Tobacco Quitline Engagement and Outcomes among Primary Care Patients Reporting Use of Tobacco or Dual Tobacco and Cannabis: An Observational Study

Abstract

Background:

Dual use of tobacco and cannabis is increasingly common, but it is unclear how it impacts individuals’ interest in or ability to stop smoking. If dual users fail to engage in treatment or have worse treatment outcomes, it would suggest that tobacco treatment programs may need to be tailored to the specific needs of those using cannabis and tobacco.

Methods:

We conducted an observational study using electronic treatment records from adults (18 years and older) who (a) were enrolled in a regional healthcare system in Washington state, (b) sought tobacco cessation treatment through an insurance-covered quitline from July 2016 to December 2018 and (c) had cannabis use frequency during the period of their quitline enrollment documented in their electronic health record (EHR) (n=1,390). Treatment engagement was defined by the total number of quitline counseling calls and web-logins completed. Point prevalent self-reported tobacco abstinence was assessed 6 months post-quitline enrollment.

Results:

Thirty-two percent of participants (n= 441) reported dual use of tobacco and any cannabis during the observation period; 9.4% (n=130) reported daily cannabis use. Among dual users reporting daily cannabis use, 13.9% had a diagnosed cannabis user disorder in the EHR. Neither engagement with quitline counseling nor long-term tobacco abstinence rates differed between those using tobacco-only and either dual-use group (i.e., persons using any cannabis or daily cannabis).

Conclusions:

Dual use of tobacco and cannabis is common among smokers seen in primary care and those enrolling in quitline care, but it may not undermine tobacco quitline engagement or smoking cessation. Opportunities exist in the US to leverage quitlines to identify and intervene with dual users of tobacco and cannabis.

1. Introduction

Tobacco use is the leading preventable cause of death and illness in the US, killing nearly half a million people annually.¹ Cannabis is the most commonly used illicit drug and the third most commonly used drug (licit or illict) among adults in the US, after tobacco and alcohol.² Given this, it is not surprising that tobacco and cannabis use are frequently co-morbid.³ In fact, in the US daily cannabis use occurs predominantly among cigarette smokers. An estimated 9% of daily smokers and 8% of non-daily smokers use cannabis daily, compared to 3% of former smokers and 1% of never smokers.⁴ Dual-use is also common among primary care patients in Washington state where adult cannabis use has been legal since 2012.⁵ Among patients offered behavioral health screening as part of routine primary care, including a single-item screen assessing the frequency of past-year cannabis use, 10% reported current tobacco use and 15% reported any cannabis use in the past year; but 32% of self-reported tobacco users endorsed past-year cannabis use and 10% reported daily use.^6,7 Moreover, 30% of patients who report daily cannabis use also report current tobacco use.

High rates of concurrent tobacco and cannabis use are concerning. Dual use can cause substantial adverse health effects due to the individual and additive impacts of each substance.^8,9 Additionally, cannabis use is associated with an increased risk of and greater levels of nicotine dependence,^10,11 which may explain why some cross-sectional and longitudinal surveys have suggested that tobacco smokers who use cannabis may have a harder time quitting smoking.^12–14 However, a systematic review of the literature concluded that dual use of tobacco and cannabis is not consistently associated with worse smoking cessation outcomes.¹⁵ Thus, dual use of these substances may not necessarily be an impediment to quitting tobacco. Whether it is may be dependent, at least in part, on the type of cessation treatment utilized or if treatment is used at all.

The current study examined whether primary care patients who report concurrent use of tobacco and cannabis have differential treatment engagement or different treatment outcomes compared to people who are not using cannabis when both groups are enrolled in a tobacco quitline through their medical insurance. This is an important issue because quitlines are a vital cornerstone of tobacco cessation services in the US, reaching hundreds of thousands of people each year through both state and commercially-funded programs.¹⁶ If differential engagement or outcomes are observed, it could suggest that quitlines should include tailored content to address the unique needs of dual tobacco and cannabis users.

More specifically, the aims of this study were to understand (a) the prevalence of cannabis use and cannabis use disorders among primary care patients who smoked and voluntarily enrolled in a commercially-funded, tobacco quitline program and (b) whether quitline treatment engagement or smoking cessation outcomes differed between individuals who were identified as dual cannabis and tobacco users compared to tobacco-only users.

2. Methods

2.1. Setting and Participants

Kaiser Permanent Washington (KPWA) is a non-profit, healthcare system which provides medical care and/or insurance coverage to approximately 700,000 Washington state residents. Participants were health plan members who enrolled in Quit for Life, an empirically-validated tobacco cessation quitline^17,18 run by Optum Health, the leading US provider of quitline services. Adult health plan members aged 18 and older can self-refer to the quitline at no out-of-pocket cost. The program includes up to 5 proactive counseling calls initiated by quitline counselors and use of a web-based program. Enrollees can also call-in for an unlimited number of additional counseling calls.

Adult health plan members were included if they: 1) enrolled in the quitline between July 2016 and December 2018; 2) had seen a health plan primary care provider from −30 to +365 days from the date of their quitline registration; and 3) completed an annual past-year cannabis use screen within this 13-month observational window. This timeframe was chosen to ensure that past-year cannabis use was concurrent with the period people were enrolled in the quitline. People were excluded if they had requested their health care records not be used for research (n = 13), were health plan employees (n = 38), died during the observational window (n = 16), or provided an invalid health plan number or enrolled using their spouse’s identification (n = 123). If an individual had more than one quitline registration associated with a cannabis screen during the observational window, one was randomly chosen for inclusion. If there was more than one cannabis screen, the one most proximal to the quitline registration date was chosen. The final analytic sample included 1,390 adults.

Routine assessment of cannabis use frequency was phased into all health plan primary care clinics from 2015 to mid-2018. Thus, some quitline registrants during the observational window had a primary care visit, but no cannabis screen. Others had no primary care visit during the observation window. Collectively, these individuals (n = 1,085) were examined as a separate cohort to assess how well the demographics and quitline outcomes among registrants with unknown cannabis use compared to those observed in the analytic cohort which included people with documented cannabis use frequency, allowing us to assess whether the analytic cohort appeared meaningfully different from other quitline callers.

2.2. Measures and Data Sources

Participants’ identification, type of tobacco used, and stage of readiness to change at quitline registration were obtained from electronic quitline service records. Participants’ demographics, mental health and cannabis use disorders (based on International Classification of Disease 9/10 codes),⁶ unhealthy alcohol use (based on AUDIT-C ≥ 3 points for women and ≥ 4 for men),¹⁹ and cannabis use were obtained from electronic health record (EHR) data. The latter was assessed via a single-item question which asked, “How often in the past-year have you used marijuana?”²⁰ with response options: “none,” “less than monthly,” “monthly,” “weekly,” and “daily or almost” as consistent with that used by the World Health Organization’s Alcohol, Smoking, and Substance Involvement Screening Test (ASSIST) questionnaire.²¹ On average, the date of the past-year cannabis use screen was assessed within 108 days of participants’ quitline registration date (SD = 112).

Study outcomes (engagement with quitline services [number of counseling calls completed, number of web-logins]) and smoking status) were obtained from quitline service records. Self-reported tobacco abstinence was assessed by phone at 6-months post-registration per Optum’s standard service protocol. Individuals were considered abstinent if they reported not using tobacco in the past 7 or 30 days (point prevalent abstinence; PPA).

All study activities were reviewed and approved by the Kaiser Permanente Washington Institutional Review Board.

2.3. Analyses

Participants were categorized into three non-exclusive groups: tobacco users only (i.e., no cannabis use), dual tobacco and any past-year cannabis use (daily and non-daily use), and dual tobacco and daily cannabis users.

Descriptive statistics were used to characterize the analytic and comparison cohort samples and unadjusted chi-squares used to compare the two groups. Chi-square and logistic regression were used to compare outcomes between tobacco users and each dual user group in the analytic sample, controlling for age and gender. Abstinence was examined among responders-only at follow-up (n = 829 [87.4%] of tobacco users) and n= 403 [91.4%] of all dual users in the analytic cohort) and among all participants, with missing values imputed as not abstinent. Web-login data was only available for participants enrolled after June 2018 (n = 296 participants in the analytic cohort and n = 194 in the comparison cohort), so analyses of this variable were limited to these subsamples.

3. Results

3.1. Participants

Participant characteristics are included in Table 1. Most were middle-aged (mean = 52.1 years), female (57.3%), and white (80.7%). The majority (68.3%) used only tobacco, with 31.7% reporting dual use of tobacco and any cannabis, and 9.4% reporting dual tobacco and daily cannabis use. Five percent of all cannabis users and 13.9% of daily users met criteria for cannabis use disorder.

Table 1.

Participant characteristics

	All Participants	Tobacco Use Only	Dual Tobacco/ Cannabis Use
			Any cannabis1	Daily cannabis2
	N = 1,390	N = 949	N = 441	N = 130
		n (%)	n (%)	n (%)
Female	796 (57.3)	565 (59.5)	231 (52.4)	65 (50.0)
Race
White	1085 (80.7)	736 (79.8)	349 (82.7)	104 (83.2)
Non-white	214(15.9)	156 (16.9)	58 (13.7)	17 (13.6)
Multi-race	45 (3.4)	30 (3.3)	15 (3.6)	4 (3.2)
Hispanic	58 (4.3)	35 (3.8)	23 (5.5)	5 (4.0)
Age [mean (SD)]	52.1 (13.5)	53.5 (13.2)	49.1 (13.8)	47.4 (14.1)
Tobacco Product Used
Cigarettes	1296 (93.2)	882 (92.9)	414 (93.9)	120 (92.3)
Other (cigar, pipe, etc.)	125 (9.0)	88 (9.3)	37 (8.4)	12 (9.2)
Unhealthy alcohol use3	375 (27.4)	219 (23.4)	156 (36.1)	44 (35.2)
Cannabis use disorder4	32 (2.3)	9 (1.0)	23 (5.2)	18 (13.9)
Mental health diagnosis4
Depression	434 (31.2)	293 (30.9)	141 (32.0)	42 (32.3)
Anxiety	362 (26.0)	233 (24.6)	129 (29.3)	47 (36.2)
Serious mental illness5	99 (7.1)	62 (6.5)	37 (8.4)	11 (8.5)
Stage of change6
Precontemplation	2 (0.1)	1 (0.1)	1 (0.2)	0
Contemplation	30 (2.2)	21 (2.2)	9 (2.0)	5 (3.9)
Preparation	1236 (88.9)	841 (88.6)	395 (89.6)	114 (87.7)
Action	50 (3.6)	35 (3.7)	15 (3.4)	2 (1.5)
Maintenance	9 (0.7)	6 (0.6)	3 (0.7)	0
Unknown/missing data	63 (4.5)	45 (4.8)	18 (4.1)	9 (6.9)

¹Tobacco users reporting any past-year cannabis use on one item screener (daily or non-daily).

²Tobacco users reporting daily cannabis use in the past year on one item screener.

³Based on AUDIT-C scores ≥ 3 points for women and ≥ 4 for men.

⁴Based on International Classification of Disease 9/10 diagnostic codes in electronic health record.

⁵Bipolar disorder, schizophrenia, or other psychosis.

⁶Stage of readiness to quit smoking tobacco.

3.2. Tobacco Cessation Program Engagement

Participants completed an average of 1.9 counseling call (SD = 1.6) calls and 0.2 (SD = 0.6) web-logins. Quitline treatment engagement did not differ between people using tobacco and those using dual tobacco and cannabis (see Table 2).

Table 2.

Quitline treatment engagement and tobacco abstinence

	Tobacco Use Only	Dual Tobacco/Cannabis Use		P1	P2
		Any Cannabis	Daily Cannabis
	n = 949	n = 441	n = 130
Treatment Engagement	Mean (SD)	Mean (SD)	Mean (SD)
Total counseling calls	1.9 (1.6)	1.8 (1.5)	1.9 (1.6)	0.28	0.44
Total Web logins3	0.18 (0.69)	0.18 (0.50)	0.31 (0.66)	0.11	0.04
Tobacco Abstinence	%	%	%
7-day PPA 4	19.3	21.3	19.5	0.58	0.96
7-day PPA 5	16.9	19.5	17.7	0.38	0.81
30-day PPA 4	18.0	18.9	17.0	0.96	0.78
30-day PPA 5	15.7	17.2	15.4	0.72	0.93

PPA = point prevalent tobacco abstinence assessed 6 months post-quitline registration.

¹Comparison of tobacco only users vs. all dual users controlling for age and gender.

²Comparison of tobacco only users to daily cannabis dual users controlling for age and gender.

³Web login analyses were limited to individuals who registered with the quitline after June 2018 and had log-in data available (Tobacco Use Only n = 213; Dual Tobacco/Any Cannabis n = 54; Dual Tobacco/Daily Cannabis n = 29).

⁴Responder-only sample limited to people who completed 6-month assessment (n = 1,232).

⁵Complete analytic cohort (n = 1,390) with missing values imputed as not abstinent from tobacco.

3.3. Tobacco Abstinence

At follow-up, 20.0% of participants in the responder only sample and 17.7% of those in the complete analytic sample (where missing values were imputed as not abstinent) reported 7-day PPA. However, abstinence rates did not differ between tobacco users and people using dual use of tobacco and cannabis for either 7 or 30-day PPA (see Table 2).

3.4. Representativeness of Analytic Cohort

Participants in the analytic cohort were similar to those in the comparison cohort in terms of race (78.1% White vs. 66.6%, P = .57), ethnicity (4.2% Hispanic vs. 3.3%, P = .27), and gender (57.3% female vs. 53.8%, P = .09), but they were older (mean age 52.1 [SD = 13.5] vs. 49.9 [SD = 13.7]; P < .0001). Since participants in the comparison cohort did not have a cannabis screen in their EHR and some had no primary care visit data during the observation window, rates of cannabis use disorders, alcohol use, or mental health diagnoses could not be compared, but there was no difference between the cohorts’ quitline engagement metrics; the average number of counseling calls completed was 1.9 (SD = 1.5) in the analytic sample vs. 1.8 (SD = 1.5) in the comparison cohort, P = .37. Mean web logins were 0.2 [SD = 0.6] in the analytic cohort and 0.3 [SD = 0.8] in the comparison cohort, P = .18. Abstinence outcomes were also similar (e.g., 7-day PPA: 20.0% in the analytic cohort vs. 19.9% in the comparison cohort (P = .97) among responders only and 17.7% vs. 18.1% (P = .81) when missing values were imputed as smoking.

4. Discussion

This study represents the first examination of real-world, clinical care data to inform the prevalence of dual cannabis and tobacco use among tobacco cessation treatment-seekers and the first to assess whether tobacco quitline engagement and cessation outcomes differ between people reporting use of tobacco only versus dual use of tobacco and cannabis. We found that a significant proportion of treatment seeking smokers reported any past-year cannabis use (31.7%) and daily use (9.4%). The observed rates are similar to those recently reported among other smokers,^4,22 including those in the same health plan population who were audited using an identical methodology for assessing cannabis use (32% any cannabis use, 10% daily use).⁶ The latter similarity suggests that primary care patients who smoke and use cannabis may be no less likely to seek tobacco cessation treatment than are those who do not use cannabis. Otherwise, one would expect cannabis use to have been lower among quitline treatment-seekers than in the general patient population from which they were sampled. While we cannot definitively assert dual users are equally interested in tobacco cessation as other tobacco users, we can conclude that many dual users are interested in quitting tobacco, despite their cannabis use. This is consistent with a recent national survey which found that 55% of people using both tobacco and cannabis had made an attempt to quit tobacco in the past year.²³

Once enrolled in treatment, we found no evidence that people using tobacco and cannabis had differential levels of treatment engagement or treatment outcomes. Both groups (tobacco users and dual tobacco and cannabis users) completed approximately 2 counseling calls, on average, and their tobacco abstinence rates were similar. In fact, abstinence rates did not differ by more than 2 percentage points between groups (Table 2), indicating the observed differences were not clinically meaningful, in addition to not being statistically significant. This is important, is consistent with a prior systematic review which concluded that smokers who use cannabis may not have a harder time quitting smoking,¹⁵ despite some suggestion to the contrary.^12–14

Finally, it is worth noting that nearly 14% of people reporting daily cannabis use and 5% of all people reporting cannabis use met diagnostic criteria for a cannabis use disorder. Taken together, the results of this study underscore the rich opportunity that tobacco quitlines have to intervene with people using both tobacco and cannabis, both to promote tobacco cessation and to address cannabis use/misuse. The latter is important given the potential additive adverse health effects of dual tobacco and cannabis use^8,9; public health concerns associated with cannabis use, such as increased risk of motor vehicle accidents²⁴; and the quitlines widespread population reach in the US.

4.1. Generalizability of Cohort and Findings

We did not randomly sample participants or assign them to treatment in this observational study. As such, there is a possibility that the sample may not be representative of tobacco users within the primary care population from which the participants were drawn; however, this does not appear to have been the case. Participants in the analytic cohort (whose cannabis use was documented, n = 1,390) had similar demographics to people in the comparison cohort (whose cannabis use was unknown, n = 1,085) and there were no differences between the cohorts’ quitline engagement metrics or abstinence outcomes. Thus, limiting the sample to people with known cannabis use status does not appear to have biased the sample in a meaningful way.

It is also theoretically possible that the analytic cohort’s treatment engagement and outcomes may not be representative of the outcomes one might observe among other quitline callers, such as those enrolled in a publicly funded quitline or recruited from other geographic regions. Unfortunately, no published studies we are aware of have examined the relation between cannabis use and tobacco cessation outcomes among quitline participants, but researchers at Optum Health (the provider of the quitline program in this study) conducted an unpublished study using data collected from callers to their state-funded quitline program in Alaska, Oregon, and Washington, DC in 2016. This study also found no difference in treatment engagement between smokers and dual tobacco and cannabis users. Among people who were eligible to receive the multi-call quitline program (n = 450), current marijuana users completed an average of 2.4 calls (SD = 1.3) compared to 2.3 calls (SD= 1.6) for non-users (P = 0.60; personal communication, Terry Bush, PhD, Optum Health, October 24, 2019). In other words, in both studies, callers completed an average of about 2 calls. We cannot compare abstinence rates, as this was not evaluated in Optum’s study.

The average number of completed counseling calls in the present study was also similar to that reported among a large, geographically diverse sample of smokers enrolled in the same quitline program (n = 11,143).¹⁷ In this nationwide sample, people completed an average of 2.1 calls (SD = 1.6) and logged into the web program once (mean = 1.1, SD = 1.9). Seven day point prevalent tobacco abstinence was 21% when missing values were imputed as smoking, compared to 18% when using a similar methodology in the current study.

4.2. Caveats, Limitations and Strengths.

Observational studies, by nature, lack the rigor to be considered confirmatory. Consequently, we cannot firmly conclude that cannabis use is not associated with differential tobacco treatment engagement or cessation, but the consistency of our findings with the prior research cited increases confidence that the present study’s results are representative and generalizable. That said, it is possible that the study findings could differ if another tobacco cessation program were being evaluated, particularly one with different engagement requirements such as ongoing, in-person participation or online interaction with no personal contact. It is very possible that the similar levels of engagement seen across groups in this study are a by-product of the proactive nature of the quitline program (i.e., follow-up counseling calls are initiated by the quitline not the participant). The lack of biochemically confirmed abstinence in this study is a limitation, but it is not inconsistent with best practice recommendations for assessing abstinence in large, geographically diverse populations and when there is no face-to-face contact ²⁵.

Finally, the pragmatic, observational design of this study is a strength. Because all cessation treatment was delivered under real-world conditions, the findings are not subject to traditional research biases (e.g., Hawthorne effect²⁶, enrollment bias) and are, therefore, more likely to generalize to real world conditions, as suggested by the similar outcomes to other quitline studies cited above.

5. Conclusion

The results of this observational study support four key conclusions. First, dual use of tobacco and cannabis is common among tobacco users in Washington state: 31.7% of those sampled in this study endorsed use of both. Second, people who report use of both tobacco and cannabis appear equally likely to seek tobacco cessation treatment as tobacco users with no cannabis use. Third, once dual substance users commit to quitline care, they do not appear to be any less engaged or have worse treatment outcomes than other tobacco users. Finally, tobacco quitlines are well-positioned to proactively identify and intervene with dual users of tobacco and cannabis, which could help reduce the health burdens caused by concomitant use of these substances and provide an opportunity to address cannabis dependence and misuse.

Highlights.

• Tobacco quitline care was equally engaging and effective among tobacco users and dual users of tobacco and cannabis

• Many daily cannabis users calling tobacco quitlines likely have a cannabis use disorder

• Tobacco quitlines can be leveraged to identify and intervene with dual users of tobacco and cannabis