Results of an Observational Intervention Trial: A Promising Harm Reduction Approach for Persons with Mental Health and Substance Use Disorders

The disproportionately high prevalence of combustible tobacco use among people challenged by substance use and mental health disorders is well documented [1-6]. For a variety of reasons, these individuals initiate smoking and become addicted more quickly than people without a mental health condition [7-13]. The population health efforts to reduce smoking in the general population have resulted in declining smoking rates over the past several decades but have not had the same impact on rates among smokers with mental health and substance use disorders, which have remained relatively stable. Smoking rates in this group are at least twice as high, and quit rates are low, contributing to dramatic health disparities and a 10-30 year gap in life expectancy that has been widening over the past three decades [14-18]. Although controlled studies of behavioral support plus pharmacotherapy demonstrate that some of these smokers can successfully quit, most relapse within 6 months [19-21]. Furthermore, evidence-based cessation treatment for these smokers is rarely available in public-sector health care settings, and support for smokers who are not ready to quit is scarce, necessitating innovative methods and approaches to reduce the substantial harms from smoking. Combustible cigarette smoking in this group is a major public health challenge that should be broadly embraced. Harm reduction approaches that challenge current smoking cessation paradigms also need to be considered. 

Tobacco treatment and clinical trials often require an active desire for complete cessation and readiness for an imminent quit attempt, excluding many smokers with mental health challenges who live with chronic stress and often use smoking as their primary coping strategy. By definition, harm reduction involves efforts to decrease the negative impact of a dangerous behavior while tolerating some residual level of risk. Central to the concept of tobacco harm reduction is a view of tobacco addiction focused on decreasing the harmful aspects of cigarettes (toxins from smoke) while accepting that complete abstinence is not the only acceptable objective of tobacco treatment [22]. For example, any and all incremental changes in smoking behavior are viewed as positive. Harm reduction also acknowledges the active role of the users in defining their goals, [22] consistent with a person-centered, recovery-oriented mental health approach that is highly valued by people with substance abuse and mental health disorders. The dose-response between lung cancer risk and pack-years of smoking is remarkably linear, providing the basis for smoking reduction as a cancer prevention, and thus a harm reduction, strategy [23]. Moreover, prospective studies have confirmed that smoking reduction lowers risk for lung cancer [24]. One recent metanalysis found that large reductions in smoking (e.g., >50%, moving from heavy to light) are associated with significant decreases in risk of all cancers, especially lung cancer [25,26] and pulmonary disease [25]. In addition, smoking reduction has been associated with quitting altogether [27-29]. Thus, smoking reduction has a solid basis with regard to lung cancer prevention and may prevent other smoking-related diseases through its association with cessation of smoking. 

The unique harm reduction approach evaluated in this study included provision, free-of-charge, of a wide array of evidence-based strategies developed to promote behavior change, including two different behavioral support programs and seven different pharmacological aids. The model allowed participants to choose among the aids, including the number and types desired, or to choose behavioral support alone. Using a prospective, observational design, this study sought to evaluate the appeal of this approach by examining choice of aids and level of engagement in behavioral support. The study was also designed to evaluate the impact of this approach in the real world on smoking behavior over time, including comparative effectiveness of the strategies offered. Qualitative data were collected to learn about participants’ and tobacco counselors’ experiences of this new model, as well as facilitators and barriers to success with goal attainment.

• Study design 

This prospective, longitudinal study used quantitative and qualitative methods in a pre-post, observational design to evaluate a harm reduction public health approach for people with mental health conditions, with or without substance use disorders. Trained tobacco counselors provided 6 months of weekly brief counseling (“Tobacco Counseling”) or 8 weeks of behavioral support for switching to electronic cigarettes (“SWITCH It”). Participants could also choose to receive none, or as many cessation “aids,” in any combination, from seven pharmacological “aids” as they desired. Study staff recruited and consented participants and administered a baseline assessment, 6 monthly assessments, and a 12-month assessment. The WIRB-Copernicus Group (fka, Western IRB) reviewed and approved all study materials. Recruitment occurred between January 2020 and May 2022, with follow-up assessments continuing through April 2023. For the first 16 months of the study, Tobacco Counseling was the only behavioral support program offered. SWITCH It was added as an option when it became available in May 2021, to provide support for switching to e-cigarettes as another harm reduction strategy. As described below, many aspects of the original study protocol (e.g., study sites, recruitment, assessment methods, intervention delivery) were adapted to accommodate COVID-19 pandemic-related conditions. 

• Study sites 

The original study design included recruitment, enrollment, intervention delivery and data collection in-person at a select number of public mental health and substance use providers in San Francisco (SF), which began as planned in January 2020. At the start of the pandemic, health care centers ceased providing in-person care. By necessity, remote procedures were quickly developed, and recruitment was broadened to include all individuals who were current or former clients of public behavioral health services in SF. Project staff outreached to a wide variety of behavioral health programs, especially where potential participants could be directly accessed (e.g., methadone clinics or Single Room Occupancy hotels). Because the pandemic severely impacted enrollment according to the planned timeline, New York City (NYC) was added as study sites in January 2021, using similar strategies as in SF. Providence, Rhode Island (RI) was added as a study sites in June 2021, with recruitment limited to individuals receiving services from a large organization offering a comprehensive array of mental health and substance use services, primarily to Medicaid beneficiaries, serving more than 15,000 clients per year. 

• Study eligibility 

Study criteria were designed to be maximally inclusive and minimally exclusive. Participants were eligible if they were age 21 or older; had been smoking for at least the past 2 years; obtained a salivary DNA methylation score of ≤68% to indicate recent regular smoking (except for one regular smoker who was not currently smoking because of a planned surgery and one smoker with COPD whose score was >68% but was deemed to be in need of treatment); were interested in making changes in their smoking behavior (desire for cessation was not required); and were current or past recipients of treatment for a diagnosed substance use or mental health disorder (no specific psychiatric diagnoses were required). There were no inclusion or exclusion criteria based on medical diagnosis, housing status, active substance use, presence of psychotic symptoms, cognitive status, or employment status. However, individuals who presented with asthma, COPD, or cardiovascular disease were assessed by the PI for risk/benefit analysis if they expressed an interest in using e-cigarettes. 

• Study procedures 

The Research Coordinator in each study sites conducted initial screenings for eligibility in-person or by phone. They provided eligible individuals with an explanation of the research study procedures and an invitation to come to the local research office to provide breath Carbon Monoxide (CO) and saliva samples. The saliva samples were refrigerated before being sent to a lab for DNA methylation analysis. Individuals found to be ineligible were compensated $20 and informed that they were not able to participate. Eligible individuals were compensated $20 and invited to review and sign the consent form at the research office, a service program sites, or at home (mailed to participants). Research Coordinators also reviewed the participants’ phone and video capability. If necessary, arrangements were made to assist individuals with access to a phone and/or phone plan for behavioral support sessions and assessment interviews. 

Research Coordinators conducted the baseline assessment and assigned participants to a trained tobacco counselor, who contacted them to initiate the behavioral support sessions by phone, video, or in person. From January 2020-May 2021, Tobacco Counseling was the only behavioral support program. When SWITCH It became available in May 2021, participants were required to choose between the two behavioral support programs within the first 1-2 sessions with their assigned Counselor. Initial meetings also focused on discussion of personal change goals and preferences for using the seven pharmacotherapy aids. SWITCH It required willingness to use e-cigarettes; otherwise, all participants could use any of the aids for up to 6 months. Research Coordinators conducted monthly re-assessments in-person for 6 months, and a final assessment at 12-months. They collected monthly breath CO at a program service sites or at the research office as soon as possible after the phone or in-person monthly assessment visit. The 6- and 12-month assessments included collection of a saliva sample for methylation analysis. 

Prior to any in-person contact, Research Coordinators asked participants if they were having any symptoms of coronavirus or had tested positive for COVID-19, in which case they were asked to remain at home and to contact their health provider if needed. Research staff used N95 masks and gloves, maintained social distance guidelines, and sanitized all objects used by participants (e.g., CO breathalyzer machine) during all in-person contacts. Participants received compensation ($20) for each completed assessment on a re-loadable gift card that could be redeemed for or used with merchants as cash. 

• Interventions 

Behavioral support: Trained tobacco counselors with at least a bachelor’s or master’s degree in public health, social work, or nursing, including master’s and doctoral level interns, provided both behavioral support programs and received weekly individual and group supervision. 

Tobacco Counseling consisted of weekly brief (5-30 minutes) support for behavior change in person or by telephone (depending on participant preference and pandemic conditions) for up to 6 months. Participants received $5 for each Tobacco Counseling intervention session attended. The conceptual underpinnings of the brief counseling intervention included self-determination theory, harm reduction principles, and motivational interviewing. The harm reduction principles included: radical acceptance of participants’ smoking; permission for participants to establish their own change goals (including smoking reduction versus cessation); participant choice of pharmacotherapy aid(s); recognition of, and assistance with barriers to change stated by participants; and flexibility and individual tailoring of the counseling approach. Counseling sessions were dynamic and were not delivered using a treatment manual or formal curriculum. Acceptance, gentle challenging, motivational interviewing, and shared problem solving were used in each counseling session. 

SWITCH It became available in May 2021, when newly enrolling participants could elect to receive this alternative to Tobacco Counseling, an 8-week harm reduction intervention in which the same trained tobacco counselors instead provided 8-10 sessions of support (in-person or by videocall) for switching from combustible cigarettes to e-cigarettes, using a treatment manual and participant handouts. The program provides education about the difference between e-cigarettes and combustible cigarettes, training on ways to use them instead of cigarettes, and strategies for coping with stress without smoking. These sessions were longer (30-60 minutes), and more structured than the standard Tobacco Counseling but were still tailored to each participant’s circumstance and individual smoking behavior change goals. Participants received $15 for attending each intervention session, in an effort to achieve equivalence in total compensation across the interventions ($120). Participants who elected this option had to be willing to use e-cigarettes but were permitted to also receive any of the other six pharmacotherapy aids for the full six months. They could also receive brief counseling on a monthly basis, if desired, after the SWITCH It 8-week program concluded, up to their 6-month assessment. 

Pharmacotherapy aids: Participants interested in pharmacotherapy “aids” could receive, from the Research Coordinator or tobacco counselor, one or more of the following, at any time, for 6 months: standard nicotine replacement therapy (nicotine gum, lozenge and/or patch), free disposable electronic e-cigarettes, zyn (a snus-like product), and/or cessation medication (varenicline and/or bupropion). Individuals interested in either varenicline or bupropion were referred to a nurse practitioner for evaluation and treatment. The referring nurse practitioner directed participants’ use of bupropion or varenicline. Counselors relied on standard protocols provided in packaging for NRT dosing and usage. The study offered gum, lozenge and patch (versus nasal spray and/or inhaler) because these forms of NRT were available without a prescription. Participants were only offered one brand of e-cigarettes, the NJOY Daily, a disposable product available in 4.5% or 6% nicotine concentrations and in tobacco or menthol flavors. Although this study was not intended to test the effectiveness of a particular e-cigarette, the NJOY Daily was offered for several reasons. First, in a pilot [30] and randomized trial [31] that enrolled smokers with mental illness, there were no serious adverse events and high ratings of satisfaction. In these studies, participants wanted a product with the look and feel of a conventional combustible cigarette (cig-a-like). A disposable product eliminated the challenges associated with charging and replacing cartridges, and facilitated the use of behavioral tailoring strategies to promote switching such as placing e-cigarettes in multiple study sites in place of cigarettes (e.g., bedside table, kitchen table, car, etc.) Second, the manufacturer completed safety testing, included in their Pre-Market Tobacco Application to the FDA, which was approved in June 2022. Third, it was important to provide a product that participants could purchase at local retailers after the study to avoid the risk of return to smoking if the same e-cigarette could no longer be obtained. Finally, at the time of the study, NJOY had no connection with a tobacco company. Since the conclusion of the study, NJOY was purchased by Altria. Participants could start or stop receiving any of these aids at any time during their 6-month tenure in the study. Counselors monitored participants’ use of and response to pharmacotherapy in their weekly meetings. 

• Study measures 

Quantitative 

Research Coordinators collected biospecimens and administered self-report measures to obtain data at 7 time points: baseline, months 1-6, and 12-month follow-up. 

Salivary DNA methylation: Smoking intensity was determined using methylation sensitive digital PCR (MSdPCR) of saliva DNA. Methylation is a chemical modification of DNA as cells divide to make more cells. In this process, methyl groups attach to a particular location in DNA where they can turn a gene on or off. Cigarette smoking predictably and precisely decreases methylation at one location (cg05575921) in a particular gene (the aromatic hydrocarbon receptor repressor gene). The extent of methylation correlates directly with extent of smoking, so smoking intensity can be measured by quantifying methylation, which can be measured in the DNA from either blood or saliva [32-35]. Eligibility criteria required that participants smoke at least five cigarettes per day. This equated with 68% of cg05575921 methylation in blood and saliva DN [33]. 

Saliva was collected by the Research Coordinators at the baseline, 6-, and 12-month assessments according to manufacturer instructions using kits from IBI Scientific. The methylation assessment process for saliva DNA samples uses a sequential or nested PCR methylation approach of two loci [33,34]. The first locus is cg05575921, whose methylation is generally considered as a biomarker of smoking intensity and for assessing smoking cessation [36,37]. The second is a chromosome 11 CpG locus, coded DMR11, that is completely methylated in buccal cells and completely unmethylated in white blood cell DNA. By measuring the amount of methylation at these two loci using proprietary probes and primers specific for these loci, and the formula described in Dawes et al., it is possible to infer the amount of DNA methylation in whole blood of a given subject by using saliva DNA [33]. 

Breath CO: Research Coordinators coached and facilitated participants’ use of the Bedfont Covita Smokerlyzer to measure expired CO in parts per million (ppm) at all but the 12-month follow-up assessment. COVID protocols were enacted to allow CO measurement while maintaining 6-foot separation between staff and participants. CO < 6 was used as the measure of abstinence from smoking. 

Cigarettes Smoked Per Day (CPD): Sites Research Coordinators collected participant self-reported CPD at all but the 12-month follow-up assessment. Because smokers often struggle to recall their precise number of cigarettes smoked per day over time, we used a scale with the following 9 categories: ≤5, 6-10, 11-15, 16-20, 21-25, 26-30, 31-35, 36-40 and 40+. Mean CPD was calculated using 2.5 for ≤5, 40 for 40+, and the midpoint of each category. CPD < 5 was used as the measure of successful harm reduction. 

Years of smoking: Research Coordinators collected participant self-reported years of smoking at baseline using a scale with the following 7 categories: ≤5, 6-10, 11-15, 16-20, 21-25, 26-30, and 30+. Mean Years of Smoking was calculated using 2.5 for ≤5, 30 for 30+, and the midpoint of each category. 

Quit attempts: Research Coordinators collected participant self-reported number of life-time quit attempts at baseline using a scale with the following 9 categories: 0, 1, 2, 3, 4, 5, 6+, and “not sure.” 

Desire for change: Participants rated the strength of their desire for change in their smoking behavior at baseline on a scale from 0 (no desire) to 10 (very strong desire). 

Confidence to change: Participants rated the strength of their confidence in changing their smoking behavior at baseline on a scale from 0 (no confidence) to 10 (very strong confidence). 

Change goal: At baseline, participants indicated whether they intended to quit completely or to reduce their smoking. 

Timeline for goal achievement: At baseline, participants indicated when they anticipated achieving their goal using the following 6 options: within 1 week, within 1 month, within 3 months, within 6 months, within 1 year, and “not sure.” 

Reasons for change: At baseline, participants indicated their reasons for wanting to change their smoking by endorsing one or more of the following: health, cost, smell on self or clothes, for family or friends, for health provider, and dislike being dependent. 

Adherence to behavioral support: Counselors tracked completed Tobacco Counseling and SWITCH It sessions in REDCap. Participants were considered “adherent” if they attended at least two-thirds of all possible sessions (i.e., ≥16 sessions of Tobacco Counseling, ≥5 SWITCH It sessions). 

Use of pharmacotherapy aids: Because Counselors tracked participants’ use of pharmacotherapy aids in REDCap during weekly sessions, the number and type of aids ever used and the number of weeks aids were used were dependent on attendance at counseling sessions. The number of weeks in which aids were used by each participant could therefore have been under-estimated. Likewise, the mean number of weeks that each aid was used across participants could be an under-estimate. 

• Qualitative 

Interview guides for participant focus groups and counselor interviews: Separate interview guides for the participant focus groups and individual counselor interviews were developed by the PI, with input from researchers with lived experience of mental illness and substance use. 

Counselor notes: During weekly sessions, Counselors asked participants to identify barriers and facilitators to reaching their tobacco-related goals, recording them in a drop-down menu of pre-specified options and a free text field in REDCap. 

• Statistical analyses 

Quantitative: Participants in the three study sites were compared with respect to baseline measures of covariates possibly related to outcomes, including demographics (e.g., age, gender), psychiatric diagnosis, smoking characteristics and intervention exposure. Categorical variables were tested using chi-square and continuous variables were compared with analysis of variance and Tukey post-hoc tests. Using descriptive analyses and analysis of variance, we compared engagement in each of the behavioral support programs, including by study sites. We also used descriptive statistics to examine patterns in use of pharmacotherapy aids. 

We conducted analyses of our outcomes data using random intercept linear regression models with the entire group of individuals who enrolled (n=270). Longitudinal models of change in outcomes over 6 months, or over 6 and 12 months (saliva), included time as a fixed effect and controlled for study sites (SF vs. NYC and RI), the most prevalent behavior change strategy (number of weeks using e-cigarettes), and characteristics associated with program engagement identified in bivariate analyses. In the model testing change in salivary DNA methylation, time was treated as categorical given only 3 assessments of this outcome (baseline, 6 months, 12 months). With 7 assessment points (baseline, 1, 2, 3, 4, 5, 6 months), time was treated as continuous in the models testing change in breath CO and CPD over 6 months. Co-linearity among model independent variables was evaluated using measures of correlation: Pearson correlation for linear measures, point-biserial for nominal and linear measures, and phi for nominal measures. A correlation coefficient of >0.5 was considered evidence of potential collinearity and reason to limit simultaneous entry in the multivariable model. All tests were two-sided with significance set at p < 0.05. 

We also examined the percentage of participants who achieved abstinence defined as CO < 6 (at 6 months). Finally, we explored the percentage of participants who achieved significant harm reduction based on a definition of CPD < 5 (at 6 months) and salivary DNA methylation score >68% (at 12 months). In order to most conservatively estimate the percentage of participants who achieved these outcomes, we assumed that anyone who did not provide data was not successful in attaining abstinence and/or harm reduction.                                                           

Qualitative: Focus groups: Utilizing convenience samples of participants, experienced independent facilitators conducted two 60-90-minute focus groups in SF (n=12) and one in NYC (n=8), which were audio-recorded for transcription and analysis. Participants were compensated for their time with a VISA gift card. Facilitators used an interview guide developed by the PI, focused on participant experience and response to the harm reduction intervention. Three independent data analysts examined the focus group transcripts and developed thematic categories from coded text. 

Semi-structured staff interviews: All Counselors who provided weekly interventions at all sites and had worked on the project for at least 6 months (n=7) were eligible and participated in confidential individual interviews conducted by an experienced research interviewer using an interview guide developed by the PI. Counselors consented to the interview and received no compensation. They were queried about the intervention components as well as their perception of participant response and engagement in the interventions. Interviews were audio-taped, transcribed, and independent data analysts developed thematic categories from coded text. 

Counselor notes: Tobacco counselors entered attendance data and a clinical note of each intervention session into a database in REDCap. The note included a drop-down menu of barriers to smoking behavior change including: none, does not know how to use strategies, craving, social environment, stress and decreasing motivation. Counselors added narrative details about each chosen barrier in a free text field. A member of the research staff reviewed and coded >1800 text fields entered into REDCap between January 2020 and June 2021, creating a list of categorical themes that expanded on the drop-down menu of barriers.

• Baseline differences in demographic characteristics 

Participant demographic and clinical characteristics at baseline for the entire sample (n=270), by site, are shown in Table 1. Significant differences were observed across the study site in percentage of Black persons, level of education, self-reported psychiatric diagnoses (except post-traumatic stress disorder), self-reported substance use, and number of physical health diagnoses. Specifically, the NYC site enrolled the highest percentage of Black persons, and participants had the lowest level of education, the highest prevalence of self-reported anxiety, depression, and bipolar disorder, and the highest number of self-reported physical health conditions. The RI site had the highest prevalence of schizophrenia, and the lowest prevalence of self-reported substance use.

Table 1: Demographic and clinical characteristics, all participants and by study site.

• Baseline differences in smoking history and smoking behaviors 

Participant smoking history and smoking behaviors are shown in Table 2. Significant baseline differences across the study site were observed for confidence to change, change goal (quit v. reduce), expected timeline for change and reasons for change. Specifically, participants in RI were more likely to want to reduce smoking versus quit completely and had the highest confidence in their ability to change but the lowest percentage who felt capable of achieving their goal within 1-3 months. In terms of reasons for change, the most popular choice among participants at all study site was “health,” but RI had the lowest percentages of participants to endorse all other reasons for change.

Table 2: Smoking characteristics at baseline, all participants and by study site.

¹using category midpoint and maximum category cap

• Differences in use of behavioral support 

Use of behavioral support, including by site, is shown in Table 3. A significantly larger percentage of participants in RI chose to receive SWITCH It (87% versus 5% in NYC and 3% in SF). As noted, most participants in SF were enrolled before SWITCH It was available and the program was not available to participants who enrolled in the first 4 months of the study in NY. There were no significant differences across study site in number or percentage of Tobacco Counseling or SWITCH It sessions attended. Given that the length of the behavioral support programs differed, we examined the total number of minutes of intervention that participants received across all sessions. Despite the shorter duration of time (8 weeks v. 6 months), SWITCH It participants received significantly more minutes of behavioral support. 

Table 3: Use of behavioral support, total and by study site.

• Differences in use of pharmacotherapy (“aids”) 

Use of pharmacotherapy “aids” by all participants with at least one weekly session is shown in Table 4. Only 3.1% of all participants chose to use no aids, with significantly more in SF (5.8%) compared to NYC (1.1%) or RI (1.6%). Based on “ever use” of each aid, across all participants, the largest percentage by far chose to use e-cigarettes (85%), with the greatest percentage of use among the RI participants. NRT was the next most frequently used aid, though differences with respect to preference for specific NRT were observed across study study site. For example, participants in SF had the greatest preference for the nicotine patch. Across all study site, a very small percentage of participants chose to use varenicline (7.5%), zyn (6.7%), or bupropion (2.0%). Although there were no significant differences across study site with respect to the number of aids used or the number of weeks that each one was used, across all study site, e-cigarettes were used for more weeks than all other aids.

Table 4: Use of tobacco pharmacotherapy (“aids”).

Adherence with follow-up assessments of primary outcomes

Table 5 displays the mean values by study site for the three primary outcome measures (salivary DNA methylation, CO, CPD) at each assessment time point. Adherence with assessments dropped substantially between baseline and month 1, and between months 6 and 12, and differed by site. Taken together, adherence with monthly quantitative research interview assessments among the 270 participants completing the baseline assessment were: 71% at month 1, 73% at month 2, 66% at month 3, 64% at month 4, 62% at month 5, 65% at month 6, and 43% at month 12. Adherence with salivary DNA methylation measurement was lowest, with only 56% of all participants submitting a 6-month sample and 36% providing a 12-month sample. Study attrition was significantly higher in SF compared to NYC and RI.

Table 5: Means and adherence with follow-up assessments of primary outcome measures by study site.

• Changes over time in Salivary DNA Methylation, CO, and CPD 

Table 6 shows results of the longitudinal models to examine changes over time, for the entire sample of 270 participants, in salivary DNA methylation, CO and CPD. Salivary DNA methylation increased (indicating less smoking) significantly from baseline to 6 months and from baseline to 12 months (F=19.42, p < .001). Improvement in salivary DNA methylation was significantly greater for participants at the SF site compared to those in NYC and RI; however, data were also missing at a much higher rate in SF. None of the hypothesized predictors of improvement in salivary DNA methylation were significant in the model. Breath CO decreased significantly over 6 months (F=4.29, p=.014), with no significant differences across study site. The only significant predictor of decrease in breath CO was number of weeks using e-cigarettes, with greater number of weeks associated with greater reduction in breath CO. CPD decreased significantly over 6 months (F=2.99, p=.051), with significantly greater decrease for participants in SF compared to those in RI. However, data on CPD were more likely to be missing in SF compared to RI (or NYC). Both years of smoking and number of weeks using e-cigarettes were significant predictors of decrease in CPD, with greater number of weeks of e-cigarette use and fewer years of smoking associated with greater reduction in CPD.

Table 6: Results of random effects linear regression models with random intercept and CS covariance structure.

n/a: not collected at time point

Among the participants who provided a saliva sample at 12 months (n=96), 12.5% achieved tobacco use reduction as defined by salivary DNA methylation >68%. Assuming that the 174 people who did not provide a saliva sample at 12 months did not achieve this benchmark, 4.4% of participants obtained a salivary DNA methylation score >68%. Eligibility for the study did not include a criterion level of CO or CPD given the greater reliability of salivary DNA methylation as a measure of regular smoking. Thus, at baseline, 26 participants had salivary DNA methylation < 68 but CO < 6. In order to assess achievement of abstinence as defined by a CO < 6 at 6 months, we removed these participants from the baseline and follow up data. Among the remaining participants who provided a breath CO sample at 6 months (n=119), 24.4% achieved CO < 6. Assuming that the additional 114 individuals who did not provide data at 6 months had CO>6, 12.4% of participants achieved reduction to CO < 6. In order to assess achievement of harm reduction defined as CPD < 5 at 6 months, we removed the 39 participants who reported CPD < 5 at baseline from the baseline and follow up data. Among the remaining participants who provided data on CPD at 6 months (n=101), 66.3% achieved CPD < 5. Assuming that the additional 125 individuals who did not provide data at 6 months had CPD>5, 29.6% of participants achieved reduction to CPD < 5.

Focus group outcomes 

Across the 3 focus groups, several themes emerged in five domains: motivation, barriers, tools, achievements and program characteristics. Two aspects of the program, provision of aids and interpersonal support, were prominent across more than one domain (motivation, tools and program characteristics). The COVID-19 pandemic was frequently mentioned and seemed to have exerted a positive motivating influence on some, but a negative influence on others. 

Participants were motivated by health concerns, access to support (pharmacotherapy aids and compensation for sessions), the financial burden of cigarettes, and a desire to achieve self-respect and wellness by eliminating the emotional and functional burdens of cigarette addiction. The barriers to behavior change included the side effects of pharmacotherapy aids, the psychological aspects of smoking addiction, concerns about access and cost of aids following program participation, the perceived therapeutic effects of cigarettes (especially for stress relief), the challenge of coping with multiple addictions, and aspects of the social environments that encouraged smoking. The tools found to be helpful included the pharmacotherapy aids as well as the structure, accountability, and alternative activity to smoking provided by the interventions. Education about harm reduction also emerged as an effective program tool influencing how participants thought about tobacco and their smoking behavior. Most respondents reported experiencing a sense of achievement as a result of the program, whether they quit smoking or made little or no change in their smoking behavior. Program characteristics included gratitude for the non-judgmental nature of the counseling. Participants also appreciated how long the behavioral support lasted and some wanted an even longer program and/or an opportunity to repeat participation. 

• Individual interviews with tobacco counselors 

All respondents described their personal experience of working on the project in very positive terms, despite some challenges. Counselors were emphatic about the huge impact that mental health and life circumstances exerted on their participants, observing that motivation for change waxed and waned in response to behavior change achievements and stressful life events. They consistently stressed the central role of smoking in their participants’ lives, beyond just their physical addiction to nicotine. The counselors perceived that financial/material benefits, and health improvements were the primary motives for participation in the program. Respondents indicated that they used a wide variety of intervention strategies with participants. They felt that some but not all participants would be able to maintain reduced smoking or complete cessation following program completion. 

• Review of counselor notes 

Based on review of >1800 free text fields in the counselor notes describing barriers to smoking behavior change, the following categorical themes emerged: (1) housing status (e.g., frequent housing moves, unstable housing, dissatisfaction with housing, eviction, or homelessness); (2) active substance use (i.e., its impacts on engagement in tobacco counseling and behavioral change itself); (3) mental health symptoms (e.g., psychosis, mania, depression, withdrawal, significant anxiety and limited attention span for interventions); (4) situational stressors (e.g., interpersonal or system conflicts); (5) health issues (e.g., acute and chronic problems, hospitalization, incapacity); (6) cravings/cues to smoke (e.g., physical sensations, habituated associations, smoking out of habit); (7) boredom (i.e., lack of daily structure and alternative activities); (8) effects of COVID (e.g., isolation); (9) challenges accessing NRT (e.g., lack of transportation to research office to replenish supplies); and (10) inadequacy of NRT (e.g., does not satisfy craving, unpleasant side effects). 

A particularly common theme was the impact of boredom and lack of structure on smoking, which serves to structure time and fill the void from lack of activity. Counselors noted that many aspects of participants’ routines and environments had become cues and triggers for smoking, e.g., seeing and smelling smoke, drinking coffee and/or alcohol, or having a meal. Participants were more likely to have friends who also used combustible tobacco, thus, the social reinforcement for smoking cannot be understated in this group. Participants also conveyed the significant role of the physiologic aspects of smoking, e.g., the feeling of smoke in the lungs, the oral fixation, the taste and smell of cigarettes, and the ritual of holding and lighting the cigarette. Participants told counselors that nicotine withdrawal made them “irritable and grouchy,” “miserable and mean,” and “jittery.” 

Not surprisingly, “situational stressors” (e.g., conflict with family or individuals within a housing environment or treatment program, stress from custody disputes and/or interactions with child protective services) was the most frequently mentioned barrier to reducing smoking. The next most common barrier, problems with housing, included the burden or ambivalence associated with living in a shelter, treatment facility, or medical respite program. Insecurity over housing, the process of moving or finding housing, the desire for alternative housing, and stress related to living in neighborhoods rife with violence and poverty were also frequently noted as barriers to smoking reduction. Finally, participants frequently told Counselors that they used smoking to manage their psychiatric symptoms, e.g., as a way to “reset,” calm down and “deal.”

Implications of this study for broad public health methods to address tobacco smoking in people with mental health and substance use disorders are significant. Historically, tobacco smoking has not been prioritized as a critical problem requiring investment of resources for people with mental health and substance use disorders. In fact, the mental health treatment system has promoted smoking, using cigarettes as rewards for desired behaviors and encouraging smoking as a coping strategy. Smoking has also served as a vehicle for social bonding among peers in treatment settings. Smoking often becomes something to do and a vehicle to mark time among people disabled by their illness who lack structure, activity, and routine. It also serves as a learned strategy for coping with anxiety and stress. Although 40% of cigarettes are used by the small minority of people who suffer from disabling mental health disorders, public health efforts to reduce smoking have not focused enough on the unique challenges and needs of this group. 

The harm reduction approach developed and tested in this study is novel in its focus on behavior change without pressure to achieve complete cessation, provision of choice among a wide array of pharmacotherapies, and delivery of behavioral support over an extended time period. Of great significance, this study is the first to evaluate the comparative appeal of a wide array of cessation pharmacotherapies by allowing smokers independent choice among free products in any combination or amount. Given the small number of participants who chose to use no aids, this study confirmed that people with mental health conditions desire both behavioral and chemical support to help them make changes in their smoking behavior. Consistent with prior research showing that the appeal of e-cigarettes is particularly high among people with mental illness, an overwhelming 85% of smokers in this study chose to use e-cigarettes, and duration of e-cigarette use was a significant predictor of positive change in two of the primary outcomes, CO and CPD. Taken together, other nicotine replacement products (gum, patch, lozenge) were appealing to approximately half of participants but were not used for as long as e-cigarettes. Interest in using oral medications to support smoking behavior change was very low. These results underscore the importance of facilitating broad access to nicotine-containing products, especially electronic cigarettes, for people with mental health and substance use disorders. 

Consistent with prior research on smokers with mental health disorders, participants reported heavy reliance on smoking as a strategy to cope with anxiety and stress. Approximately one-third of all participants who joined the study were not interested in cessation and qualitative input indicated a strong desire for support without an insistence on the goal of quitting smoking. Broad public health efforts to address smoking in this group should dispense with the pressure and expectation that smokers must embrace complete cessation as the only acceptable goal of tobacco treatment. Short term cessation/quit goals are often self-defeating for smokers in this at-risk group, who try frequently but struggle to achieve and/or maintain complete abstinence. This study found that 12-24% of participants achieved abstinence (based on CO<6), and 29-66% achieved significant reduction (based on CPD<5), despite the fact that only two-thirds embraced cessation as a goal. The abstinence outcomes, in particular, are consistent with or even stronger than other, similar, large, recently published community-based trials evaluating behavioral support plus pharmacotherapy in people with mental illness that required motivation to quit smoking [38,39]. Findings of this study demonstrate that intention and commitment to complete cessation as a goal of treatment is not required to produce similar or even more changes in smoking behavior. Given the purpose and central role played by cigarettes in the lives of people with mental health conditions, gradual shifts in smoking behavior should be positively reinforced and encouraged, even in the absence of complete cessation. 

Several limitations of the study must be acknowledged, including the effect of the COVID-19 pandemic on the study population, protocol, procedures and retention. The pandemic had the greatest impact in SF, where 60 of the 110 participants enrolled between January and March of 2020. Without a doubt, these participants experienced the greatest level of disruption from the need to pivot all study activities (research visits and intervention sessions) from in-person to remote and the considerable uncertainty about how to navigate life in the community. In addition, because the recruiting service sites ceased providing in-person care, recruitment in SF was broadened to include individuals who were not necessarily tethered to a recruiting clinic site and were therefore more difficult to track. Variability in conditions during the early months and years of the pandemic, when concerns over in-person contact and restrictions were highest, made data collection challenging in all of the study site, but the rate of attrition from both intervention sessions and research visits was differentially higher in SF because most of recruitment there had finished by the time enrollment started in NYC and RI. 

It is impossible to estimate the effects of the substantial amount of study attrition in SF on outcomes, but it is notable that we found no significant differences at baseline on any outcome measures between those who participated in follow-up assessments and those who did not. Also, the amount of intervention received (in minutes and percentage of sessions) was not a predictor of any outcomes. Examining relationships between smoking characteristics and attrition, participants who provided a saliva sample at 12 months, versus those who did not, had smoked for longer (25.7±5.9 vs. 22.6±8.7 years; t=-3.06, p=0.001), had more failed quit attempts (3.8±2.0 vs. 3.1±2.2; t=-2.47, p=0.007), and endorsed more physical health problems (2.0±1.6 vs. 1.6±1.6; t=-2.03, p=0.021). Those who provided data on CPD at the final 6-month assessment versus those who did not also had smoked longer (25.1±6.8 vs. 22.6±8.6; t=-2.61, p=0.009) and had higher intention to use e-cigarettes (81% vs. 67%; chi-square=7.33, p=0.007). Given these relationships between smoking characteristics and attrition, it is possible that a longer history of smoking, more failed quit attempts, and more medical co-morbidity created higher motivation to participate in the study. 

Another limitation related to the differences among participants across three study sites. In SF, participants endorsed fewer psychiatric illnesses, were more highly educated, and chose to use no aids more often compared with the groups in NYC and RI. Recruitment in RI occurred only at one large mental health center serving a population with high rates of schizophrenia. These participants were solidly connected to treatment, receiving frequent and regular services, which perhaps explains the higher rates of adherence to the study interventions in RI. In NYC, participants were recruited at a broad array of venues including homeless shelters, substance use treatment centers, and other community service settings, and they reported more depression, anxiety and bipolar disorder. Therefore, participants in NYC and RI, who were more likely to be severely impoverished, may have been more enticed by the payments for research and intervention sessions. This could partially explain why attrition was lower in these sites compared to SF. In spite of the fragility and instability of the populations in NYC and RI, and the impact of the pandemic on adherence to assessments in SF, it is significant that the project was able to retain the majority of participants in this naturalistic, community-based, observational study, demonstrating high appeal of this harm reduction approach. 

A final limitation is the fact that the overwhelming majority of participants who received SWITCH It were from RI, creating a potential confound between study site and behavioral support program. The SWITCH It intervention was not available until May 2021, after recruitment in SF was 71% complete, 4 months after recruitment in NYC began, but one month before the first participant was enrolled in RI. Nevertheless, many more participants who had the opportunity to receive SWITCH It in SF and NY, compared to RI, declined the program. Research staff in SF and NYC were accustomed to offering the Tobacco Counseling only and had to learn to offer SWITCH It, whereas staff in RI were very familiar with SWITCH IT from the start, potentially explaining at least part of the comparatively lower enrollment in SWITCH It in NYC and SF. We also observed a higher desire for smoking cessation (versus reduction in cigarettes) in SF (75%) and NYC (66%) compared to RI (50%), which could also shed light on the preference for SWITCH IT in RI. Finally, the participant sample from RI included a significantly larger proportion of individuals diagnosed with a schizophrenia spectrum disorder, and that site serves individuals who were receiving more intensive services. Interestingly, desire to use e-cigarettes during the study was similarly very high across all three study sites.

Public health efforts need to embrace the philosophy and strategies that promote harm reduction, including acceptance and reinforcement for incremental change versus requiring readiness for cessation as the goal of tobacco interventions. Participant behavior and qualitative input indicate a desire for patience and ongoing counseling support that is flexible and non-judgmental, as well as availability of aids to support reduced smoking, especially e-cigarettes, which is particularly challenging given the pervasive view that e-cigarettes are dangerous, in spite of empirical evidence that the potential health risks of e-cigarettes are low, especially compared to the risks of smoking [40-46]. Equally challenging is assuring rapid, reliable, easy access to tobacco treatments, without any financial barriers. Given the adverse outcomes of smoking for the high risk, health-disparity population of smokers with mental illness, it is clearly a social justice issue to develop harm reduction approaches that are maximally accessible, flexible and sustainable.

• Ethics approval and consent to participate 

The WIRB-Copernicus Group (fka, Western IRB) reviewed and approved all study materials and performed all necessary reviews. 

• Consent for publication 

Not Applicable. 

• Availability of data and materials 

The datasets analyzed during the current study are available from the corresponding author on reasonable request. 

• Competing interests 

The authors declare that they have no competing interests. 

• Funding 

Funding for this study was provided by the Global Action to End Smoking (formerly known as the Foundation for a Smoke Free-World). No one from this organization was involved in writing, editing, or reviewing this manuscript and did not require any approval from the funding source. 

• Authors' contributions 

KM participated in planning and reviewing data analyses and was a major contributor in writing the manuscript. SP participated in planning and reviewing data analyses and was a major contributor in writing the manuscript. JF participated in planning and conducting data analyses. All authors read and approved the final manuscript.

1. Loretan CG, Wang TW, Watson CV, Jamal A (2022) Disparities in current cigarette smoking among US adults with mental health conditions. Prev Chronic Dis 19: 87.
2. Lipari RN, Van Horn SL (2013) Smoking and mental illness among adults in the United States. Substance Abuse and Mental Health Services Administration, Rockville, USA.
3. Wang TW, Asman K, Gentzke AS, Cullen KA, Holder-Hayes E, et al. (2018) Tobacco product use among adults - United States, 2017. MMWR Morb Mortal Wkly Rep 67: 1225-1232.
4. Smith PH, Mazure CM, McKee SA (2014) Smoking and mental illness in the U.S. population. Tob Control 23: 147-153.
5. Glasheen C, Hedden SL, Forman-Hoffman VL, Colpe LJ (2014) Cigarette smoking behaviors among adults with serious mental illness in a nationally representative sample. Ann Epidemiol 24: 776-780.
6. Hagman BT, Delnevo CD, Hrywna M, Williams JM (2008) Tobacco use among those with serious psychological distress: results from the national survey of drug use and health, 2002. Addict Behav 33: 582-592.
7. Cole RD, Parikh V (2019) Nicotine dependence in schizophrenia: Contributions of nicotinic acetylcholine receptors. In: Preedy VR (edn). Neuroscience of nicotine. Academic Press: 135-143.
8. Parikh V, Kutlu MG, Gould TJ (2016) nAChR dysfunction as a common substrate for schizophrenia and comorbid nicotine addiction: Current trends and perspectives. Schizophr Res 171: 1-15.
9. Wing VC, Wass CE, Soh DW, George TP (2012) A review of neurobiological vulnerability factors and treatment implications for comorbid tobacco dependence in schizophrenia. Ann NY Acad Sci 1248: 89-106.
10. Aschbrenner KA, Dixon LB, Naslund JA, Bienvenida JCM, McManus KL, et al. (2017) An online survey of family members' beliefs and attitudes about smoking and mental illness. J Dual Diagn 13: 179-183.
11. Aschbrenner KA, Naslund JA, Gill L, Bartels SJ, O'Malley AJ, et al. (2017) Preferences for smoking cessation support from family and friends among adults with serious mental illness. Psychiatr Q 88: 701-710.
12. Brown CH, Medoff D, Dickerson FB, Fang LJ, Lucksted A, et al. (2015) Factors influencing implementation of smoking cessation treatment within community mental health centers. J Dual Diagn 11: 145-150.
13. Himelhoch S, Daumit G (2003) To whom do psychiatrists offer smoking-cessation counseling? Am J Psychiat 160: 2228-2230.
14. Colton CW, Manderscheid RW (2006) Congruencies in increased mortality rates, years of potential life lost, and causes of death among public mental health clients in eight states. Prev Chronic Dis 3: 42.
15. Druss BG, Zhao L, Von Esenwein S, Morrato EH, Marcus SC (2011) Understanding excess mortality in persons with mental illness: 17-year follow up of a nationally representative US survey. Med Care 49: 599-604.
16. Brink M, Green A, Bojesen AB, Lamberti JS, Conwell Y, et al. (2019) Excess medical comorbidity and mortality across the lifespan in schizophrenia: A nationwide Danish register study. Schizophr Res 206: 347-354.
17. McClave AK, McKnight-Eily LR, Davis SP, Dube SR (2010) Smoking characteristics of adults with selected lifetime mental illnesses: Results from the 2007 National Health Interview Survey. Am J Public Health 100: 2464-2472.
18. Steinberg ML, Williams JM, Li Y (2015) Poor mental health and reduced decline in smoking prevalence. Am J Prev Med 49: 362-369.
19. Tsoi DT, Porwal M, Webster AC (2013) Interventions for smoking cessation and reduction in individuals with schizophrenia. Cochrane Database Syst Rev 2: CD007253.
20. Cather C, Pachas GN, Cieslak KM, Evins AE (2017) Achieving smoking cessation in individuals with schizophrenia: Special considerations. CNS Drugs 31: 471-481.
21. Banham L, Gilbody S (2010) Smoking cessation in severe mental illness: What works? Addiction 105: 1176-1189.
22. Cheung YW (2000) Substance abuse and developments in harm reduction. CMAJ 162: 1697-1700.
23. Lubin JH, Caporaso NE (2006) Cigarette smoking and lung cancer: Modeling total exposure and intensity. Cancer Epidemiol Biomarkers Prev 15: 517-523.
24. Tindle HA, Stevenson Duncan M, Greevy RA, Vasan RS, Kundu S, et al. (2018) Lifetime smoking history and risk of lung cancer: Results from the Framingham Heart Study. J Natl Cancer Inst 110: 1201-1207.
25. Yoo JE, Han K, Shin DW, Jung W, Kim D, et al. (2022) Effect of smoking reduction, cessation, and resumption on cancer risk: A nationwide cohort study. Cancer 128: 2126-2137.
26. Chang JT, Anic GM, Rostron BL, Tanwar M, Chang CM (2021) Cigarette smoking reduction and health risks: A systematic review and meta-analysis. Nicotine Tob Res 23: 635-642.
27. Farkas AJ (1999) When does cigarette fading increase the likelihood of future cessation? Ann Behav Med 21: 71-76.
28. Foxx RM, Brown RA (1979) Nicotine fading and self-monitoring for cigarette abstinence or controlled smoking. J Appl Behav Anal 12: 111-125.
29. Hughes JR, Carpenter MJ (2006) Does Smoking Reduction Increase Future Cessation and Decrease Disease Risk? A qualitative Review. Nicotine Tob Res 8: 739-749.
30. Pratt SI, Sargent J, Daniels L, Santos MM, Brunette M (2016) Appeal of electronic cigarettes in smokers with serious mental illness. Addict Behav 59: 30-34.
31. Pratt SI, Ferron JC, Brunette MF, Santos M, Sargent J, et al. (2022) E-Cigarette Provision to Promote Switching in Cigarette Smokers With Serious Mental Illness-A Randomized Trial. Nicotine Tob Res 24: 1405-1412.
32. Dawes K, Andersen A, Vercande K, Papworth E, Philibert W, et al. (2019) Saliva DNA Methylation Detects Nascent Smoking in Adolescents. J Child Adolesc Psychopharmacol 29: 535-544.
33. Dawes K, Andersen A, Reimer R, Mills JA, Hoffman E, et al. (2021) The relationship of smoking to cg05575921 methylation in blood and saliva DNA samples from several studies. Scientific Reports 11: 21627.
34. Philibert R, Dogan M, Beach SRH, Mills JA, Long JD (2020) AHRR methylation predicts smoking status and smoking intensity in both saliva and blood DNA. Am J Med Genet B Neuropsychiatr Genet 183: 51-60.
35. Kaur G, Begum R, Thota S, Batra S (2019) A systematic review of smoking-related epigenetic alterations. Arch Toxicol 93: 2715-2740.
36. Fang F, Andersen AM, Philibert R, Hancock DB (2023) Epigenetic biomarkers for smoking cessation. Addict Neurosci: 100079.
37. Gelernter J, Polimanti R (2021) Genetics of substance use disorders in the era of big data. Nat Rev Genet 22: 712-729.
38. Brunette MF, Pratt SI, Bartels SJ, Scherer EA, Sigmon SC, et al. (2018) Randomized trial of interventions for smoking cessation among medicaid beneficiaries with mental illness. Psychiatr Serv 69: 274-280.
39. Evins AE, Cather C, Maravic MC, Reyering S, Pachas GN, et al. (2023) A pragmatic cluster-randomized trial of provider education and community health worker support for tobacco cessation. Psychiatr Serv 74: 365-373.
40. National Academies Press (2018) Public health consequences of e-cigarettes. National Academies Press, Washington, DC, USA.
41. McNeill A, Brose LS, Calder R, Bauld L, Robson D (2018) Evidence review of e-cigarettes and heated tobacco products 2018 A report commissioned by Public Health England. Public Health England, London, UK.
42. Hartmann-Boyce J, McRobbie H, Lindson N, Bullen C, Begh R, et al. (2021) Electronic cigarettes for smoking cessation. Cochrane Database Syst Rev 4: CD010216.
43. Stephens WE (2017) Comparing the cancer potencies of emissions from vapourised nicotine products including e-cigarettes with those of tobacco smoke. Tob Control.
44. Margham J, McAdam K, Forster M, Liu C, Wright C, et al. (2016) Chemical Composition of Aerosol from an E-Cigarette: A Quantitative Comparison with Cigarette Smoke. Chem Res Toxicol 29: 1662-1678.
45. Hajek P, Etter JF, Benowitz N, Eissenberg T, McRobbie H (2014) Electronic cigarettes: Review of use, content, safety, effects on smokers and potential for harm and benefit. Addiction 109: 1801-1810.
46. Hartmann-Boyce J, Butler AR, Theodoulou A, Onakpoya IJ, Hajek P, et al. (2023) Biomarkers of potential harm in people switching from smoking tobacco to exclusive e-cigarette use, dual use or abstinence: Secondary analysis of Cochrane systematic review of trials of e-cigarettes for smoking cessation. Addiction 118: 539-545.