Tobacco Cessation Services in Addiction Treatment: What Do Clients Say?

Twenty-four publicly funded addiction programs participating in the National Institute on Drug Abuse Clinical Trials Network (CTN) were selected in 2013, including seven outpatient, 10 residential, and seven methadone programs. Publicly funded programs receive more than 50% of funding from government sources and serve low-income populations. Methadone programs provide daily on-site methadone administration for opioid use disorder. Nonmethadone outpatient and residential programs typically treat all substance use disorders. Outpatient programs involve one or two visits per week, whereas residential programs involve structured daily treatment with clients living on site (8). Our team visited each program in 2015 and 2016, recruiting up to 50 clients per program per year for an anonymous online survey. The 2015 sample (N=1,125) represented 17% of all 6,801 active clients in these programs when site visits were conducted. The 2016 survey asked participants if they recalled taking the survey in the previous year, to remove probable duplicate cases. All who provided informed consent also completed the survey (N=2,119) and received a $20 gift card. Program directors were interviewed by phone about program tobacco policies. Program selection, participant recruitment, and coding of tobacco-free grounds are reported elsewhere (2). Study procedures were approved by the University of California, San Francisco, Institutional Review Board.

In addition to indicating demographic information, participants provided information for a number of independent variables, including reporting the number of days in the past month when their physical health or mental health was not good (9). Four categories were defined by <14 days of “physical and mental distress” (i.e., low health distress), ≥14 days of “physical health distress,” ≥14 days of “mental health distress,” and ≥14 days of both “physical and mental health distress.” Current smoking was defined as having smoked at least 100 cigarettes in one’s lifetime, and self-reporting as a current smoker (https://www.cdc.gov/nchs/nhis/tobacco/tobacco_glossary.htm). Current smokers reported number of cigarettes smoked per day (CPD) and whether they made a quit attempt in the past year, were thinking of quitting in the next 30 days (10), and wanted help with quitting. Smokers estimated the chance (0%−100%) they would get lung cancer, have trouble catching their breath, or have a heart attack; the mean of these percentages represented perceived health risk of smoking (11). Each survey was coded as to whether the participant’s program had tobacco-free grounds or not.

These independent variables were selected because quit attempts are associated with both readiness to quit and receipt of cessation services in addiction treatment (5), because persons with health concerns are more likely to quit smoking while in addiction treatment (12), and because implementation of tobacco-free grounds is associated with increased cessation services reported by clients (2).

For tobacco service outcomes, clients reported whether any staff member had asked about their smoking or advised them on how to quit smoking. To assess how smokers were assisted with quitting, we used four measures: any referral, any counseling, any medication, and counseling and medication, as recommended by Clinical Practice Guidelines (7). Receipt of a referral to a smoking cessation specialist (yes or no) or to a smoking quitline (yes or no) was coded as having received any referral. Participants who reported having attended a cessation support group (yes or no) or that their counselor encouraged them to quit smoking or arranged an appointment to discuss quitting (responses to both items were never, occasionally, often, very often, or always, dichotomized as no for never or occasionally and yes for the others) were coded as having received any counseling. Receipt of nicotine replacement therapy or other cessation medication (bupropion, varenicline) was coded as having received any medication. Last, we combined the counseling and medication measures to assess the proportion who received this recommended combination of services (7).

Comparing the 2015 and 2016 samples showed few differences on demographic, smoking behavior, or tobacco service measures. Thus, we collapsed responses across time and, using multiple logistic regression, tested factors associated with tobacco service outcomes. We collapsed data across years, and after removing 145 possible duplicate cases and 14 with indeterminate smoking status, we derived a final sample size of 2,119.

The first outcome, whether clients were asked about smoking, included smokers (N=1,633) and nonsmokers (N=486). Smoking variables were not available for nonsmokers, so the model included current smoking status, demographic characteristics (age, gender, race or ethnicity, education, time in treatment, treatment type, and health distress), and tobacco-free grounds status. Analyses accounted for nesting of participants within the program via generalized estimating equation models for correlated data. For the health distress measure, low health distress was used as the referent. The remaining five outcomes applied to smokers only. Each model was the same as the one already described, but without smoking status and with the addition of smoking characteristics (CPD, quit attempt in the past year, thoughts of quitting, health risk perception, wanted help quitting). The rate of missing data was <1% for each independent variable except for health distress (3%) and ≤5% for all multivariate models. Missing data were not expected to affect the results. Analyses were conducted with SAS version 9.4.

Participants (N=2,119) had a mean±SD age of 38.2±11.9, 47% (N=989) were women, and 79% (N=1,666) had a high school diploma or GED. The sample included white (57%, N=1,199), African-American (17%, N=351), and Hispanic (13%, N=282) participants, with fewer persons of Native American (5%, N=102), Asian or Pacific Islander (2%, N=47), or other races or ethnicities (7%, N=138). Most (69%, N=1,425) reported low health distress, whereas fewer reported mental health distress (17%, N=341), physical health distress (4%, N=91), or both (10%, N=197). Participants were recruited from outpatient (31%, N=647), residential (40%, N=854), and methadone (29%, N=618) programs.

Most respondents (77%, N=1,633) were current smokers, consuming 13.2±8.5 CPD. Half (50%, N=878) had made a quit attempt in the past year, and 26% (N=416) were thinking of quitting in the next 30 days. Mean perceived health risk from smoking was 47%±25%. Many (40%, N=714) wanted help with quitting smoking, and 28% (N=584) were in programs that had tobacco-free grounds. The 2016 sample included fewer methadone clients (χ²=7.3, df=2, p=0.026) and reported shorter time in treatment than in 2015 (64.3±146.1 days vs. 86.4±162.8 days; t=3.3, df=2,112, p=0.001), and these variables were included in the analysis.

Most clients (76%, N=1,602) had been asked their smoking status, and among smokers, 53% (N=870) had been advised on how to quit. Smokers reported receiving any referral (46%, N=753), any counseling (41%, N=671), any medication (26%, N=424), and both counseling and medication (17%, N=284).

Participants with both mental and physical health distress had lower odds of referral to tobacco cessation services (odds ratio [OR]=0.69; Table 1). Smokers thinking of quitting in the next 30 days had greater odds of receiving cessation counseling (OR=1.43). As seen in Table 1, smokers who wanted help with quitting, compared with those who did not, had greater odds of receiving every tobacco-related service (Table 1). Similarly, participants in programs with tobacco-free grounds, compared with participants in programs without, had greater odds of receiving three of the five services measured (Table 1).

These data also may be interpreted by using predicted probabilities, or the probability of receiving a service when the value of the independent variables is known (data not shown). For a smoker who did not want help with quitting and who was not in a program with tobacco-free grounds, the probability of receiving advice on how to quit was 49%. This probability rose to 60% if the smoker wanted help with quitting and to 72% if the same smoker was in a program with tobacco-free grounds. Similarly, the probabilities of receiving tobacco-related counseling were 32% for smokers who did not want help with quitting and who were not in a program with tobacco-free grounds, 46% for those who wanted help with quitting, and 62% if the same smoker attended a program with tobacco-free grounds. Last, the probabilities of receiving tobacco cessation medication in the same three instances were 14%, 29%, and 68%, respectively.

Clients who wanted help with quitting smoking reported higher odds of receiving tobacco cessation services. It is possible that clients request cessation services, which are then provided. Increasing client interest in quitting might be done through motivational interviewing (13) or patient empowerment interventions (14), both of which are shown to be effective among smokers younger than 50 who, as in our sample, may not have serious health concerns. Smokers in programs with tobacco-free grounds also had higher odds of being advised to quit and of receiving counseling and medication. This finding is consistent with literature showing increased use of tobacco-related services and greater clinician support of client smoking cessation in addiction treatment programs with tobacco-free grounds (2). Availability of tobacco cessation services may also be greater in programs with tobacco-free grounds. Barriers to tobacco-free grounds policies include staff misconceptions about smoking cessation during addiction treatment. However, staff training may address these barriers and increase client receipt of tobacco services (6). Increasing client motivation to quit and increasing adoption of tobacco-free grounds are feasible low-technology and low-cost strategies to address smoking in addiction treatment.

Study limitations include generalizability because of lack of information about nonrespondents and because programs recruited through the CTN are shown to differ from non-CTN programs (15). The program sample included only publicly funded programs, and participant samples were either U.S. Census based (in residential programs) or convenience samples (in other programs). The cross-sectional design did not permit causal interpretation. The survey did not collect data on psychiatric diagnoses, which may be associated with receipt of tobacco services. Receipt of tobacco-related services was based on self-report and not corroborated by chart review. The study did not collect data on availability of tobacco services in participating programs.

Among clients recruited from addiction programs within a national research network, smoking behavior occurred at an epidemic rate. Robustly associated with receipt of tobacco services were whether the client wanted to quit smoking and whether the program had tobacco-free grounds. Treatment programs may address client smoking in part through interventions that increase motivation to quit. Agencies that fund, license, and regulate such programs should require tobacco-free grounds policies. These steps are necessary to reduce smoking, and related morbidity and mortality, among those who have sought health care in addiction treatment systems.

The authors gratefully acknowledge the support of directors who agreed that their program could participate in the study, the program staff who coordinated site visits and data collection, and the clients who gave their time to complete study surveys.