A Daily Assessment Study of Smoking Cessation After a Head and Neck Cancer Diagnosis

Abstract

Introduction

This intensive longitudinal study describes key events in the process of smoking cessation after a new head and neck cancer (HNC) diagnosis. Prior longitudinal studies show some cancer patients quit, while others continue to smoke, but details about the pattern in which these discrete outcomes arise are scarce. This study is meant to help rectify this gap in the literature.

Aims and Methods

Participants were 42 HNC patients who reported current smoking at enrollment. Participants were recruited from an outpatient oncology clinic and completed a baseline questionnaire prior to begin a 30-day daily assessment.

Results

Few participants (9.52%) achieved 30-day continuous abstinence from smoking. On average, participants reported 9.64 ± 11.93 total days of abstinence. Nearly, all (94.44%, n = 34) participants made at least one quit attempt, with an average of 16.94 ± 11.30 quit attempt days. Fewer participants were able to achieve a 24-hour quit attempt (52.78%, n = 19), with a corresponding average of 5.50 ± 8.69 24-hour days. The median time to first 24-hour quit attempt was 13 days after enrollment. Based on smoking behavioral patterns, participants were categorized into five groups, the most common being “persistent attempters,” which involved unsuccessful quit attempts throughout the study. Only 45% of participants (n = 19) used evidence-based treatment, the most common being cessation medication.

Conclusions

This intensive longitudinal study found that cancer diagnosis can spur a lot of efforts to quit smoking. Unfortunately, this study suggests that many quit attempts are short lived, possibly a result of an absence or insufficient use of evidence-based treatments.

Implications

For adults who are current smokers at the time of cancer diagnosis, there is a high likelihood of persistent cigarette smoking and use of other tobacco products in the weeks and months after a cancer diagnosis. Furthermore, this study shows that while a lot of quit attempts may occur, few are successful, which may be partly attributable to the low use of evidence-based tobacco treatment. Future research with cancer patients should aim to identify predictors of quit attempts and abstinence as well as treatment utilization.

Introduction

Head and neck cancer (HNC) disease sites include the larynx, nasal cavity, oral cavity, pharynx, and salivary glands.¹ In the United States, HNC comprises 4% of all new cancers, with more than 60 000 new primary diagnoses each year.² The overall 5-year survival rate for HNC is now roughly 60%, an outcome that has improved in recent decades.^3,4 Alongside human papillomavirus infection, the chief causes of HNC are tobacco use, and to a lesser extent, alcohol use,⁵ which makes HNC one of the most preventable cancers.

For some HNC patients, cigarette smoking is a prominent aspect of their cancer experience, present at the point of diagnosis, throughout curative treatment, and into long-term follow-up.⁶ For example, one longitudinal study found that 25% of HNC patients smoked after diagnosis, and half of the patients who had quit at diagnosis did not remain abstinent.⁷ Similarly, another longitudinal study of HNC and lung cancer patients found that 66% of patients reported smoking during the week before surgery, and of those patients, 60% reported smoking a year after surgery.⁸ Furthermore, a systematic literature review found that 33% of all HNC and lung cancer patients reported current smoking, with 54% of those who smoked at diagnosis continuing to do so.⁹ The clinical ramifications of smoking after a cancer diagnosis are profound and include a higher risk for all-cause mortality, cancer-specific mortality, a new primary cancer, cancer recurrence, or progression, and poor treatment response.^10–13 Additionally, smoking after a cancer diagnosis is associated with lower quality of life.¹⁴ For all these reasons, an evidence-based treatment-assisted quit attempt should occur immediately upon receipt of HN or any other cancer diagnosis.

Motivation to quit smoking may be strong due to the “teachable moment” of a cancer diagnosis,¹⁵ but there are a host of other variables at play (eg, psychological distress, fatalistic beliefs, changes in social network, and support) that may interfere with cancer patients’ ability to make a quit attempt and/or maintain abstinence. To design interventions that effectively capitalize on the “teachable moment,” it is imperative to appreciate fully the dynamic changes in smoking behavior after a cancer diagnosis. If there are naturally occurring quit attempts or other efforts to reduce tobacco use, then perhaps interventions could be designed to augment predictable spikes in motivation or bolster the success or otherwise haphazard quit attempts with “stand by” medication or behavioral support. Research that informs innovations in treatment remains important because prior clinical trials for smoking cessation in cancer patients have yielded only modest success^16,17 and patient engagement has been challenging.¹⁸

The current study is based on the premise that an intensive longitudinal design yields the most comprehensive and detail-oriented description of smoking behavior in the immediate aftermath of a cancer diagnosis. Overall, prior studies in this area are subject to many limitations, including a strong reliance on cross-sectional designs⁹ or only a handful of assessments that are spaced far apart in time,⁸ fraught with recall bias and limited rigor for drawing any firm conclusions about characteristic patterns of behavior change. Another limitation found in many studies is a sole focus on cigarette smoking at the near exclusion of other tobacco use (for an exception, see ^19–21), which raises questions about the extent to which cancer patients engage in poly-tobacco use as seen in the general adult population.²² Finally, studies that report on rates of abstinence without mention of quit attempts or treatment utilization²³ make it hard to understand the naturalistic process by which individuals strive to achieve tobacco cessation after a cancer diagnosis. For this study, and in the context of a new HNC diagnosis, patients who smoke report their experiences related to smoking (cessation) and other tobacco use (cessation) during a 30-day period of daily assessment. The sole study objective is to describe the nature of key events in the process of smoking cessation after a new HNC diagnosis.

Materials and Methods

Participants

Participants were adults recently diagnosed with HNC, who were cigarette smokers (ie, smoked at least one cigarette in the past 30 days) at the time of study enrollment. Exclusion criteria included: (1) prior cancer diagnosis other than nonmelanoma skin cancer; (2) unreliable phone access; (3) serious cognitive or psychiatric impairment; (4) inability to speak, read, or write in English; and (5) pregnant or planning to become pregnant in the next 6 months.

Procedure

Study procedures, including recruitment and enrollment strategies, are described in detail elsewhere.²⁴ Briefly, patients were prospectively recruited for this observational study from a university-based HNC outpatient clinic in the southeast United States. Apart from the last two participants, there was no standardized or population-based tobacco use screening or treatment program in place at the time of their recruitment in the clinic. Patients were screened for eligibility, and if appropriate, they were informed about study procedures (which included an electronic medical record review) and asked to provide written informed consent and HIPAA authorization. Participants completed a baseline questionnaire via paper form or phone interview and were then scheduled to begin a 30-day daily assessment via a phone-based interactive voice response system, at no financial cost to participants. All participants completed the baseline assessment, and on average, participants completed 22.45 ± 7.94 of the daily assessments. For participants who reported 7-day point prevalence abstinence at the end of the daily assessment, biochemical verification of smoking status was achieved through cotinine analysis. Participants whose cotinine level was ≤15 mg/mL were defined as abstinent, and self-reported abstinence was biochemically confirmed in 92.3% of cases. Participants received a $20 check for the baseline assessment, up to $80 total for the daily assessments ($2 per day plus a $5 bonus for each week with complete data) and $10 for the saliva sample. This study was approved by the University of Kentucky Institutional Review Board (#44421).

Measures

Baseline Assessment

At baseline, participants reported their race, ethnicity, education level, age, relationship status, employment status, and household income on standard items used in survey research.²⁵ Participants also responded to questions regarding their tobacco use and quit attempt history, with questions consistent with those commonly employed in survey research and clinical research with cancer patients (as examples, see ^20,26,27).

Daily Assessment

Six items assessed participants’ tobacco-related behavior, with the first question requiring a numeric response and all other questions involving yes/no responses. Participants were first asked to report the number of cigarettes they smoked on the prior day. Participants who reported abstinence (ie, 0 cigarettes) were then asked if their abstinence was in the context of a successful quit attempt (“Is the reason you didn’t smoke yesterday because you made an attempt to stop smoking for good?”) while participants who reported smoking (ie, 1–35 cigarettes) were asked if their smoking was in the context of an unsuccessful quit attempt (“Did you make an attempt to stop smoking for good yesterday but found yourself smoking anyway?”). After this series of questions, participants were asked if they used any of the following on the prior day: (1) electronic cigarettes, (2) other tobacco products, (3) nicotine replacement therapy, or other medication for tobacco cessation, and (4) behavioral treatments such as quitline services or brief advice for tobacco cessation, with the order of questions, randomly determined each day. Prior studies that employ interactive voice response) technology or ecological momentary assessment of tobacco use have relied on questions like those here, with evidence of strong validity and reliability.^14,28,29

Data Analysis

Descriptive data analysis focused on the prevalence and/or frequency of behavioral events in the process of smoking cessation, including cigarettes per day on smoking days, a 50% reduction in the number of cigarettes smoked per day, any attempt to quit for good (regardless of the length of abstinence that followed), any attempt to quit that last ≥24 hours, 7-day floating abstinence and 30-day point prevalence abstinence.^30–33 Additionally, time to event analysis considered the number of days until these outcomes first occur: Any quit attempt, 24-hour quit attempt, and lapse, and relapse after the first 24-hour quit attempt.³⁴ Finally, descriptive data analysis addressed outcomes related to non-cigarette tobacco use and tobacco treatment use.

Results

Sample Characteristics

Table 1 details the 42 participants’ demographic and clinical characteristics at baseline. Participants’ average age was 57.64 ± 7.34. Participants were predominately male (71.43%, n = 30). With one exception, participants were White non-Hispanic (97.61%, n = 41). Nearly half of the participants were married or partnered (45.24%, n = 19). Half of the participants reported that they could not work due to “disability” (50.0%, n = 21). Many participants reported an annual household income of less than $20 000 (64.29%, n = 27). The most common level of educational attainment was high school or its equivalent (42.86%, n = 18) followed by some college or technical school (28.57%, n = 12). The most common cancer site was larynx (42.86%, n = 18) followed by pharynx (26.19%, n = 11). Over half the sample received a diagnosis of stage IV or metastatic cancer (52.38%, n = 22). Most of the sample received two or more cancer treatment modalities (eg, radiation and chemotherapy) (61.91%, n = 26). On average, participants were diagnosed 1.21 ± 1.69 months prior to enrollment, with a range from 0 to 8 months.

Table 1.

Sample Demographic and Clinical Characteristics at Baseline (n = 42)^a

Variable	% (n)
Race
White, non-Hispanic	97.61 (42)
Other	2.39 (1)
Sex
Male	71.43 (30)
Female	28.57 (12)
Relationship status
Married or partnered	45.24 (19)
Unpartnered	54.76 (23)
Education level
Less than high school	26.20 (11)
High school/GED	42.86 (18)
Some college/technical school	28.57 (12)
College	2.38 (1)
Employment status
Employed	23.81 (10)
Unemployed	11.90 (5)
Disabled	50.00 (21)
Retired	14.29 (6)
Annual household income
Less than $10 000	40.48 (17)
$10 000–19 999	23.80 (10)
$20 000–34 999	19.05 (8)
$35 000 or more	16.67 (7)
Cancer stage
0 or in situ	4.76 (2)
I	9.52 (4)
II	21.43 (9)
III	9.52 (4)
IV or metastatic	52.38 (22)
Treatment type
Surgery only	21.43 (9)
Radiation only	11.90 (5)
Radiation and chemotherapy	33.33 (14)
Radiation and surgery	14.29 (6)
Surgery, radiation, and chemotherapy	14.29 (6)
Missing	4.76 (2)
Age in yearsb	57.64 ± 7.34

Data are frequencies (percentages) unless otherwise noted.

Data are means ± SD.

Tobacco Use Behavior at Baseline

Nearly all participants (90.48%, n = 38) were current smokers at baseline, as defined by 30-day point prevalence; in the lag between enrollment and baseline, the tobacco use status of four participants changed from current smoker to former smoker. Current smokers reported 12.35 ± 11.27 cigarettes per day, with 28.57% (n = 12) smoking their first cigarette within 5 minutes of waking (indicative of high nicotine dependence) and 30.95% (n = 13) smoking their first cigarette between 6 and 30 minutes of waking. Just over one-third (38.10%, n = 16) of participants reported use of at least one non-cigarette tobacco product in the past month, with one participant reporting use of ≥2 non-cigarette tobacco products. This reflects a 30-day point prevalence rate of 28.57% (n = 12) for electronic cigarettes and 9.52% (n = 4) for other products (eg, chew, cigars, snus). Intention and confidence to abstain from all tobacco use for the next 30 days were in the moderate-to-high range (7.20 ± 3.84 and 5.79 ± 3.83, respectively). In the past year, 54.76% (n = 23) of participants reported at least one 24-hour quit attempt, with 50.00% (n = 21), making a 24-hour quit attempt since cancer diagnosis. Finally, 60.87% (n = 14) of participants used medication (eg, nicotine replacement therapy, varenicline) and 78.26% (n = 18) of participants used behavioral treatment (eg, Quitline, group classes) to aid the past year’s quit attempts.

Tobacco Use during the Daily Assessment

Overview

The text below summarizes the outcomes for smoking, other tobacco use, and treatment use.

Smoking

While very few participants reported 30-day continuous abstinence from cigarettes (9.52%, n = 4), a sizeable portion (35.41%, n = 15), reported 7-day floating abstinence. However, only a few participants (19.05%, n = 8), reported 7-day point prevalence abstinence at day 30. On average, participants reported 9.64 ± 11.93 total abstinence days. On smoking days, participants reported 12.35 ± 8.99 cigarettes per day. However, there was great variability in this outcome both across and within participants, as illustrated in Figure 1. For current smokers at baseline, 82.14% (n = 23), achieved at least one day of 50% reduction in cigarettes per day with a 50% reduction observed on 8.21 ± 8.57 of the 30 days. Regarding quit attempts of any duration, nearly all (94.44%, n = 34), participants made at least one quit attempt, with an average of 16.94 ± 11.30 (essentially trying to quit every other day). Fewer participants were able to achieve a 24-hour quit attempt (52.78%, n = 19), with a corresponding average of 5.50 ± 8.69. Figure 2 depicts survival curves for both quit attempt outcomes, showing a median of 1 and 13 days, respectively, prior to the event.

Figure 1.

Cigarettes per day for each individual participant (n = 42).

Figure 2.

Day-to-day smoking behavior after a head and neck cancer diagnosis, with participants ordered by abstinence (n = 42). Each row depicts the outcomes for an individual participant across the 30 days of assessment. White squares indicate smoking abstinence; gray squares indicate smoking on a day of a planned quit attempt; black squares indicate smoking on a day with no planned quit attempt; Xs indicate a day with missing data.

Figure 3 illustrates each participant’s smoking behavior across the 30-day daily assessment. Participants’ day-to-day behavior could be described as falling into one of 5 patterns: First, some participants exhibited “continuous abstaining” (eg, ID 14, 20, 59, and 93), reflective of abstinence every single day. Second, a few participants showed a pattern of “lapsing” (eg, ID 55, 64, and 76), characterized by periods of abstinence, followed by one or more days of smoking, and then a return to abstinence. Third, the most common pattern was “persistent attempting” (eg, ID 16, 23, 24, and 44). This is evidenced by participants reporting a quit attempt every day, or nearly every day, together with the report of smoking every day, or nearly every day—a pattern of frequent, unsuccessful quit attempts. Fourth, some participants demonstrated “continuous smoking,” where they smoked every day with either no quit attempts or only a handful of unsuccessful quit attempts (eg, ID 37, 09, and 07, 63 respectively). Finally, some participants had significant missing data, which could be considered a behavioral pattern itself (eg, ID 10, 19, 25, and 57).

Figure 3.

Survival curves for time to first quit attempt and time to first 24-hour quit attempt.

Other Tobacco Use

Fifty-eight percent of participants reported 30-day continuous abstinence from electronic cigarettes (n = 23), with 19.30 ± 10.63 total abstinence days at the sample level. Sixty-three percent of participants achieved 30-day continuous abstinence from other tobacco use (n = 25), with 21.28 ± 10.05 total abstinence days at the sample level. Considered jointly with the cigarette smoking outcomes, 9.52% (n = 4) of participants achieved 30-day continuous abstinence from all tobacco products and 23.81% (n = 10) of participants achieved 7-day floating abstinence from all tobacco products. On average, participants were abstinent from all tobacco products on 7.74 ± 11.01 days.

Tobacco Cessation Treatment

Less than half the sample reported use of any evidence-based tobacco treatment during the 30-day daily assessment (45.24%, n = 19). Forty percent (n = 17), used medication at least once, with an average of 6.94 ± 7.31 days of use among users. Even fewer participants engaged in behavioral treatment (23.81%, n = 10), with an average of 3.10 ± 4.15 days of use among users. The combination of both medication and behavioral treatment on the same day was rare (11.90%, n = 5; 1.00 ± 3.20 days).

Discussion

The objective of this observational study was to provide a fine-grained, thorough depiction of smoking and other tobacco use after a HNC diagnosis. This study is among the first to employ a daily diary or intensive longitudinal design to explore the dynamic process of cigarette smoking and other tobacco cessation after a cancer diagnosis.¹⁴ Three major findings emerge from this study.

First, continued cigarette smoking following diagnosis was high in terms of both prevalence and amount. Few participants achieved 30-day continuous abstinence from smoking, and on average, participants smoked more than a half pack of cigarettes on smoking days. While the occasional finding of a 50% smoking reduction day may reflect participants’ efforts toward abstinence, as appeared to be the case in a study with adult smokers who intended to quit,³⁵ it also possible this was an unintentional, temporary consequence of the arduous nature of HNC treatment. Electronic cigarette use and non-cigarette tobacco use were also somewhat common in this study, with 38%–43% of participants using these products during the daily assessment. Other recent studies show 3%–25% of cancer patients use electronic cigarettes^19,21 and 1%–4% use other non-cigarette tobacco products.^18,36 It is possible that non-cigarette tobacco products are used as smoking cessation aids, to circumvent smoking restrictions, for harm reduction, or for other, yet unknown reasons.³⁷ As is, this study’s findings on the extent of tobacco product use in the weeks after a cancer diagnosis (and for many, while undergoing cancer treatment) underscore the importance of healthcare providers continuously assessing for any and all forms of tobacco use among cancer patients and being prepared to treat those who are current users.³⁸

Second, nearly all (94%) of the HNC patients in this study made at least one quit attempt, but few (53%) achieved a 24-hour quit attempt, and even fewer (35%) achieved 7 or more days of continuous abstinence. For comparison, the results of two 28-day daily assessment studies with adult smokers in the general population who intended to quit in the next 3 months showed 75%–84% of participants made a quit attempt, 48%–69% of quit attempts lasted less than 24 hours, and 56%–72% of quit attempts were not preceded by an intention to quit.^35,39 Unfortunately, the most common behavioral pattern among participants is that of “persistent quitting,” defined by a series of unsuccessful quit attempts. It appears that frequent, perhaps spontaneous or unplanned, unassisted quit attempts are common among smokers in both cancer and noncancer healthy samples.^14,35 The high number of quit attempts found here could mean participants needed practice. However, it is far more likely that the manner in which participants tried to quit in addition to other, psychosocial factors were impediments to their ability to quit and maintain abstinence.^20,40

The third and final major finding is that only 40% of participants reported any evidence-based treatment to aid their attempts. The occurrence of largely unassisted quit attempts is a common finding, both in cancer patient samples^41,42 and in the general population of smokers.^43,44 Less than half the sample used medication (eg, nicotine replacement therapy), and less than one-quarter used behavioral treatment (eg, Quitline, group counseling), with prevalence rates that resemble those from a recent population-based study of cancer patients (38% and 8%, respectively).⁴² Further, the HNC patients who used treatment did so only sparingly, with treatment occurring on average only 7 of 30 days. Low treatment use may be accounted by strong determination to quit on one’s own; low health literacy, education, or lack of appreciation for the safety and effectiveness of treatment; insufficient insurance coverage coupled with the high, albeit short-term cost of some treatment, and system-level issues related to trouble navigating the healthcare system, a lack of standardized assessment and intervention, and inequities in treatment offers and accessibility.^45,46 Indeed, past research indicates that less than half of oncologists assess willingness to quit, assist with tobacco cessation, or arrange for follow-up of their patients.⁴⁷ Importantly, the National Cancer Institute (NCI) is working to address some of these barriers to care through its Cancer Center Cessation Initiative, wherein 52 NCI-designated cancer centers received funding to create or expand upon existing tobacco treatment programs, with the immediate goal of providing cancer patients access to evidence-based tobacco treatment.⁴⁸

All study outcomes must be considered alongside the study’s major limitations. First, this study observed a higher incidence of missing data than in daily assessment or other intensive longitudinal studies with nonclinical samples,^35,49 most likely due to the physical and mental health demands (eg, fatigue, emotional distress) and practical issues (eg, hospitalization) associated with a cancer diagnosis and its treatment. Second, social desirability may have biased participant responses. If this were the case, however, one would expect a higher rate of abstinence, greater occasions of smoking reduction, and more comprehensive treatment utilization. Given the overall behavioral pattern, social desirability is possible, but not absolute. Third, this study does not provide any insight into linkages between cancer patients’ tobacco use and treatment regiments, even though some cancer treatments could have a direct impact on tobacco use behavior. Fourth, this study focuses exclusively on HNC patients. Much of the cancer survivorship research on this topic centers on HNC and/or lung cancer patients,⁹ and while there is no reason to believe the results would not generalize to other cancer patients, the lack of heterogeneity in the disease site is a limitation. Finally, as participants were recruited from a clinical setting that did not have a robust or standardized tobacco treatment program, it is possible the study results do not generalize to other clinical settings where the standard of care is different.

Even with its limitations, this rigorous intensive longitudinal study gives new insight into the dynamic nature of tobacco use among patients recently diagnosed with HNC. Results indicate that many cancer patients may continue to smoke cigarettes and/or use non-cigarette tobacco products during and shortly after curative treatment. Furthermore, although quit attempts were common in this sample, few participants were able to maintain abstinence for more than 24 hours, which may be partly attributable to minimal use of evidence-based treatment. Due to the adverse effects of persistent tobacco use post-cancer diagnosis, it is imperative cancer patients quit immediately and stay tobacco-free.⁶ Thus, it is imperative that there is an increase in the delivery of evidence-based tobacco treatment to HNC patient populations. While increasing access to evidence-based tobacco treatment may spur increases in motivation to quit and increase the number of quit attempts made, it remains important to understand fully and intervene properly on any psychosocial impediments to maintaining abstinence after a quit attempt, which is an area ripe for future research.

Supplementary Material

A Contributorship Form detailing each author’s specific involvement with this content, as well as any supplementary data, are available online at https://academic.oup.com/ntr.

Funding

Research reported in this publication was supported by the National Cancer Institute of the National Institutes of Health under award number K07 CA181351, the Patient-Oriented and Population Sciences Shared Resource Facility, and Biostatistics and Bioinformatics Shared Resource Facility of Markey Cancer Center under award number P30 CA177558, and the National Center for Advancing Translational Sciences of the National Institutes of Health under award number UL1 TR001998. The content of this publication is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Declaration of Interests

None declared.

Acknowledgments

The authors would like to acknowledge the assistance of Caitlin Dunworth, MPH and Joan Kahl, MS who helped with study coordination.

Data Availability

Deidentified data used and analyzed during the current study are available from the corresponding author upon reasonable request.