Abstract
Introduction:
Receiving a healthcare provider recommendation to screen is an important predictor for whether individuals at high risk for lung cancer undergo lung cancer screening. Although sociodemographic and socioeconomic characteristics are associated with differential screening participation, it is unknown whether those characteristics are associated with receiving a healthcare provider recommendation for lung cancer screening.
Methods:
This cross-sectional study used Facebook-targeted advertising to recruit a national sample of lung cancer screening-eligible adults (N=515) who completed questionnaires on sociodemographic information (age, gender, race, marital status), socioeconomic characteristics (income, insurance status, education, rurality of residence), smoking status, and receiving a healthcare provider recommendation to screen. Pearson’s chi-square tests and independent samples t-tests evaluated whether sociodemographic, socioeconomic, and smoking-related characteristics were associated significantly with receiving a healthcare provider recommendation to screen.
Results:
Higher household income, having insurance coverage, and being married were associated significantly with receiving a healthcare provider recommendation to screen (all p < .05). Age, gender, race, education, rurality of residence, and smoking status were not associated significantly with receiving a recommendation to screen.
Discussion:
Particular subgroups of individuals at high risk for lung cancer—including those with lower income, without insurance coverage, and who are not married—are less likely to receive a recommendation to screen from their healthcare provider, despite being at high risk for lung cancer and eligible for screening. Future research should test whether differential screening participation and low screening uptake could be addressed by clinician-focused interventions that encourage ubiquitous discussion and recommendation to undergo screening for people at high risk for lung cancer.
Keywords: lung, cancer, screening, healthcare, provider, recommendation
Income, insurance coverage, and marital status predict why some screening-eligible adults receive a recommendation from a healthcare provider to undergo lung cancer screening, whereas others don’t.
1. Introduction
In the U.S., there are an estimated 350 deaths per day from lung cancer—the leading cause of cancer-related mortality—accounting for nearly 25% of all cancer deaths [1]. Among those at high risk for lung cancer, annual screening via low-dose computed tomography (LDCT) of the chest produced a 39% reduction in lung cancer mortality [1], providing empirical support for the lung cancer screening recommendation issued by the U.S. Preventive Services Task Force (USPSTF) in 2013 [2]. However, lung cancer screening rates have remained low among eligible individuals with only a modest increase from 3% to 5% from 2010 to 2018 [1]. Understanding how to promote lung cancer screening among screening-eligible individuals is crucial for facilitating earlier diagnosis and improving outcomes.
The pathway to engage in lung cancer screening differs across countries. In the U.S., it is recommended that healthcare providers—including primary care providers such as physicians and nurse practitioners or specialists such as pulmonologists—use age and smoking history to conduct a risk assessment with their patients to determine whether the individual is eligible for lung cancer screening (i.e., aged 50–80 years old with a 20 pack-year smoking history who is either currently smoking or quit within the past 15 years) [2]. Then, screening-eligible individuals engage in a process of shared decision-making with their healthcare provider to discuss the risks and benefits of screening prior to undergoing LDCT of the chest [2]. It is possible that healthcare providers engage in lung cancer risk assessment as part of a patient’s annual wellness or preventive care visit or the recommendation could result from the patient experiencing symptoms (e.g., shortness of breath) that would bring them to a specialty visit with a pulmonologist. For those without insurance coverage in the U.S., risk assessment and LDCT of the chest is offered at some free clinics, state health departments, and federally qualified health centers, if the individual meets the lung cancer screening eligibility criteria.
Theory and research suggest that one of the most important facilitators of lung cancer screening is receiving a healthcare provider recommendation to screen [3]. Indeed, the majority of screening-eligible patients report a willingness to undergo lung cancer screening if recommended by their provider [4]. However, screening-eligible patients commonly report that no healthcare provider has recommended or discussed lung cancer screening [5].
Sociodemographic and socioeconomic factors such as age, gender, living alone, insurance coverage, and rural residence are associated significantly with differential lung cancer screening participation [6,7]. However, less is known about the characteristics that are associated with receiving a healthcare provider recommendation to screen for lung cancer. It is possible that disparities in lung cancer screening are explained, in part, by differences in whom healthcare providers discuss and recommend screening.
Therefore, the goal of the current cross-sectional study was to test whether sociodemographic characteristics, socioeconomic information, and smoking status were associated with receiving a recommendation to screen for lung cancer from a healthcare provider.
2. Method
2.1. Participants and Procedure
A national sample of participants was recruited using Facebook-targeted advertisement. Using Facebook, recruitment advertisements can “target” potential participants by demographic characteristics and information listed in users’ profiles. For this study, targeting criteria included country of residence (U.S.), language spoken (English), people whose age on Facebook was greater than 55, and those who indicated an “interest” of smoking within their profile. Eligibility criteria reflected the 2013 USPSTF lung cancer screening eligibility criteria: 1) 55 to 80 years old; 2) ≥30 pack-year smoking history; 3) currently smoked or quit smoking within the past 15 years; and 4) never diagnosed with lung cancer. Of note, the USPSTF lung cancer screening eligibility criteria changed in 2021 to reduce the pack-year smoking history to 20 instead of 30; the 30 pack-year smoking history study eligibility criterion aligns with the screening criteria that were current at the time data were collected. Participants provided informed consent and completed questionnaires via a secure online survey. Participants were compensated with a $15 gift card. All procedures were approved by the Institutional Review Board at Indiana University. Additional information on the sample and study have been reported previously [3].
2.2. Measures
Participants reported their age, gender, race, education, annual household income, marital status, insurance status, zip code, and smoking history via self-report. Participants responded to a dichotomous (yes, no) item that assessed whether the participant had received a healthcare provider recommendation to screen for lung cancer. Rurality of residence was coded using the publicly available rural-urban commuting area code that corresponded with the participant’s zip code; scores ranged from 1–10 with higher scores indicating greater rurality, based on the area’s geography and work-related commuting flows [8].
2.3. Analytic Strategy
Descriptive statistics were computed for all variables. Pearson’s chi-square tests and independent samples t-tests were conducted to test whether sociodemographic information (age, gender, race, marital status), socioeconomic characteristics (income, insurance status, educational attainment, rurality of residence), and smoking status were associated significantly with receiving a recommendation for LDCT screening from a healthcare provider. Data were analyzed with Stata Statistical Software (Release 17). Across all study variables, missing data was minimal, ranging between 0–3.7%. Notably, participants with complete data did not differ from participants with missing data on any study variable, indicating that the data was missing at random. As such, we used listwise deletion to address missing data.
3. Results
In total, there were N = 1156 people who clicked the Facebook advertisement. Of those, n = 826 completed the screening survey, n = 598 were eligible for the study, n = 532 enrolled in the study (clicked ‘yes’ to participation after reading through the informed consent), and n = 515 completed the study and comprise the analyzed sample in the current report.
Participants (N = 515) were between 55 and 80 years of age (Mean [M] = 61.41, standard deviation [SD] = 5.41); most were female (n = 334, 64.9%), White/Caucasian (n = 435, 84.5%), currently smoking (n = 324, 62.9%), covered by health insurance (n = 463, 89.9%), and not married or living as married (n = 281, 54.6%) Participants were diverse in educational attainment (n = 39, 7.6% less than high school; n = 157, 30.5% high school graduate; n = 184, 35.7% some college; n = 135, 26.2% college degree or higher) and annual household income (n = 223, 43.3% less than $25K/year; n =175, 34.0% $25–50K/year; n = 117, 22.7% more than $50K/year). On average, participants tended to live in areas that were considered more urban than rural (M = 2.13, SD = 2.34). See Table 1 for a display of these characteristics. Additional sociodemographic and health status characteristics are reported elsewhere [3].
Table 1:
Characteristics of N = 515 Participants
| Characteristic | n (%) |
|---|---|
| Age (mean, SD) | 61.41 (5.41) |
| Rural-Urban Commuting Area Code (mean, SD) | 2.13 (2.34) |
| Gender | |
| Male | 179 (34.8) |
| Female | 334 (64.9) |
| Other | 2 (0.4) |
| Race | |
| Black or African American | 53 (10.3) |
| White or Caucasian | 435 (84.5) |
| Other race | 14 (2.7) |
| Multi | 13 (2.5) |
| Education | |
| Less than high school | 39 (7.6) |
| High school graduate | 157 (30.5) |
| Some college | 184 (35.7) |
| College degree or higher | 135 (26.2) |
| Annual Household Income | |
| <$25,000 | 223 (43.3) |
| $25,000 – $50,000 | 175 (34.0) |
| >$50,000 | 117 (22.7) |
| Married or living as married | |
| Yes | 231 (44.9) |
| No | 281 (54.6) |
| Did not report | 3 (0.6) |
| Covered by health insurance | |
| Yes | 463 (89.9) |
| No | 52 (10.1) |
| Currently smoking | |
| Yes | 324 (62.9) |
| No | 191 (37.1) |
| Received a healthcare provider recommendation for lung cancer screening | |
| Yes | 326 (63.3) |
| No | 189 (36.7) |
Note: The Rural-Urban Commuting Area Code is a designation based on U.S. census information that takes into account population density, urbanization, and daily commuting times to generate a score for how urban vs. rural an area is based on zip code. Scores can range from 1–10 with higher scores indicating greater rurality of residence.
Among this sample of screening-eligible adults, most participants (n = 326, 63.3%) reported never having received a recommendation from their healthcare provider to screen for lung cancer. Higher household income (χ2 = 8.22, p = .016), having insurance coverage (χ2 = 4.62, p = .032), being or living as married (χ2 = 5.49, p = .019) were significantly associated significantly with receiving a healthcare provider recommendation to screen for lung cancer. See Table 2 for summary of results. Age, gender, race, education, rurality of residence, and smoking status were not associated significantly with receiving a healthcare provider recommendation to screen.
Table 2:
Comparison of individuals at high risk for lung cancer who received a healthcare provider recommendation to screen, compared to those who did not receive a recommendation to screen
| Characteristic | Received a recommendation to screen, n(%) | Did not receive a recommendation to screen, n(%) | χ2 (p-value) |
|---|---|---|---|
| Annual Household Income | 8.22 (.016) | ||
| <$25,000 | 76 (40.2) | 147 (45.1) | |
| $25,000 – $50,000 | 57 (30.2) | 118 (36.2) | |
| >$50,000 | 56 (29.6) | 61 (18.7) | |
| Married or living as married | 5.49 (.019) | ||
| Yes | 98 (51.9) | 133 (41.2) | |
| No | 91 (48.2) | 190 (58.8) | |
| Covered by health insurance | 4.62 (.032) | ||
| Yes | 177 (93.7) | 286 (87.7) | |
| No | 12 (6.4) | 40 (12.3) |
Note: The sample consisted of N = 515 adults in the U.S. who were eligible for lung cancer screening (i.e., 55 to 80 years old; ≥30 pack-year smoking history; currently smoked or quit smoking within the past 15 years; and never diagnosed with lung cancer).
We conducted a post-hoc multivariable logistic regression analysis to evaluate the unique predictive utility of income, marital status, and insurance coverage in explaining the variation in receiving a healthcare provider recommendation to screen for lung cancer. Results from this analysis demonstrated that none of the characteristics was associated independently with receiving a healthcare provider recommendation to screen (individual p-values ranged from .103–.542). This could be explained, in part, by the high degree of association between income and marital status (χ2 = 8.22, p < .001).
4. Discussion
This study demonstrated that among a national sample of screening-eligible adults, the majority of participants did not receive a recommendation to screen for lung cancer from their health care provider, consistent with and extending qualitative research [5]. This concerning finding highlights one possible explanation for why uptake of lung cancer screening has remained low since the 2013 release of the USPSTF recommendation [1], particularly given how influential health care provider recommendations are in patients’ decision making process to screen [4]. Furthermore, this study demonstrates that higher household income, having insurance coverage, and being or living as married are associated significantly with receiving a healthcare provider recommendation to screen for lung cancer. This is a novel finding that extends prior research demonstrating that similar sociodemographic and socioeconomic factors are associated with differential participation in lung cancer screening [6,7]. However, it is notable that none of these variables were associated independently with the outcome when analyzed as simultaneous predictors, highlighting the overlapping and co-occurring nature of these characteristics (particularly between income and marital status in the current sample).
It is important to acknowledge that providers must believe in the utility of lung cancer screening in order to raise the topic with their eligible patients. Prior research has found that willingness to recommend lung cancer screening may be influenced by its Grade B recommendation for some providers, suggesting a tension between the recommendation and the provider’s belief in the recommendation [9]. In addition, lack of health care provider recommendation may also stem from their lack of knowledge related to cost and reimbursement. While lung cancer screening is a covered preventive service, both providers and screening-eligible patients may not be aware of this coverage.
Participants who were not married (or living as married) were less likely to receive a screening recommendation by their healthcare provider, consistent with research in colorectal cancer screening [10]. As we consider ways to support providers in primary and secondary prevention efforts, awareness of a potential implicit bias with patients presenting without an advocate may be warranted. In addition, given the risk of incidental findings, false positives or false discovery rates, it is possible that healthcare providers may hesitate to refer patients for screening without knowledge of their social support (which may be more visible among patients who are married). Qualitative interviews may help elucidate screening referral patterns in relation to a provider’s perception of a patient’s social support network. Alternatively, spouses or partners may advocate for screening during a medical appointment by asking questions to the healthcare provider or by sending their partner with questions.
This study has important limitations, including the lack of longitudinal assessment for how recently or frequently participants visit or communicate with their healthcare provider as well as the lack of information about comorbidity status, which should be measured and evaluated in future research to better understand under for whom and under what conditions a healthcare provider provides a recommendation for lung cancer screening. Given the cross-sectional study design and lack of corresponding medical chart information (e.g., comorbidity status), these results should be interpreted with caution and conceptualized as hypothesis-generating (rather than confirming) that guide future research efforts to further understand low uptake of lung cancer screening and related disparities. Additionally, the sample is not representative of the broader population of adults eligible for lung cancer screening in the U.S. Although the sample is diverse with regard to income and educational attainment, the same is predominantly White and comprised of women and those with health insurance coverage, highlighting the need for increased efforts to recruit screening-eligible adults from diverse characteristics and backgrounds that reflect the broader screening-eligible population. Finally, similar research efforts could be strengthened by leveraging multi-method assessment of medical and smoking information (e.g., via medical chart review) in addition to characteristics provided via self-report. This study also has notable strengths, including a large national sample of screening-eligible individuals and the use of rural-urban continuum codes to assess ecological characteristics. Clinician-focused interventions that facilitate shared decision making might be useful for encouraging ubiquitous recommendation of lung cancer screening for eligible individuals.
5. Conclusions
This study identified subgroups of individuals at risk of developing lung cancer—including individuals with lower household incomes, without insurance coverage, and who are not married—who are less likely to receive a recommendation to participate in LDCT screening from their healthcare provider, despite being at high risk for lung cancer and eligible for screening. Receiving a recommendation to screen is just one part of the lung cancer screening process and does not necessarily equate to undergoing lung cancer screening. Despite this, these results highlight differences in receiving screening recommendations based on key demographic variables that merit further research to better understand how they may be relevant for patients’ engagement in the lung cancer screening process. Future research should investigate the extent to which disparities in lung cancer screening participation are explained, in part, by differences in whom healthcare providers discuss and recommend screening.
Highlights.
Not all lung cancer screening-eligible adults receive a recommendation to screen
Income, insurance coverage, and marital status predicted screening recommendations
Disparities in lung cancer screening may be related to providers’ recommendations
Acknowledgements:
The authors thank those who participated in the research.
Funding:
This work was supported in part by grants from the National Cancer Institute (T32CA009461; P30CA008748; R15CA208543; K99CA256351; R00CA256351). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. There are no other financial disclosures.
Footnotes
Declarations of Interest: None
- Timothy J. Williamson: Conceptualization, Formal Analysis, Methodology, Resources, Visualization, Writing—original draft, Writing—review & editing; Leah E. Walsh: Conceptualization, Writing—review & editing; Susan M. Rawl: Conceptualization, Writing—review & editing; Lisa Carter-Bawa: Conceptualization, Funding Acquisition, Investigation, Methodology, Project Administration, Resources, Supervision, Writing—review & editing.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
References
- [1].Siegel RL, Miller KD, Fuchs HE, Jemal A, Cancer Statistics, 2022, CA. Cancer J. Clin 72 (2022) 7–33. 10.3322/caac.21708. [DOI] [PubMed] [Google Scholar]
- [2].Krist AH, Davidson KW, Mangione CM, Barry MJ, Cabana M, Caughey AB, Davis EM, Donahue KE, Doubeni CA, Kubik M, Landefeld CS, Li L, Ogedegbe G, Owens DK, Pbert L, Silverstein M, Stevermer J, Tseng CW, Wong JB, Screening for Lung Cancer: US Preventive Services Task Force Recommendation Statement, JAMA. 325 (2021) 962–970. 10.1001/JAMA.2021.1117. [DOI] [PubMed] [Google Scholar]
- [3].Carter-Harris L, Slaven JE, Monahan PO, Draucker CB, Vode E, Rawl SM, Understanding lung cancer screening behaviour using path analysis, J. Med. Screen (2019). 10.1177/0969141319876961. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [4].Jonnalagadda S, Bergamo C, Lin JJ, Lurslurchachai L, Diefenbach M, Smith C, Nelson JE, Wisnivesky JP, Beliefs and attitudes about lung cancer screening among smokers, Lung Cancer. 77 (2012) 526–531. 10.1016/j.lungcan.2012.05.095. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [5].Simmons VN, Gray JE, Schabath MB, Wilson LE, Quinn GP, High-risk community and primary care providers knowledge about and barriers to low-dose computed topography lung cancer screening, Lung Cancer. 106 (2017) 42–49. 10.1016/j.lungcan.2017.01.012. [DOI] [PubMed] [Google Scholar]
- [6].Hirsch EA, Barón AE, Risendal B, Studts JL, New ML, Malkoski SP, Determinants Associated With Longitudinal Adherence to Annual Lung Cancer Screening: A Retrospective Analysis of Claims Data, J. Am. Coll. Radiol 18 (2021) 1084–1094. 10.1016/J.JACR.2021.03.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [7].Hestbech MS, Siersma V, Dirksen A, Pedersen JH, Brodersen J, Participation bias in a randomised trial of screening for lung cancer, Lung Cancer. 73 (2011) 325–331. 10.1016/J.LUNGCAN.2010.12.018. [DOI] [PubMed] [Google Scholar]
- [8].Hart LG, Larson EH, Lishner DM, Rural Definitions for Health Policy and Research, Am. J. Public Health. 95 (2005) 1149–1155. 10.2105/AJPH.2004.042432. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [9].Lafata JE, Brown RF, Pignone MP, Ratliff S, Shay LA, Primary care physicians’ support of shared decision making for different cancer screening decisions, Journals.Sagepub.Com. 37 (2017) 70–78. 10.1177/0272989X16660547. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [10].Ye J, Xu Z, Aladesanmi O, Provider recommendation for colorectal cancer screening: Examining the role of patients’ socioeconomic status and health insurance, Cancer Epidemiol 33 (2009) 207–211. 10.1016/J.CANEP.2009.07.011. [DOI] [PubMed] [Google Scholar]