Remission from nicotine dependence among people with severe mental illness who received help/services for tobacco/nicotine use

Abstract Objectives A growing body of evidence supports pharmacological interventions to assist smoking cessation in people with severe mental illness (SMI); that is, lifetime major depressive disorder, bipolar disorders, or schizophrenia. Little is known about whether behavioral services are also associated with high probability of remission from nicotine dependence as compared to other types of help/services received (pharmacological, behavioral, or both). Methods A sample of 726 American lifetime adult smokers with SMI and a history of nicotine dependence, who received help/services for tobacco/nicotine use, were identified. These data came from a limited public use dataset, the 2012–2013 NESARC‐III. Survival analysis was used to compare the probability of remission from nicotine dependence and the time needed for full remission from nicotine dependence by type of help/services received for tobacco/nicotine use. Results Remission was more frequent among those who received behavioral services. In addition, the average time from onset of nicotine dependence until full remission from nicotine dependence was shorter among those who received behavioral services. Conclusions The current study suggests a clinical need for behavioral interventions to promote the probability of remission from nicotine dependence among smokers with SMI. Health care providers could play a role in educating and encouraging smokers with SMI to seek and utilize behavioral services.

enhance remission from nicotine dependence are important for people with SMI.
Several studies have examined factors associated with remission from nicotine dependence in the general and psychiatric populations.
These studies have shown being female, older, white, married, and with higher income, higher education, and later onset of tobacco/nicotine use are associated with higher probability of nicotine dependence remission (Goodwin, Pagura, Spiwak, Lemeshow, & Sareen, 2011;Lopez-Quintero et al., 2011;Peters, Schwartz, Wang, O'Grady, & Blanco, 2014;Segal, Esan, Burns, & Weinberger, 2017). Lopez-Quintero et al. (2011) reported that for the general population, the average time (in years) from nicotine dependence onset until full remission, differs by racial/ethnic groups. Nicotine dependence remission occurs approximately 24 years after dependence onset among whites, compared to 35 years among blacks and 16 years among Hispanics. Yet, little understanding exists about whether the time from onset of nicotine dependence until full remission from nicotine dependence in people with SMI differs by type of help/services received for tobacco/nicotine use.
Our current study extends the work of prior research by estimating the time and probability of full remission from nicotine dependence in people with SMI by type of help/services received for tobacco/nicotine use. This study contributes to the literature by providing data to enhance our understanding about nicotine dependence remission and improve efforts to promote more effective smoking cessation programs aimed at increasing the likelihood of nicotine dependence remission in people with SMI.
To avoid selection biases that might occur in samples recruited from clinical settings, we used a population-based sample to estimate the time and probability of nicotine dependence remission people with SMI. The specific aims for this study were: (a) to estimate whether the time from nicotine dependence onset until full nicotine dependence remission differs by type of help/services received for tobacco/nicotine use (pharmacological, behavioral, and both); (b) to estimate differences in the probability of nicotine dependence remission by type of help/services received, controlling for sociodemographic characteristics, comorbidity with another mental illness and smoking-related factors. We hypothesized that the estimated time from onset of nicotine dependence until full remission and the probability of nicotine dependence remission differs by type of help/services received for tobacco/nicotine use.  (Grant et al., 2014). Briefly, the national sample consisted of 36,309 noninstitutionalized U.S. adults from 18 to 99 years old. Participants were selected through multistage probability sampling, and data were obtained through face-to-face and computer assisted interviews, with an overall response rate of 60.1%. Sample weights were developed to adjust for sampling biases and to ensure representativeness of the U.S. population. The Census Bureau and the U.S. Office of Management and Budget confirmed the ethics protocol for this survey. All participants provided informed consent, and they received $90.00 for their participation. Interview Schedule-5 (AUDADIS-5) that has been shown to have concordance with clinical evaluations (κ = 0.24-0.59; Hasin et al., 2015).
A variable created by the NESARC study team was used to indicate if a participant had lifetime major depressive or bipolar disorder. Lifetime schizophrenia was identified through an item asking, "Did a doctor or other health professional ever tell you that you had schizophrenia or a psychotic illness or episode?" We combined these items into a single variable for lifetime SMI (yes or no).
The AUDADIS-5 also contained a module that assessed tobacco/ nicotine use (Grant et al., 2015). This module included items capturing nicotine dependence criteria with follow-up questions to determine if it occurred before or within the past 12 months (with nicotine dependence defined as meeting more than 2 of the 11 criteria in a single year). Reliability of the DSM-5 nicotine dependence (k = 0.50-0.87, intraclass correlation coefficient [ICC] = 0.83-0.84) has been found to be fair to excellent (Grant et al., 2015). A single item assessed if the participant had ever seen anyone for help related to tobacco/nicotine use. Figure 1 is a consort diagram illustrating the sample selection for this study's final sample (n = 726).

| Remission
Nicotine dependence remission was assessed using a variable indicating whether or not the participant had a current nicotine dependence (past year nicotine dependence). Individuals in the past year, who did F I G U R E 1 Consort diagram not meet any DSM 5 criteria other than craving, also were classified as having remitted.

| Help/services received for tobacco/ nicotine use
This was measured by asking if participants had ever gone (in their entire lifetime) anywhere or seen anyone to get help that was in any way to their use of tobacco or nicotine, or if they did anything else to help them quit or cut down on tobacco or nicotine use. Those responding "yes" were then asked a series of questions listing types of assistance received. Study participants were categorized into three groups based on the type of help/services they received for tobacco/ nicotine use (pharmacological only, behavioral only, or both). Those who responded "yes" to any of the following were classified as obtaining pharmacological service: "use nicotine patches, lozenges, or gum," or "have a health professional prescribe any medicine or drug," whereas those who responded "yes" to any of the following were classified as obtaining behavioral services: "receive acupuncture, acupressure, e-therapy or meditate," "go to counseling, family services, or other social services," "go to a support group," or "use any other methods to help you quit or cut down." Those who received both pharmacological and behavioral services were classified into a third group named "both pharmacological/behavioral treatments."

| Tobacco/nicotine use-related factors
Nicotine use-related factors that were included in this study are selfreported age at tobacco/nicotine first use ("the youngest age reported to questions on age when first used cigarettes, cigars, pipe, or snuff"), age at onset of daily smoking ("age started using tobacco/nicotine every day"), type of tobacco/nicotine product used (e.g., cigarettes, cigars, pipe, and snuff/chewing tobacco), quantity of smoking ("usual quantity when used tobacco/nicotine every day [number of times used tobacco/nicotine every day]"), duration of daily smoking in years ("duration [years] when used tobacco/nicotine every day"), and age at tobacco/nicotine last use ("age when last used tobacco/nicotine").

| Mental and substance use comorbid illnesses
Comorbidity with lifetime DSM-5 mental and substance use comorbid illnesses was also included. The mental comorbid illnesses included dysthymia (persistent chronic depression), any anxiety disorder (e.g., panic, agoraphobia, social phobia, specific phobia, and generalized anxiety), posttraumatic stress disorder, and any personality disorder (e.g., schizotypal, borderline, antisocial). Substance use disorders included alcohol use disorder and any drug use disorder (e.g., sedative/tranquilizer, cannabis, stimulant, cocaine, heroin, hallucination, and inhalant use disorders). Various AUDADIS-5 modules were used to identify participants meeting criteria for mental and substance use comorbid disorders. The psychometric properties of the modules that were used to identify those disorders ranged from fair to excellent (κ = 0.36-0.87, ICCs >0.68; Chou et al., 2016;Grant et al., 2015;Hasin et al., 2015).

| Data analysis
Statistical analysis was conducted using SPSS version 25, taking into account the complex sampling design of the NESARC. Weight and stratification variables were applied to ensure that the data were representative of the U.S. population. Descriptive statistics were used to describe all variables by type of help/services received for tobacco/ nicotine use. Means and standard deviations were computed for continuous variables, and absolute and relative frequencies were calculated for categorical variables. All variables were compared by type of help/services received for tobacco/nicotine use, using ANOVA-tests for continuous variables and Chi-square tests for categorical variables.
The pattern of missing data was checked across all variables, and missing data ranged from 0 to 0.4%. Missingness was randomly distributed with no monotonicity indications, and the results of Little's missing completely at random (MCAR) test were not statistically significant (Chi-square = 6.762, df = 6, p = .347). Hence, we concluded that missing data were probably MCAR, and missingness was ignorable.
Survival analysis was conducted to estimate the time and probability of nicotine dependence remission by type of help/services received, controlling for other potential covariates and confounders.
The outcome of interest was "remission from nicotine dependence." The difference in age at full nicotine dependence remission minus the age at onset of nicotine dependence defined the length of time participants had nicotine dependence. Participants with nicotine dependence without remission were defined as censored and censoring time was their age at interview minus age at nicotine dependence onset.
The Kaplan-Meier (KM) method was used to obtain univariate descriptive statistics, including the median time from age at onset of nicotine dependence until age at full nicotine dependence remission by type of help/services received. The log-rank test was used to examine whether the length of time of having nicotine dependence differed by type of help/services received. Then a Cox proportional hazard model was used to analyze the association between the services received and the probability of nicotine dependence remission, controlling for potential confounders and covariates. The Cox proportional hazard model assumptions were checked including linearity of the continuous independent variables, multicollinearity, and proportionality of hazards. The value of the hazard ratio indicated the probability of remission from nicotine dependence with each additional year having nicotine dependence.

T A B L E 1
Characteristics of individuals diagnosed with severe mental illness and lifetime nicotine use disorder by type of help/services received for tobacco/nicotine use (NESARC-III, N = 663)
The proportion with some college was higher among those who received both types of services for tobacco/nicotine use; whereas the proportion between 45 and 65 years of age was higher among those who received behavioral services. Personal income under $35,000 was also reported more often by those receiving behavioral services.
T A B L E 2 Remission from nicotine dependence by mental comorbid illnesses, and smoking-related factors among individuals with a history of lifetime severe mental illness who received help/services for tobacco/nicotine use (NESARC-III, N = 663)

| Mental comorbid illnesses
Comorbid mental illnesses were common in the sample. Table 2 compares comorbid disorders by type of help/services received for tobacco/nicotine use. Only comorbidity with a substance use disorder differed across the type of help/services received for tobacco/nicotine use (χ 2 = 1.9, p < .05). Comorbidity with a substance use disorder was more common among those who only received behavioral services (81.8%) than those receiving pharmacological or both services.

| Smoking-related factors
Factors, such as age at first tobacco/nicotine use (F = 0.4, p < .05), duration of daily smoking (F = 1.7, p < .01), and age at smoking cessation (F = 2.6, p < .01) also differed across the type of help/services received for tobacco/nicotine use. Those who received both services reported earlier age at first tobacco/nicotine use and older age at smoking cessation compared to those who received pharmacological and behavioral services. On the other hand, those who received behavioral reported less duration of daily smoking compared to those who only received pharmacological or both services (see Table 2).

| Remission from nicotine dependence by type of help/services
Essentially, out of people with SMI who remitted from nicotine dependence (n = 828), only 19.1% had received help/services for tobacco/nicotine use. Out of those who received help/services (n = 663), 21.7% had remitted from nicotine dependence. Remission from nicotine dependence differed by type of help/services received for tobacco/nicotine use (χ 2 = 9.3, p < .05). The proportion that remitted with a history of receiving pharmacological services was 17.6%.
Remission was more frequent among those who received behavioral services (28.5%) or when both types of services were received (19.6%).
The cumulative probability of nicotine dependence remission differed by type of help/services received (log rank = 7.8, p < .05). The Kaplan-Meier plot examined the probability of remission from nicotine dependence over time (see Figure 2). The curve of those who received behavioral services only was higher than for those who received pharmacological services or those who received both, indicating that those who received behavioral services had the highest probability of nicotine dependence remission. The average time from onset of nicotine dependence until full nicotine dependence remission among those who received behavioral services was 35 years (95% CI: 32.2, 37.6), compared to 37 years (95% CI: 32.1, 42.3) among those who received pharmacological services and 47 years (95% CI: 43.9, 49.8) for those who received both.
The associations among type of help/services received and the probability of nicotine dependence remission were initially tested using univariate Cox proportional hazard models (see Table 3). Then the potential associations between type of help/services and the probability of remission from nicotine dependence was tested using three multivariate models adjusted for sociodemographic characteristics. Model (1) included SMI type and comorbidity with another mental illness; Model (2) contained smoking-related factors; and Model (3) combined Model 1 and 2 covariates. All potential correlates reaching statistical significance at 0.15 level in the univariate models were included in the third (final) multivariate model. Table 3 shows the results of the association between each type of help/services received for tobacco/nicotine use and the probability of remission from nicotine dependence. This association was consistent for Models 2 and 3, and differed for Model 1. In Model 1, the probability of nicotine dependence remission was higher among those who received one type of service compared to those who received both, whereas the probability of nicotine dependence remission in F I G U R E 2 Probability of remission from nicotine dependence over time by type of help/service received for tobacco/ nicotine use Models 2 and 3 was higher among those who received behavioral or both services compared to those who received pharmacological services. The inferiority of combined services in the crude model and Model 1 suggests that smoking factors might have influenced the delivery of combined services. In Model 3, the value of hazard ratio (HR = 1.95, 95% CI: 1.93, 1.97) for those who received behavioral services means that for each additional year of nicotine dependence, the probability of nicotine dependence remission was 1.95 times that of those who received pharmacological services, after adjustment for other potential covariates and confounders.
Additionally, the findings showed how SMI type, smoking-related factors, comorbidity with other mental illnesses were significantly associated with the probability of nicotine dependence remission (see Table 3).
T A B L E 3 Cox proportional hazards model of the association between type of help/services received for tobacco/nicotine use and the probability of remission from nicotine use disorder (NESARC-III, N = 663)

| DISCUSSION
The current study sought to estimate the time and probability of remission from nicotine dependence by type of help/services received for tobacco/nicotine use in people with SMI. We observed that those who received behavioral services had a higher probability of remission from nicotine dependence compared to those who received pharmacological or both services. Interestingly, the probability of remission from nicotine dependence among those who only received behavioral services was higher than those who received both. The finding in some of our analyses that combined services were associated with a lower likelihood of remission suggests that smoking factors and the severity of nicotine dependence might influence the delivery of combined services.
Our findings were consistent with previous studies that found an association between receiving services for tobacco/nicotine use and nicotine dependence remission (Aldi et al., 2018;Annamalai et al., 2015;Rüther et al., 2014;Stubbs, Vancampfort, Bobes, De Hert, & Mitchell, 2015). In addition, the current study extends the prior research with population-based data.
Our study findings differed from studies that found pharmacological services were more helpful in people with SMI compared to behavioral services (Aldi et al., 2018;Secades-Villa et al., 2017). However, our findings were consistent with research that has found a combination of pharmacological services and psychoeducation is more Our study suggests that smokers with SMI would benefit from receiving behavioral services for tobacco/nicotine use to improve the probability of nicotine dependence remission. Hence, smokers with SMI should be encouraged to seek and use behavioral services to enhance their odds of nicotine dependence remission. According to Evins et al. (2014), receiving pharmacological services for at least 1 year has shown promise for long-term abstinence among smokers with SMI. Our study is also congruent with the recent U.S. Public Health Service guidelines for smoking cessation in the general population that recommend receiving a combination of pharmacological services and counseling for smoking cessation as it is more effective than received pharmacological services alone (Maglione et al., 2017).
The study findings should be interpreted while considering common limitations in most studies based on data from a national survey.
First, information on nicotine dependence and remission from nicotine dependence was based on self-report measures and not confirmed by objective measures. Second, data on smoking related factors and receiving services for tobacco/nicotine use might be subject to recall bias and cognitive impairment associated with mental illness. The help/service categories are crudely assessed and additional information on the timing (when attempt made, sequencing of services, etc.) and how often quit attempts were made would provide greater insight into the service utilization patterns of those attempting to quit.
Third, NESARC survey excludes institutionalized populations, who might have a higher probability of having nicotine dependence.
Fourth, differential service delivery was not considered in this study due to data limitation. Fifth, the data for this study were collected 6-7 years ago, and therefore might not reflect current practices. Nonetheless, our study used a comprehensive definition of behavioral services by including psychoeducation or counseling, along with family or other social services; support groups; acupuncture, acupressure, etherapy or meditation; and any other methods. In addition, study strengths include the large sample size allowing for subgroups analysis, representativeness and relative generalizability, and the careful assessment methods.

| CONCLUSION AND IMPLICATIONS
The current study suggests a clinical need for behavioral interventions to promote the probability of nicotine dependence remission among smokers with SMI. Health care providers could educate and encourage smokers with SMI to seek and use behavioral services for tobacco/nicotine use to improve the probability of their remission from nicotine dependence and facilitate prolonged abstinence. Future research should take into account differential service delivery to examine associations between receiving behavioral services and remission from nicotine dependence. In addition, it would be beneficial to compare the effectiveness of different behavioral services and evaluate them for smokers with SMI.

DECLARATION OF INTEREST STATEMENT
No conflicts of interest were declared by any author.

ACKNOWLEDGMENTS
The National Epidemiologic Survey on Alcohol and Related Conditions-III (NESARC-III) is funded by the National Institute on Alcohol Abuse and Alcoholism (NIAAA) with the supplemental support from the National Institute on Drug Abuse, National Institute of Health.
Sponsors and funders of the NESARC-III had no role in the design and conduct of the study, management, analysis, and interpretation of the data, and preparation, review, and approval of the manuscript.

DISCLAIMER
The view and opinions expressed in this report are those of the authors and should not be constructed to represent the views of any of the sponsoring organization or agencies or the U.S. government.