Outcomes Among People With Schizophrenia Participating in General-Population Smoking Cessation Treatment: An Observational Study

Abstract

French

Objective

People with schizophrenia are much more likely than others to smoke tobacco, raising risks of disease and premature mortality. These individuals are also less likely to quit successfully after treatment, but the few existing clinical and observational studies have been limited by small sample sizes, and have generally considered specialized treatment approaches. In this analysis, we examine outcomes, service use, and potential explanatory variables in a large sample of people with schizophrenia treated in a general-population cessation program.

Method

Our sample comprised 3,011 people with schizophrenia and 77,790 controls receiving free nicotine replacement therapy through 400 clinics and health centres. We analysed self-reported 7-day abstinence or reduction at 6-month follow-up, as well as the number of visits attended and self-reported difficulties in quitting. We adjusted for demographic, socioeconomic, and health variables, and used multiple imputation to address missing data.

Results

Abstinence was achieved by 16.2% (95% confidence interval [CI], 14.5% to 17.8%) of people with schizophrenia and 26.4% (95% CI, 26.0% to 26.7%) of others (absolute difference = 10.2%; 95% CI, 8.5% to 11.9%; P < 0.001). After adjustment, this difference was reduced to 7.3% (95% CI, 5.4% to 9.3%; P < 0.001). Reduction in use was reported by 11.8% (95% CI, 10.3% to 13.3%) and 12.5% (95% CI, 12.2% to 12.8%), respectively; this difference was nonsignificant after adjustment. People with schizophrenia attended more clinic visits (incidence rate ratio [IRR] = 1.15, 95% CI = 1.12% to 1.18%, P < 0.001) and reported more difficulties related to “being around other smokers” (odds ratio [OR] = 1.28; 95% CI, 1.11% to 1.47%; P = 0.001).

Conclusion

There is abundant demand for tobacco cessation treatment in this population. Outcomes were substantially poorer for people with schizophrenia, and this difference was not explained by covariates. Cessation remained much better than for unaided quit attempts, however, and engagement was high, demonstrating that people with schizophrenia benefit from nonspecialized pharmacological treatment programs.

Keywords

tobacco nicotine replacement therapy treatment engagement

Introduction

Cigarette smoking is highly prevalent among people with schizophrenia. In North America, approximately 62% of schizophrenia patients smoke tobacco, compared to 15% of the population without a psychiatric condition.^1,2 People with schizophrenia also tend to smoke more cigarettes per day and to be more dependent on tobacco.³ Premature mortality among people with schizophrenia is very high, and this is partly due to a heavy burden of tobacco-related disease.^4-8

Explanations of the high prevalence of tobacco use in this group include biological, psychosocial, and systemic factors. People with schizophrenia experience dysfunction in neural networks involving dopamine, glutamate, and γ-aminobutyric acid, which may be partially remediated by activation of nicotine acetylcholine receptors during smoking.⁹ Tobacco smoke also expedites the metabolism of antipsychotic medications and people who smoke therefore have a reduced blood concentration of antipsychotic medications. This may reduce antipsychotics’ side effect burden, but also often necessitates an increase in medication doses.¹⁰ Due to illness-related functional difficulties and gaps in educational attainment, people with schizophrenia are also overrepresented among low-wage workers, who have a higher prevalence of tobacco use.¹¹ Many also report that smoking acts as a “social lubricant,” facilitating interpersonal interactions in work and leisure settings; smoking may thus be linked to a sense of belonging and quitting to fear of isolation,^12,13 especially in light of the stigma and marginalization that affect people with schizophrenia.

Mental health systems have also long had smoking-permissive cultures, due to the historical use of cigarettes to control inpatients,¹⁴ efforts by the tobacco industry to promote smoking among people with mental illness,¹⁵ misconceptions regarding potential mental health benefits of smoking,¹⁴ and clinicians’ doubts about patients’ ability to quit.¹⁶ Diagnostic overshadowing of tobacco use by what seem to be more pressing mental health concerns may also lead clinicians to assign smoking cessation a lower priority for people with severe mental illness.^13,17 Treatment is also often difficult to afford or access for people with schizophrenia, who often have low incomes and sometimes struggle to obtain stable housing.^18-20 This also means, however, that the financial burden of smoking is often acute for these individuals, and that cessation offers potential additional improvements in quality of life.²⁰

A recent meta-analysis of observational studies showed a lower probability of cessation among people with schizophrenia than among either healthy controls (odds ratio [OR] = 0.45) or people with other psychiatric conditions (OR = 0.79).² The authors, however, noted significant publication bias, a small number of included studies, small sample sizes, and a significant association between reported cessation rate and study quality. Difficulties in quitting may be related to more severe symptoms of nicotine withdrawal, which include irritability, loss of focus, anxiety, and impatience.¹³ Given the effects of smoking on the metabolism of antipsychotic medications,¹⁰ increased medication side effects may also occur if smoking is reduced. Finally, the same factors that contribute to the increased prevalence of smoking in this population may also account for some differences in treatment outcome. These include low levels of income, education, and occupational attainment, and high levels of comorbid mental health and substance-related conditions,^18,21-23 all of which have been shown in general-population cohorts to be associated with lower probabilities of successful cessation after treatment.^24,25

Evidence supports the use of pharmacological treatments for smoking cessation among people with schizophrenia.^12,26,27 These treatments include the prescription medications varenicline and bupropion, as well as nicotine replacement therapy (NRT), which is generally available over the counter. Varenicline and bupropion have been shown not to cause important neuropsychiatric side effects among people with psychiatric conditions,²⁸ although evidence for people with psychotic conditions specifically is less clear. NRT has a long history of use and a well-established safety profile, even if used while continuing to smoke.²⁹ Limited evidence from trials suggests that behavioural therapy and lifestyle changes may be less effective in this population,^12,30 although this question remains somewhat uncertain, due to heterogeneity across studies and the widespread practice of incorporating both psychosocial and pharmacological treatments into interventions, which continues to be the standard of care.³¹

In this study, we consider outcomes after NRT-based smoking cessation treatment among people with schizophrenia. NRT alleviates withdrawal symptoms and reduces cravings by replacing nicotine from cigarettes. It is delivered via long-acting transdermal nicotine patches and/or shorter-acting forms such as lozenges, gum, inhalers, and nasal sprays.³² NRT does not require a prescription, and can thus be provided by nurses and counsellors in community health centres, addictions agencies, and other contexts. Given the cost, and sometimes scarcity, of physician treatment, this makes it an important option for public health programs, and may be particularly valuable in the case of people with schizophrenia, for whom nonpharmacological interventions have shown unclear effectiveness.

In the general population, NRT increases the probability of quitting successfully by 50%, more when the patch is paired with a short-acting form.^32,33 Few clinical trials of NRT have been conducted among people with schizophrenia, and sample sizes have been very small.^34,35 The one larger clinical trial of patients with psychotic disorders (57% with schizophrenia) that exists failed to detect a significant effect on cessation, but reported an effect estimate consistent with that in the general population (OR = 1.72), and observed a significant effect on smoking reduction.

Our primary goal in the present analysis was to measure cessation and reduction outcomes among people with schizophrenia and to compare these outcomes to those among other patients. Our secondary goal was to evaluate factors that may account for some differences in outcomes. By using the largest cohort to date of people with schizophrenia treated for smoking cessation, we aimed to measure differences and evaluate potential effect modifiers with greater statistical power than has previously been available.

Methods

Sample Population

We drew data from the Smoking Treatment for Ontario Patients (STOP) program. STOP provides up to 26 weeks of free NRT and behavioural counselling, and operates through approximately 400 partnered clinics, which include team-based primary care (family health teams, community health centres, and nurse practitioner-led practices) and addiction treatment agencies. The STOP program recommends that providers supply both the nicotine patch and one short-acting form, with doses adjusted for the heaviness of smoking and types supplied according to patient preference. Participants are encouraged to set a quit date, but the long treatment period supports the use of NRT before cessation and for reduction goals, as well as for relapse or repeated quit attempts. Most clinicians providing care through the STOP program are nurses, pharmacists, social workers, or counsellors. STOP is funded by the Ontario Ministry of Health, with ethics approval from the Research Ethics Board at the Centre for Addiction and Mental Health (protocol numbers 058-2011 and 154-2012). Neither participants nor participating sites receive financial incentives to participate.

STOP enrolls 15,000 to 20,000 people each year, and participants report, on average, lower levels of income and education and a higher prevalence of health conditions than the general population.³⁶ Participants are typically seen every 2 to 4 weeks. At each contact, providers, supply more NRT as needed and collect visit-specific data. Cessation outcomes are assessed through interviews conducted 3, 6, and 12 months after the initial baseline visit. The program uses the 6-month interview as its primary outcome and makes the most intensive efforts to obtain data at this time. Some follow-up assessments are completed during a clinical appointment, but most participants are contacted by email or phone. Given the program size and nature of follow-ups, biochemical confirmation was not feasible. However, the validity of self-reported smoking status has been shown to be good in the general population^37,38 and among people with psychotic disorders.²⁶

We included enrolments between April 11, 2016, and November 30, 2021. This start date was chosen because the baseline questionnaire was expanded at this time to include important covariates, while the end date allows for inclusion of all participants who were eligible for a 6-month outcome assessment at the time of analysis. There were 118,014 enrolments within the study period. Of these, we removed 2,288 (1.9%) without signed consent to use the data for research purposes, 620 (0.5%) that had incomplete baseline surveys, 8,556 (7.2%) who were not daily smokers at baseline, 1,038 (0.9%) who had no recorded visits, and 195 (0.2%) who did not report their baseline smoking level. Participants can also reenroll in STOP after one year. In order to ensure that observations were independent, we used probabilistic deduplication to include only the first enrolment for each person. This meant removing a further 19,958 (16.9%) enrolments, leaving an analysis sample of 85,359. We identified people with schizophrenia in the cohort if they reported that they had “ever been diagnosed with” schizophrenia. All others were assigned to the control/comparison group.

Outcomes

For our primary outcome, we created a 3-level outcome variable using data from the 6-month (6M) follow-up: self-reported 7-day tobacco abstinence (a response of “no” to the question, “have you smoked a cigarette, even a puff, in the last 7 days?”); reduction (decrease of 50% or more in cigarettes per day or a transition from daily to occasional smoking); and continued smoking (all others). We then examined changes in the measured association between these outcomes and schizophrenia after adjustment for different blocks of covariates. To more fully address our secondary research question, we also examined difficulties experienced during the quit attempt. Participants were asked whether they had experienced any of these difficulties during the 6M follow-up interview. We created one outcome by summing positive responses to items asking about “depression,” “anxiety,” “stress,” and “boredom”; and separately analysed the presence or absence of difficulties related to “being around other smokers.” Finally, to more closely examine differences in service use, we compared the number of clinical visits attended in the 6M following enrolment.

Analysis

To examine the importance of site-level differences in outcomes, we first fit a multilevel model including a random intercept for the clinical site. As the intercept-level variance was small, we used single-level models in our analysis. For our primary outcome (quit vs. reduced vs. continued smoking), we used multinomial logistic regression. In this type of model, coefficients for other levels of the outcome reflect comparisons to the base level and thus depend on the proportion of subjects at that level. We therefore calculated average adjusted predictions—adjusted probabilities of quitting, reducing, and continuing to smoke—for people with and without schizophrenia. As these models yield a coefficient for each level of the outcome, we also used postestimation tests to obtain an overall p-value for the schizophrenia indicator variable.

The analysis for our primary outcome had two aims: to measure the association between schizophrenia and outcome independent of potentially important covariates and to gauge the importance of other variables in modifying this association. For the latter purpose, we fit separate models for each of the 6 blocks of variables, as well as unadjusted and fully adjusted models. In Model 1, we included only age and gender. In subsequent models, we included age and gender plus a set of additional variables: (1) smoking history and severity of tobacco dependence (age began daily smoking, cigarettes smoked per day at baseline, time to first cigarette after waking); (2) engagement in treatment and care received (type of organization attended, type of NRT initially prescribed, total clinical visits); (3) motivation and planning (whether a quit date was specified, self-rated importance of quitting and confidence in the ability to quit); (4) health and substance use (body mass index [BMI] category, physical health diagnosis, nonschizophrenia mental health diagnosis); (5) substance-related diagnosis (alcohol or drug dependence), past-month cannabis use; and (6) socioeconomic status (SES; education level, household income, and employment status). Finally, we fit a full-adjusted model including all covariates. We used fractional polynomial models to examine continuous variables for linearity. Based on these results, we used a cubic transformation of patient age and an inverse transformation of the age of initiation. As 89% (2,524 of 2,842 who completed the item) of STOP participants with schizophrenia reported current medication use, and we lacked more detailed data, we did not include this variable in our analysis. We selected covariates a priori based on relevant literature and prior STOP research.

Outcome data from the 6M follow-up interview were available for 50,934 (60%) of the participants. This proportion was somewhat lower among people with schizophrenia (n = 1,608, 53%). Our dataset also had sporadic missing data for baseline variables (see Table 1), including the schizophrenia item itself, which was not answered by 4,558 (5.3%) of participants. We retained these patients in our analyses, on grounds that nonresponse to this item may not have been random. To adjust analyses for missingness, we used multiple imputation (MI) with chained equations. We generated 20 imputed datasets. In the imputation process, we included all variables from our substantive models, as well as a set of auxiliary variables associated with our primary outcome: quit status at 3-month (3M) and 12-month (12M) follow-ups (where available), and whether the participant had quit smoking at their last clinical visit before the 6M interview. As with other missing variables, participants who did not respond to the schizophrenia item were randomly classified within each MI dataset based on their calculated probability of having this diagnosis. To understand the possible implications of this decision, we conducted a brief sensitivity analysis by rerunning the fully adjusted model for our primary outcome including only people who responded to the question about schizophrenia diagnosis.

Table 1.

Descriptive Statistics.^a

	Schizophrenia diagnosis
	No (n = 77,790)	Yes (n = 3,011)
Aged (mean, SD) (missing = 39, 0.04%)	50.4 (14.2)	45.6 (13.0)
Cigarettes per day (m = 138, 0.2%)	19.7 (11.0)	24.5 (14.7)
Age began daily smoking (m = 542, 0.7%)	17.0 (5.2)	17.6 (6.3)
Gender (m = 92, 0.1%)
Male	38,099 (49%)	1,990 (66.2%)
Female	39,436 (50.8%)	1,006 (33.5%)
Other	169 (0.2%)	9 (0.3%)
Time to the first cigarette after waking (m = 490, 0.6%)
Within 5 min	31,016 (40.1%)	1,896 (63.6%)
6 to 30 min	30,401 (39.3%)	801 (26.9%)
31 to 60 min	9,343 (12.1%)	167 (5.6%)
More than 60 min	6,570 (8.5%)	117 (3.9%)
Organization type (m = 0)
Family health team	49,111 (63.1%)	1,043 (34.6%)
Addiction agency	14,049 (18.1%)	1,172 (38.9%)
Community health centre	12,869 (16.5%)	656 (21.8%)
Nurse practitioner-led clinic	1,761 (2.3%)	140 (4.6%)
Type of NRT initially provided (m = 0)
None	1,024 (1.3%)	40 (1.3%)
Short-acting only	16,428 (21.1%)	715 (23.7%)
Patch only	7,927 (10.2%)	350 (11.6%)
Patch + short-acting	52,411 (67.4%)	1,906 (63.3%)
Response to quit date questions (m = 0)
Refused to set a date	8,255 (10.6%)	527 (17.5%)
Unsure/no response	10,427 (13.4%)	500 (16.6%)
Willing to set a date	17,328 (22.3%)	513 (17%)
Knew their quit date	41,780 (53.7%)	1,471 (48.9%)
BMI category (m = 7,983, 9.9%)
Normal or underweight	26,499 (37.8%)	755 (28.3%)
Overweight	23,037 (32.8%)	830 (31.1%)
Obese	20,610 (29.4%)	1,087 (40.7%)
Physical health condition^a (m = 5,724, 7.1%)
No	42,690 (59.0%)	1,587 (58.8%)
Yes	29,689 (41.0%)	1,111 (41.2%)
Nonschizophrenia MH diagnosis^b (m = 3,606, 4.5%)
No	36,114 (48.5%)	676 (24.8%)
Yes	38,354 (51.5%)	2,051 (75.2%)
Substance use disorder^c (m = 1,976, 2.4%)
No	60,525 (79.7%)	1,920 (67.6%)
Yes	15,461 (20.3%)	919 (32.4%)
Past-month cannabis use (m = 1,309, 1.6%)
No	52,922 (69.1%)	2,031 (68.8%)
Yes	23,620 (30.9%)	919 (31.2%)
Education (m = 4,287, 5.3%)
Below high school	18,258 (24.8%)	1,038 (36.5%)
High school diploma	19,364 (26.3%)	741 (26.0%)
Some postsecondary	12,736 (17.3%)	566 (19.9%)
Postsecondary graduate	23,309 (31.6%)	502 (17.6%)
Employment status (m = 2,305, 2.9%)
Not working	27,185 (36.0%)	1,854 (63.8%)
Employed	32,970 (43.6%)	452 (15.5%)
Unable to work	6,631 (8.8%)	473 (16.3%)
Aged 65+	8,803 (11.6%)	128 (4.4%)
Annual household income (m = 27,718, 34.3%)
≤$20,000	18,859 (37.1%)	1,831 (82%)
$20,001 to $60,000	19,082 (37.5%)	336 (15%)
>$60,000	12,909 (25.4%)	66 (3.0%)
Importance of quitting, 1 to 10 rating (m = 497, 0.6%)
1 to 5	2,711 (3.5%)	205 (6.9%)
6 or 7	6,960 (9.0%)	393 (13.2%)
8	13,650 (17.7%)	534 (17.9%)
9	11,677 (15.1%)	377 (12.7%)
10	42,331 (54.7%)	1,466 (49.3%)
Confidence in ability to quit, 1 to 10 (m = 707, 0.9%)
1 to 5	16,705 (21.7%)	839 (28.3%)
6 or 7	20,708 (26.8%)	822 (27.7%)
8	17,296 (22.4%)	548 (18.5%)
9	8,129 (10.5%)	245 (8.3%)
10	14,291 (18.5%)	511 (17.2%)

Note. NRT = nicotine replacement therapy; MH = malignant hyperthermia. ^aTable excludes 4,558 people who did not respond to the schizophrenia status question at baseline. ^bBipolar disorder, depression, or anxiety. ^cAlcohol use disorder or non–tobacco substance–related disorder.

For each of our two secondary outcomes, we fit unadjusted and fully adjusted models. For the “quit difficulties” outcome, initial models appeared to violate the proportional odds assumption for ordinal regression. We therefore fit generalized ordinal logistic models using the gologit2 package for Stata. We used logistic regression for the “other smokers” outcome and negative binomial regression for the number of clinical visits attended. We performed all analyses using Stata 16.³⁹

Results

Of our total sample of 85,359, 3,011 (3.5%) self-reported a lifetime diagnosis of schizophrenia, 77,790 (91.1%) did not, and 4,558 (5.3%) did not respond. Including only people with complete data, the prevalence of schizophrenia was 3.7% (95% confidence interval [CI], 3.6% to 3.9%). Descriptive characteristics are shown in Table 1. Compared to controls, people with schizophrenia were more often male, had much lower levels of income, had a higher prevalence of obesity and comorbid mental health and substance use conditions, and were more dependent on tobacco.

After MI, but before statistical adjustment, 16.2% (95% CI, 14.5% to 17.8%) of people with schizophrenia reported being abstinent from tobacco at 6M follow-up, and another 10.7% (95% CI, 9.4% to 12.1%) had reduced their use. For other participants, 26.0% (95% CI, 26.0% to 26.7%) were abstinent and 12.6% (95% CI, 12.3% to 12.8%) had reduced their use. The absolute group differences were 10.2% (95% CI, 8.5% to 11.9%; P < 0.001) for cessation and 1.8% for reduction (95% CI, 0.5% to 3.2%; P < 0.001). After the addition of all covariates, predicted group differences in cessation probabilities (Table 2) became 7.3% (95% CI, 5.1% to 9.3%) for cessation and 0.7% (95% CI, −0.9% to 2.2%; nonsignificant) for reduction. Our sensitivity analysis, which included only people who provided their schizophrenia status, changed the coefficients reflecting associations of this variable with cessation and reduction by under 1.5%.

Table 2.

Adjusted Average Predictions by Model for Participants with and Without Schizophrenia.

	Continue to smoke (95% CI)	Reduce smoking (95% CI)	Quit (95% CI)
Unadjusted
Schizophrenia	73.1% (71.2% to 75.0%)	10.7% (9.4% to 12.1%)	16.2% (14.5% to 17.8%)
Control	61.1% (60.7% to 61.4%)	12.6% (12.3% to 12.8%)	26.0% (26.0% to 26.7%)
Group diff.	12.0% (10.1% to 13.9%)	−1.8% (−3.2% to −0.5%)	−10.2% (−11.9% to −8.5%)
Demographics
Schizophrenia	72.9% (71.0% to 74.8%)	10.5% (9.2% to 11.9%)	16.6% (14.9% to 18.2%)
Control	61.1% (60.7% to 61.5%)	12.6% (12.3% −12.8%)	26.4% (26.0% to 26.7%)
Group diff.	11.8% (9.9% to 13.7%)	−2.0% (−3.4% to −0.7%)	−9.8% (−11.5% to −8.1%)
Smoking history
Schizophrenia	70.6% (68.6% to 72.6%)	11.7% (10.3% to 13.1%)	17.7% (15.9% to 19.5%)
Control	61.2% (60.8% to 61.6%)	12.5% (12.3% to 12.8%)	26.3% (26.0% to 26.6%)
Group diff.	9.4% (7.4% to 11.4%)	−0.8% (−2.2% to 0.7%)	−8.6% (−10.4% to −6.8%)
Care received/engagement
Schizophrenia	74.2% (72.3% to 76.0%)	10.2% (8.9% to 11.5%)	15.6% (14.0% to 17.2%)
Control	61.0% (60.7% to 61.4%)	12.6% (12.3% to 12.8%)	26.4% (26.1% to 26.7%)
Group diff.	13.1% (11.3% to 15.0%)	−2.4% (−3.7% to −1.0%)	−10.8% (−12.4% to −9.1%)
Motivation/planning
Schizophrenia	71.4% (69.4% to 73.3%)	11.0% (9.6% to 12.3%)	17.7% (15.9% to 19.4%)
Control	61.2% (60.8% to 61.5%)	12.5% (12.3% to 12.8%)	26.3% (26.0% to 26.6%)
Group diff.	10.2% (8.3% to 12.1%)	−1.6% (−3.0% to −0.2%)	−8.6% (−10.4% to −6.8%)
Health
Schizophrenia	71.4% (69.4% to 73.3%)	11.0% (9.7% to 12.4%)	17.6% (15.8% to 19.3%)
Control	61.1% (60.8% to 61.5%)	12.5% (12.3% to 12.8%)	26.3% (26.0% to 26.6%)
Group diff.	8.0% (6.0% to 10.1%)	−0.7% (−2.2% to 0.9%)	−7.3% (−9.3% to −5.4%)
Socioeconomic status
Schizophrenia	69.9% (67.8% to 71.9%)	11.0% (9.6% to 12.4%)	19.2% (17.2% to 21.1%)
Control	61.2% (60.8% to 61.6%)	12.5% (12.3% to 12.8%)	26.2% (25.9% to 26.6%)
Group diff.	8.6% (6.6% to 10.7%)	−1.6% (−3.0% to −0.1%)	−7.1% (−9.0% to −5.1%)
Full
Schizophrenia	69.3% (67.2% to 71.3%)	11.8% (10.3% to 13.3%)	18.9% (17.0% to 20.8%)
Control	61.3% (60.9% to 61.6%)	12.5% (12.2% to 12.8%)	26.2% (25.9% to 26.6%)
Group diff.	8.0% (6.0% to 10.1%)	−0.7% (−2.2% to 0.9%)	−7.3% (−9.3% to −5.4%)

Model coefficients are shown in Table 3. In all models, joint group differences across the 3 outcome levels were highly significant (P < 0.001), and people with schizophrenia were less likely than others to quit successfully (all P < 0.001). The model predictions and their CIs, however, do not show clear differences across sets of variables, with the possible exception of the comparison of the baseline model to that including SES variables.

Table 3.

Relative Risk Ratios (RRRs) for Associations Between Schizophrenia and Smoking Cessation Treatment Outcomes Across Models.

	Smoking reduction		Smoking cessation
	RRR (95% CI)	P-value	RRR (95% CI)	P-value
Unadjusted	0.71 (0.62 to 0.82)	<0.001	0.51 (0.45 to 0.58)	<0.001
Demographics	0.70 (0.61 to 0.81)	<0.001	0.52 (0.46 to 0.59)	<0.001
Smoking history	0.80 (0.70 to 0.92)	0.002	0.58 (0.51 to 0.65)	<0.001
Care received/engagement	0.64 (0.56 to 0.74)	<0.001	0.45 (0.39 to 0.52)	<0.001
Motivation/planning	0.74 (0.64 to 0.85)	<0.001	0.56 (0.50 to 0.64)	<0.001
Health	0.75 (0.65 to 0.86)	<0.001	0.56 (0.50 to 0.64)	<0.001
Socioeconomic status	0.76 (0.66 to 0.88)	<0.001	0.64 (0.56 to 0.72)	<0.001
Full model	0.80 (0.69 to 0.93)	0.004	0.59 (0.51 to 0.68)	<0.001

Results for our secondary outcomes are shown in Table 4. People with schizophrenia reported more difficulties related to self-reported depression, anxiety, boredom, and stress, but this difference was not significant after adjustment (overall P = 0.14). However, we did find a significant but small difference in difficulties related to “being around other smokers.” People with schizophrenia also attended more clinical visits than other patients, with an average of 3.9 (median = 3, SD = 3.5), compared to 3.5 (median = 2, SD = 2.8) among other participants. After adjustment, this difference remained highly significant (P < 0.001).

Table 4.

Associations Between Schizophrenia and Secondary Outcomes.

	OR/IRR (95% CI)	P-Value
Quit difficulties related to “being around other smokers”
Unadjusted	1.48 (1.29 to 1.69)	<0.001
Adjusted	1.28 (1.11 to 1.47)	0.001
Number of other difficulties
Unadjusted—0 quit difficulties	1.27 (1.07 to 1.51)	0.006
Unadjusted—1 quit difficulties	1.29 (1.14 to 1.46)	<0.001
Unadjusted—2 quit difficulties	1.47 (1.34 to 1.62)	<0.001
Unadjusted—3 quit difficulties	1.64 (1.47 to 1.83)	<0.001
Unadjusted—4 quit difficulties	Ref.	Ref.
Adjusted—0 quit difficulties	0.91 (0.76 to 1.08)	0.281
Adjusted—1 quit difficulties	0.89 (0.78 to 1.01)	0.073
Adjusted—2 quit difficulties	0.95 (0.86 to 1.05)	0.322
Adjusted—3 quit difficulties	1.05 (0.93 to 1.19)	0.427
Adjusted—4 quit difficulties	Ref.	Ref.
Total visits
Unadjusted	1.13 (1.10 to 1.16)	<0.001
Adjusted	1.15 (1.12 to 1.18)	<0.001

Note. OR = odds ratio; IRR = incidence rate ratio; CI = confidence interval.

Discussion

In this community-based smoking cessation program, patients with schizophrenia were less likely to quit successfully than other participants, but were about equally likely to reduce their smoking. Notably, our reference group consisted of all other program enrollees, and thus included a substantial proportion of people with psychiatric conditions other than schizophrenia. Model results suggest that lower SES and more prevalent comorbid health problems among people with schizophrenia explain a portion of the group difference in outcomes, while their longer engagement in treatment probably reduced it slightly. Differences in self-reported barriers to quit were not substantial, except for a small potential effect of exposure to other smokers.

We estimated that 16% of people with schizophrenia were abstinent from smoking at 6M. This proportion is slightly higher than those reported in clinical trials, but is not inconsistent with them. In the recent Smoking Cessation Intervention for Severe Mental Ill Health Trial (SCIMITAR+) trial, personalized interventions were provided, and the biochemically verified quit proportion was 14%; however, only 55% of patients received pharmacological treatments. Importantly, our outcomes are considerably better than those for unaided quit attempts, which are successful approximately 5% of the time.⁴⁰ Consistent with previous work, we also found that people with schizophrenia were only marginally less likely than others to reduce the amount they smoked. Although this is not as desirable as cessation, reduction does have health benefits⁴¹ and including these individuals means that over 25% of this group saw some meaningful improvement in their use of tobacco over a single episode of treatment.

Difficulties in quitting smoking in this population have been linked to more severe withdrawal symptoms,¹³ attention,⁴² affect,⁴³ executive function,^42,44 and interactions with antipsychotic drugs.⁹ Circumstances and conditions associated with cessation success in the general population, including comorbid illnesses, substance use, severity of dependence, and lower SES, are also highly prevalent among people with schizophrenia.^18,21-25 Given the number of variables we considered, we did not attempt to test specific causal pathways. However, we were able to adjust for most of the factors above, with the exception of medication use and specific psychiatric symptoms, and our secondary analysis showed that self-reported barriers to cessation were only slightly more common among people with schizophrenia. Although residual confounding cannot be excluded, our results imply that explanations for the remaining disparity in outcomes should probably be sought in issues arising from schizophrenia itself. The reported role of attention, affect, and executive function in prediction cessation within samples of people with psychotic illnesses suggests that psychiatric symptoms might similarly underlie differences between these individuals and others. It is also possible that competing needs, arising, for example, from psychiatric crises or other personal difficulties, might interfere with cessation attempts for some individuals.

At 3.7%, the prevalence of schizophrenia among STOP participants is considerably higher than in the general population.⁴⁵ This will be due in part to the association between schizophrenia and smoking,^9,46 and also to the nature of the program, which reaches people with lower levels of SES, as well as those who are in regular contact with the healthcare system. The numbers also demonstrate, however, that there is abundant demand for smoking cessation treatment among people with schizophrenia.

Although people with schizophrenia are less likely than others to receive adequate care for physical health conditions,^47,48 STOP participants with schizophrenia attended slightly more clinical visits than others, and this difference was not accounted for by any of the covariates we considered. This may be because their quit attempts are protracted, or because they continue to use NRT for longer than other participants after quitting. It is also possible, however, that this is an advantage of integrated treatment. Participating STOP clinics also provide other forms of treatment or support, and people can thus access NRT and counselling in the same place they receive other care. In any case, this result shows that people with schizophrenia are able and willing to engage successfully in this form of care. Combined with the high prevalence of schizophrenia in our program, this speaks to the desire of many of these individuals to quit and contradicts any beliefs that may remain about the secondary importance of physical and preventive healthcare in this population. It is possible that integrating smoking cessation treatment into specialized psychiatric care would provide further benefits. However, our results demonstrate that many people with schizophrenia benefit from enrolment in nonspecialized, low-cost treatment programs.

Strengths and Limitations

Due to the nature of observational data, we are unable to identify clear causal associations. Our data also do not include detailed information on medication use and rely on self-reports both for the diagnosis of schizophrenia and for smoking cessation outcomes. Data on income, employment, and other variables tend to support the validity of data on diagnosis, and, given the stigma attached to it, we believe that false-positive self-reports will have been uncommon. Biochemical confirmation of smoking status was also not practical, given the STOP program's size and the nature of the follow-ups. As we have noted, however, self-reported cessation outcomes agree well with biochemically verified smoking status in this population. We were also able to adjust only fairly crudely for some variables, notably substance-related problems, and it is possible that more detailed measures would make it possible to better understand outcome differences.

Although the follow-up rate in STOP is equal to or better than those of any observational datasets of comparable size,^49,50 our results are also limited by missing outcomes. Our use of MI should have mitigated any resulting biases, but this approach is entirely accurate only when missingness is a function of observed values (i.e., when data are “missing at random”), which is not a verifiable assumption. An important strength of this study, however, is that we were able to analyse a very large sample of people with schizophrenia, which provided the statistical power to constrain effects and explore potential explanatory variables. Finally, the real-world nature of the data also means that results should be broadly generalizable.

Conclusion

Our results confirm that outcomes following NRT-based smoking cessation treatment are poorer among people with schizophrenia, and suggest that this is not due to differences in treatment, SES, or other health variables. However, cessation and reduction remain far more common than in unaided quit attempts, and the high levels of participation and engagement show that many people with schizophrenia are committed to smoking cessation and can successfully engage with, and benefit from, NRT-based smoking cessation programs targeted at the general population—an approach that is advantageous because of its wide reach and low cost, relative to specialized mental health care.

Footnotes

Data Access

STOP data include sensitive personal and medical information, and cannot be made publicly available. Access to deidentified data can be arranged on reasonable request by contacting the primary author.

Declaration of Conflicting Interests

The author(s) have no conflict of interest to declare. However, Peter Selby has the following general disclosures to report: grants and/or salary and/or research support from the Centre for Addiction and Mental Health, Health Canada, Ontario Ministry of Health and Long-term Care, Canadian Institutes of Health Research, Canadian Centre on Substance Use and Addiction, Public Health Agency of Canada, Ontario Lung Association, Medical Psychiatry Alliance, Extensions for Community Healthcare Outcomes, Canadian Cancer Society Research Institute, Cancer Care Ontario, Ontario Institute for Cancer Research, Ontario Brain Institute, McLaughlin Centre, Academic Health Sciences Centre, Workplace Safety and Insurance Board, National Institutes of Health, and the Association of Faculties of Medicine of Canada. PS also reports receiving funding and/or honoraria from the following commercial organizations: Pfizer Inc., Canada, Shoppers Drug Mart, Bhasin Consulting Fund Inc., Patient-Centered Outcomes Research Institute, ABBVie, and Bristol-Myers Squibb. Further, PS reports receiving consulting fees from Pfizer Inc., Canada, Evidera Inc., Johnson & Johnson Group of Companies, Medcan Clinic, Inflexxion Inc., V-CC Systems Inc., MedPlan Communications, Kataka Medical Communications, Miller Medical Communications, Nvision Insight Group, and Sun Life Financial. Through an open tender process Johnson & Johnson, Novartis, and Pfizer Inc. are vendors of record for providing smoking cessation pharmacotherapy, free or discounted, for research studies in which PS is the principal investigator or coinvestigator.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Ontario Ministry of Health and Long-Term Care (grant number HLTC5047FL).

ORCID iDs

Scott Veldhuizen

Osnat Melamed

Peter Selby

References

Dickerson

Schroeder

Katsafanas

, et al. Cigarette smoking by patients with serious mental illness, 1999–2016: an increasing disparity. Psychiatr Serv. 2018;69(2):147-153. doi:10.1176/appi.ps.201700118

Zeng

Zong

Zhang

, et al. Worldwide prevalence of smoking cessation in schizophrenia patients: a meta-analysis of comparative and observational studies. Asian J Psychiatry. 2020;54:102190. doi:10.1016/j.ajp.2020.102190.

de Leon

Diaz

. A meta-analysis of worldwide studies demonstrates an association between schizophrenia and tobacco smoking behaviors. Schizophr Res. 2005;76(2-3):135-157. doi:10.1016/j.schres.2005.02.010.

Callaghan

Veldhuizen

Jeysingh

, et al. Patterns of tobacco-related mortality among individuals diagnosed with schizophrenia, bipolar disorder, or depression. J Psychiatr Res. 2014;48(1):102-110. doi:10.1016/j.jpsychires.2013.09.014.

Caponnetto

Polosa

. Approved and emerging smoking cessation treatments for people with schizophrenia spectrum disorders: a narrative review. Health Psychol Res. 2020;8(2):9237. doi:10.4081/hpr.2020.9237.

Dregan

McNeill

Gaughran

, et al. Potential gains in life expectancy from reducing amenable mortality among people diagnosed with serious mental illness in the United Kingdom. PLoS One. 2020;15(3):e0230674. doi:10.1371/journal.pone.0230674.

Gatov

Rosella

Chiu

Kurdyak

. Trends in standardized mortality among individuals with schizophrenia, 1993–2012: a population-based, repeated cross-sectional study. CMAJ. 2017;189(37):E1177-E1187. doi:10.1503/cmaj.161351

Newman

Bland

. Mortality in a cohort of patients with schizophrenia: a record linkage study. Can J Psychiatry Rev Can Psychiatr. 1991;36(4):239-245. doi:10.1177/070674379103600401.

Lucatch

Lowe

DJE

Clark

Kozak

George

. Neurobiological determinants of tobacco smoking in schizophrenia. Front Psychiatry. 2018;9:672. doi:10.3389/fpsyt.2018.00672.

10.

Tsuda

Saruwatari

Yasui-Furukori

. Meta-analysis: the effects of smoking on the disposition of two commonly used antipsychotic agents, olanzapine and clozapine. BMJ Open. 2014;4(3):e004216. doi:10.1136/bmjopen-2013-004216.

11.

Lawrence

Hafekost

Hull

Mitrou

Zubrick

. Smoking, mental illness and socioeconomic disadvantage: analysis of the Australian national survey of mental health and wellbeing. BMC Public Health. 2013;13:462. doi:10.1186/1471-2458-13-462.

12.

Cather

Pachas

Cieslak

Evins

. Achieving smoking cessation in individuals with schizophrenia: special considerations. CNS Drugs. 2017;31(6):471-481. doi:10.1007/s40263-017-0438-8.

13.

Lum

Skelton

Wynne

Bonevski

. A systematic review of psychosocial barriers and facilitators to smoking cessation in people living with schizophrenia. Front Psychiatry. 2018;9:565. https://www.frontiersin.org/article/10.3389/fpsyt.2018.00565.

14.

Lawn

Campion

. Achieving smoke-free mental health services: lessons from the past decade of implementation research. Int J Environ Res Public Health. 2013;10(9):4224-4244. doi:10.3390/ijerph10094224

15.

Prochaska

Hall

Bero

. Tobacco use among individuals with schizophrenia: what role has the tobacco industry played?*. Schizophr Bull. 2008;34(3):555-567. doi:10.1093/schbul/sbm117.

16.

Smith

McNeill

Kock

Shahab

. Exploring mental health professionals’ practice in relation to smoke-free policy within a mental health trust: a qualitative study using the COM-B model of behaviour. BMC Psychiatry. 2019;19(1):54. doi:10.1186/s12888-019-2029-3.

17.

Montejo

. The need for routine physical health care in schizophrenia. Eur Psychiatry. 2010;25:S3-S5. doi:10.1016/S0924-9338(10)71699-0.

18.

Buckley

Miller

Lehrer

Castle

. Psychiatric comorbidities and schizophrenia. Schizophr Bull. 2009;35(2):383-402. doi:10.1093/schbul/sbn135

19.

Hammett

Lando

Taylor

, et al. The relationship between smoking cessation and binge drinking, depression, and anxiety symptoms among smokers with serious mental illness. Drug Alcohol Depend. 2019;194:128-135. doi:10.1016/j.drugalcdep.2018.08.043.

20.

Steinberg

Williams

Ziedonis

. Financial implications of cigarette smoking among individuals with schizophrenia. Tob Control. 2004;13(2):206-206.

21.

Hunt

Large

Cleary

Lai

HMX

Saunders

. Prevalence of comorbid substance use in schizophrenia spectrum disorders in community and clinical settings, 1990–2017: systematic review and meta-analysis. Drug Alcohol Depend. 2018;191:234-258. doi:10.1016/j.drugalcdep.2018.07.011

22.

Ramsay

Stewart

Compton

. Unemployment among patients with newly diagnosed first-episode psychosis: prevalence and clinical correlates in a U.S. Sample. Soc Psychiatry Psychiatr Epidemiol. 2012;47(5):797-803. doi:10.1007/s00127-011-0386-4.

23.

Saraceno

Barbui

. Poverty and mental illness. Can J Psychiatry Rev Can Psychiatr. 1997;42(3):285-290. doi:10.1177/070674379704200306.

24.

Hiscock

Bauld

Amos

Fidler

Munafò

. Socioeconomic status and smoking: a review. Ann N Y Acad Sci. 2012;1248:107-123. doi:10.1111/j.1749-6632.2011.06202.x.

25.

Reid

Hammond

Boudreau

Fong

Siahpush

. Socioeconomic disparities in quit intentions, quit attempts, and smoking abstinence among smokers in four western countries: findings from the international tobacco control four country survey. Nicotine Tob Res. 2010;12(Suppl 1):S20-S33. doi:10.1093/ntr/ntq051.

26.

Gilbody

Peckham

Bailey

, et al. Smoking cessation for people with severe mental illness (SCIMITAR+): a pragmatic randomised controlled trial. Lancet Psychiatry. 2019;6(5):379-390. doi:10.1016/S2215-0366(19)30047-1.

27.

Tsoi

Porwal

Webster

. Interventions for smoking cessation and reduction in individuals with schizophrenia. Cochrane Database Syst Rev. 2013(2). doi:10.1002/14651858.cd007253.pub3.

28.

Anthenelli

Benowitz

West

, 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 Lond Engl. 2016;387(10037):2507-2520. doi:10.1016/S0140-6736(16)30272-0.

29.

Moore

Aveyard

Connock

Wang

Fry-Smith

Barton

. Effectiveness and safety of nicotine replacement therapy assisted reduction to stop smoking: systematic review and meta-analysis. Br Med J. 2009;338:b1024. doi:10.1136/bmj.b1024.

30.

Stubbs

Vancampfort

Bobes

De Hert

Mitchell

. How can we promote smoking cessation in people with schizophrenia in practice? A clinical overview. Acta Psychiatr Scand. 2015;132(2):122-130. doi:10.1111/acps.12412.

31.

Selby

Zawertailo

. Tobacco addiction. N Engl J Med. 2022;387(4):345-354. doi:10.1056/NEJMcp2032393.

32.

Zwar

Mendelsohn

Richmond

. Supporting smoking cessation. BMJ. 2014;348:f7535. https://www.jstor.org/stable/26513141.

33.

Stead

Perera

Bullen

, et al. Nicotine replacement therapy for smoking cessation. Cochrane Database Syst Rev. 2012;11:CD000146. doi:10.1002/14651858.CD000146.pub4.

34.

Chou

Chen

Lee

. The effectiveness of nicotine-patch therapy for smoking cessation in patients with schizophrenia. Int J Nurs Stud. 2004;41(3):321-330. doi:10.1016/j.ijnurstu.2003.07.001.

35.

Dharshanram

Iyer

Balasubramanium

Madankumar

. Effectiveness of smoking cessation program using nicotine chewing gums and transdermal nicotine patches among patients with schizophrenia in Chennai City: an interventional study. Ann Indian Psychiatry. 2019;3:105-109. doi:10.4103/aip.aip_5_19.

36.

Government of Canada. 2016 census of Canada topic based tabulations; 2016 [accessed 2022 August 16]. https://www12.statcan.gc.ca/census-recensement/2016/dp-pd/dt-td/Index-eng.cfm.

37.

Wong

Shields

Leatherdale

Malaison

Hammond

. Assessment of validity of self-reported smoking status. Health Rep. 2012;23(1):47-53.

38.

Patrick

Cheadle

Thompson

Diehr

Koepsell

Kinne

. The validity of self-reported smoking: a review and meta-analysis. Am J Public Health. 1994;84(7):1086-1093. doi:10.2105/ajph.84.7.1086.

39.

Stata Statistical Software: Release 16 [computer program]. Published online 2020.

40.

Hughes

Keely

Naud

. Shape of the relapse curve and long-term abstinence among untreated smokers. Addiction. 2004;99(1):29-38. doi:10.1111/j.1360-0443.2004.00540.x.

41.

Chang

Anic

Rostron

Tanwar

Chang

. Cigarette smoking reduction and health risks: a systematic review and meta-analysis. Nicotine Tob Res. 2021;23(4):635-642. doi:10.1093/ntr/ntaa156.

42.

Culhane

Schoenfeld

Barr

, et al. Predictors of early abstinence in smokers with schizophrenia. J Clin Psychiatry. 2008;69(11):15068.

43.

Dutra

Stoeckel

Carlini

Pizzagalli

Evins

. Varenicline as a smoking cessation aid in schizophrenia: effects on smoking behavior and reward sensitivity. Psychopharmacology (Berl). 2012;219(1):25-34.

44.

Steinberg

Williams

Gandhi

Foulds

Epstein

Brandon

. Task persistence predicts smoking cessation in smokers with and without schizophrenia. Psychol Addict Behav. 2012;26:850-858. doi:10.1037/a0028375.

45.

Statistics Canada. Health state descriptions for Canadians: Section G - schizophrenia; 2015 [accessed 2022 August 16]. https://www150.statcan.gc.ca/n1/pub/82-619-m/2012004/sections/sectiong-eng.htm.

46.

Parikh

Kutlu

Gould

. nAChR dysfunction as a common substrate for schizophrenia and comorbid nicotine addiction: current trends and perspectives. Schizophr Res. 2016;171(1-3):1-15. doi:10.1016/j.schres.2016.01.020.

47.

Crawford

Jayakumar

Lemmey

, et al. Assessment and treatment of physical health problems among people with schizophrenia: national cross-sectional study. Br J Psychiatry. 2014;205(6):473-477. doi:10.1192/bjp.bp.113.142521.

48.

Mitchell

Malone

. Physical health and schizophrenia. Curr Opin Psychiatry. 2006;19(4):432-437. doi:10.1097/01.yco.0000228767.71473.9e

49.

Kerkvliet

Wey

Fahrenwald

. Cessation among state quitline participants with a mental health condition. Nicotine Tob Res. 2015;17(6):735-741. doi:10.1093/ntr/ntu239.

50.

Nair

Brady

O’Connor

Bell

. Factors predicting client Re-enrollment in tobacco cessation services in a state quitline. Prev Chronic Dis. 2018;15:180144. doi:10.5888/pcd15.180144.