Abstract
Background
Implementation of effective smoking cessation interventions in lung cancer screening has been identified as a high-priority research gap, but knowledge of current practices to guide process improvement is limited due to the slow uptake of screening and dearth of data to assess cessation-related practices and outcomes under real-world conditions.
Objective
To evaluate cessation treatment receipt and 1-year post-screening cessation outcomes within the largest integrated healthcare system in the USA—the Veterans Health Administration (VHA).
Design
Observational study using administrative data from electronic medical records (EMR).
Patients
Currently smoking Veterans who received a first lung cancer screening test using low-dose CT (LDCT) between January 2014 and June 2018.
Main Outcomes
Tobacco treatment received within the window of 30 days before and 30 days after LDCT; 1-year quit rates based on EMR Smoking Health Factors data 6–18 months after LDCT.
Key Results
Of the 47,609 current smokers screened (95.3% male), 8702 (18.3%) received pharmacotherapy and/or behavioral treatment for smoking cessation; 531 (1.1%) received both. Of those receiving pharmacotherapy, only one in four received one of the most effective medications: varenicline (12.1%) or combination nicotine replacement therapy (14.3%). Overall, 5400 Veterans quit smoking—a rate of 11.3% (missing=smoking) or 13.5% (complete case analysis). Treatment receipt and cessation were associated with numerous sociodemographic, clinical, and screening-related factors.
Conclusions
One-year quit rates for Veterans receiving lung cancer screening in VHA are similar to those reported in LDCT clinical trials and cohort studies (i.e., 10–17%). Only 1% of Veterans received the recommended combination of pharmacotherapy and counseling, and the most effective pharmacotherapies were not the most commonly received ones. The value of screening within VHA could be improved by addressing these treatment gaps, as well as the observed disparities in treatment receipt or cessation by race, rurality, and psychiatric conditions.
Supplementary Information
The online version contains supplementary material available at 10.1007/s11606-021-07011-0.
Key Words: tobacco, nicotine, smoking, veterans, cancer screening tests
INTRODUCTION
As lung cancer screening (LCS) is increasingly adopted by healthcare organizations and screening volume increases, major knowledge gaps remain around implementation. A key question is how to best provide cessation support for current smokers[1–3], who comprise approximately half of the screening population[4–6]. Although it is widely acknowledged that supporting smoking cessation and encouraging the use of evidence-based tobacco treatment as a means of doing so is a high priority[7], little is known about long-term cessation rates or tobacco treatment receipt among screening patients seen in real-world clinical settings. Such information could guide the development of processes and policies to maximize the value of screening.
Large clinical trials[8–10] and cohort studies[11, 12] of LCS have produced estimates of long-term (1 to 2 year) quit rates in the range of 10–17%. However, these estimates could be influenced by selection bias, as many of the characteristics associated with self-selecting into research participation (e.g., age, socioeconomic status, and health status[13]) are also correlates of smoking cessation[14, 15]. Cessation data extracted from healthcare administrative data are not subject to self-selection bias and consequently provide a valuable alternative source of information about smoking cessation rates following LCS. The use of administrative data also allows for analysis of individual predictors of cessation to inform efforts to improve cessation outcomes for LCS recipients. Only one previous study used electronic medical record data, in combination with self-report data, to assess 1-year cessation outcomes in a clinical sample of patients undergoing LCS[16]. The cessation rate among current smokers in that study (21%) was higher than in the trials and cohort studies cited previously, although the analysis was restricted to patients who completed a follow-up CT exam—a sample which is likely not representative of all patients who initiate LCS.[16]
One factor that is likely to influence cessation outcomes is the receipt of evidence-based treatment. Prior studies have not examined cessation treatment receipt among LCS patients or its relationship with cessation outcomes in a real-world screening setting, and most have not evaluated patient characteristics associated with receiving tobacco treatment. This information is important given the unique aspects of the LCS context. For example, smokers eligible for LCS are older than the general population of smokers, as the minimum recommended age for screening is 50.
To address these gaps in the literature on smoking cessation in the context of LCS, the first aim of the present study was to describe the receipt of both behavioral treatment and pharmacotherapies for smoking cessation at the time of screening as well as 1-year post-screening cessation rates among Veterans receiving LCS through the Veterans Health Administration (VHA). The VHA is the nation’s largest integrated healthcare system, where nearly one million patients (13% of the 6.7 million total patients served) meet screening criteria.[17] The second aim of the study was to identify factors associated with cessation treatment receipt and smoking abstinence at 1 year, including demographic, health, and screening-related factors.
METHODS
Participants
The individuals included in this analysis were 47,609 currently smoking Veterans who received a first LCS test using low-dose CT through the VHA between January 2014 and June 2018. To be included, Veterans had to have at least 18 months of post-screening medical care (through December 2019) documented in the electronic medical record. We excluded any patients who died within 12 months of screening.
Procedures
For this retrospective observational study, we used administrative VHA electronic medical record (EMR) data to identify Veterans who received a first LCS test at 79 VHA sites nationwide between January 1, 2014, and June 30, 2018. LCS tests were identified through CPT code G0297, code S8032 which was used from 2014 to 2016, or flags used by the VHA’s lung cancer screening demonstration project [17]. We focus on the first LCS test only, although patients may have had more than one test during the time period. A total of 82,732 LCS exams were identified: 73,676 (89%) by CPT code G0297; 6335 (8%) by demonstration project flags; and 2721 (3%) by temporary CPT code S8032.
Smoking status at the time of screening was extracted from the screening record using Health Factors smoking status data prior to the screening test to categorize Veterans as either a current smoker, former smoker, or unknown status using the categorization methods described by McGinnis and colleagues[18]. A Veteran was considered a current smoker if the most proximal smoking Health Factor prior to screening indicated current smoking. Only current smokers were included in the present analysis (47,609 of 82,732 total lung cancer screening participants; 58%).
Age, gender, race, ethnicity, marital status, co-pay status, and geographic information including region of the US and rurality of the Veteran’s residence were also extracted from the records. Rurality was classified as urban, rural, and highly rural residence corresponding with the following rural-urban commuting area, or RUCA, codes: urban=1.0–1.1; rural=1.2–9.9, 10.1–10.6; highly rural=10.0)[19]. Race/ethnicity was categorized as non-Hispanic Black, Hispanic, non-Hispanic White, and other/unknown. Financial and disability categories were defined using a 4-category variable based on VA copay requirements (VA copay required due to means, no copay required due to disability, no copay required due to means/other, and not assigned). Twenty-two pre-existing chronic medical conditions (Charlson comorbidity index)[20, 21] as well as 6 co-occurring mental health conditions were assessed via inpatient and outpatient diagnostic codes (ICD-9-CM and ICD-10-CM codes) in the 12 months prior to screening (Supplementary Table S1). We originally planned to include non-nicotine substance use disorders as a seventh category of mental health conditions but found that the number of individuals with this diagnosis was too low for meaningful analyses (n=69), so we excluded this category.
Quit rates 1 year following screening were calculated based on updated smoking Health Factors during a designated time window (i.e., 6–18 months) following the first LCS. This window was chosen based on knowledge of VA procedures for LCS scheduling and clinical reminders. Specifically, clinical reminders for tobacco use are to be completed annually, and ongoing or follow-up LCS should occur 6–12 months after the first screening, depending on whether any abnormal findings were detected. Thus, a 12-month window—between 6 and 18 months after the index LCS—includes both the expected window for tobacco use re-assessment as well as the expected window for the next LCS procedure. Nonetheless, not all patients had updated Smoking Health Factors (16% were missing),
In addition to quit rates, we examined receipt of behavioral counseling and pharmacotherapy for tobacco cessation around the time of screening, between 30 days prior to LCS and 30 days following LCS, regardless of whether treatment was initiated prior to that time window. The rationale for including tobacco treatment received up to 30 days prior to screening is that VA (and other national) guidelines recommend that cessation counseling occur during the initial discussion with a provider about getting screened, which may occur several weeks before the actual date of the LDCT test due to scheduling delays. Behavioral counseling was identified using VA stop codes 707 and 708[22], or CPT codes consistent with VA’s internal reporting methods[23] (see Supplementary Table 2). Pharmacotherapy was categorized into single-formulation NRT, bupropion (150 mg) alone, bupropion + NRT, combination NRT, and varenicline (with or without any other FDA-approved medication). The protocol was approved by the Institutional Review Board of the VA Puget Sound Health Care System (ID: 00819).
Statistical Analysis
Descriptive analyses summarized Veteran demographic and clinical characteristics in the full sample of currently smoking Veterans (n=47,609) (see Table 1). Multilevel logistic regression with receipt of any cessation support (medication, behavioral counseling, or both) within 30 days of the initial LCS as the primary outcome of interest was used to evaluate potential correlates of receipt of any cessation support. Quit rates at one year were calculated two ways: (1) using complete case analysis (i.e., limiting the sample to Veterans with EMR smoking Health Factors data available), and (2) using a missing=smoking imputation (i.e., the Russell standard[24]; using the full sample and considering Veterans with missing EMR smoking Health Factors data as continuing smokers). Since observed quit rates can vary considerably depending on the method of handling missing data and each method has limitations, we report both complete-case and missing=smoking as co-primary estimates of the 1-year quit rate. A multilevel logistic regression model was used to evaluate correlates of smoking abstinence at one year, using the missing=smoking assumption as the primary analysis and conducting a supplemental sensitivity analysis using only the complete case data (see Supplementary Figure 1). A random intercept for facility was used in all models to account for potential differences in recruitment to screening or other associations that may have varied across the 79 facilities. Outcomes were reported as odds ratios (ORs). All analyses were conducted in Stata Version 15 and R version 3.6.1[25].
Table 1.
Characteristics of Currently Smoking Veterans Who Underwent Initial Lung Cancer Screening Between 2014 and 2018: Overall and by Receipt of Any Smoking Cessation Treatment (N=47,609)
| No cessation treatment (n=38,907) | Cessation treatment (n=8702) | Overall (n=47,609) | |
|---|---|---|---|
| Age | |||
| 50–59 | 7280 (18.7%) | 2140 (24.6%) | 9420 (19.8%) |
| 60–69 | 22,966 (59.0%) | 5098 (58.6%) | 28,064 (58.9%) |
| 70–79 | 8661 (22.3%) | 1464 (16.8%) | 10,125 (21.3%) |
| Gender | |||
| Male | 37,201 (95.6%) | 8187 (94.1%) | 45,388 (95.3%) |
| Female | 1706 (4.4%) | 515 (5.9%) | 2221 (4.7%) |
| Race | |||
| White non-Hispanic | 28,298 (72.7%) | 6094 (70.0%) | 34,392 (72.2%) |
| Black non-Hispanic | 6955 (17.9%) | 1832 (21.1%) | 8787 (18.5%) |
| Hispanic | 997 (2.6%) | 241 (2.8%) | 1238 (2.6%) |
| Other | 594 (1.5%) | 133 (1.5%) | 727 (1.5%) |
| Unknown | 2063 (5.3%) | 402 (4.6%) | 2465 (5.2%) |
| Marital status | |||
| Married | 16,363 (42.1%) | 3363 (38.6%) | 19,726 (41.4%) |
| Never married/single | 4714 (12.1%) | 1215 (14.0%) | 5929 (12.5%) |
| Separated/divorced | 15,604 (40.1%) | 3661 (42.1%) | 19,265 (40.5%) |
| Widowed | 2088 (5.4%) | 434 (5.0%) | 2522 (5.3%) |
| Unknown | 138 (0.4%) | 29 (0.3%) | 167 (0.4%) |
| Copay | |||
| No copay (means/disability) | 32,099 (82.5%) | 7367 (84.7%) | 39,466 (82.9%) |
| Copay required | 6793 (17.5%) | 1332 (15.3%) | 8125 (17.1%) |
| Missing | 15 (0.0%) | 3 (0.0%) | 18 (0.0%) |
| Count of pre-existing medical conditions | |||
| 0 | 11,851 (30.5%) | 2593 (29.8%) | 14,444 (30.3%) |
| 1 | 11,701 (30.1%) | 2772 (31.9%) | 14,473 (30.4%) |
| 2 | 5965 (15.3%) | 1373 (15.8%) | 7338 (15.4%) |
| 3 | 4102 (10.5%) | 828 (9.5%) | 4930 (10.4%) |
| 4+ | 5288 (13.6%) | 1136 (13.1%) | 6424 (13.5%) |
| Co-occurring mental health conditions | |||
| Trauma/stress disorder | 6176 (15.9%) | 1892 (21.7%) | 8068 (16.9%) |
| Anxiety disorder | 4435 (11.4%) | 1286 (14.8%) | 5721 (12.0%) |
| Unipolar mood disorder | 9177 (23.6%) | 2833 (32.6%) | 12,010 (25.2%) |
| Personality disorder | 954 (2.5%) | 306 (3.5%) | 1260 (2.6%) |
| Bipolar disorder | 1371 (3.5%) | 432 (5.0%) | 1803 (3.8%) |
| Psychotic disorder | 1384 (3.6%) | 382 (4.4%) | 1766 (3.7%) |
| Region | |||
| South | 19,166 (49.3%) | 3813 (43.8%) | 22,979 (48.3%) |
| Midwest | 8388 (21.6%) | 1970 (22.6%) | 10,358 (21.8%) |
| Northeast | 7525 (19.3%) | 1920 (22.1%) | 9445 (19.8%) |
| West | 3654 (9.4%) | 954 (11.0%) | 4608 (9.7%) |
| Island territories | 162 (0.4%) | 41 (0.5%) | 203 (0.4%) |
| Missing | 12 (0.0%) | 4 (0.0%) | 16 (0.0%) |
| Rurality | |||
| Urban | 26,396 (67.8%) | 6329 (72.7%) | 32,725 (68.7%) |
| Rural | 12,123 (31.2%) | 2293 (26.4%) | 14,416 (30.3%) |
| Highly rural | 375 (1.0%) | 76 (0.9%) | 451 (0.9%) |
| Missing | 13 (0.0%) | 4 (0.0%) | 17 (0.0%) |
| Year of lung cancer screening | |||
| 2018 | 9334 (24.0%) | 2193 (25.2%) | 11,527 (24.2%) |
| 2017 | 16,780 (43.1%) | 3735 (42.9%) | 20,515 (43.1%) |
| 2016 | 10,386 (26.7%) | 2124 (24.4%) | 12,510 (26.3%) |
| 2015 | 1570 (4.0%) | 439 (5.0%) | 2009 (4.2%) |
| 2014 | 837 (2.2%) | 211 (2.4%) | 1048 (2.2%) |
RESULTS
Characteristics of the Sample
The majority of the sample was male (95.3%), non-Hispanic White (72.2%), of lower socioeconomic status (i.e., 47.5% had no VA co-pay due to low income and 35.4% had no co-pay due to disability), and lived in an urban area (68.8%). Because the number of screens conducted annually in the VA has been substantially increasing year over year since 2014, most of the Veterans (67.3%) were screened in 2017 or the first half of 2018. Medical and mental health conditions were common: 23.8% had 3 or more comorbid medical conditions on the Charlson index, 25.2% had a past-year unipolar mood disorder diagnosis, 16.9% had a past-year trauma/stress disorder diagnosis, and 12.0% had a past-year anxiety disorder diagnosis (see Table 1).
Receipt of Cessation Treatment
Observed Rates of Receiving Cessation Treatment
Of the 47,609 current smokers screened, 8702 (18.3%) received either behavioral treatment, pharmacotherapy, or both for smoking cessation: 7806 (16.4%) received pharmacotherapy only, 365 (0.7%) received behavioral treatment only, and 531 (1.1%) received both pharmacotherapy and behavioral treatment. Of the 8337 (17.5%) Veterans who received any pharmacotherapy, the majority received single-formulation NRT (n=4440; 53.3%), followed by combination NRT (n=1189; 14.3%), bupropion (n=1166; 14.0%), varenicline (n=860; 10.3%), bupropion + NRT (n=536; 6.4%), varenicline + NRT (n=113; 1.4%), varenicline + bupropion (n=28; 0.3%), and varenicline + bupropion + NRT (n=5; <0.1%). Varenicline, either alone or in combination with other medications, was used by 0.12% (n=1006) of Veterans overall and 12.1% of those Veterans who used any medication. Proportions of LCS participants using each type of FDA-approved medication over time (see Fig. 1) show that (a) single-formulation NRT is the most-prescribed treatment each year from 2014 to 2018 and (b) after a drop in use in 2015, receipt of varenicline has rebounded to a level similar to other treatment regimens (with the exception of single-formulation NRT).
Figure 1.
Type of medication received among currently smoking Veterans who underwent lung cancer screening and received cessation medication between 2014 and 2018 (n=8337).
Correlates of Receiving any Cessation Treatment
In the multivariable logistic regression models controlling for the effects of facility (see Table 2), significant correlates of receiving behavioral treatment and/or pharmacotherapies for smoking cessation included younger age, past-year trauma/stress disorder, past-year anxiety disorder, past-year unipolar mood disorder, bipolar disorder, and non-rural residence.
Table 2.
Odds Ratios for Predictors of Cessation Treatment Receipt from Multivariable Logistic Regression Models
| Variable | OR | Lower 95% CI | Upper 95% CI | P value |
|---|---|---|---|---|
| Age | 0.83 | 0.81 | 0.85 | < 0.001 |
| Gender (female) | 1.11 | 1.00 | 1.24 | 0.048 |
| Race | ||||
| White | Reference | --- | --- | --- |
| Black non-Hispanic | 1.06 | 0.99 | 1.13 | 0.095 |
| Hispanic | 0.94 | 0.80 | 1.11 | 0.478 |
| Other | 0.96 | 0.79 | 1.17 | 0.679 |
| Unknown | 0.93 | 0.83 | 1.04 | 0.185 |
| Marital status | ||||
| Married | Reference | --- | --- | --- |
| Never married/single | 1.03 | 0.95 | 1.11 | 0.497 |
| Separated/divorced | 1.03 | 0.97 | 1.08 | 0.320 |
| Widowed | 1.01 | 0.90 | 1.13 | 0.914 |
| Unknown | 0.93 | 0.61 | 1.43 | 0.743 |
| Copay required | 0.96 | 0.90 | 1.03 | 0.289 |
| Count of pre-existing medical conditions (0–22) | 1.01 | 0.99 | 1.02 | 0.375 |
| Co-occurring mental health conditions | ||||
| Trauma/stress disorder | 1.29 | 1.21 | 1.38 | < 0.001 |
| Anxiety disorder | 1.09 | 1.01 | 1.17 | 0.022 |
| Unipolar mood disorder | 1.36 | 1.28 | 1.44 | < 0.001 |
| Personality disorder | 1.04 | 0.90 | 1.20 | 0.581 |
| Bipolar disorder | 1.15 | 1.03 | 1.30 | 0.016 |
| Psychotic disorder | 1.03 | 0.92 | 1.17 | 0.583 |
| Region | ||||
| South | Reference | --- | --- | --- |
| Midwest | 0.98 | 0.84 | 1.14 | 0.788 |
| Northeast | 1.14 | 0.97 | 1.33 | 0.103 |
| West | 1.16 | 0.98 | 1.38 | 0.086 |
| Island territories | 1.42 | 0.79 | 2.55 | 0.245 |
| Rurality | ||||
| Urban | Reference | --- | --- | --- |
| Rural | 0.90 | 0.85 | 0.96 | <0.001 |
| Highly rural | 0.97 | 0.75 | 1.25 | 0.823 |
| Year of lung cancer screening | ||||
| 2018 | Reference | --- | --- | --- |
| 2017 | 0.99 | 0.93 | 1.05 | 0.782 |
| 2016 | 0.95 | 0.89 | 1.02 | 0.154 |
| 2015 | 1.02 | 0.89 | 1.17 | 0.757 |
| 2014 | 0.88 | 0.74 | 1.05 | 0.147 |
Note: A random intercept for facility was used to account for potential differences in recruitment to screening or other associations that may have varied across the 79 VA facilities
Smoking Cessation 12 Months Following an Initial LCS
Observed Rate of Smoking Cessation
Overall, 5400 Veterans quit smoking 1 year post-screening for a quit rate of 13.5% (5400/39,941) under a complete case analysis and 11.3% (5400/47,609) under the missing=smoking assumption.
Correlates of Cessation
In multivariable logistic regression models controlling for facility effects and using the missing=smoking assumption to estimate 1-year cessation rates for the 16% of Veterans who had missing EMR Smoking Health Factors data (see Table 3), characteristics associated with greater odds of quitting were older age; greater medical comorbidity, as indicated by a higher Charlson score; living in the Northeast; receipt of varenicline compared to no medications; receipt of bupropion compared to no medications; and receipt of bupropion plus NRT compared to no medications. Lower odds of quitting were associated with being never married, separated/divorced, or widowed, relative to being married; having a past-year psychotic disorder; being screened prior to 2018; and being non-Hispanic Black. As a sensitivity analysis, we also re-ran this model using the complete-case method (i.e., excluding the 16% of Veterans with missing smoking Health Factors data), and results were essentially unchanged (see Supplementary Figure 1).
Table 3.
Odds Ratios for Predictors of 1-Year Smoking Cessation Using Multivariable Logistic Regression Models (N=47,609)
| Variable | OR | Lower 95% CI | Upper 95% CI | P value |
|---|---|---|---|---|
| Age | 1.22 | 1.19 | 1.26 | < 0.001 |
| Gender (female) | 1.00 | 0.86 | 1.16 | 0.971 |
| Race | ||||
| White | Reference | --- | --- | --- |
| Black non-Hispanic | 0.91 | 0.83 | 0.99 | 0.026 |
| Hispanic | 0.91 | 0.74 | 1.12 | 0.397 |
| Other | 1.03 | 0.81 | 1.30 | 0.839 |
| Unknown | 0.91 | 0.80 | 1.04 | 0.187 |
| Marital status | ||||
| Married | Reference | --- | --- | --- |
| Never married/single | 0.73 | 0.66 | 0.80 | < 0.001 |
| Separated/divorced | 0.70 | 0.66 | 0.75 | < 0.001 |
| Widowed | 0.78 | 0.68 | 0.89 | < 0.001 |
| Unknown | 0.83 | 0.49 | 1.40 | 0.481 |
| Copay required | 0.95 | 0.88 | 1.03 | 0.182 |
| Count of pre-existing medical conditions (0–22) | 1.04 | 1.03 | 1.06 | < 0.001 |
| Co-occurring mental health conditions | ||||
| Trauma/stress disorder | 0.95 | 0.88 | 1.04 | 0.270 |
| Anxiety disorder | 1.07 | 0.98 | 1.18 | 0.134 |
| Unipolar mood disorder | 0.95 | 0.88 | 1.02 | 0.156 |
| Personality disorder | 0.92 | 0.76 | 1.13 | 0.436 |
| Bipolar disorder | 1.07 | 0.92 | 1.25 | 0.382 |
| Psychotic disorder | 0.77 | 0.65 | 0.92 | 0.005 |
| Region | ||||
| South | Reference | --- | --- | --- |
| Midwest | 1.04 | 0.86 | 1.24 | 0.707 |
| Northeast | 1.27 | 1.05 | 1.53 | 0.012 |
| West | 0.87 | 0.70 | 1.08 | 0.207 |
| Island territories | 1.29 | 0.65 | 2.57 | 0.469 |
| Rurality | ||||
| Urban | Reference | --- | --- | --- |
| Rural | 0.98 | 0.91 | 1.05 | 0.522 |
| Highly rural | 0.90 | 0.66 | 1.23 | 0.521 |
| Year of lung cancer screening | ||||
| 2018 | Reference | --- | --- | --- |
| 2017 | 0.71 | 0.66 | 0.76 | < 0.001 |
| 2016 | 0.60 | 0.55 | 0.65 | <0.001 |
| 2015 | 0.78 | 0.66 | 0.93 | 0.006 |
| 2014 | 0.75 | 0.60 | 0.94 | 0.011 |
| Cessation medication | ||||
| No cessation medication | Reference | --- | --- | --- |
| Bupropion + NRT | 1.79 | 1.41 | 2.26 | < 0.001 |
| Bupropion only | 1.51 | 1.28 | 1.79 | <0.001 |
| Combination NRT | 0.99 | 0.81 | 1.21 | 0.935 |
| Single NRT | 1.01 | 0.91 | 1.13 | 0.784 |
| Varenicline | 1.38 | 1.15 | 1.67 | 0.001 |
| Behavioral counseling | 1.05 | 0.85 | 1.29 | 0.672 |
Note: A random intercept for facility was used to account for potential differences in recruitment to screening or other associations that may have varied across the 79 VA facilities. NRT, nicotine replacement therapy
DISCUSSION
In this report, we examine 1-year smoking cessation outcomes, receipt of evidence-based cessation treatment, and individual-level correlates of treatment receipt and smoking cessation among Veterans undergoing a first LCS test through the VHA. Quit rates at 1 year (11.3% using missing=smoking assumption and 13.5% using complete case analysis) were on the lower end of the range of estimated quit rates in previous clinical trials[8–10] and cohort studies[11, 12] of lung cancer screening, which have produced 10–17% quit rates 1 to 2 years following screening. This is not surprising, as research participants differ in many ways from clinical populations[13]. In the VA setting in particular, the clinical population may have elevated risk factors for cessation failure, such as low socioeconomic status and mental health conditions[26]. Notably, the quit rates for a 1-year window 6 to 18 months after the initial LCS were very close to the recently reported annual quit rate of 12.0% among 754,504 current tobacco users in the VHA, extracted from 2009 administrative data using methods similar to this study[27].
Another key finding was that tobacco cessation treatment was underutilized by current smokers undergoing LCS, with 82% of screening participants receiving neither counseling nor pharmacotherapy, 16% receiving pharmacotherapy alone, and only 1% of smokers receiving the recommended, most effective form of treatment: behavioral treatment plus pharmacotherapy[28]. This estimate of tobacco treatment receipt around the time of a first LDCT is substantially lower than annual estimates in the general population, outside of the LCS context. One prior study in VA identified that 26% of current tobacco users in 2013 received some form of pharmacotherapy in the prior year[29], and VHA’s Academic Detailing dashboard (October 2019) reports that over 28% of current smokers were provided some form of pharmacotherapy in the prior year[23]. In our analysis of VHA LCS data, pharmacotherapy was used at a much higher rate than behavioral treatment (8337 vs. 896, or 9:1), which is better than prior research showing a pharmacotherapy to behavioral therapy receipt of 20:1 in the US population.[30] Rates of behavioral treatment receipt, with or without medication (896/47,609, or 1.9%) were nonetheless lower than the national rate of 5.2% reported in VHA’s Academic Detailing dashboard[23] (October 2019) and a prior VA study showing that only 3.8% of Veterans received behavioral treatment over a 1-year period in 2012 [22]. While our period of evaluation was considerably shorter (i.e., a 60-day window anchored to the screening test rather than a 12-month window anchored to a calendar year), screening participants are taking preventive action by participating in screening and are supposed to be receiving cessation support as part of screening implementation. Consequently, this low rate of tobacco treatment receipt is concerning, regardless of how it compares to the annual rates of tobacco treatment receipt in the broader VA population.
Among those who used medications, single-formulation NRT (53.3%) was prescribed three to four times more often than combination NRT (14.4%) or varenicline (12.1%). These proportions are similar to the VHA’s Academic Detailing dashboard in which 10.2% of medication users used combination NRT and 11.6% used varenicline[23]. Given the robust evidence from efficacy trials that single-formulation NRT is not the most efficacious medication option available,[28, 31] the comparatively lower use of other medication approaches that may have greater effectiveness warrants efforts to understand and modify contributing factors at the patient, provider, and system levels. These factors likely differ by medication type and may be context-dependent. For example, varenicline is significantly less accessible given that it is a second-line treatment within the VA system, requiring a number of criteria to be met in order to meet prescribing guidelines (e.g., failed trial of a first-line medication in the previous year, no current unstable/untreated mental health conditions). Nonetheless, prescriptions of each medication type by year (see Fig. 1) showed varenicline use increasing over time since 2015 and approaching a level similar to the other medications, except single-formulation NRT. This finding is consistent with the last few years of data by Ignacio et al [29], which examined treatment patterns between 2004 and 2013.
A surprising medication-related finding was that receipt of varenicline, bupropion alone, and bupropion plus NRT was associated with better quit rates than no medication, but single-formulation NRT and combination NRT were not. Caution is warranted when interpreting these findings since they are observational. However, given that the analyses controlled for numerous potential confounds that may have influenced selection into medication groups (e.g., the possibility that patients with more severe mental health conditions, who are less likely to quit smoking, may have been more likely to receive NRT due to concerns about neuropsychiatric adverse events from varenicline and bupropion), the findings are provocative and warrant an additional evaluation in the context of a randomized, controlled trial that evaluates effectiveness for cessation as well as cost-effectiveness.
There are at least two patterns among the demographic and clinical factors associated with tobacco treatment and/or cessation that warrant comment. One is that quit rates are higher in recent years than in the earlier years of screening. This suggests that the increased availability of LCS does not discourage screening-eligible Veterans from quitting. Another noteworthy pattern is the evidence for disparities in treatment receipt and/or cessation. Rural residents were less likely to receive cessation support, and quit rates were lower among Black Veterans and those with past-year psychotic disorders. Our finding that there were no differences in receipt of tobacco treatment by race, but that non-Hispanic Black Veterans had lower quit rates than White Veterans, was unexpected based on previous indications that lower quit rates among Black tobacco users may be driven by differential use of evidence-based treatment [32]. These results suggest that other factors may have been responsible for lower quit rates observed among Black Veterans in this context of lung cancer screening. For example, Black Veterans may be more likely to smoke menthol cigarettes, which can make quitting more difficult [33]. Because the reasons for such disparities in treatment receipt and outcomes are complex, multi-level interventions are needed to address them effectively,[34] including both individual and healthcare system-level efforts.
This study has a number of limitations, primarily stemming from the limitations of administrative datasets. Accuracy of smoking status indicators for specific time points post-screening or for detecting changes in smoking status over time are unknown, although prior work has shown that our method for classifying smoking status using EMR Health Factors in the broader VA patient population has 84–95% sensitivity and 79–91% specificity for categorizing patients into current vs. not current smokers[18, 35]. A new EMR Health Factor to identify tobacco users was added in October 2018, so more recent data may have even greater sensitivity and specificity. Treatment receipt data derived from the EMR may be incomplete due to patients using over-the-counter medications, treatments prescribed outside of the VA system, or behavioral treatments that aren’t captured in the medical record, like telephone quitlines or digital interventions. Additionally, VA is not a fee-for-service setting, so coding for cessation counseling may be underreported. Notably in 2019, quality measure data[36] identified that 96% of Veterans identified in primary care as smokers were at least briefly counseled about cessation, so our findings regarding the behavioral treatment codes are likely more reflective of low receipt of more intensive behavioral interventions rather than of brief assessment and advice to quit. Finally, the quality of the behavioral intervention provided cannot be extracted from the medical record, omitting a variable that is potentially important for understanding the relationship between treatment receipt and cessation outcomes. Regarding the interpretation of our findings, we did not distinguish between treatment initiated during the 30 days before and 30 days after the LDCT versus treatment that had been initiated prior to that time point and continued into that period, so it is not possible to state conclusively that receipt of tobacco treatment was causally linked to receipt of LCS. It is also not possible to determine reasons for low rates of treatment receipt—particularly the combination of pharmacotherapy and behavioral treatment—which may stem from patient-, provider-, and/or system-level factors. An important direction for future research is to understand provider behaviors and system-level characteristics that optimize treatment receipt and cessation outcomes, which may be highly variable even within the 79 sites included in this sample. The test of medication and behavioral treatment receipt as predictors of abstinence is also limited by selection bias (i.e., people may self-select into treatment based on having greater difficulty quitting, as noted previously); the relatively small number of participants who received behavioral counseling, resulting in a large confidence interval for this variable and inability to test for interactions with pharmacotherapy use; and absence of information on actual use of the medication (e.g., consistency and duration of use), which is not available in the medication records and is known to influence cessation outcomes.[37] Additionally, although we excluded bupropion dosing regimens that would be more indicative of mental health treatment than tobacco cessation treatment (i.e., we only considered receipt of bupropion 150 mg with twice-daily dosing), we cannot rule out the potential for this regimen being used in the treatment of mental health conditions rather than smoking cessation. Thus, the estimated receipt of bupropion for smoking cessation may be overestimated. Other limitations are the uncertain generalizability of the findings outside of the VA system, which (a) serves a predominantly male and low-income population; (b) makes all FDA-approved cessation pharmacotherapies available to Veterans, along with behavioral interventions in multiple formats; and (c) implements multiple system-level methods of prompting providers to document and treat tobacco use (e.g., clinical reminders integrated into the medical records system). Finally, we were unable to include the LDCT screening result in the analyses of predictors of treatment receipt and smoking cessation due to large amounts of missing data and inconsistent coding of results. While the evidence is not universally consistent, several studies have demonstrated that an abnormal test may increase the likelihood of making a quit attempt and/or successful quitting[8, 38].
In spite of these limitations, the study has many strengths. As one of the largest providers of LCS in the USA, with close to 1 million patients estimated to be screening-eligible [17], the Veterans Health Administration maintains a clinical database that is unique in its size and geographic reach, providing an invaluable opportunity to understand how tobacco treatment is being implemented in the context of LCS at a national level. Our data suggest that cessation support could be improved through efforts to increase the use of any cessation treatment and, among those to engage in the treatment, to increase the use of more effective pharmacotherapies and concomitant behavioral treatment. Given the evidence for some disparities in treatment receipt and/or outcome by race/ethnicity, rurality, and medical and mental health conditions, targeted efforts to address these disparities are warranted. These findings provide clear direction for improved implementation of LCS across the VA and point toward missed opportunities to support smoking cessation that may be occurring in other healthcare systems, as well.
Supplementary Information
(DOCX 27 kb)
Author Contribution
Dr. Heffner drafted the manuscript. Dr. Coggeshall, Dr. Wheat, and Ms. Johnson conducted the statistical analyses. All authors made substantive contributions to the design, conduct, and/or analysis of the study from which the data were derived.
Funding
This work was supported by the Department of Veterans Affairs Office of Research and Development, Health Services Research and Development Service, (PI: Zeliadt, PPO-14-130 and IIR-16-117), and is a product of a Prevention Research Center supported by Cooperative Agreement Number (14-011) from the Centers for Disease Control and Prevention.
Declarations
Disclaimer
The findings and conclusions in this article are those of the author(s) and do not necessarily represent the official position of the Department of Veterans Affairs or the Centers for Disease Control and Prevention.
Conflict of Interest
Dr. Heffner has received non-financial support from Pfizer, outside of the work presented in the paper. Dr. Feemster reports grants from VA HSR&D and the American Lung Association and personal fees from the National Committee for Quality Assurance and Annals of American Thoracic Society, outside the work presented in the paper. None of the other authors have conflicts to disclose.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Kathuria H, Detterbeck FC, Fathi JT, et al. Stakeholder research priorities for smoking cessation interventions within lung cancer screening programs. An official American Thoracic Society research statement. Am J Respir Crit Care Med.e. 2017;196(9):1202–1212. doi: 10.1164/rccm.201709-1858ST. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Mazzone PJ, Silvestri GA, Patel S, et al. Screening for lung cancer: CHEST guideline and expert panel report. Chest. 2018;153(4):954–985. doi: 10.1016/j.chest.2018.01.016. [DOI] [PubMed] [Google Scholar]
- 3.Joseph AM, Rothman AJ, Almirall D, et al. Lung cancer screening and smoking cessation clinical trials. SCALE (Smoking Cessation within the Context of Lung Cancer Screening) collaboration. Am J Respir Crit Care Med. 2018;197(2):172–182. doi: 10.1164/rccm.201705-0909CI. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Aberle DR, Adams AM, Berg CD, et al. Reduced lung-cancer mortality with low-dose computed tomographic screening. New Eng J Med. 2011;365(5):395–409. doi: 10.1056/NEJMoa1102873. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Jackson GL, King HA, McNeil RB, et al. Evaluation of the VA Lung Cancer Screening Clinical Demonstration Project – report submitted to the Veterans Affairs Office of the Under Secretary for Health. Veterans Affairs Center for Health Services Research in Primary Care and National Center for Health Promotion and Disease Prevention.
- 6.Tammemagi MC, Berg CD, Riley TL, Cunningham CR, Taylor KL. Impact of lung cancer screening results on smoking cessation. J Natl Cancer Inst. 2014;106(6):dju084. doi: 10.1093/jnci/dju084. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Tanner NT, Kanodra NM, Gebregziabher M, et al. The association between smoking abstinence and mortality in the National Lung Screening Trial. Am J Respir Crit Care Med. 2016;193(5):534–541. doi: 10.1164/rccm.201507-1420OC. [DOI] [PubMed] [Google Scholar]
- 8.Ashraf H, Tonnesen P, Holst Pedersen J, Dirksen A, Thorsen H, Dossing M. Effect of CT screening on smoking habits at 1-year follow-up in the Danish Lung Cancer Screening Trial (DLCST) Thorax. 2009;64(5):388–392. doi: 10.1136/thx.2008.102475. [DOI] [PubMed] [Google Scholar]
- 9.van der Aalst CM, van den Bergh KA, Willemsen MC, de Koning HJ, van Klaveren RJ. Lung cancer screening and smoking abstinence: 2 year follow-up data from the Dutch-Belgian randomised controlled lung cancer screening trial. Thorax. 2010;65(7):600–605. doi: 10.1136/thx.2009.133751. [DOI] [PubMed] [Google Scholar]
- 10.Park ER, Gareen IF, Jain A, et al. Examining whether lung screening changes risk perceptions: National Lung Screening Trial participants at 1-year follow-up. Cancer. 2013;119(7):1306–1313. doi: 10.1002/cncr.27925. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Townsend CO, Clark MM, Jett JR, et al. Relation between smoking cessation and receiving results from three annual spiral chest computed tomography scans for lung carcinoma screening. Cancer. 2005;103(10):2154–2162. doi: 10.1002/cncr.21045. [DOI] [PubMed] [Google Scholar]
- 12.Styn MA, Land SR, Perkins KA, Wilson DO, Romkes M, Weissfeld JL. Smoking behavior 1 year after computed tomography screening for lung cancer: Effect of physician referral for abnormal CT findings. Cancer Epidemiol Biomarkers Prev. 2009;18(12):3484–3489. doi: 10.1158/1055-9965.EPI-09-0895. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Martinson BC, Crain AL, Sherwood NE, Hayes MG, Pronk NP, O’Connor PJ. Population reach and recruitment bias in a maintenance RCT in physically active older adults. J Phys Act Health. 2010;7(1):127–135. doi: 10.1123/jpah.7.1.127. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Kale D, Gilbert HM, Sutton S. Are predictors of making a quit attempt the same as predictors of 3-month abstinence from smoking? Findings from a sample of smokers recruited for a study of computer-tailored smoking cessation advice in primary care. Addiction. 2015;110(10):1653–1664. doi: 10.1111/add.12972. [DOI] [PubMed] [Google Scholar]
- 15.Ferguson J, Bauld L, Chesterman J, Judge K. The English smoking treatment services: one-year outcomes. Addiction. 2005;100(Suppl 2):59–69. doi: 10.1111/j.1360-0443.2005.01028.x. [DOI] [PubMed] [Google Scholar]
- 16.Borondy Kitts AK, McKee AB, Regis SM, Wald C, Flacke S, McKee BJ. Smoking cessation results in a clinical lung cancer screening program. J Thoracic Dis. 2016;8(Suppl 6):S481–487. doi: 10.21037/jtd.2016.03.11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Kinsinger LS, Anderson C, Kim J, et al. Implementation of lung cancer screening in the Veterans Health Administration. JAMA Intern Med. 2017;177(3):399–406. doi: 10.1001/jamainternmed.2016.9022. [DOI] [PubMed] [Google Scholar]
- 18.McGinnis KA, Brandt CA, Skanderson M, et al. Validating smoking data from the Veteran’s Affairs Health Factors dataset, an electronic data source. Nicotine Tob Res. 2011;13(12):1233–1239. doi: 10.1093/ntr/ntr206. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Rural Health Research Center Rural-Urban Commuting Areas. https://depts.washington.edu/uwruca/ruca-uses.php Accessed June 18, 2021.
- 20.Charlson M, Szatrowski TP, Peterson J, Gold J. Validation of a combined comorbidity index. J Clin Epidemiol. 1994;47(11):1245–1251. doi: 10.1016/0895-4356(94)90129-5. [DOI] [PubMed] [Google Scholar]
- 21.D’Hoore W, Bouckaert A, Tilquin C. Practical considerations on the use of the Charlson comorbidity index with administrative data bases. J Clin Epidemiol. 1996;49(12):1429–1433. doi: 10.1016/S0895-4356(96)00271-5. [DOI] [PubMed] [Google Scholar]
- 22.Kelly MM, Sido H, Rosenheck R. Rates and correlates of tobacco cessation service use nationally in the Veterans Health Administration. Psychol Serv. 2016;13(2):183–192. doi: 10.1037/ser0000076. [DOI] [PubMed] [Google Scholar]
- 23.VHA Academic Detailing Service. Tobacco Use Disorder (TUD) dashboard, patient report and priority panel report. https://vaww.portal2.va.gov/sites/ad. Accessed August 19, 2019.
- 24.West R, Hajek P, Stead L, Stapleton J. Outcome criteria in smoking cessation trials: proposal for a common standard. Addiction. 2005;100(3):299–303. doi: 10.1111/j.1360-0443.2004.00995.x. [DOI] [PubMed] [Google Scholar]
- 25.R Core Team. R: A language and environment for statistical computing. R Foundation for Statistical Computing. https:/www.R-project.org/. Published 2017. Accessed February 17, 2021.
- 26.Edwin SW, Virginia W, Chuan-Fen L, Paul LH, Matthew LM. Do Veterans Health Administration enrollees generalize to other populations? Med Care Res Rev. 2015;73(4):493–507. doi: 10.1177/1077558715617382. [DOI] [PubMed] [Google Scholar]
- 27.Barnett PG, Chow A, Flores NE, Sherman SE, Duffy SA. Changes in Veteran tobacco use identified in electronic medical records. Am J Prev Med. 2017;53(1):e9–e18. doi: 10.1016/j.amepre.2017.01.009. [DOI] [PubMed] [Google Scholar]
- 28.Fiore MC, Jaén CR, Baker TB, et al. Treating tobacco use and dependence: 2008 update. In: Clinical Practice Guideline. Rockville, MD: U.S. Department of Health and Human Services, Public Health Service; 2008. [Google Scholar]
- 29.Ignacio RV, Barnett PG, Kim HM, et al. Trends and patient characteristics associated with tobacco pharmacotherapy dispensed in the Veterans Health Administration. Nicotine Tob Res. 2018;20(10):1173–1181. doi: 10.1093/ntr/ntx229. [DOI] [PubMed] [Google Scholar]
- 30.Cokkinides VE, Ward E, Jemal A, Thun MJ. Under-use of smoking-cessation treatments: results from the National Health Interview Survey, 2000. Am J Prev Med. 2005;28(1):119–122. doi: 10.1016/j.amepre.2004.09.007. [DOI] [PubMed] [Google Scholar]
- 31.Anthenelli RM, Benowitz NL, West R, et al. Neuropsychiatric safety and efficacy of varenicline, bupropion, and nicotine patch in smokers with and without psychiatric disorders (EAGLES): a double-blind, randomised, placebo-controlled clinical trial. Lancet. 2016;387:2507–20. doi: 10.1016/S0140-6736(16)30272-0. [DOI] [PubMed] [Google Scholar]
- 32.Centers for Disease Control and Prevention Quitting smoking among adults--United States, 2001-2010. MMWR. 2011;60(44):1513–1519. [PubMed] [Google Scholar]
- 33.Smith PH, Assefa B, Kainth S, Salas-Ramirez KY, McKee SA, Giovino GA. Use of mentholated cigarettes and likelihood of smoking cessation in the United States: A meta-analysis. Nicotine Tob Res. 2020;22(3):307–316. doi: 10.1093/ntr/ntz067. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.US National Cancer Institute. A socioecological approach to addressing tobacco-related health disparities: National Cancer Institute Tobacco Control Monograph 22. Bethesda, MD 2017. NIH Publication No. 17-CA-8035A.
- 35.Calhoun PS, Wilson SM, Hertzberg JS, et al. Validation of Veterans Affairs electronic medical record smoking data among Iraq- and Afghanistan-Era Veterans. J Gen Intern Med. 2017;32(11):1228–1234. doi: 10.1007/s11606-017-4144-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Anhang Price R, Sloss EM, Cefalu M, Farmer CM, Hussey PS. Comparing quality of care in Veterans Affairs and non-Veterans Affairs settings. J Gen Intern Med. 2018;33(10):1631–1638. doi: 10.1007/s11606-018-4433-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Siahpush M, Shaikh RA, McCarthy M, Sikora Kessler A, Tibbits M, Singh GK. Association between duration of use of pharmacotherapy and smoking cessation: findings from a national survey. BMJ Open. 2015;5(1):e006229. doi: 10.1136/bmjopen-2014-006229. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Ostroff JS, Buckshee N, Mancuso CA, Yankelevitz DF, Henschke CI. Smoking cessation following CT screening for early detection of lung cancer. Prev Med. 2001;33(6):613–621. doi: 10.1006/pmed.2001.0935. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
(DOCX 27 kb)