Research Priorities for Interventions to Address Health Disparities in Lung Nodule Management: An Official American Thoracic Society Research Statement

Abstract

Background

Lung nodules are common incidental findings, and timely evaluation is critical to ensure diagnosis of localized-stage and potentially curable lung cancers. Rates of guideline-concordant lung nodule evaluation are low, and the risk of delayed evaluation is higher for minoritized groups.

Objectives

To summarize the existing evidence, identify knowledge gaps, and prioritize research questions related to interventions to reduce disparities in lung nodule evaluation.

Methods

A multidisciplinary committee was convened to review the evidence and identify key knowledge gaps in four domains: 1) research methodology, 2) patient-level interventions, 3) clinician-level interventions, and 4) health system–level interventions. A modified Delphi approach was used to identify research priorities.

Results

Key knowledge gaps included 1) a lack of standardized approaches to identify factors associated with lung nodule management disparities, 2) limited data evaluating the role of social determinants of health on disparities in lung nodule management, 3) a lack of certainty regarding the optimal strategy to improve patient–clinician communication and information transmission and/or retention, and 4) a paucity of information on the impact of patient navigators and culturally trained multidisciplinary teams.

Conclusions

This statement outlines a research agenda intended to stimulate high-impact studies of interventions to mitigate disparities in lung nodule evaluation. Research questions were prioritized around the following domains: 1) need for methodologic guidelines for conducting research related to disparities in nodule management, 2) evaluating how social determinants of health influence lung nodule evaluation, 3) studying approaches to improve patient–clinician communication, and 4) evaluating the utility of patient navigators and culturally enriched multidisciplinary teams to reduce disparities.

Contents

• Overview

  • Key Conclusions and Recommendations

• Introduction

• Methods

• Results

  • Considerations Related to Defining Disparities in Lung Nodule Evaluation and Research Methodology

  • Patient-Level Interventions

  • Interventions at the Clinician Level

  • Interventions at the Level of the Health System

• Discussion

Overview

Lung nodules are common findings on diagnostic and screening computed tomography (CT) scans of the chest. National guidelines recommend algorithms to evaluate lung nodules because a subset are malignant. Although guideline-concordant lung nodule evaluation rates are low overall, the risk of suboptimal follow-up is higher for minoritized racial and ethnic groups. This research statement summarizes the existing evidence, identifies research gaps, and reports a formal consensus development process to prioritize future research questions in four domains related to interventions to mitigate health disparities in lung nodule evaluation: 1) considerations regarding research methodology, 2) interventions at the patient level, 3) interventions at the clinician level, and 4) interventions at the health-system level. This statement offers a research agenda to inform investigators and funding agencies on the considerations to generate high-priority, high-quality research on interventions to mitigate inequities in the evaluation of lung nodules.

Key Conclusions and Recommendations

Key knowledge gaps

• •There is no standard approach to conducting studies identifying factors that contribute to disparities in lung nodule management and limited understanding of the underlying mechanisms of these disparities.
• •There are few data evaluating how social determinants of health (SDOH), communication and language barriers, or coexisting medical conditions contribute to disparities in lung nodule management and the optimal approaches to addressing these barriers.
• •Effective strategies to improve patient–clinician communication and address education barriers, limited English proficiency, and barriers related to low health literacy or numeracy to reduce disparities in lung nodule management are unclear.
• •The impact of interventions such as culturally enriched multidisciplinary teams and patient navigators on lung nodule evaluation disparities is unknown.
• •The optimal approach to tracking and arranging the evaluation of incidental lung nodules is unclear, particularly in high-risk settings for lack of follow-up, such as emergency departments.

Prioritized research questions and interventions to reduce disparities in lung nodule management

• 1.What research methods should be used to better document and define disparities in lung nodule management?
• 2.How should we evaluate the effectiveness of appointment-level interventions and patient navigation?
• 3.What is the effect of various communication strategies on information transmission and retention related to lung nodule management and the impact of these interventions on disparities in lung nodule evaluation?
• 4.What is the impact of strategies that allocate resources such as navigators, interpreters, or transportation assistance to individuals at highest risk for not attending lung nodule follow-up visits?
• 5.What are the optimal approaches for tracking lung nodule follow-up, and do these methods address disparities in lung nodule management?

Recommendations

• 1.Study design and analytic strategies should be selected to explicitly state assumptions, ensure adequate control for confounding, improve transparency, and facilitate comparison of results across studies.
• 2.A national research agenda should include studies that better define disparities and evaluate the underlying mechanisms of disparities in lung nodule management, test the effectiveness of appointment-level interventions and patient navigation, assess resource allocation strategies, and evaluate innovative approaches to improving access to subspecialty expertise in resource-limited settings on disparities in lung nodule management.

Introduction

Lung nodules are common findings on diagnostic and screening CT scans of the chest (1). Approximately 5% of individuals with lung nodules will ultimately receive diagnoses of lung cancer (1, 2), the leading cause of cancer death in men and women in the United States (3, 4) and globally (5). Furthermore, lung cancer survival strongly depends on the stage at diagnosis (6), and thus timely evaluation of lung nodules is key to achieving good long-term outcomes (1, 2).

Health disparities in lung nodule follow-up and treatment are a major problem. Black patients have a higher risk of delayed follow-up for lung nodules (7), and Latino/a individuals have lower rates of guideline-concordant follow-up for lung nodules (8). In lung cancer screening, delayed or absent follow-up of screen-detected nodules occurred more frequently in Black individuals (9–12), those with mental health conditions or substance use disorders (9), and those with lower incomes (9). Delayed follow-up of lung nodules, which may represent lung malignancy, can disproportionately affect the health of communities with higher lung cancer rates (13–17) and contribute to known health disparities in lung cancer stage and survival (13, 18). Although it is unclear if disparities in lung cancer similarly extend to evaluation of lung nodules, such care gaps could contribute to the lower rates of surgical intervention or complete resection and the higher lung cancer mortality in individuals who are Black, Latino/a, or of lower socioeconomic status (19–25).

This research statement discusses the methodology and development of a research priority agenda that identifies interventions and implementation strategies to mitigate health inequities in evaluating and managing lung nodules detected on diagnostic and screening CT scans. We used a health equity implementation framework (26) and the cascade of care model (7) to inform our recommendations regarding research priorities.

Methods

This American Thoracic Society (ATS) Assembly on Thoracic Oncology project was approved by the ATS Program Review Subcommittee. The chair (K.S.) and cochair (J.W.) convened a multidisciplinary, international committee of members representing pulmonary medicine, medical oncology, thoracic surgery, primary care, the patient experience, government funding agencies, urban and rural healthcare settings, academic institutions, and the Veterans Health Administration, as well as an unaffiliated clinician and a patient representative (Table 1). Committee members had expertise in health disparities, lung nodule evaluation, tobacco use, and lung cancer screening. Potential conflicts of interest were disclosed and managed per the policies and procedures of the ATS.

Table 1.

Committee Members and Areas of Expertise

Committee Member	Institution	Role	Area of Expertise
Katrina Steiling, M.D., M.Sc.	Boston University Chobanian & Avedisian School of Medicine	Chair, member, clinician workgroup, Writing Committee	Lung cancer screening health disparities, implementation perspective
Juan Wisnivesky, M.D., Dr.P.H.	Icahn School of Medicine at Mount Sinai	Cochair, member, health systems workgroup, Writing Committee	Health disparities, primary care perspective
David E. Ost, M.D., M.P.H.	MD Anderson Cancer Center	Moderator, patient workgroup, Writing Committee	Interventional pulmonary, disparities in lung cancer survival
M. Patricia Rivera, M.D.	University of Rochester Medical Center	Moderator, health systems workgroup, Writing Committee	Disparities in lung cancer screening
Nichole T. Tanner, M.D., M.S.C.R.	Medical University of South Carolina and Ralph H. Johnson VA Medical Center	Moderator, clinician workgroup	Disparities in lung cancer care
Abbie Begnaud, M.D.	University of Minnesota	Member, patient workgroup	Implementation of lung cancer screening in Native American communities, improving access to care in rural communities
Juan C. Celedón, M.D., Dr.P.H.	University of Pittsburgh	Member, health systems workgroup	Role of racial ancestry and genetics in pathogenesis of airway diseases
Marjory Charlot, M.D. M.P.H., M.Sc.	University of North Carolina Lineberger Comprehensive Cancer Center	Member, patient workgroup	Addressing disparities in cancer care, patient engagement, medical oncology perspective
Frank Dietrick, B.A.	Unaffiliated	Member, patient workgroup	Patient perspective
Narjust Duma, M.D.	Dana-Farber Cancer Institute	Member, clinician workgroup	Cancer care equity, medical oncology perspective
Chidiebere Peter Echieh, M.B. B.Ch., F.W.A.C.S.	University of Calabar Teaching Hospital	Member, health systems workgroup, Writing Committee	Thoracic surgery, international perspective
Kwun M. Fong, M.B. B.S., Ph.D.	The Prince Charles Hospital and University of Queensland	Member, health systems workgroup	Lung cancer, international perspective
Jean G. Ford, M.D.	Unaffiliated	Member, clinician workgroup	Cancer screening disparities, minority participation in research
Michael K. Gould, M.D., M.S.	Kaiser Permanente Bernard J. Tyson School of Medicine	Member, health systems workgroup	Pulmonary nodule evaluation, health services research
Fernando Holguin, M.D.	University of Colorado School of Medicine	Member, clinician workgroup	Latino health disparities
Hasmeena Kathuria, M.D.	Boston University Chobanian & Avedisian School of Medicine	Member, patient workgroup, Writing Committee	Tobacco treatment interventions, health disparities, ATS Tobacco Action Committee
Eliseo J. Pérez-Stable, M.D.	NIMHD, NIH	Member, health systems workgroup	Health disparities; tobacco smoking; Latino health care; director, NIMHD
Carey Conley Thomson, M.D., M.P.H.	Mount Auburn Hospital; Havard Medical School	Member, health systems workgroup	Gender disparities in lung cancer, lung nodule evaluation
Renda Soylemez Wiener, M.D., M.P.H.	VA Boston Healthcare System; Boston University Chobanian & Avedisian School of Medicine	Member, health systems workgroup	Lung nodule evaluation, VA Healthcare System perspective

Definition of abbreviations: ATS = American Thoracic Society; NIMHD = National Institute on Minority Health and Health Disparities; VA = Veterans Health Administration.

The chair and cochair developed an overview of current knowledge gaps on disparities in nodule management. Because of the limited data available related to interventions to address this topic, the committee also reviewed studies on interventions in the context of lung cancer diagnosis and treatment, lung cancer screening, and related to the management and screening for other cancers (e.g., breast, colorectal, prostate).

The committee held two virtual meetings on April 19 and May 12, 2021, consisting of presentations and discussions focused on identifying existing evidence related to inequalities in lung nodule evaluation and interventions to address these disparities. A cascade-of-care framework (7) was used to represent a patient’s journey from lung nodule detection to completion of recommended evaluation and to facilitate identification of areas in which care disruption could lead to management delays (Figure 1). This approach led to the identification of three key areas in a patient’s journey from lung nodule detection to completion of recommended evaluation where knowledge gaps existed and potential interventions could be targeted: the patient, clinician, and health-system levels. Virtual breakout sessions aligned with each of these focus areas were held on May 20, May 24, and May 26, 2021, led by designated moderators. Committee members were assigned to and attended specific workgroups (Table 1) but had the option to attend additional workgroups and view recordings of other workgroups asynchronously. Each of the three workgroups, led by a designated moderator (Table 1), discussed knowledge gaps and research questions related to interventions to address health disparities in the focus area. All three workgroups identified knowledge gaps related to research methodology and defining disparities in lung nodule evaluation. All breakout sessions discussed methodology challenges, which were combined into a fourth area of research questions. The chairs and moderators then drafted and refined research questions along four topic areas: 1) research methodology, 2) patient-level factors, 3) clinician-level factors, and 4) health system–level factors (Figure 2).

Figure 1.

Cascade-of-care framework for identifying potential areas for intervention to reduce disparities in lung nodule evaluation. This figure shows each step of a patient’s path from lung nodule detection through completion of recommended follow-up. Delays of care could potentially be introduced at each step. Interventions to mitigate disparities in lung nodule evaluation might target each of these discrete steps at the patient, clinician, and/or system level. SDOH = social determinants of health.

Figure 2.

Modified Delphi process for identifying research priorities. The chairs and designated moderators drafted research questions related to the four key topic areas. Then, research questions in each topic area were distributed to committee members as a survey with a seven-point Likert scale, with modifications during each subsequent round on the basis of the survey results. For the third and final survey round, committee members were asked to score the questions using a seven-point Likert scale and rank the questions in order of importance within each key topic area. Research questions were prioritized on the basis of the average Likert score from the final survey round, with ties broken by average ranking.

Thirty-three questions were initially drafted (Table 2). A modified Delphi process was used for the whole committee to prioritize each question and to achieve consensus through subsequent online surveys (Figure 2). In the first round, committee members were asked to rate research questions from 1 (extremely unimportant) to 7 (extremely important). In round 1, most research questions were rated highly (>80% ranked questions >5, where 1 = extremely unimportant, 2 = unimportant, 3 = somewhat unimportant, 4 = neither important nor unimportant, 5 = somewhat important, 6 = important, and 7 = extremely important). Although round 1 indicated that the committee was in consensus that most of the research questions were at least somewhat important (Table 2), having most questions rated highly hindered the committee’s ability to prioritize the research questions. Thus, we modified the degrees of the Likert scale in round 2 to better delineate the degree of importance (1 = unimportant, 2 = neither important nor unimportant, 3 = maybe important, 4 = somewhat important, 5 = important, 6 = very important, and 7 = extremely important). Questions rated as important, very important, or extremely important were carried forward to the third round of voting. During this round, committee members were asked to rank the research questions for each topic area hierarchically and rate on the basis of importance using the same seven-point scale as in round 1. The final prioritization of the questions was generated on the basis of the average Likert score, with the average rank used to break ties.

Table 2.

Prioritization of Research Questions

	First-Round Likert Score (Mean [SD])	Second-Round Likert Score (Mean [SD])	Third-Round Likert Score (Mean [SD])	Third-Round Average Rank
Defining disparities and research methodology
What are the disparities in lung nodule management?	6.5 (0.7)	6.2 (1.1)	6.6 (0.70)	1.3
What systems need to be in place to identify patient factors that contribute to disparities in lung nodule management?	6.1 (1.0)	5.6 (1.5)	6.0 (1.54)	2.6
How can we best identify, document, and measure the degree of influence of patient (e.g., stigma, unmet social needs, language barriers) and clinician-level factors (e.g., implicit bias, cultural sensitivity) that may affect disparities in lung nodule management?	5.7 (1.6)	5.5 (1.1)	5.9 (1.20)	3.7
What are the optimal approaches to measuring patient-level and clinician-level factors that are associated with disparities in lung nodule management?	6.1 (0.7)	5.6 (1.3)	5.7 (1.16)	3.3
How do patient values affect lung nodule management? How can we account for differences in patient preferences and values when measuring adherence to lung nodule management recommendations?	6.0 (0.8)	5.5 (1.2)	5.7 (0.95)	4.1
What are patient values surrounding lung nodule management, and how much do they vary among individuals and among groups of individuals?	5.6 (0.9)	5.0 (1.4)	Dropped
What are the optimal approaches for measuring the quality of patient–provider discussion and documentation of shared decision making for lung nodule follow-up?	5.4 (1.5)	5.00(1.1)	Dropped
Are the types and extent of disparities observed in lung nodule management similar to or different than those observed in early-stage lung cancer and lung cancer screening?	5.4 (1.3)	4.5 (1.5)	Dropped
Patient-level factors
How do patients’ social determinants of health influence lung nodule follow-up and/or provider interactions?	6.1 (0.8)	6.1 (1.0)	6.3 (0.67)	2.9
What appointment-level interventions (e.g., telehealth, same-day evaluation, mobile imaging units) are most effective in improving disparities in timely lung nodule management? How does this vary across different settings (e.g., urban vs. rural)?	6.5 (0.7)	6.2 (1.1)	6.3 (0.82)*	3.1
Does patient navigation affect disparities in lung nodule management?	6.3 (0.8)	5.9 (1.1)	6.3 (0.82)*	3.4
What are the optimal approaches to addressing communication and language barriers?	6.4 (0.7)	5.2 (1.2)	6.3 (0.82)*	4.1
How do communication and language barriers affect lung nodule management?	6.1 (0.9)	5.6 (0.8)	6.1 (0.88)	3.3
How do patients’ comorbidities influence lung nodule follow-up? How do they contribute to disparities in lung nodule management?	5.8 (0.8)	5.1 (0.8)	5.8 (1.03)	4.2
Does the use of community health workers or community nurse navigators improve patient–provider communication?	6.1 (0.7)	5.1 (1.2)	Dropped
Does patient outreach via text messaging affect disparities in lung nodule evaluation?	5.8 (1.2)	5.1 (1.3)	Dropped
How does immigration and documentation status influence lung nodule follow-up?	5.7 (0.8)	4.8 (1.6)	Dropped
Clinician-level factors
What are the strategies and/or interventions to improve patient–clinician communication (e.g., closed-loop communication), reduce disparities in lung nodule follow-up, and increase the accuracy of information transmission and information retention?	6.6 (0.6)	6.2 (1.0)	6.6 (0.52)	1.6
How does implicit bias affect communication, time spent with patients, patient agreement to lung nodule follow-up, and adherence to lung nodule follow-up?	6.1 (0.8)	5.5 (1.5)	6.0 (1.41)	2.5
What are the clinician characteristics (e.g., time after training, years of education, practice setting, race, and gender) that affect disparities in nodule follow-up? How do gender and perceived race concordance affect patient–clinician interactions (satisfaction, lung nodule follow-up)?	5.8 (0.9)	5.2 (1.3)	5.4 (0.70)	3.0
Do interventions such as implicit bias training and/or culturally sensitive language training improve lung nodule follow-up? Do they improve disparities in lung nodule management? Do they improve patient/clinician interactions?	5.9 (1.5)	5.2 (1.9)	5.4 (1.51)	2.9
How does hospital or clinic setting (e.g., academic vs. non-academic) affect disparities in lung nodule evaluation (e.g., access to diagnostic procedures, diagnostic biomarker testing, expert opinions)?	5.8 (1.0)	5.0 (0.9)	Dropped
System-level factors
What interventions (e.g., culturally enriched multidisciplinary teams, nurse navigators, patient navigators) improve lung nodule evaluation and reduce disparities in lung nodule follow-up?	6.6 (0.6)	5.9 (1.2)	6.3 (0.82)	5.9
What types of resource allocation strategies (e.g., machine learning, formulas, EHR) ensure that resources (e.g., patient navigators, translation services, transportation assistance) are allocated to individuals who have the highest risk of nonadherence to lung nodule follow-up?	5.9 (1.0)	5.8 (0.9)	6.2 (1.03)*	3.6
What is the optimal method to track lung nodule follow-up?	6.1 (1.0)	5.9 (1.4)	6.2 (1.03)*	3.9
How can we track and arrange follow-up for patients with incidental lung nodules found in settings more prone to loss of follow-up (e.g., emergency department, inpatient units)?	6.2 (1.0)	6.2 (0.9)	6.2 (0.79)	4.7
What are interventions (e.g., tele–tumor board), particularly at low-resource centers, that improve access to diagnostic procedures and subspecialty consultation to mitigate disparities in lung nodule management?	6.4 (0.9)	5.9 (1.1)	6.1 (0.74)*	4.9
What communication strategy (e.g., letter, telehealth, in person) is most effective to ensure individuals with low health literacy understand results and the importance of returning for follow-up?	6.5 (0.8)	5.9 (1.0)	6.1 (0.74)*	6.3
What system-level changes (e.g., simplify algorithms for follow-up [similar to LCS], tracking) are most effective in overcoming delays in lung nodule care?	6.5 (0.7)	5.9 (1.2)	6.0 (0.94)	7.1
How do system-level interventions such as after-visit summaries and automated reminders affect disparities in lung nodule follow-up?	5.6 (0.8)	5.6 (1.2)	5.9 (0.99)	4.4
What are the optimal approaches to measuring and quantifying cost-effectiveness and added value for infrastructure and personnel for improving disparities in lung nodule management?	5.5 (1.0)	5.0 (1.4)	5.5 (0.85)	6.4
How can we best develop educational content that is understandable for patients to improve lung nodule follow-up?	5.9 (1.0)	5.4 (1.1)	5.6 (0.84)	7.8
Who (e.g., radiologist, technician, patient navigator, clinician) should discuss nodule results with socially disadvantaged individuals?	5.5 (0.9)	5.2 (1.3)	Dropped

Definition of abbreviations: EHR = electronic health record; LCS = lung cancer screening.

Round 1 rating scale: 1 = extremely unimportant, 2 = unimportant, 3 = somewhat unimportant, 4 = neither unimportant nor important, 5 = somewhat important, 6 = important, 7 = extremely important.

Round 2 rating scale: 1 = unimportant, 2 = neither unimportant nor important, 3 = maybe important, 4 = somewhat important, 5 = important, 6 = very important, 7 = extremely important.

Round 3 rating scale: 1 = extremely unimportant, 2 = unimportant, 3 = somewhat unimportant, 4 = neither unimportant nor important, 5 = somewhat important, 6 = important, 7 = extremely important.

^*Ties were broken by using the SD of the Likert scale rating. In the case of ties of both the average Likert score and its SD, the average ranking of the question was used to break the tie. For the patient domain (six questions in the final voting round, where the question ranked first was rated most important), questions related to appointment-level interventions were ranked 3.1, patient navigation 3.4, and communication or language barriers 4.1. For the health system domain (11 questions in the final voting round, where the question ranked first was rated most important), questions related to resource allocation strategies were ranked 3.6, methods for tracking follow-up 3.9, interventions for low-resource settings 4.9, and communication strategies for individuals with low literacy 6.3.

The chair and cochair drafted the initial version of the manuscript with assistance from a writing committee (C.P.E., H.K., D.E.O., and M.P.R.). The manuscript was then circulated to the full committee and iteratively revised. The ATS Board of Directors approved the final document.

Results

We divide this section into four key topic areas for which research questions were identified: 1) considerations related to defining disparities in lung nodule evaluation and research methodology to study this problem, 2) interventions that act at the patient level, 3) interventions that act at the clinician level, and 4) interventions that act at the health-system level. In each of these four areas, we first summarize the evidence and knowledge gaps identified during the virtual breakout sessions and literature review. We then highlight our findings from the modified Delphi process used to prioritize the research questions.

Considerations Related to Defining Disparities in Lung Nodule Evaluation and Research Methodology

Summary of evidence and knowledge gaps related to defining disparities in lung nodule evaluation

Identifying the specific disparities and their underlying contributors could better inform the study of interventions to address these gaps in care. Although there are several studies describing disparities in lung cancer screening, lung cancer care, and lung cancer mortality, there are no current studies that address whether the same disparities exist in the evaluation of pulmonary nodules. For example, compared with non-Hispanic White individuals, minoritized groups are disproportionately affected by lung cancer and have worse outcomes (13, 14), and Black individuals experience higher lung cancer rates than White individuals for equivalent cigarette smoke exposure (15–17). Similar disparities in lung cancer rates and survival exist for patients in rural locations (27, 28). In addition, minoritized patient groups, including Black, American Indian and Alaskan Native individuals, tend to receive diagnoses of more advanced lung cancer, are less likely to undergo guideline-concordant treatment, and experience worse long-term survival (13, 18). Black and Latino men are less likely to undergo any surgical intervention or complete resection for lung cancer than White men (19–21), raising the question of whether similar disparities exist in invasive diagnostic evaluations for suspicious lung nodules.

Although less studied in the published literature, disparities in lung nodule evaluation and lung cancer care may also affect other groups historically discriminated against because of age (29), ethnicity (19–25), disability status (30), sexual orientation and gender minority status (31–34), HIV status (35), rural location (36), and severe mental illness (37–40). For example, transgender individuals are less likely to undergo lung nodule evaluation (34). Adults with severe mental illness such as schizophrenia are more likely to smoke cigarettes (37, 38), less likely to quit smoking (39), and experience higher lung cancer mortality associated with delays in diagnosis and treatment (40). Women are more likely to undergo limited resection for localized-stage lung cancer than White men (19). Understanding the causes of such disparities could inform interventions to improve lung nodule management.

Furthermore, the underlying mediators of disparities in lung nodule evaluation are unclear. For example, disparities in lung cancer screening, lung cancer care, or lung cancer mortality have been reported for historically discriminated groups other than racial and ethnic minoritized populations, including women (19, 29), transgender individuals (34), older adults, and adults with severe mental illnesses (41–43). Some of these inequalities may be explained by disparities in lung nodule evaluation. However, little is known about the causes of these disparities, making it challenging to determine which interventions might best support the clinical care of these patients.

Summary of evidence and knowledge gaps related to research methodology for studying disparities in lung nodule evaluation

All three workgroups identified and discussed factors related to research methodology that are critical to ensuring accurate, valid, and generalizable results from studies evaluating specific approaches to improve disparities in lung nodule evaluation. Although the existing literature does not recommend specific methods for designing studies to identify patient-, clinician-, or system-level factors contributing to lung nodule management disparities, there was consensus among the committee about the importance of defining these factors, determining the optimal research approaches to document them, and evaluating the best methods for measuring the influence of these factors on disparities in lung nodule evaluation.

The committee reviewed the importance of identifying direct and indirect mediators of health disparities related to lung nodule evaluation. Causal inference is particularly challenging in this setting given that randomization cannot be used except when evaluating interventions to mitigate disparities in care. However, observational data may be prone to confounding and other biases (44), particularly without sufficiently large sample sizes, prospective data collection with standardized measures, and adequate retention. Therefore, rigorous methodology to study health disparities is warranted, including consistently specifying the underlying causal models being tested, precisely stating the study hypotheses, and more accurately describing the population being investigated. Advances in scientific, statistical, and epidemiologic inference (44), such as causal models with directed acyclic graphs or cluster randomization, may be of particular utility to researchers. Use of these approaches in studying health disparities might be particularly advantageous by enabling the identification of key mediators of the disparities and enabling targeted interventions. This approach may also facilitate a more judicious allocation of limited healthcare resources to interventions likely to have a large impact.

The committee also discussed the information infrastructure required to carry out high-quality national disparities research. There was a consensus that the existing information infrastructure is insufficient, most notably because of fractured and inaccessible data across systems, making larger scale studies difficult. Improvements to the information infrastructure, including electronic health records (EHRs) and standardized measures of SDOH and other factors, will be required if more meaningful research on health disparities related to lung nodule management is conducted.

Finally, the committee discussed the potential impact of patient values and preferences on assessing lung nodule management disparities and effective communication about risks and benefits. Although there was a consensus that it would be important to account for patient preferences and cultural values, the optimal approach to achieve this is unclear. Thus, additional studies are needed in this area.

Prioritization of research questions

Voting results indicated that committee members highly ranked questions on defining and documenting specific disparities in lung nodule evaluation, including studies evaluating whether the disparities noted in lung cancer screening, diagnosis, and treatment outcomes similarly pertain to evaluation of lung nodules (average Likert score 6.6) (Table 2). Committee members also highly ranked questions related to identifying the strategies necessary to document and measure factors contributing to disparities in lung nodule management (average Likert score 6.0) (Table 2).

Patient-Level Interventions

Summary of the evidence and knowledge gaps

Various factors at the patient level may influence the completion of lung nodule follow-up testing, including SDOH (45–47), cultural beliefs and attitudes (48), health literacy and numeracy (49, 50), English language proficiency (51, 52), ease of accessing and affording medical care and follow-up imaging (53–55), and ability to navigate the health system (56–61). However, very few studies have examined whether interventions to address patient-level factors improve disparities in lung nodule evaluation.

SDOH include the socioeconomic factors that influence health (such as household income, educational attainment, and employment), other individual determinants including demographics, and structural determinants such as the physical environment, access to health care, and economic opportunity. Studies of SDOH have shown that residence in disadvantaged neighborhoods is associated with higher incidence of lung cancer (45) and that public insurance or uninsured status and lower income are linked to lower rates of resection (including use of minimally invasive surgery), increased risk for postoperative mortality, and lower long-term survival in localized-stage lung cancer (46). Moreover, structural racism may underlie disparities in lung cancer screening. Black race was associated with lower rates of low-dose CT screening in neighborhoods where Black residents were historically denied home loans (redlining) (47). Additional studies are needed to evaluate how SDOH contribute to lung nodule evaluation disparities and which interventions could improve outcomes.

Health literacy is also an important determinant of healthcare access and guideline-concordant care. Many minoritized patient groups have low literacy and numeracy, hindering their ability to assess the risk of lung cancer for a newly detected lung nodule or the potential risk/benefit ratio of the recommended workup. Prior research has demonstrated that language barriers are associated with difficulty navigating health systems and disparities in care (49, 50). However, there is a paucity of studies evaluating whether language and communication barriers contribute to disparities in lung nodule evaluation and which interventions may effectively address these factors. In the context of smoking cessation, culturally tailored counseling programs for patients with limited English proficiency can increase abstinence rates (51, 52). Although there is limited research specific to disparities in the evaluation of lung nodules, limited English proficiency, stigma related to lack of English proficiency, or inadequate cultural sensitivity may explain some of the gaps in care for Latino/a patients and other minoritized populations, such as recent immigrants.

Racial and ethnic health disparities in several chronic conditions are well established, including obesity (62), diabetes (44), hypertension (63), cardiovascular disease (64), and chronic kidney disease (65), and these factors may worsen disparities in lung nodule evaluation. Although the mechanisms underlying these associations are unclear, potential mediators include prioritization of the discussion of lung nodule evaluation in the context of competing medical priorities, frequency of interaction with the healthcare system, the impact of comorbidities on survival that may influence the appropriateness of ongoing lung nodule evaluation, and potential limitations in patient functional status related to comorbidities that make it more difficult to attend lung nodule follow-up appointments (66). Indeed, comorbidities may contribute to endometrial and breast cancer mortality (67, 68), and higher comorbidity is associated with lower rates of guideline-concordant care for breast cancer. Moreover, patients with multiple comorbidities are less likely to be included in clinical trials (69). Patients with severe mental illness including schizophrenia more likely to experience delays in diagnosis and treatment for lung cancer, and individuals with HIV infection are less frequently treated for lung cancer than HIV-uninfected counterparts (70). Older individuals are less likely to receive guideline-concordant treatment for lung cancer (71), raising the possibility that comorbidities may mediate this association given the higher occurrence of comorbidities in these patient groups (72). People at increased risk for lung cancer have higher tobacco-related comorbidities, including chronic obstructive pulmonary disease (72), highlighting the importance of accounting for comorbidities when evaluating disparities in lung nodule evaluation. In particular, researchers should consider accounting for comorbidities that significantly affect survival, affect functional status, or are related to cigarette smoking and may influence the risk of lung cancer.

Patient navigation, defined as the use of trained personnel to guide patients through their medical care, could help address patient-level factors contributing to disparities in lung nodule management. Navigators or community health workers familiar with cultural values of minoritized groups and community resources can bridge communication gaps and assist patients through complex health systems (56). Patient navigation in urban, minoritized, and low-income populations has been shown to improve mammography completion for breast cancer screening, follow-up for abnormal mammography results (57–59, 73), and breast cancer survival (57). Among patients undergoing breast, cervical, colorectal, or prostate cancer screening, patient navigation improves time to diagnostic resolution and treatment initiation (60, 74), with the most significant benefit at centers with greater delays in follow-up (60). For incidental lung nodules, an administrative navigator within the health system who contacted clinicians of patients flagged as lacking imaging follow-up improved the completion of lung nodule surveillance (61). These findings warrant further study, as they suggest that patient navigation may mitigate disparities in lung nodule evaluation by improving follow-up for patients at the highest risk for delays.

Appointment-level interventions can be targeted to improve healthcare access or completion of recommended care. Interventions such as telehealth and e-consults to assess lung nodule risk offer opportunities to address geographic barriers to lung nodule evaluation but have not been systematically studied. Telehealth has emerged as a pragmatic approach to providing quality health services to diverse populations (75) and may have particular utility for rural and geographically isolated populations (76). However, telemedicine alternatives have not been studied in the context of lung nodule evaluation. Targeted telephone navigation for patients at highest risk for appointment no-shows effectively improved attendance to cancer care clinic appointments (77). However, minoritized groups, low-income populations, and older adults may have limited access to digital communication technologies, generating concerns that appointment-level interventions dependent on these systems may aggravate care disparities. In addition, these interventions, including the optimal intensity and method of reminders, have not been systematically evaluated in the context of lung nodule evaluation.

Appointment reminders are a pragmatic intervention used in cancer screening but have not been systematically studied in lung nodule evaluation. However, studies of tailored mammography reminders to patients (including written letters and phone calls) have yielded conflicting results for breast cancer screening of Black women in the United States (78, 79). In Australia’s culturally and linguistically diverse patient groups, translated and culturally tailored phone reminders increased screening mammography scheduling rates, whereas written letters did not (80). Given that language and health literacy barriers are associated with care gaps (50), additional studies are needed to determine whether summaries tailored to patients with low literacy or limited English proficiency are more effective.

Prioritization of research questions

Committee members scored highly on questions defining how SDOH influence disparities in lung nodule evaluation and evaluating the impact of appointment-level interventions, patient navigation, and strategies to address communication and language barriers. Questions related to defining how SDOH affect disparities in lung nodule evaluation were ranked most highly (average Likert score 6.3, average rank 2.9). This was followed by questions related to the effectiveness of appointment-level interventions and evaluation of how their effectiveness varies across different practice settings (average Likert score 6.3, average rank 3.1), questions related to the impact of patient navigation (average Likert score 6.3, average rank 3.4), and questions related to the optimal approach to addressing communication and language barriers (average Likert score 6.3, average rank 4.1) (Table 2). The committee noted that without first defining the impact of comorbidities on disparities in lung nodule evaluation, there would be a high probability of confounding, as the comorbidities would be associated with the exposure and outcome of interest.

Interventions at the Clinician Level

Summary of the evidence and knowledge gaps

Clinician recommendations to patients are a critical intermediate step to increasing appropriate follow-up for lung nodules. Yet there has been little research on clinician-level factors and even fewer studies on interventions to improve lung nodule evaluation. In addition, several factors may influence clinical recommendations for lung nodule follow-up, including patient–clinician communication, clinician demographic characteristics, and other factors such as years of training, implicit biases, and cultural stereotypes.

Patient–clinician communication may contribute to disparities in managing nodules, as these interactions influence patients’ understanding of the recommended testing and the importance of follow-up. However, there is a lack of research examining this question, given that communication is not well reflected in the EHR and that patients may not accurately recall and/or report their clinical interactions. In addition, it is challenging and costly to record, transcribe, and code patient–clinician interactions to directly analyze the quality of communication. Prior studies have shown that patient–clinician communication influences patients’ perceptions of lung nodules and related distress (81). A survey study showed that patients with lung cancer who perceived communication with their clinicians as warm and friendly and reported receiving high-quality information about lung cancer management were more likely to undergo treatment (82). In this study, other communication domains, such as physician support and patient symptoms or needs, were not independent predictors of treatment (82). Oral communication rather than communication by letter was preferred by patients in a qualitative study evaluating communication about lung cancer screening results (83). However, communication preferences were not stratified by race or ethnicity. Specific strategies to reduce disparities by improving patient–clinician communication, such as closed-loop and written versus oral communication, have not been examined in lung nodule management.

Racial or ethnic concordance between patients and their healthcare providers is associated with higher patient experience scores (84). It is unclear whether and how clinician characteristics such as gender, race, ethnicity, years of education, specialty, years since completion of training, or practice setting affect disparities in lung nodule evaluation. The committee acknowledged that practice setting might affect access to diagnostic procedures, diagnostic biomarker testing, and expert review and input regarding management.

Little is known about the role of implicit bias or stereotypes in disparities in lung nodule evaluation. In lung cancer screening, women may be less likely than men to have discussed lung cancer screening with their clinicians (85), suggesting an implicit gender bias. In mammography screening for breast cancer, in which abnormal imaging results are common, and false-positive results are common, perceived discrimination was associated with significant diagnostic delays (86). Studies on patients with advanced cancer showed that higher degrees of oncologist implicit racial bias were associated with shorter patient interactions, less patient-centered and supportive communication, and more patient difficulty remembering the contents of the interaction (87). In primary care, physicians’ implicit racial bias was associated with more frequent use of first-person pronouns, which may suggest a sense of social dominance, and anxiety-related words such as “worry” and “nervous,” which highlights the potential role of anxiety in racially discordant interactions (88). Studies of the impact of implicit bias and linguistic patterns, reducing implicit bias, or implementing cultural sensitivity training on reducing disparities in lung nodule evaluation are needed.

Prioritization of research questions

Committee members ranked questions on patient–clinician communication and the role of implicit bias very highly. Questions related to interventions to improve patient–clinician communication to reduce disparities in lung nodule evaluation and increase the quality of information transmission and retention were highly prioritized with slight variation (average Likert score 6.6). Questions on how implicit bias may affect communication, time spent with patients, and completion of lung nodule follow-up were ranked second on average (average Likert score 6.0). The committee noted that concerns about the feasibility of investigating implicit bias and its effect on disparities in lung nodule evaluation and patient outcomes were a factor in the rankings (Table 2). The committee also acknowledged that interventions targeting clinician-level factors would require concomitant delivery with system-focused interventions such as EHR reminders and inclusion of non-physician clinical care team members (e.g., nurse coordinators, patient service representatives, medical assistants) that provide initial face-to-face contact with patients.

Interventions at the Level of the Health System

Summary of the evidence and knowledge gaps

A mismatch between patient needs and health system resources may contribute to disparities in lung nodule management. This section summarizes knowledge gaps and evidence for interventions that address access to health care, information technology resources, and programmatic structure that may affect disparities in lung nodule evaluation.

Access to health care

Lack of adequate health insurance affects people of all racial and ethnic backgrounds and individuals living in rural areas (53). Before the Patient Protection and Affordable Care Act was implemented, half the U.S. population could not afford recommended care, including cancer screening (53). Moreover, even relatively small copayments decreased screening mammography rates (54). Inequities in resource availability and allocation within health systems primarily serving non-Hispanic White patients may further exacerbate disparities in care (55). These inequities may be worsened during health emergencies constraining resources such as the coronavirus disease (COVID-19) pandemic (89).

Patients evaluated in the emergency department may or may not have primary care physicians documented for communication of incidental findings such as lung nodules. Although there are no studies evaluating lung nodule follow-up rates stratified by whether the patient has a primary care provider (PCP) documented in the EHR, many radiology notification systems and interventions rely on communicating the result with a patient’s PCP or ordering clinician (90). However, one in four patients in the United States does not have a PCP, and minoritized groups and patients without insurance are less likely to have a regular source of care (91). In addition, new patients in a health system (e.g., those undergoing CT scans in emergency departments) may not have PCPs listed in the EHR (90). These factors may further exacerbate disparities in lung nodule evaluation.

Factors such as residential segregation, geographic distance to healthcare facilities, and access to transportation also contribute to health disparities (76, 92). There are significant knowledge gaps in this area related to lung nodule management. In the context of lung cancer screening, access to programs and follow-up care has significant geographic variability. Rural residents are less likely to have access to a lung cancer screening center within 30 miles or a 30-minute drive (93). Geospatial cluster analysis showed that the distribution of comprehensive lung cancer screening programs was lowest in the southeastern United States, with the highest lung cancer burden (28). Even in urban areas, where screening programs are more tightly clustered, individuals of low socioeconomic status are less likely to undergo screening despite a higher risk of lung cancer (94). Disparities between the number of individuals at increased risk for lung cancer and the availability of accredited lung cancer screening programs in geographic regions suggest similar disparities for lung nodule evaluation and highlight the risk that access to subspecialty care and specialized diagnostic procedures, such as transthoracic needle biopsy or bronchoscopy, may be similarly limited.

Various approaches may reduce disparities in lung nodule management because of geographic barriers and access to care. For example, telehealth and e-consults can bring quality health services to rural and other underresourced clinical settings, including remote interpretation of CT scans by qualified radiologists to help maintain lung nodule reporting consistency. Multidisciplinary telehealth consultative services might also broaden evaluation options for geographically isolated patients and facilitate shared decision making in lung cancer screening (95). Another approach to reducing geographic barriers to obtaining follow-up imaging is mobile CT units (76). For example, the Manchester Lung Health Check program provided lung cancer screening to low-income communities, achieving high uptake and 90% adherence to annual follow-up among eligible individuals (96). In addition, 75% of individuals screened through this program indicated that location was essential to their choice to be screened, and 23% reported that they would be less likely to seek out screening in a hospital-based program (96).

Information technology resources

In the context of lung cancer screening, appropriate tracking and management of abnormal low-dose CT scan findings, including lung nodules, has been identified as a significant challenge at federally qualified health centers (97). Investment in information technology resources, such as lung nodule tracking software, can facilitate abnormal result follow-up (98). Other approaches to improving incidental lung nodule tracking and follow-up have included multidisciplinary teams tasked with reviewing CT examinations and notifying the ordering clinician and the patient (61), automated processing (e.g., natural language processing) of free-text CT reports (90), computerized registries (99), and use of standardized templates for reporting lung nodule results and recommendations (100). However, the optimal strategies to implement technologies to improve management of lung nodules are unclear. In addition, whether improved tracking reduces disparities in lung nodule evaluation is unknown.

There has been little research on resource allocation strategies for lung nodule management. One approach would entail measuring and quantifying the cost-effectiveness and value-added for infrastructure and personnel. Using machine learning or the EHR, a complementary approach might use automated or semiautomated resource allocation strategies to identify individuals at highest risk for loss to follow-up. For example, one study evaluating factors predictive of lung nodule follow-up demonstrated that a combination of physician–patient communication variables (e.g., whether and how the finding was communicated), patient variables (e.g., demographics), and radiology report variables (e.g., nodule size, whether follow-up recommendations were given) were significant predictors of patient follow-up for lung nodule evaluation (101). This observation suggests that data analytic approaches can optimize lung nodule follow-up and have potential utility for allocating resources to patients at highest risk for delayed workup. Alternatively, system-level interventions such as the implementation of physician reminders, direct outreach to patients in need of follow-up, and audit feedback might improve disparities in lung nodule management. This approach has effectively improved breast, cervical, and colorectal cancer screening (102) and reduced racial disparities in lung cancer treatment (103).

The panel reached a consensus about the importance of literacy matched educational content and communication about lung nodules found on CT and their required follow-up in reducing disparities in lung nodule management. However, the optimal approach to achieving this important goal is unknown. For example, it is unclear whether system-level interventions such as written after-visit summaries and automated reminders improve lung nodule management in patients with low literacy. In addition, compared with in-person notifications, phone calls or letters about mammography results can lead to diagnostic delays and an incomplete understanding of mammography results (86). It is unclear whether the method of communication of lung nodule follow-up recommendations affects disparities in care.

Program structure

The role of centralized lung nodule management protocols, as opposed to decentralized approaches that rely solely on the ordering or primary care clinician, may be of particular relevance to patients who receive diagnoses of incidental pulmonary nodules in the context of emergency department visits. Studies have shown greater adherence to annual lung cancer screening in centralized screening programs, and improved disparities in follow-up rates for Black individuals (10). The impact of centralized versus decentralized lung nodule follow-up programs on disparities in lung nodule evaluation is not known.

Variability in radiology reporting of CT findings, including vague language (104, 105) and lack of inclusion of the finding in the final impression of the report, has been associated with decreased completion of lung nodule follow-up recommendations (106). Prior studies have shown that follow-up recommendations can vary if an individual’s risk factors for lung cancer are unknown to the radiologist (90). The complexity of Fleischner Society guidelines for incidental lung nodule management (107), which incorporate multiple clinical and radiographic risks, may be an additional barrier to lung nodule tracking and standardization of recommendations. A standardized lexicon may decrease variability in radiology reports, but it is unclear if this would reduce disparities in lung nodule management.

Prioritization of research questions

Questions related to specific interventions, such as patient navigators or culturally trained multidisciplinary teams, to improve lung nodule evaluation were ranked most highly (average Likert score 6.3). This was followed by questions related to resource allocation strategies and methods for tracking lung nodule follow-up (average Likert score 6.2); questions about the use of machine learning, algorithms, and the EHR to allocate resources to patients at highest risk for lack of or delayed follow-up to lung nodule follow-up (average Likert score 6.2); and questions on the management of patients with incidental lung nodules detected during emergency department or inpatient visits (average Likert score 6.2) (Table 2).

Discussion

Our committee identified significant gaps in knowledge regarding interventions to address disparities in lung nodule evaluation. There is currently no standardized approach to identifying factors associated with lung nodule management disparities and a limited understanding of the factors that mediate these associations. Furthermore, there are few data evaluating the role of SDOH, limited health literacy and numeracy, or coexisting medical conditions on disparities in lung nodule management. The optimal approaches to addressing these barriers also remain unclear. The optimal strategy to improve patient–clinician communication and information transmission and/or retention is unknown, as is the impact of culturally enriched multidisciplinary teams.

Using a formal consensus development process, the committee identified several areas in which further research would significantly affect decision making regarding resources and strategies to address disparities in lung nodule management. Regarding research methodology, committee members gave highest ratings to questions better defining the disparities in lung nodule management. Within the patient domain, research questions related to how SDOH influence lung nodule evaluation and affect clinician interactions were rated most highly. For the clinician domain, the need for studies to evaluate strategies to improve patient–clinician communication to reduce disparities in lung nodule evaluation was ranked highest. Finally, in the health systems domain, committee members indicated that research on the effectiveness of specific interventions such as patient navigators and culturally trained multidisciplinary teams in reducing disparities in lung nodule management was the highest priority.

This research statement establishes a framework for guiding future studies to address knowledge gaps in lung nodule evaluation and a roadmap for governmental and nongovernmental funding agencies to generate high-priority and high-quality research related to interventions addressing disparities in lung nodule management.