Lung cancer screening in rural primary care practices in Colorado: time for a more team-based approach?

Lung cancer is the leading cause of cancer-related death in both men and women in the United States (U.S.), averaging around 150,000 deaths per year [1‐3]. The American Cancer Society estimated the number of new cases in 2020 was 228,820 and estimated the number of deaths at 135,720 [4]. Rural areas are disproportionately affected by smoking and incidences of lung cancer in comparison to urban areas [5]. Americans living in rural areas are more likely to die from lung cancer than their urban counterparts with a 20% higher lung cancer mortality rate [6]. The U.S. Preventive Services Task Force (USPSTF) recommends patients ages 50–80 with a 20 pack-years smoking history who currently smoke or quit within the past 15 years receive low-dose computed tomography (LDCT) lung cancer screening (LCS) annually [7, 8].

In the United States, primary care practices have not widely implemented these guidelines, especially in comparison to other evidence-based screening guidelines for other cancers [9, 10]. Of the LCS eligible Americans, only 5% to 6% have been screened as of 2020 [11, 12]. While utilization varies significantly across states, people who are uninsured are less likely to undergo LCS [13]. LCS is higher among those with chronic respiratory conditions, who were divorced, separated, or widowed, who had previous cancer diagnoses, and aged 65 to 74 [13].

Various barriers to LCS exist for rural patients, clinicians, and health care facilities [14]. Rural patients who are most in need of LCS tend to have lower levels of formal education, are less trusting of doctors and health care, have inadequate insurance coverage, and face geographic access barriers [15, 16]. Rural patients also are more likely to have limited access to primary care physicians who address the LCS recommendation and referral, and to specialty care [3, 10]. LCS via LDCT in the U.S. is a covered test by Medicare and some Medicaid programs (CMS); however, Medicaid in many states does not cover it, and some state Medicaid programs do not cover it as well as some private insurances in the U.S [17]. Around 50% of those eligible for LDCT per the recommendations of USPSTF are uninsured or Medicaid insured [18, 19].

Compared with clinicians of other specialties, primary care clinicians serve an important role of ensuring appropriate screenings are recommended to eligible patients [20]. Studies of primary care clinicians have reported barriers including lack of knowledge of the current guidelines (e.g., incorrectly ordering chest X-rays over LDCT scans), implicit biases based on sex, race, ethnicity, and smoking history that hinder recommendations for LCS to patients, and lack of time, shared decision-making tools, or ability to facilitate an effective conversation on LCS [10, 21]. Disparities in access to both information and screening facilities is more severe for rural than urban patients [10].

Despite all that is known about issues related to LCS, we still need to understand more about how perspectives of different key players (e.g., physicians, other staff, patients) may influence different aspects of implementation as it relates to the context of rural primary care. With many processes in primary care, especially around cancer screenings, clinical and administrative staff play a key role in identifying eligible patients, supporting the clinician in shared decision making about the screening choice, and coordinating screening activities with providers of services and reimbursement with payers. Thus, their perspective is an important one to capture. Likewise, understanding the patient’s perspective adds insight on why patients may or may not agree to the LDCT or follow through with test completion.

We have previously used the RE-AIM framework and more recently its contextual expansion to PRISM in our implementation research [22] and determined that it might provide unique insights into different aspects of implementation, especially across the various roles of important members in the system. RE-AIM is an acronym for reach and effectiveness at the patient level, and adoption, implementation and maintenance at the setting level [23, 24]. Examining multiple key players’ views through a RE-AIM lens might reveal more about how LCS eligibility and referral for services can be conducted in rural primary care. Thus, the purpose of this paper is to understand issues related to performance of LCS activities in rural primary care practices in Colorado. We assessed the conduct of various aspects of LCS including patient identification, tracking, shared decision making, use of decision aids, smoking cessation advice and counseling, referral and follow-up from the perspectives of clinicians, practice staff, and patients and using the RE-AIM model to elucidate these issues in the hopes of overcoming potential barriers to LCS implementation in practice.

We conducted this qualitative study as part of the Colorado Implementation Science Center for Cancer Control (https://​coisc3.​org); the study was approved by the Colorado Multiple Institutional Review Board (COMIRB) for research with human subjects (COMIRB #: 19–1706; date: April 27, 2020). We used methods in accordance with relevant guidelines and regulations. Recruitment occurred June – November, 2020; interviews occurred July – December, 2020. Analysis occurred throughout the interview period and ended June, 2021.

Table 2 provides the descriptive characteristics of the nine practices that participated. Thirty-two patients were recruited and 23 completed an interview. Most were female (n = 19; 82.6%), all were White race (predominant in rural Colorado), and some reported Hispanic ethnicity (n = 4; 17.3%). Patients ranged in age from 53–74 (mean of 64.3) years. About half currently smoked (n = 13; 56.6%) with the rest reported a history of smoking. Among individuals who previously smoked (n = 10), about half had quit within the past year (n = 6; 26%). About half (n = 12; 51.7%) recalled being told about LCS (with about half of those deciding to undergo LCS). All were eligible for LCS based on recruitment guidelines for LCS age and smoking history eligibility.

Although previous studies have examined the challenges of consistently conducting shared decision making for LCS in rural primary care, this study is unique in its use of qualitative methods, an implementation framework (RE-AIM) for examining implementation issues specifically, and assessment of multiple participants in the primary care setting including patients in order to triangulate viewpoints. Key new findings for the understanding of LCS implementation include the lack of systematic processes that integrate clinic staff for shared decision making for LCS (as compared to smoking cessation and other cancer screenings [29]), which is further complicated by contingencies based on patient responses and preferences and stigma around the relationship of smoking to lung cancer. A main recommendation is further study to investigate if systematizing the process and involving more practice team members results in better uptake of LCS.

Some of our results replicate those found in prior research about LCS in rural settings [30, 31]. These include lack of geographic access to LDCT screening programs and limited or ineffective clinician-patient communication [19, 32]. Although previous studies assessing patients’ perceptions of cancer screening found that fatalistic belief systems contribute to patients’ declining LCS [33], this factor seemed especially strong among patients in our study.

Further research is needed to investigate what might be done to overcome this challenge, but it may be due in part to the combination of long term smoking, stigma regarding smoking [34, 35], sociocultural beliefs about lung cancer having a poor prognosis [36, 37], and social determinants of health challenges [38, 39] since both rural patients and people who smoke tend to have more social needs challenges than other groups [40, 41].

As with other research [39, 42] there was also confusion related to insurance coverage and costs associated with LCS. These perceptions run contrary to established policies and regulations dictating coverage for Medicare beneficiaries, Affordable Care Act policies given the USPSTF B grade for service recommendation, and coverage of LCS for Medicaid beneficiaries in Colorado [43, 44]. This barrier was identified following the initial launch of LCS while payers adapted their coverage and data management systems, but issues have substantially declined with more years of experience. It is important to explore this barrier more to determine if insurance issues are linked with inappropriate coverage refusals by payers or if initial coverage challenges contributed to sustained misperceptions regarding LCS policy.

This paper further contributes to the existing literature because of its focus on implementation within the practice. Of note, we found smoking identification, counseling, and referral to resources and assistance were systematic processes involving staff with follow-up support by the clinician with patient decision making; whereas, these processes for LCS were largely absent. Although the shared decision making may still rest with the clinician, other parts of the process – such as calculating pack years – could be added to staff responsibilities and be systematically collected. As found by Slatore, et al. [45] in their survey of practitioners in Oregon, the process was essentially left up to the clinician. They also found that registry and EMR systems lacked support for this effort [45]. This was especially true for rural settings in their study. This finding aligns with research showing that a centralized LCS program may be better suited to manage annual and follow-up screening [46], while maintaining communication with referring primary care clinicians regarding patient management. Other research describes how to set up a team-based approach for other cancer screenings and recommends that the same be done for LCS [47].

In examining perspectives of different participants in this process, we found similar barriers across groups regarding the practical difficulties of completing LCS and the variable response among patients to shared decision making regarding LCS. Patients were able to explain in detail their reasoning for accepting or declining LCS. Some wanted and took the opportunity to get LCS because it might help catch cancer early and prevent death, and others had belief systems that were inconsistent with screening such as it being not possible to change (“it’s my time when it’s my time to go”) or not necessary (“deal with it when something comes up”). Shared mental models about what something is, how it works, and why it is important are an important implementation concern [48]. One potential option is to systematically elicit patient perceptions as part of the shared decision making process, such as through a pre-visit questionnaire or having the patient watch a pre-visit video [49, 50]. Knowing their perceptions would illuminate to what extent patient values and perspectives are influencing the process of getting LCS versus logistical or financial concerns and help the clinician focus their consultation time on correcting misperceptions or problem-solving barriers to screening.

We found the RE-AIM model was helpful to conceptualize and categorize factors related to LCS [51, 52]. Using RE-AIM made it evident that some dimensions were discussed less often than others; for example, practice staff paid less attention to the effectiveness of LCS and the maintenance of continuing LCS. This has significant implications for implementation and points to areas for intervention in terms of clinical consultations regarding the process and value of LCS. In particular, the cascade effect of the interaction among several sequential factors in the RE-AIM model (e.g., if a patient was identified; if so, then approached for discussion, etc.) was particularly salient.

Our findings have several implications for implementation of LCS initiatives directed at rural primary care practices and their patients. First, context is a critically important factor, which we define broadly: in addition to physical settings and available resources and workflows, there are also the more subjective contextual issues such as history and patient and clinician values. Patient perspectives and preferences may be different and require different strategies. For example, in our study, some patients voiced concerns with getting health care interventions, consistent with a minimizer perspective [53], which may be more common in rural areas [26]. Second, processes for implementing LCS in rural settings need to be pragmatic, not overly time consuming, fit into existing workflows, and tap available resources in rural primary care as opposed to those applicable in large integrated care settings where much of the research on LCS has been conducted. This may be particularly challenging in rural settings that have less access to centralized LCS programs that take a more active role in managing the full LCS process and helping patients navigate the complexities. One important workflow option could involve clinical staff assessing smoking status and offering smoking cessation, including calculating eligiblility for LCS, while clinicians maintain responsibility for the LCS discussion with the patient. Third, outcomes could be enhanced by educating staff regarding 1) current eligibility/and reimbursement details; 2) differences between screening for high-risk patients versus diagnostic follow-up of symptomatic patients; and 3) importance of implementing of high quality shared decision making and smoking cessation counseling and not just “checking the box” [54, 55]. Finally, many rural settings do not have state of the art EHR systems or other technologies; enhancing automated identification and prompting systems for identification and follow up with LCS eligible patients, and conducting ongoing audit and feedback would likely enhance success.

Limitations of this study include the relatively small sample size in Colorado, although we did find thematic saturation and feel the sample was sufficient for this qualitative exploration. Our findings need replication – especially as they were obtained during COVID-19 and at a time when policy and reimbursement issues around LCS remain poorly understood. This as well as the relatively new and somewhat changing and uncertain specifics around requirements for reimbursement might well produce findings that could vary over time. Our study also has strengths and unique contributions, especially the use of an implementation science approach to LCS with use of RE-AIM and multiple perspectives for the analysis [56].

In conclusion, there are multiple contextual factors that affect implementation of LCS and performance of shared decision making in rural settings. Several of the perceived challenges were shared across different types of participants. Future research should attempt to replicate and expand our findings in different settings and evaluate interventions based on proposed recommendations regarding more robust assessment of smoking by staff and integrating additional tools to educate candidates for LCS.