3) conducted in the US, and 4) discussed socioeconomic and race-based LCS outcomes. Eligible studies were assessed for risk of bias.

Results:

Of 3,721 studies screened, 21 were eligible. Eligible studies were evaluated, and their findings were categorized into three themes related to LCS disparities faced by Black and socioeconomically disadvantaged individuals: (1) eligibility; (2) utilization, perception, and utility; and (3) post-screening behavior and care.

Conclusion:

Disparities in LCS exist along racial and socioeconomic lines. There are several steps along the LCS pipeline in which Black and socioeconomically disadvantaged individuals miss the potential benefits of LCS, resulting in increased mortality. We identified potential sources of inequity that require further investigation. We recommend the implementation of prospective trials that evaluate eligibility criteria for underserved groups and the creation of interventions focused on improving utilization and follow-up care in order to decrease LCS disparities.

Introduction

Non-small cell lung cancer (NSCLC) is the leading cause of cancer deaths, as most NSCLC patients present with stage IV disease, when cure is unlikely.^1,2 However, lung cancer screening (LCS) using low-dose chest computed tomography (LDCT)^3,4 enables the detection of NSCLC at earlier stages, leading to increased rates of resection for cure.^5,6 The National Lung Screening Trial (NLST), a US-based randomized controlled trial, showed that three annual screenings using LDCT reduced NSCLC mortality by 20% and all-cause mortality by 6%.^7,8 Thus, LCS is a critically important tool to decrease mortality from NSCLC. Accordingly, in 2013, the US Preventive Services Task Force (USPSTF) recommended annual LCS for 55–80-year-old (1) current smokers with a 30 pack-year smoking history or (2) former, heavy smokers who have quit within the past 15 years.⁹ Subsequently, the Dutch-Belgian LCS randomized controlled trial [Nederlands–Leuvens Longkanker Screenings Onderzoek (NELSON)],¹⁰ which used different eligibility criteria than the NLST, found an even greater mortality benefit of LCS. Based on these new results, the USPSTF is updating its LCS guidelines to recommend annual screenings for 50–80-year-old (1) current smokers with a 20 pack-year smoking history or (2) former, heavy smokers who have quit within the past 15 years.¹¹

Although the results from the NLST and NELSON trial were exciting, neither focused on attaining significant racial or socioeconomic diversity among its participants. Indeed, 89.6% of participants in the NLST were White and of higher socioeconomic status (SES) than the general population,^8,12 and these important demographic factors were not reported for the NELSON trial.¹⁰ As such, the results from these landmark trials may not be generalizable to other groups, such as Black and socioeconomically disadvantaged individuals, who have: 1) a higher incidence of NSCLC,^13–18 2) higher mortality from NSCLC,^{15–17,19–21} 3) lower rates of treatment,^22–30 and 4) a greater chance of being diagnosed at advanced stages of the disease.^31,32

85% of lung cancers result from tobacco smoking,³³ as such the fact that such strong tobacco marketing efforts were historically directed at members of marginalized and underserved communities must be highlighted,³⁴ especially as these same groups have the highest incidence and mortality from the disease. Therefore, assurance that members of these marginalized groups obtain life-saving LCS is a true restorative justice issue. In order to extend the benefits of LCS to all high-risk individuals, regardless of race/ethnicity or income, it is essential that we identify and rectify all potential sources of inequity in the LCS pipeline, from eligibility to follow-up care. For example, eligibility criteria based on a majority White, high-SES population are likely to exclude the marginalized groups at greater risk for NSCLC. There are also considerable barriers that contribute to the low participation rates (up to 16%) among individuals who are eligible for LCS, such as low referral rates by uninformed providers and uncertainty regarding the benefits of LCS.^35,36 These barriers may be magnified in the care of marginalized groups. Furthermore, support and follow-up care for individuals who are screened may differ based on patient demographics.

Thus, each step from eligibility to follow-up represents an opportunity to reduce disparities in LCS and, ultimately, NCSLC outcomes. In this systematic review, we aim to critically appraise all available studies conducted in the US that have examined the associations of LCS outcomes with race and SES to identify known disparities, as well as potential sources of inequity that require further investigation to be fully mitigated.

Methods

Search Strategy

We conducted this systematic review according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.³⁷ Two professional medical librarians (JD and KL) searched three online databases: PubMed, Ovid MEDLINE, and CINAHL Plus. The librarians worked with the reviewers (AA, GD, and LE) to develop a list of keywords, MESH terms, and subject headings to find all publications from January 1, 2010, to February 27, 2020, that examined the impact of race and/or SES on LCS. Keywords included smoking habit, cancer screening, socioeconomic status, lung cancer, race, and disparities. A complete list of search terms is available (Supplementary Appendix 1). The literature search matrix used for PubMed was adapted for use in the other databases. No other limitations were applied to the search to avoid unintended exclusion of relevant documents.

Article Review Process

Once the articles were retrieved from each database, they were exported into a reference management software (Covidence; Veritas Health Innovation Ltd) for study selection. This tool allowed all reviewers (AA, GD, LE, and ES) to review the articles independently and simultaneously. In the initial screening stage, the title of each article was reviewed for topic relevance, and the abstract was reviewed for further clarification as necessary. During this screening, each article was reviewed by only two reviewers. Any article approved by both reviewers was deemed eligible for full-text review, whereas articles with two “no” votes were eliminated. Any article with one “yes” and one “no” vote was reviewed and discussed by all reviewers collectively until consensus was reached on whether it was eligible or ineligible for full-text review.

Data Abstraction

In the full-text review stage, each reviewer (two per article) independently assessed the original research studies for eligibility according to the predefined criteria (Table 1). The same voting system used to determine full-text review eligibility was used to determine article eligibility for the systematic review. The reviewers abstracted publication, participant, and study characteristics (author, title, study type, sample size, key findings, outcomes, and themes) from each article.

Table 1.

Inclusion and exclusion criteria used for the systematic review.

Inclusion Criteria	Exclusion Criteria
Peer-reviewed literature	Gray literature (editorials, news, letter to the editor )
Published between January 1, 2010, and February 27, 2020,	Published before January 1, 2010, or after February 27, 2020
Written in English	Not written in English
Study conducted in the US only	Study conducted outside of the US
Humans subjects only	Animal study
Participants 45–80 years of age	Participants only <45 or >80 years of age
Current and former smokers	Does not include current or former smokers
Evaluates socioeconomic and/or race-based outcomes relating to LCS	Does not evaluate socioeconomic and/or race-based outcomes relating to LCS
Study design: Cohort studies, randomized control trials, and observational studies	Case control studies, case control studies
Setting: Clinical or healthcare setting	Non-clinical or health care settings

Study Quality Assessment

Risk of bias was assessed using the National Heart, Lung, and Blood Institute (NHLBI) Quality Assessment Tool for Observational Cohort and Cross-Sectional Studies.³⁸ This tool includes questions for evaluating potential flaws in study methods or implementation that may lead to bias and lower study quality and has been used in previously published systematic reviews.³⁹ Each question could be answered with a “yes,” “no,” or “not reported/not applicable.” Based on the answers, the reviewers then rated individual studies as being of “good,” “fair,” or “poor” quality. Studies rated “good” were considered to have the least risk of bias and to possess valid results. Studies rated “fair” were considered to have some bias, but not enough to invalidate the results. Studies with a “poor” rating were considered to have a significant risk of bias. The quality of each study was rated independently by two reviewers (ES and LE), who then met to discuss their ratings. Discrepancies were discussed by the reviewers until consensus was reached.

Results

We identified 5,169 potentially relevant citations, including 1,447 duplicates, leaving 3,721 studies to be screened for eligibility. Of these, 98 were selected for full-text review, and 21 were deemed eligible using our predefined criteria (Table 1; Figure 1). The reasons for exclusion are documented in Table 2. Detailed study characteristics are provided in Table 3.

Figure 1. Screening process for articles included in the systematic review.

This diagram shows each step in the selection process for articles included in the systematic review.

Table 2.

Reasons for excluding articles from the systematic review.

# of Articles Excluded	Reason
28	Wrong study design
17	Does not evaluate socioeconomic and/or race-based outcomes relating to LCS
13	Wrong outcomes
12	Wrong patient population
5	Gray literature (editorials, news, letter to the editor)
1	Study not conducted in the US
1	Wrong setting

Table 3.

Individual study characteristics for articles included in the systematic review.

Publication Details	Participant Characteristics				Study Characteristics
Author and -Publication year	Age	Gender	Race/Ethnicity	Geographic Location	Data Source	Sample Size	Study Design
Aldrich et al., 201946	40–79	Male and female	Black White	Southern United States	NLST Southern Community Cohort Study	48,364	Cohort study
Annangi et al., 201945	40–85+	Male and female	Black White	United States	Surveillance, Epidemiology, and End Results (SEER) Program from NCI US cancer registries Centers for Disease Control (CDC) NLST	486,403	Cross-sectional study
Balekian et al., 201963	55–65	Male and female	Black White	United States	NLST	723	Cohort study
Carter-Harris et al., 2018a51	55–79	Male and female	Black White	United States	Collected through community-based recruitment methods in the state of Indiana	438	Cross-sectional study
Carter-Harris et al., 2018b61	55–77	Male and female	Black White	United States	Collected through community-based recruitment methods in the state of Indiana	159	Cross-sectional study
Cattaneo et al., 201857	55–80	Male and female	Black White	Maryland, United States	NLST Veterans’ Health Administration Ann Arundel Medical Center patient data	1,241	Cohort study
Fiscella et al., 201544	55–77	Male and female	Black Hispanic White	United States	National Health and Nutrition Examination Survey PLCO	2,562	Cohort study
Guichet et al., 201859	50–78	Male and female	Asian Black Hispanic White	California, United States	NLST University of Southern California	275	Cohort study
Hall et al., 201855	55–74	Male and female	Black White	United States	Massachusetts General Hospital NLST	169	Cross-sectional study
Han et al., 202040	50–80	Male and female	Asian Black Hispanic White	United States	National Health Interview Survey Cancer Intervention and Surveillance Modeling Network Smoking History Generator USPSTF PLCO US Census Southern Community Cohort Study	100,000	Cohort study
Japuntich et al., 201847	55–80	Male and female	Black White	Rhode Island, United States	Miriam Hospital USPSTF	200	Cohort study
Kumar et al., 201660	50–74	Male and female	Black White	United States	NLST CDC	6,813	Cohort study
Li et al., 201948	55–80	Male and female	Black White	United States	2014 Health and Retirement Study USPSTF	7,348	Cross-sectional study
Pasquinelli et al. 201858	55–74	Male and female	Black White	Illinois, California	NLST University of Illinois at Chicago	500	Cohort study
Percac-Lima et al., 201956	50–79	Male and female	Hispanic Non-Hispanic	Massachusetts, United States	Massachusetts General Hospital-affiliated community health centers NLST	894	Cross-sectional study
Richmond et al., 202052	55+	Male and female	Black White	North Carolina, United States	US Census Lung Cancer Screening Registry University of North Carolina at Chapel Hill	262	Cohort study
Ryan et al., 201649	33–91	Male and female	Black White	Maryland, United States	NCI-Maryland Lung Cancer Study NLST Centers for Medicare and Medicaid Services USPSTF	1,658	Cohort study
Sesti et al., 201962	55–74	Male and female	Black White	United States	NLST	14,000	Cohort study
Steiling et al., 202053	55–80	Male and female	Asian Black Hispanic White	Massachusetts, United States	Boston Medical Center Lung Cancer Screening Program Boston Medical Center Clinical Database Warehouse NLST USPSTF	1,325	Cohort study
Su et al., 201854	50–80	Male and female	Black Hispanic White	New York, United States	Albert Einstein College of Medicine USPSTF	855	Cohort study
Tanner et al., 201512	55–74	Male and female	Black White	United States	NLST	53,452	Cohort study

The 21 eligible studies were evaluated, and their findings were categorized into three overarching themes related to steps in the LCS pipeline in which racial and socioeconomic disparities are known or are likely to exist (Table 4). The disparities at each step in the pipeline negatively impact outcomes and ultimately result in disparate mortality rates (Figure 2).

Table 4.

Identified/Potential Sources of Racial and Socioeconomic Disparities in the Lung Cancer Screening Pipeline.

	Eligibility	Utilization, Perception, and Utility			Post-Screening Behavior and Care
		LCS Behavior/Utilization	Perceptions/Understanding	NSCLC Stage at Diagnosis	Smoking Behavior & Smoking Cessation	Follow-Up	Surgical Treatment	Mortality
Aldrich et al., 201946	●
Annangi et al., 201945	●			●
Balekian et al., 201963							●
Carter-Harris et al., 2018a51		●
Carter-Harris et al., 2018b61					●
Cattaneo et al., 201857		●		●		●
Fiscella et al., 201544	●				●
Guichet et al., 201859				●		●
Hall et al., 201855			●
Han et al., 202040	●
Japuntich et al., 201847	●	●
Kumar et al., 201660					●
Li et al., 201948	●
Pasquinelli et al. 201858				●
Percac-Lima et al., 201956			●
Richmond et al., 202052
Ryan et al., 201649	●
Sesti et al., 201962						●
Steiling et al., 202053		●
Su et al., 201854		●		●				●
Tanner et al., 201512					●			●

Figure 2. Lung cancer screening pipeline.

This schematic presents all reported findings captured by this systematic review. Statements in italics indicate contentious or ambiguous results. Any associations not shown were not reported or captured by our search and thus require further study.

LCS Eligibility:

Not Eligible

Seven studies (33%) investigated eligibility for screening according to the 2013 USPSTF LCS guidelines. Several studies found that Black smokers, with or without a diagnosis of NSCLC, were less likely to be eligible for LCS than their White counterparts. Han et al.⁴⁰ evaluated the characteristics of smokers in younger (50–54-year-old) and older (71–80-year-old) individuals who were missed by the USPSTF guidelines but were selected as high-risk by the PLCO_m2012 model, a validated risk-based screening model that was shown to be more sensitive than the USPSTF criteria for lung cancer detection.^41–43 Han et al. found that high proportions of younger and older Black individuals were ineligible for screening by USPSTF criteria despite being high-risk according to the PLCO_m2012 model. This rate of ineligible but high-risk Black individuals was significantly greater than the rate of ineligible but high-risk White individuals.⁴⁰ In a similar study, Fiscella et al. found that PLCO_m2012 criteria resulted in a statistically significant increase in LCS eligibility for Black men compared to Medicare criteria, which adopted the USPSTF smoking criteria but changed the age criterion to 55–77 years.⁴⁴ Annangi et al. used the Surveillance, Epidemiology, and End Results (SEER) database to determine the frequency of early-onset NSCLC in the 45–54-year age range.⁴⁵ Black individuals had a significantly greater frequency of early-onset NSCLC compared to White individuals.⁴⁵ However, these patients with early-onset NSCLC were not eligible for LCS because they were younger than the age range defined by USPSTF guidelines.

Additionally, Aldrich et al. found that the percentage of individuals eligible for LCS according to USPSTF guidelines was significantly lower among Black smokers diagnosed with NSCLC than among White smokers with NSCLC.⁴⁶ Three other studies also found that Black current and former smokers were significantly less likely to be eligible for LCS according to USPSTF guidelines because they had a shorter smoking history or longer time since quitting than required for eligibility under USPSTF guidelines.^47–49 Another study applied USPSTF screening criteria to 2,000 patients diagnosed with NSCLC and found eligibility differences based on race, as well.⁵⁰ These findings highlight the shortcomings of current USPSTF guidelines. Although the proposed updates to the USPSTF criteria may improve eligibility rates for Black individuals at high-risk for lung cancer, potential gaps in eligibility remain that may limit screening for Black and other underrepresented minority individuals (Table 5). Further, innovative research methods, such as rigorous data modeling, that incorporate adequate numbers of overlooked groups are needed to quickly and inexpensively address remaining gaps and improve our understanding of appropriate eligibility criteria for all.

Table 5.

USPSTF Recommendation Updates and Remaining Gaps

	Current USPSTF eligibility recommendations	New proposed USPSTF recommendations	Possible Remaining Gaps after updated guidelines
Population	• Aged 55 to 80 years • 30 pack-year smoking history • Currently smoke or have quit smoking within the past 15 years.	• Aged 50 to 80 year • 20 pack-year smoking history • Currently smoke, or have quit smoking within the past 15 years	• Annangi found that 7.2% of Black NSCLC patients were less than 50 years of age compared to only 4.3% of White NSCLC patients; therefore, the new guidelines will still miss more Black vs White high-risk smokers.45 • Lung cancer risk remains elevated after 15 yeas of cessation;73–75 therefore, Black smokers who are more likely to have quit times of greater than 15 years may be excluded from LCS at higher rates secondary to this criteria .47,72,94
Recommendation	• Annual screening for lung cancer with LDCT • Screening should be discontinued once a person has not smoked for 15 years.	• Annual screening for lung cancer with LDCT • Screening should be discontinued once a person has not smoked for 15 years.

The association between education and screening eligibility was evaluated in two studies. In their study on high-risk individuals who were missed by the USPSTF guidelines, Han et al. found that, compared to those with higher education, a significantly higher proportion of high-risk individuals with a high school education or less were ineligible because they were younger than the recommended age range (p<.001).⁴⁰ In contrast, Li et al. found that, compared to those with a college education or higher, individuals with a high school education or less were more likely to be eligible for LCS (odds ratio [OR]=1.8, 95% confidence interval [CI]=1.5–2.3). Interestingly, they also found that higher household income was associated with greater eligibility.⁴⁸ This finding and the disparate results regarding the association between education and LCS eligibility require further investigation.

LCS Utilization, Perception, and Utility:

LCS Not Offered or Utilized

Five studies (24%) evaluated whether race and/or SES impact the utilization of LCS. Three studies reported race-based differences in LCS utilization by screening-eligible patients. They found that Black participants were less likely to have been screened or to intend to be screened,^51,52 and eligible non-Black patients were 2.8 times more likely to be screened than eligible Black patients (30% vs. 12%).⁴⁷ Additionally, a higher annual income was significantly associated with the completion of or intention to receive screening.⁵¹ A fourth study conducted in a diverse community setting (41.4% Black individuals) found that Black participants had a lower screening rate, comprising only 37.6% of the screen population, whereas White participants made up 46% of the screened group (p<.001). They also found that unscreened patients had a lower annual household income than those who were screened.⁵³ Specifically, the study noted that 68.4% of the screened population had an annual household income greater than $50,000, whereas only 59.6% of the unscreened population had the same income (p=.022).⁵³ Conversely, another study evaluated 855 patients with lung cancer who were eligible for screening (46% Black, median income by home zip code was $20,009) and did not find significant differences based on race, ethnicity, or SES, as measured by median income, between those who completed screening versus those who did not.⁵⁴ Nevertheless, there is overall compelling evidence that race- and SES-based disparities exist in LCS utilization.

Negative Perceptions/Understanding of LCS

Two studies (10%) examined patient beliefs about LCS. One study did not find any correlations between sociodemographic variables and patients’ perceptions about the accuracy of LCS; however, the study did find a correlation between patients’ education levels and their understanding of why they were referred for LCS (p=.01).⁵⁵ Another study reported that, compared to non-Latinx patients, Latinx patients were more likely to believe that NSCLC can be prevented (74.6% vs. 48.2%; OR=3.07, 95% CI=1.89–5.01), were less worried about developing NSCLC (34.8% vs. 50.3%; OR=0.44, 95% CI=0.27–0.72), and had a greater willingness to be screened when educated about screening (90.7% vs. 67%; OR=4.79, 95% CI=2.31–9.05)].⁵⁶

Late Stage Diagnosis

Five studies (24%) investigated the impact of race and/or SES on the stage of NSCLC in those diagnosed via LCS. In an LCS community program in which 10.2% of participants were Black, 21 of the 29 cancers identified in 26 patients were stage I.⁵⁷ Additionally, in a screening program that, compared to the NLST, served a greater proportion of Black participants (69.6% vs. 4.5%) and current smokers (72.8% vs. 48.1%), more than 50% of NSCLC cases were detected at stage I.⁵⁸ Su et al. found that screening impacted NSCLC stage at diagnosis, with detection at an earlier stage for patients who were screened compared to those who were not; however, they found no difference in stage at diagnosis based on race, ethnicity, or SES.⁵⁴ Through analysis of the SEER database, Annangi et al found that a significantly lower proportion of Black lung cancers was diagnosed at early stages (IA, IB, IIA, IIB) compared to Whites across all age groups (p<0.05).⁴⁵Finally, in an LCS program with a majority Black patient population (n= 231 of 275 [84%] who underwent screening), no early-stage disease was found; instead, two patients were diagnosed with advanced-stage lung cancer.⁵⁹ Across studies, there were mixed findings regarding whether Black patients are more likely than White patients to have more advanced disease detected by LCS; however, like the NLST, most LCS studies with a higher proportion of Black participants found that early-stage disease was prevalent.

Post-Screening Behavior and Care:

Failure to Quit Smoking

Three studies (14%) examined smoking behavior and cessation differences in LCS participants by race/ethnicity. Both Hispanic ⁴⁰ and Black participants reported fewer pack-years of smoking compared to White participants.^44,60 Black participants were also less likely to report previous alternative tobacco use than White participants (32.6% vs. 39.0%, p=.007) but more likely to report menthol cigarette use (58.8% vs. 20.8%, p<.001).⁶⁰

Kumar et al. found that, compared to White participants, Black participants in the NLST had higher rates of smoking cessation for periods of 24 hours (52.7% vs. 41.2%, p<.001) and 7 days (33.6% vs. 27.2%, p=.002).⁶⁰ In addition, Black participants had slightly higher rates of participation in smoking cessation programs (7.0% vs. 4.5%, p=.03). Higher incomes were also associated with a higher likelihood of 24-hour and 7-day quit attempts; however, race and income were not significantly predictive of smoking cessation success.⁶⁰

Racial differences were also found in current smokers’ preferences for receiving smoking cessation information, with White participants being four times more likely to report a preference for digital support (social media, Internet, web-based programs, and/or text messages) than face-to-face support, telephone support, or printed materials. This relationship was not impacted by income.⁶¹

Lack of Follow-Up Care/Loss to Follow-Up

Three studies (14%) reported on the completion of follow-up care after LCS. One study defined follow-up as undergoing either a diagnostic, invasive procedure or having imaging compared with historical images as a result of a positive screening exam.⁶² The study found lower rates of follow-up in Black patients compared to White patients (82.8 % vs. 89.6%, p<0.05).⁶² Another study found that, among 511 patients who had undergone LCS and had Lung CT Screening Reporting and Data System (Lung-RADS) category 1 or 2 scores (suggesting benign findings), neither race nor insurance status had an impact on the rates of annual adherence (defined as returning for imaging within 1 year + 90 days) or any follow-up—both of which were low overall.⁵⁷ Another study that implemented an LCS program in a low-SES, predominantly Black community reported a 75% follow-up rate among patients with positive baseline LDCT screens but did not analyze differences by race, ethnicity, or SES.⁵⁹

No Surgery

Only one study evaluated the impact of race on surgical treatment in patients with NSCLC detected by LCS. The authors found that Black men were 28% less likely to undergo a surgical procedure to treat NSCLC than White men.⁶³ No studies evaluated the impact of SES on surgical treatment.

Post-LCS Mortality

Of the two studies (10%) that investigated the impact of race on lung cancer mortality among patients at high risk for lung cancer, one evaluated the racial differences in outcomes within the NLST. The study found that Black current smokers had a two-fold higher risk of lung cancer-specific mortality than White smokers. In addition, the study found that screening with LDCT reduced lung cancer mortality in all racial groups but more so in Black individuals (hazard ratio=0.61 vs. 0.86).¹² Similarly, all-cause mortality was 1.35 times higher in Black individuals than in White individuals; however, Black and other non-White individuals had statistically significant reductions in all-cause mortality after screening with LDCT that were not observed in White participants.¹² Su et al. found that screened patients had a lower lung cancer mortality rate than unscreened patients, but they did not find survival differences by race.⁵⁴ High-risk smokers in the NLST with a college education or higher had significantly lower lung cancer-specific mortality than those with a high school degree or less.¹²

Study Quality Assessment

Of the 21 studies, 18 were rated as having a low risk of bias (“good”), and three were rated as having some risk of bias (“fair”). Details on the risk of bias ratings are provided in Figure 3.

Figure 3. Risk of Bias Assessment.

This figure shows the risk of bias determination for each article included in the systematic review.

*Due to the nature of the studies included in the review, the repeated exposure measurement assessment criterion is not applicable and was thus not considered in the quality rating decision.

Discussion

Racial and socioeconomic disparities are prevalent in NSCLC, as manifested by higher incidence and mortality in marginalized groups.^17,18,21 As a screening tool, LDCT has been shown to significantly decrease lung cancer mortality in high-risk patients; however, its universal underutilization limits its impact.^35,36 It is feared that the barriers to LCS utilization may be magnified by racial and socioeconomic inequities, thus further increasing the NSCLC mortality gap. It is therefore critical that we identify and address all potential sources of inequity in the LCS pipeline.

Our systematic review revealed three primary themes related to the steps of the LCS pipeline in which racial and socioeconomic disparities prevail: 1) eligibility; 2) utilization, perception, and utility; and 3) post-screening behavior and care. Achieving equity in lung cancer care and outcomes will require efforts to fix obvious “leaky” steps in this pipeline. It is also essential that we identify and interrogate areas where leakage is possible but there is not yet enough data to confirm or determine how to address the problem.

Eligibility

The eligibility criteria outlined by the 2013 USPSTF guidelines are based on the landmark NLST, in which 89.6% of participants were White and of higher socioeconomic status than the general population.⁷ It is therefore no surprise that these criteria lack the ability to appropriately detect those who would most benefit from LCS among racial minorities and individuals with fewer socioeconomic resources.^40,46,48 Compared to White men, Black men are diagnosed with NSCLC at an earlier age, after smoking fewer cigarettes and with longer quit times .^7,60 Consequently, significantly more Black men with NSCLC will be ineligible for LCS under current guidelines.^40,45,46 There is also some evidence that patients with lower education and lower household income are also less likely to be eligible for screening.^40,48 The upcoming changes to the USPSTF guidelines¹¹ will hopefully help narrow the noted disparities in LCS eligibility; however, there remains a need for focused studies to evaluate the impact and determine the suitability of the criteria for racial minority groups and groups of low socioeconomic position.

Utilization, Perception, and Utility

Unfortunately, even if high-risk Black individuals and individuals of lower socioeconomic position are eligible for LCS, they are still less likely to obtain the screening.^47,51,53 It is possible that a lack of knowledge among patients and their providers may impact these lower rates.⁴⁷ Raz et al. found that 80% of LCS eligible smokers had never heard of LCS and that a significant proportion of smokers expressed shame and stigma as being important barriers to screening.⁶⁴ Whether these barriers are more significant in marginalized communities is understudied. In addition, a lack of awareness of USPSTF guidelines among primary care physicians (PCPs) has been associated with lower utilization of LDCT for screening.⁶⁵ Providers in underserved communities may experience more barriers to adherence to LCS recommendations.

Perceptions about LCS may also impact screening rates; however, Percac-Lima et al. found that Hispanic/Latinx participants had an increased willingness to be screened (i.e., a more positive perception) after they were educated on lung cancer and screening options.⁵⁶ Importantly, Hall et al. reported that patients with lower education were less likely to understand the reason for their referral to LCS,⁵⁵ and this lack of clarity is likely to lead to lower utilization in this group.

We also acknowledge that access is a powerful barrier that limits underserved groups from obtaining screening, as seen in other cancer studies.^66–69 Financial barriers may impact LCS utilization in marginalized groups, as Medicaid does not universally cover LCS. Many individuals of low socioeconomic position are uninsured or underinsured and therefore face financial hurdles that prevent them from obtaining annual LCS. Insured individuals of low SES may also experience barriers related to transportation or the inability to miss work for screening during work hours. Despite these difficulties, Pasquinelli et al reveal that in the setting of Federally Qualified Health Centers (FQHC), which serve the most marginalized and under-resourced communities, LCS can have favorable outcomes. ⁵⁸ Therefore, LCS should be intentionally implemented in FQHCs in order to ensure utilization in groups with the highest risk.

Lastly, the goal of LCS is to diagnose NSCLC at an early stage, when it is curable; however, if individuals from marginalized groups are diagnosed at later stages, as some studies suggest,^43,59 then members of these groups will not experience the full benefit of screening. As this trend would lead to lower utility of LCS among racial minority and lower SES individuals, it is critical to perform studies to further investigate this issue.

Post-Screening Behavior and Care

In general, Black smokers smoke fewer cigarettes than White smokers,⁶⁰ contributing to fewer pack-years. There are also race-based differences in the experiences and behaviors of high-risk patients after receiving LCS. In the NLST, Black individuals had higher rates of short-term smoking cessation than White individuals after LCS, but sustained cessation rates were similarly low among both groups. In the same study, low-income smokers had fewer quit attempts than those with more resources after LCS, suggesting that they benefit less from smoking cessation efforts. Therefore, there remains a need to understand which efforts are most beneficial to the most vulnerable groups.

Rates of post-LCS follow-up care may also differ by race and/or SES. Adherence represents a crucial stage in the LCS process, as failure to adhere can completely undermine the benefit of LCS. One study reported lower rates of follow-up among Black participants,⁶² whereas another study found no race or SES-based differences in follow-up.⁵⁷ Overall, there is a lack of clarity regarding trends in LCS adherence due to a lack of studies exploring this topic. It is therefore paramount that we undertake long-term studies to evaluate this issue and identify barriers to adherence, as well as potential interventions.

In the single study that evaluated the impact of race on surgical resection of stage I disease, Black men were less likely than White men to obtain surgery.⁶³ This disparity undoubtedly contributes to the significantly higher lung cancer-specific mortality rate faced by Black NSCLC patients who are diagnosed by LCS.¹² It is also important to reiterate that Black individuals have a significantly greater reduction in lung cancer-specific and all-cause mortality after screening with LDCT.¹² This underscores the absolute need to ensure that high-risk Black individuals, in particular, obtain this life-saving examination.

Future Recommendations

The articles reviewed here have revealed multiple stages in the LCS pipeline in which marginalized patients may be excluded from the benefits of LCS, gradually worsening the disparities in NSCLC mortality among high-risk patients. Although additional studies are needed to confirm or further elucidate the potential sources of inequity in the LCS pipeline, we can begin to address some of the more obvious disparities now. We propose the following recommendations:

1. Address gaps in eligibility: Increase studies that evaluate the suitability of LCS criteria in underrepresented minority populations and communities of low socioeconomic position.

   Several studies have shown the limits of current USPSTF LCS guidelines in capturing all individuals who could benefit from screening.^{40,44–48,70} Specifically, smoking practices (i.e., pack-years) and age requirements disqualify many high-risk Black individuals, as they typically have fewer pack-years and are diagnosed with lung cancer at younger ages than the current USPSTF pack-year and age criteria.^60,71 As such, lowering the age and pack-year requirements could increase recommendations and referrals for screening among racial minority and low-income populations that are currently missed by current requirements. This approach is already garnering support, and USPSTF guidelines are being updated to include a wider age range (50 to 80 years) and lower pack-year smoking history (20 pack-years).¹¹ However, this update was based on the NELSON study, which was not based on a racially diverse cohort. Furthermore, Black smokers have longer quit times than White smokers,^47,72 and the risk of lung cancer does not drop significantly for quit times >15 years.^73–75 This difference becomes important because former smokers lose eligibility once they have quit smoking for greater than 15 years. This particular disparity is not addressed by the upcoming updates to the USPSTF guidelines. As such, now that RCTs have established that LCS with LDCT leads to mortality-reduction, pragmatic designs and rigorous data modeling are required in order to determine which LCS criteria are most appropriate and will provide the greatest benefit to minority and low SES groups. Biomarker evaluation and the assessment of risk prediction models should also be incorporated into early detection trials. In addition, there is not enough information available to determine if lower education negatively impacts eligibility rates. Studies and modeling focused on individuals with lower education are also required to identify individual-level factors that must be addressed to promote eligibility among high-risk populations.

2. Address gaps in utilization and utility: Implement interventions that increase PCP knowledge and adherence to guidelines and encourage discussion about screening with patients.

   Despite recent efforts, only 16% of eligible patients undergo LCS, therefore, LCS underutilization remains a major problem.³⁶ There are likely several patient-, provider-, and system-level barriers that need to be addressed to reduce underutilization by disadvantaged groups. Efforts to improve provider adherence to LCS recommendations must be prioritized. Raz et al. found that only half of PCPs were familiar with USPSTF LCS recommendations, and only 12% had referred a patient to an LCS program over the prior 12 months. The respondents explained that they had uncertainty regarding the benefits and harms of LCS, as well as concerns regarding insurance coverage.⁷⁶ Thus, to increase utilization by eligible individuals, we must support the PCPs who refer patients to LDCT. Specifically, there is a need to improve providers’ understanding of the LCS guidelines, the benefits of LDCT for high-risk patients, and the harms and costs of LDCT. Once this understanding is achieved, it can be passed on to patients. This lack of awareness is undergirded by the absence of endorsement of the USPSTF LCS recommendations by the American Association of Family Physicians (AAFP), which has supported almost all of the other USPSTF guidelines.⁷⁷ It will be difficult for LCS utilization to increase without more engagement from the primary care community. Additionally, even among primary care physicians who are aware and support the USPSTF guidelines, the requirement set by the Center for Medicare and Medicaid Services of a separate shared informed decision making process prior to performing LDCT screening, is a known barrier to utilization that may be even more of a challenge in already underserved communities where clinicians may have even more time constraints.^78,79

   PCP-focused educational efforts, incentives for complying with USPSTF recommendations, and measurable quality metrics on the screening of high-risk patients may need to be implemented, especially in under-resourced communities where many underserved patients obtain care. In addition, discussions (e.g., shared decision-making) between patients and physicians can provide a space for patients to ask questions and become more comfortable with LCS, ultimately contributing to a greater likelihood of participation. This approach has been successful in the context of colorectal cancer screenings among racial and ethnic minorities.^80–82

   Creating programs and interventions that directly target barriers to LCS can also help increase screening rates. The Community Preventive Services Task Force (CPSTF) endorses multicomponent interventions, which have been shown to increase the utilization of breast, colorectal, and cervical cancer screening services in underserved communities.⁸³ Multicomponent interventions increase demand for and access to cancer screening by combining two or more intervention approach strategies. Interventions relevant to LCS include (1) increasing community demand for screening (e.g., group education or mass media efforts), (2) increasing community access to screening (e.g., addressing transportation barriers, assisting with appointment scheduling, or providing language translation services), and as mentioned above, (3) increasing provider delivery of screening services (e.g., through assessments of and incentives for provider compliance).

   Several community-based interventions to increase LCS uptake have already proven successful^54,57,59 and can serve as models for future interventions to ensure that the benefits of screening reach a larger population, and specifically, those negatively impacted by NSCLC disparities. In addition to targeting underrepresented populations and providing screening opportunities, these interventions should incorporate patient navigation.⁸⁴ Efforts to enhance statewide screening have been successful in overcoming participation of underserved communities by leveraging networks of nurse navigators in states, like Delaware. This cost effective model resulted in reduced late stage diagnosis and mortality and is thus worthy of consideration.⁸⁵ Screening opportunities alone may be ineffective if patients do not have the resources to traverse the unfamiliar LCS environment.

   Lastly, there is a significant gap in knowledge regarding the utility of LCS among Black and low-SES patients. In particular, it is unclear if these groups are more likely to have later-stage disease upon diagnosis from LCS. Thus, there is a need for prospective trials that over-recruit vulnerable communities to better answer this question.

3. Address gaps in follow-up care: Encourage provider discussion of follow-up options, including smoking cessation and surgery, to support patients through their treatment process.

   Lower rates of smoking cessation attempts among low-SES smokers may be mitigated by multilevel interventions such as “Quit Happens,”⁸⁶ an evidence-based smoking cessation intervention for low-income populations that uses a combination of behavioral counseling and pharmacotherapy. The program has proven successful among high-risk populations and includes standardized training for program staff, allowing program adopters to adjust the program to their own needs. The program also includes several points of intervention for smokers and offers support from traditional and behavioral health providers, providing multiple avenues for patients to succeed. This approach ensures that patients have the support to continue their quit attempts if they fall short at any stage. Smoking cessation interventions should be incorporated into annual screening-follow ups of LCS programs, which provide an opportunity to personalize and enhance smoking cessation guidance and can increase quitting rates.⁸⁷

   In addition to comprehensive interventions, we must conduct additional studies to fully understand the impact of race and SES on follow-up after LCS. Current studies have had mixed findings, with some reporting lower rates of follow-up among Black patients,⁶⁰ and others reporting no race or SES-based differences in follow-up.⁵⁷ Because survival benefit is only obtained after three annual LDCT screenings, it is important to rectify any hurdles that may negatively impact follow-up. To better understand where the gaps are in follow-up, long-term prospective studies that include marginalized groups are necessary. We also recommend patient navigation and social support to help underserved patients navigate transportation, work, or family issues that may prevent compliance.

   Although some studies have identified racial disparities in lung cancer patients’ access to surgery, ^63,88 few solutions have been offered to improve rates of surgery among marginalized patients. Efforts to address surgical disparities in other cancers have proposed the use of both multidisciplinary management and medical advocates throughout the course of a patient’s treatment .⁸⁹ These advocates can ensure proper education and help with communication between the clinical staff and the patient. Physicians and surgeons must also acknowledge that their own implicit biases can prevent patients from obtaining important care. The fact that Black patients receive lower rates of surgery “suggests that there is a difference in how physicians manage cancer that is based on a patient’s race, regardless of other attributes, and that the consequence of these lapses in care is reduced survival among blacks.”⁹⁰ Cultural competency training and implicit/explicit bias coursework should be mandatory for every provider in medical school and as a part of continued medical education. In addition, we need a renewed focus on strengthening patient-physician relationships and communication, especially when the patient and provider are from different backgrounds.

Strengths and Limitations

We acknowledge that this study has limitations. The topics and articles included in this systematic review were limited by a paucity of literature available on LCS disparities. As such, although we acknowledge that LCS disparities may exist for other minority populations, given the scarcity of literature on other racial and ethnic minorities, we focused primarily on disparities between Black and White individuals. This study also did not focus on the disparities that women face regarding LCS. In light of the higher incidence of lung cancer in women compared to men,⁹¹ it is remarkable that the NELSON trial included accrual of so few women in its design. While the focus of sex-based disparities in LCS was beyond the scope of this review, the fact that women face higher incidence despite smoking fewer cigarettes⁹² and are diagnosed at a lower age⁹³ supports the need for more inclusive research to understand LCS eligibility, utilization and outcome differences in women. Additionally, due to limited publications on this topic, the majority of our analysis relied on observational and secondary analysis studies. Similarly, we could only conduct a qualitative analysis due to a lack of randomized controlled trials and insufficient rigor and uniformity to perform quantitative meta-analyses.

Despite these limitations, our systematic review also has several strengths. First, to our knowledge, this is the first systematic review to investigate racial and socioeconomic disparities in LCS. Second, we intentionally analyzed and included studies from a wide spectrum of domains at each level of screening (including pre- and post-screening). Third, most articles included in this systematic review were of good quality, as indicated by their low risk of bias. Finally, many of the articles included in this review provide templates for potential interventions and form the basis for potential policy changes to overcome disparities in LCS and, ultimately, NSCLC.

Conclusion

Our systematic review of the literature has revealed that disparities are prevalent in LCS eligibility; utilization, perception, and utility; and post-screening behavior and care. These disparities collectively cause significant NSCLC survival gaps along racial and socioeconomic lines, and greater awareness of their underlying sources is critical if we are to combat them. We recommend efforts to increase prospective trials and rigorous data modeling to evaluate the suitability of LCS eligibility criteria for underserved groups. We also believe that multicomponent interventions that increase community demand and access, as well as provider engagement, will help decrease the disparities that are currently associated with this life-saving tool.