Abstract
Introduction
Smoking denormalization has been paralleled by reduced smoking prevalence, but smoking rates among socioeconomically disadvantaged populations remain high. The social unacceptability of smoking has also led to increased perceptions of smoking-related stigma. By examining how smoking stigma influences cessation intervention effectiveness, we can better tailor interventions to socioeconomically disadvantaged smokers.
Aims and Methods
Data are from a randomized controlled trial evaluating the effectiveness of a proactive cessation intervention on abstinence. Current smokers enrolled in Minnesota Health Care Programs were randomized to proactive outreach (n = 1200) or usual care (n = 1206). The intervention included mailings, telephone outreach, counseling, and access to free cessation treatments. Using baseline measurements, groups with lower (n = 1227) and higher (n = 1093) perceived stigma were formed. Intervention, stigma, and their interaction term were added to a logistic regression modeling abstinence at 12 months.
Results
Lower perceived smoking-related stigma was associated with less support for quitting, lower rates of physician quitting advice, and less motivation for quitting. A logistic regression modeling abstinence found a significant intervention × stigma interaction. The proactive intervention was more effective among smokers with lower perceived smoking-related stigma (odds ratio 1.94, 95% confidence interval, 1.29 to 2.92) than those with higher perceived smoking-related stigma (odds ratio 1.04, 95% confidence interval, 0.70 to 1.55).
Discussion
Smokers with lower perceived smoking-related stigma had social environments that were conducive to smoking, received less physician advice to quit, and were less motivated to quit than higher stigma smokers. Despite these barriers, the intervention was more effective for lower stigma smokers, suggesting that proactive outreach is an efficient treatment for these hard-to-reach smokers.
Implications
Smoking denormalization has led to increased perceptions of smoking-related stigma among many smokers; however, little is known about how this stigma influences the cessation process. In the present study, smokers with lower levels of perceived smoking-related stigma lived in social environments that were more conducive to smoking and were less motivated to quit than higher stigma smokers. Despite these barriers, our proactive outreach cessation intervention was more effective for lower stigma smokers, suggesting that interventions which utilize proactive outreach to stimulate interest in quitting and offer facilitated access to free cessation treatments are an effective treatment approach for these hard-to-reach smokers. These strategies may be particularly effective for motivating smokers enrolled in government-subsidized health insurance programs to take advantage of cessation resources.
Introduction
The past several decades have seen a dramatic decline in the rate of smoking in the United States, with a current prevalence of less than 17%.1 One proposed mechanism for this reduction in smoking prevalence is the process of smoking denormalization,2 which aims to reduce smoking rates by making it more difficult for individuals to continue to smoke. Examples of smoking denormalization include media campaigns expounding the dangers of smoking, the proliferation of clean indoor air laws, and employer-based incentives for quitting smoking.2–5 Smoking denormalization has been associated with clear public health benefits,6–8 but policy approaches to smoking reduction can have the unintended effect of instituting a form of structural discrimination that can impinge on the social and economic opportunities available to continuing smokers.3,9–11 This can lead to stigmatization when smokers internalize feelings of shame or emotional discomfort for engaging in a behavior that is perceived to be socially unacceptable, irresponsible, or immoral.11
Although the overall prevalence of smoking has decreased, the smoking disparity between populations of high and low socioeconomic status continues to grow. Indeed, the prevalence of smoking among those below the poverty level is greater than 26%, compared to 15% among those at or above the poverty level.1 One possible mechanism for the widening socioeconomic disparity in smoking prevalence are the stringent antismoking policies associated with smoking denormalization, policies which have increased the geographic and cultural segregation of socioeconomically disadvantaged smokers.9,11 The creation of these “smoking islands”9,11 can have the counterintuitive effect of increasing the social acceptability of smoking within socioeconomically disadvantaged populations,9,11 thereby reducing perceptions of smoking-related stigma and contributing to the perpetuation of smoking.2,7,12 Thus, while smoking denormalization is associated with heightened perceptions of smoking-related stigma among those of higher socioeconomic status, it appears to have the opposite effect among socioeconomically disadvantaged populations.
Many health-related stigmas, including those associated with conditions like AIDS and mental health disorders, give rise to negative psychosocial consequences including perceived discrimination, social rejection, and lowered self-esteem.13,14 However, the effects of smoking-related stigma are more complex.10 Research has shown that smoking-related stigma is associated with deleterious health behaviors, including a reluctance to seek cessation assistance and a tendency to conceal smoking behaviors from one’s health care provider.15 Conversely, it has also been shown that increases in smoking-related stigma are associated with lowered smoking prevalence at the state level and increased intention to quit at the individual level.12 Furthermore, perceptions of the social unacceptability of smoking are associated with lowered smoking prevalence at the state level12 and increased likelihood of abstinence at the individual level.2 Thus, in addition to smoking denormalization policies that make it more difficult to be a smoker, perceptions of smoking-related stigma may also contribute to reductions in smoking prevalence.
Socioeconomically disadvantaged smokers who are motivated to quit experience considerable barriers to achieving cessation. At the psychosocial level, socioeconomically disadvantaged smokers tend to view smoking as a normative behavior9,11 and have lower self-efficacy,16 factors which are strongly associated with a reduced likelihood of cessation.17,18 At the health care provider level, many physicians lack the time or motivation to discuss cessation treatments with socioeconomically disadvantaged smokers19,20 and often do not discuss these treatments in a culturally sensitive manner.21 As such, proactive cessation strategies, which promote heightened contact with smokers and facilitate access to free cessation treatment, may be a particularly effective approach for minimizing psychosocial and provider-based barriers to treatment among socioeconomically disadvantaged smokers. Furthermore, proactive outreach interventions have previously been shown to be effective in other socioeconomically disadvantaged populations22 and among Veterans,23 suggesting that population-based outreach strategies are a promising approach to smoking cessation for disadvantaged and vulnerable populations.
In this secondary analysis of a randomized controlled smoking cessation trial for socioeconomically disadvantaged smokers, our aims were twofold. First, we sought to better understand smoking-related stigma by examining the correlates of this stigma with respect to mental health, social environment characteristics, health care provider experiences, smoking history, and psychosocial beliefs. Next, we evaluated whether baseline perceptions of smoking-related stigma impact the effectiveness of a proactive cessation intervention.
To date, no studies have explored how smokers’ perceptions of smoking-related stigma influence the effectiveness of cessation interventions. On one hand, smokers who perceive greater stigma might benefit more from a proactive cessation intervention as they may have more extrinsic motivation to quit and would make better use of the opportunity for assisted cessation. On the other hand, smokers who perceive lower levels of stigma might benefit more from a proactive smoking cessation intervention as it is designed to facilitate access to cessation treatments, thus making the prospect of cessation seem less arduous for smokers who may be less ready to quit. As such, our primary aim was to examine the role of baseline perceptions of smoking-related stigma as a potential moderator of the effectiveness of a proactive outreach intervention with respect to rates of 6-month prolonged abstinence at 12-month follow-up.
Methods
Study Design
The data presented here were obtained from a two-group randomized controlled trial that demonstrated the effectiveness of a proactive care tobacco cessation outreach intervention, compared to usual care, for increasing population-level 6-month prolonged abstinence at 12-month follow-up.24 The study surveyed adults enrolled in Minnesota Health Care Programs (MHCP) and randomized responding smokers to intervention arms. MHCP is a state-funded health insurance plan for low-income Minnesota residents composed of two publicly subsidized health care assistance programs: Medicaid and MinnesotaCare. Baseline survey stigma measure data were used to categorize respondents into lower stigma (n = 1227) and higher stigma (n = 1093) groups.
Proactive Care Intervention
The proactive care intervention was designed to overcome the health care provider and psychosocial barriers to smoking cessation treatment experienced by low-income smokers and had two primary components. The first component involved a system of proactive outreach in the form of tailored mailings and telephone calls. An invitation packet describing the tobacco treatment services available to MHCP enrollees was sent to those in the intervention arm. Participants then received an outreach call from a counselor trained in motivational interviewing and cessation treatment. The call was meant to (1) provide motivational advice on how to quit smoking; (2) promote self-efficacy; (3) encourage individuals to participate in cessation treatment; and (4) provide information on the safety, efficacy, and benefits of nicotine replacement therapy.
The second component involved facilitated access to free, comprehensive cessation treatment. An 8-week course of nicotine replacement therapy was mailed directly to participants before their target quit date. Participants were also offered intensive, telephone-based behavioral counseling. This counseling involved a motivational interviewing and cognitive-behavioral approach to addressing substance abuse.
Usual Care
Participants in usual care had access to the existing smoking cessation care structure provided by the MHCP system. All MHCP enrollees have a primary care provider and may consult with their provider to receive cessation treatment. Thus, participants assigned to usual care were able to receive similar cessation treatment options as those offered to the intervention group. However, participants in usual care were not specifically invited to receive cessation treatment or provided with facilitated access to these treatments and had to self-initiate contact with their provider in order to obtain treatment.
Measures
Abstinence
A self-report measure assessed 6-month prolonged abstinence at 12-month follow-up.25 Participants who reported smoking at least once on seven consecutive days or at least once on two consecutive weekends in the 6-month period prior to the follow-up survey were considered continuing smokers. As per the Society for Research on Nicotine and Tobacco workgroup recommendations,25 a prolonged abstinence measure was chosen because the not-even-a-puff criterion “is overly stringent in that it counts as failures long-term abstainers who have a single slip”. Furthermore, at the time of study implementation 6-month prolonged abstinence was considered the gold standard for smoking cessation trials.
Stigma
Perceptions of smoking-related stigma were assessed using a measure adapted by Stuber et al. 11 from the Mental Health Consumers’ Experience of Stigma Scale developed by Link et al.26,27 This measure used a mean score of seven items, with each item assessed on a scale from 1 to 5 and higher scores indicating greater perceptions of smoking-related stigma. Examples of items include, “I have worried that others view me unfavorably because I smoke” and “I have been treated as less intelligent by others because I smoke” (see Supplementary Figure 1).
Demographics
Race/ethnicity, education, employment status, and income, in addition to the stratification variables of sex, age category, and insurance program were assessed at baseline from administrative records and survey.
Mental Health
The anxiety measure was created by taking the sum of seven items from the Patient-Reported Outcomes Measurement Information System instrument. Items were assessed on a scale from 1 to 5, with higher scores indicative of greater anxiety. The depression measure was created by taking the sum of two items from the Patient-Reported Outcomes Measurement Information System instrument.28 Items were assessed on a scale from 1 to 4, with higher scores indicating greater depressive thoughts.
Smoking History
Standard questions from the California Tobacco Survey29 and the CDC Behavioral Risk Factor Surveillance System30 assessed smoking history, including lifetime duration of smoking, time until first cigarette, and past-year quit attempts.
Health Care Provider
Access to Health Care
A composite variable measuring access to health care was created by taking the sum of five items assessing participants’ perceptions of their transportation-related difficulties, difficulties obtaining a prompt appointment, pre-existing work or family responsibilities, difficulties with the cost of care, and difficulties with medication costs. Items were assessed on a three-point scale, with higher scores indicating greater health care provider barriers. Analyses run on the resulting summary measure for these items produced a Cronbach’s alpha reliability score of 0.71. Participants were also asked if they had a regular doctor.
Health Care Provider Cessation Advice
Healthcare Effectiveness Information and Data Set tobacco performance measures31 assessed participants’ past-year experiences with their health care providers, including receipt of any cessation-related care, advice to quit, to use cessation products, or ways of quitting other than products.
Health Care Provider Bias/Cultural Competence
A composite variable to describe potential provider bias was created by taking a mean score from three items on the Physician Bias and Interpersonal Cultural Competence Measures Scale.32 These items assessed the participants’ perceptions of being treated with respect by their doctor, their doctor’s understanding of the participant’s background and values, and the perception that the doctor looks down on the participant’s way of life. Items were assessed on a scale from 1 to 5, with higher values indicating greater physician bias. Analyses run on the resulting summary measure for these items produced a Cronbach’s alpha reliability score of 0.74.
Social Environment
Social Support
A composite variable measuring perceived social support for cessation was created by taking the mean of two support-related variables.29 This variable was assessed on a scale from 1 to 5, with higher scores indicating greater support. Analyses run on the resulting summary measure for these items produced a Cronbach’s alpha reliability scores of 0.53.
Social Norms
Participants reported the proportion of their close friends and family who smoked.29
Home Environment
Participants indicated whether they lived with a child under the age of 18 and whether they lived with another smoker. Another measure assessed smoking rules within the home, categorized as “Smoking is not allowed anywhere,” “Smoking is allowed in some places or at some times,” or “Smoking is allowed everywhere”.29
Cessation Beliefs
Self-Efficacy
Self-efficacy for quitting was assessed on a scale from 1 to 10, with higher values indicating greater confidence in quitting.33
Readiness to Quit
Readiness to quit on a scale from 1 to 10 was assessed using the Contemplation Ladder, with higher values indicating greater confidence in quitting.34
Cessation Treatment Utilization
Participants indicated whether they had used cessation medication or behavioral counseling in the past year.
Statistical Analysis
A median split of the baseline smoking-related stigma measure was used to separate participants into stigma categories. Although the original stigma measure is based on items assessed on a five-point scale, using this measure in a more categorical fashion does have precedent in the literature as Stuber et al. have previously used this measure to separate participants into “high, medium, and low” categories.11 As such, participants with a mean score of 2 or lower were classified as having lower perceived stigma (n = 1227); those with a mean score of greater than 2 were classified as having higher perceived stigma (n = 1093). The lower and higher stigma groups were then compared across baseline demographics, smoking history, health care provider characteristics, social environment characteristics, and cessation-related psychosocial beliefs using two-sample t tests and Pearson chi-square tests.
A logistic regression modeled 6-month prolonged abstinence at the 12-month follow-up using intervention condition (proactive outreach vs. usual care), stigma group (lower vs. higher), and an intervention by stigma interaction term as predictors. This model also controlled for the stratification variables of sex, age, and insurance program.
Given the significant intervention by stigma interaction, two simple effects models were estimated for the lower stigma and higher stigma groups, respectively. These logistic regressions modeled 6-month prolonged abstinence at 12-month follow-up with intervention condition as the predictor, while controlling for sex, age, and insurance program. Adjusted odds ratios (aORs) and 95% confidence intervals (CIs) are reported for the predictors in these models.
Missing Data
In the regression models performed, there were 252 (19%) and 249 (23%) missing observations, respectively, for the lower and higher stigma groups due to either missing stigma measures or missing abstinence outcomes. As part of the larger study, an iterative regression model process was employed to impute 10 complete imputed versions of the baseline survey data.35 The relative efficiencies of this finite imputation process, using 10 imputations, were greater than 0.98; similarly, across the different survey measures, the increase in variance stemming from the missing data ranged from near 0 to 0.15 with a median of 0.02 and 90% less than 0.07. Importantly, for the stigma measure considered here these measures were 0.997 and 0.026, respectively.
To address the potential for informative nonresponse bias in the analyses of our abstinence outcome, we fit a series of selection model analyses. These analyses jointly modeled the conditional distributions of prolonged abstinence and of observation of the prolonged abstinence outcome. The conditional distribution of the abstinence outcomes was modeled using a logistic regression model assuming the log odds of abstinence depends on age, gender, health care program, study intervention received, stigma, and an interaction between intervention and stigma as in the analysis described above; however, here this model was fit to all participants as discussed below. In this sensitivity analysis, the conditional distribution for whether we observed a participant’s abstinence outcome was modeled using each of the following four different logistic regression models:
*Nonignorable Missing Response Model 1. log odds of observing outcome depends on age, gender, insurance program, the value of the abstinence outcome, intervention received, and interactions between age and abstinence, program and abstinence, and intervention and abstinence,
Nonignorable Missing Response Model 2. added stigma and an interaction between the value of the abstinence outcome and stigma to Model 1,
Nonignorable Missing Response Model 3. added stigma and an interaction between intervention and stigma to Model 1,
Nonignorable Missing Response Model 4. added stigma, an interaction between stigma and intervention, and an interaction between stigma and abstinence.
The first of the models was chosen based on the analyses outlined in the study by Fu et al.36 where this model was one of the better performing selection models for analyzing the study outcomes. These models were fitted by including two observations for those with missing outcomes, one where the outcome is abstinence and one where the outcome is continued use; these observations were initially equally weighted and the weights then updated iteratively to find the maximum expected likelihood estimates for the joint model parameters using the expectation-maximization algorithm of Ibrahim and Lipsitz.37 The selection models were fitted to each of the imputed data sets, and the results were aggregated using standard methods for multiple imputation as described by Rubin.38
Results
Study Setting and Participants
Baseline surveys were mailed to 21,181 prospective participants aged 18–64 who were MHCP clients. The baseline response rate was 44%, with 9362 surveys returned. Eligible participants (1) had a valid home address, (2) were proficient in English, (3) and reported being a current cigarette smoker (having smoked a cigarette in the past 30 days, even a puff). Among the respondents, 6826 individuals did not meet study inclusion criteria, 130 declined to participate, and the remaining 2406 were enrolled and randomized.
Overall, rates of employment within our socioeconomically disadvantaged sample were low (51%), and the majority (68%) of participants’ annual household incomes were $20,000 or less.
Baseline Comparison of Lower and Higher Perceived Smoking-Related Stigma Groups
Participants in the lower perceived smoking-related stigma group were more likely to be male, less educated, and to have lower incomes than those in the higher stigma group. Participants in the lower stigma group also reported lower levels of anxiety and depression (all p < .001) (Table 1).
Table 1.
Baseline Demographic, Mental Health, Health Care Provider, and Social Environment Characteristics of Lower Versus Higher Stigma Smokers
| Characteristic | Lower stigma N = 1227 |
Higher stigma N = 1093 |
p value |
|---|---|---|---|
| No. (%) or M ± SD (median) | |||
| Stigma | |||
| Smoking-related stigma | 1.5 ± 0.3 (1.4) | 2.8 ± 0.6 (2.7) | <.001 |
| Demographics | |||
| Insurance type | |||
| MinnesotaCare | 332 (27.1) | 303 (27.7) | .720 |
| Medicaid | 895 (72.9) | 790 (72.3) | |
| Gender | |||
| Female | 841 (68.5) | 808 (73.9) | .004 |
| Male | 386 (31.5) | 285 (26.1) | . |
| Age | |||
| 18–24 | 271 (22.1) | 215 (19.7) | .342 |
| 25–34 | 414 (33.7) | 387 (35.4) | . |
| 35–64 | 542 (44.2) | 491 (44.9) | . |
| Race/ethnicity | |||
| White | 913 (74.4) | 908 (83.1) | <.001 |
| Black | 155 (12.6) | 86 (7.9) | . |
| American Indian | 104 (8.5) | 60 (5.5) | . |
| Hispanic | 26 (2.1) | 14 (1.3) | . |
| Asian | 29 (2.4) | 25 (2.3) | . |
| Education | |||
| Grade 11/lower | 190 (15.8) | 117 (10.9) | <.001 |
| HS grad/GED | 455 (37.8) | 301 (27.9) | . |
| Some college | 445 (36.9) | 514 (47.7) | . |
| College grad/higher | 115 (9.5) | 146 (13.5) | . |
| Employment | |||
| Employed/self-employed | 619 (51.1) | 556 (51.5) | .239 |
| Student | 72 (6.0) | 88 (8.2) | . |
| Out of work | 165 (13.6) | 130 (12.1) | . |
| Unable to work/disabled | 284 (23.5) | 248 (23.0) | . |
| Homemaker | 71 (5.9) | 57 (5.3) | . |
| Yearly income | |||
| Less than $10k | 473 (40.7) | 348 (32.8) | <.001 |
| $10,001–$20k | 348 (29.9) | 354 (33.4) | . |
| $20,001–$40k | 224 (19.3) | 256 (24.1) | . |
| More than $40k | 118 (10.2) | 103 (9.7) | . |
| Mental health | |||
| Anxiety | 52.8 ± 10.7 (52.6) | 57.8 ± 10.2 (58) | <.001 |
| Depression | 1.4 ± 1.6 (1) | 2.0 ± 1.9 (2) | <.001 * |
| Health care provider | |||
| Access | |||
| Regular physician | 915 (75.3) | 816 (75.8) | .799 |
| Problems w/ access | 7.0 ± 2.3 (6) | 7.8 ± 2.4 (8) | <.001 |
| Advice | |||
| Physician advised to quit | 796 (74.3) | 763 (79.1) | .011 |
| Physician discussed meds | 479 (44.8) | 468 (48.7) | .083 |
| Physician discussed other | 443 (41.6) | 418 (43.4) | .412 |
| Receipt of cessation care | 838 (70.1) | 801 (75.8) | .003 |
| Cultural Competence | |||
| Physician bias | 1.9 ± 0.9 (1.7) | 2.1 ± 1.0 (2) | <.001 * |
| Social environment | |||
| Social support | |||
| Support of important others for quitting | 4.4 ± 0.8 (4.5) | 4.5 ± 0.7 (5) | <.001 |
| Social norms | |||
| Friends/family smokers | |||
| Almost all | 284 (23.3) | 176 (16.2) | <.001 |
| Over half | 260 (21.3) | 239 (22.0) | . |
| About half | 320 (26.2) | 291 (26.8) | . |
| Less than half | 185 (15.2) | 169 (15.6) | . |
| Very few | 153 (12.5) | 189 (17.4) | . |
| Almost all | 19 (1.6) | 21 (1.9) | . |
| Home Environment | |||
| Other smoker in home | 683 (56.0) | 541 (49.9) | .003 |
| Child in home | 681 (56.5) | 599 (55.4) | .594 |
| Home smoking rules | |||
| Smoking is not allowed | 588 (48.2) | 572 (52.62) | .002 |
| Smoking is allowed at times | 314 (25.7) | 298 (27.41) | . |
| Smoking is allowed | 318 (26.1) | 217 (20.0) | . |
*Satterthwaite test.
Those in the lower stigma group were less likely to report that their physician had advised them to quit smoking and were less likely to have received cessation-related care from their health care provider. Those in the lower stigma group reported lower levels of physician bias than the higher stigma group (Table 1).
Participants in the lower perceived stigma group reported lower levels of social support for quitting, had greater proportions of friends and family who were smokers, were more likely to live with another smoker, and had more lax home smoking rules than those in the higher perceived stigma group (Table 1).
The stigma groups did not differ greatly on measures of nicotine dependence. However, the lower stigma group had a lower proportion of participants who had made a quit attempt in the past year and were less likely to have used cessation treatment in the past year than the higher stigma group (Table 2).
Table 2.
Baseline Smoking History, Treatment Utilization, and Cessation Belief Characteristics of Lower Versus Higher Stigma Smokers
| Characteristic | Lower stigma N = 1227 |
Higher stigma N = 1093 |
p value |
|---|---|---|---|
| No. (%) or M ± SD (median) | |||
| Smoking history | |||
| Cigs/day | 13.4 ± 9.1 (10) | 13.9 ± 9.3 (12) | .168 |
| Duration | 20.1 ± 13.2 (17) | 20.3 ± 12.9 (17) | .780 |
| Time until first cig (minutes) | |||
| ≤5 | 314 (25.8) | 276 (25.6) | .773 |
| 6–15 | 342 (28.1) | 328 (30.4) | . |
| 16–30 | 185 (15.2) | 157 (14.6) | . |
| 31–60 | 150 (12.3) | 130 (12.1) | . |
| >60 | 227 (18.6) | 187 (17.4) | . |
| Quit attempt (past year) | 595 (49.1) | 655 (60.6) | <.001 |
| Treatment utilization | |||
| Medication (past year) | |||
| Counseling (past year) | |||
| Phone | 21 (1.9) | 33 (3.3) | .032 |
| Group | 7 (0.6) | 5 (0.5) | .723 |
| One-on-one | 17 (1.5) | 20 (2.0) | .369 |
| Any cessation treatment used (past year) | 285 (31.6) | 350 (43.1) | <.001 |
| Cessation beliefs | |||
| Quitting self-efficacy | 4.9 ± 2.8 (7) | 5.1 ± 2.8 (7) | .195 |
| Contemplation Ladder | 5.9 ± 2.8 (7) | 6.7 ± 2.8 (7) | <.001 |
Prolonged Smoking Abstinence by Stigma Group
We observed a significant interaction between intervention arm and perceived stigma group (p = .027).
In the lower perceived stigma group, participants in the proactive outreach condition had significantly higher odds of prolonged abstinence than participants in the usual care condition (16.4% vs. 9.4%, aOR = 1.94, 95% CI: 1.29 to 2.92, p = .001) (Table 3). In the higher perceived stigma group, treatment condition was not associated with prolonged abstinence (15.1% vs. 14.8%, aOR = 1.04, 95% CI: 0.70 to 1.55, p = .846) (Table 3).
Table 3.
Prolonged Smoking Abstinence by Stigma Group
| Stigma group | Outreach | Usual care | Odds ratio* (95% CI) |
|---|---|---|---|
| No. (%) | |||
| Lower stigma | 69/617 (16.4) | 44/610 (9.4) | 1.94 (1.29 to 2.92) |
| Higher stigma | 56/540 (15.1) | 65/553 (14.8) | 1.04 (0.70 to 1.55) |
*Adjusted for sex, age, and insurance program.
Selection Model Analyses
Four selection model analyses examined potential informative nonresponse bias in the analyses of our abstinence outcome. Although the estimates for the interaction between intervention condition and stigma group were no longer statistically significant, for each of the selection model analyses, the estimated abstinence ORs and associated CIs for the higher and lower stigma groups as well as the estimated interaction effect were similar to those obtained in the original data analyses (Table 4).
Table 4.
Summary of Selection Model Analysis: Simple Effects Comparing Outreach to Usual Care in Each Stigma Group and Model Estimated Interaction Between Intervention and Stigma
| Logistic regression models | Estimate | Standard error | t-statistic | p value | Odds ratio | Odds ratio confidence interval | |
|---|---|---|---|---|---|---|---|
| Lower | Upper | ||||||
| Complete data only | |||||||
| Model 1 | |||||||
| Outreach versus usual care, low stigma | 0.63 | 0.18 | 3.57 | 1.87 | 1.33 | 2.64 | |
| Outreach versus usual care, high stigma | 0.17 | 0.17 | 1.00 | 1.18 | 0.85 | 1.64 | |
| Intervention by stigma interaction | −0.46 | 0.24 | −1.87 | .06 | 0.63 | 0.39 | 1.02 |
| Nonignorable missing response | |||||||
| Model 2 | |||||||
| Outreach versus usual care, low stigma | 0.63 | 0.18 | 3.57 | 1.87 | 1.33 | 2.64 | |
| Outreach versus usual care, high stigma | 0.17 | 0.17 | 1.00 | 1.18 | 0.85 | 1.64 | |
| Intervention by stigma interaction | −0.46 | 0.24 | −1.87 | .06 | 0.63 | 0.39 | 1.02 |
| Model 3 | |||||||
| Outreach versus usual care, low stigma | 0.63 | 0.19 | 3.35 | 1.88 | 1.30 | 2.72 | |
| Outreach versus usual care, high stigma | 0.21 | 0.17 | 1.25 | 1.24 | 0.89 | 1.72 | |
| Intervention by stigma interaction | −0.42 | 0.25 | −1.65 | .10 | 0.66 | 0.40 | 1.08 |
| Model 4 | |||||||
| Outreach versus usual care, low stigma | 0.62 | 0.17 | 3.63 | 1.86 | 1.33 | 2.61 | |
| Outreach versus usual care, high stigma | 0.18 | 0.17 | 1.08 | 1.19 | 0.87 | 1.65 | |
| Intervention by stigma interaction | −0.45 | 0.24 | −1.86 | .06 | 0.64 | 0.40 | 1.03 |
| Model 5 | |||||||
| Outreach versus usual care, low stigma | 0.63 | 0.19 | 3.37 | 1.87 | 1.30 | 2.69 | |
| Outreach versus usual care, high stigma | 0.19 | 0.17 | 1.15 | 1.21 | 0.87 | 1.69 | |
| Intervention by stigma interaction | −0.43 | 0.25 | −1.72 | .08 | 0.65 | 0.40 | 1.06 |
Discussion
The first aim of the present study was to explore differences between smokers who perceive lower versus higher levels of smoking-related stigma. Consistent with past research,2,11 lower stigma smokers had lower incomes, less education, and were more likely to be non-White than their higher stigma counterparts. With respect to our primary aim, a significant intervention by perceived stigma interaction was observed. In the lower stigma group, participants in the proactive outreach condition had significantly higher rates of prolonged abstinence than participants in the usual care condition. We did not observe a significant intervention effect in the higher stigma group.
In addition to replicating the socioeconomic correlates of smoking-related stigma, the present study adds to the existing literature by exploring associations between perceived smoking-related stigma and mental health, health care provider experiences, and social environment characteristics. Consistent with research on other forms of health-related stigma,13,14 lower smoking stigma was associated with better mental health within our sample. Specifically, lower stigma smokers reported lower levels of anxiety and depression than higher stigma smokers. These results suggest that the discrimination, prejudice, and moral condemnation associated with engaging in this stigmatized behavior may be deleterious to mental health, although the cross-sectional nature of our data makes it difficult to assert a causal relationship.
With respect to health care provider experiences, lower perceived stigma smokers were less likely to report that their physician had advised them to quit or that they had received any cessation-related care in the past year than higher stigma smokers. Furthermore, lower stigma smokers reported lower levels of physician bias than higher stigma smokers. These results may reflect the lower readiness to quit observed among lower stigma smokers, as well as social norms that are more supportive of smoking.
In terms of social environment, lower stigma smokers reported less social support for quitting, had larger proportions of family and friends who were smokers, were more likely to reside with another smoker, and had less restrictive home smoking rules. These findings suggest that lower stigma smokers live in environments that are more amenable to smoking and less supportive of quit attempts or abstinence. Within the context of the “smoking islands” phenomenon,9 the lower perceptions of stigma among socioeconomically disadvantaged smokers may be due to the geographic and cultural isolation of these populations. Even within our predominantly low-income sample, lower stigma smokers belonged to social groups that contained more smokers, where smoking is a more socially acceptable behavior, and quitting is not supported.
Surprisingly, lower and higher perceived stigma smokers exhibited few differences with respect to nicotine dependence. This may relate to the similar age distribution of these groups, a factor which is strongly associated with nicotine dependence.39 However, lower stigma smokers were much less likely to have engaged in a quit attempt or to have used any cessation treatment in the past year than higher stigma smokers. Taken in conjunction with the lower readiness to quit observed in this group, it appears that lower stigma smokers are less ready to quit and less likely to take steps toward achieving abstinence.
These baseline comparisons indicate that the lower perceived stigma smokers in our sample were less likely to have received cessation-related care from their health care providers and tended to occupy social environments that were more conducive to smoking. With respect to cessation-related beliefs and behaviors, lower stigma smokers exhibited less readiness to quit and were less likely to have made a quit attempt or to have used cessation treatment in the past year.
In light of these baseline differences, the primary aim of the present study was to explore the role of perceived smoking-related stigma as a potential moderator of the effectiveness of the proactive cessation intervention with respect to 6-month prolonged abstinence at 12-month follow-up. We observed a significant intervention condition by stigma group interaction. Within the lower stigma group, the proactive outreach group reported significantly higher rates of prolonged abstinence than the usual care group (16.4% and 9.4%). We did not observe a significant intervention effect within the higher perceived stigma group, with both the proactive outreach and usual care group reporting relatively high rates of prolonged abstinence (15.1% vs. 14.8%). The high abstinence rates across both proactive outreach and usual care suggest a heightened underlying propensity for engaging in successful quit attempts among higher stigma smokers, a finding which is consistent with past research.12
The significant treatment effect observed among the lower perceived stigma smokers is particularly promising when viewed within the context of cessation treatment availability for smokers enrolled in publically subsidized health care programs. Conventional wisdom suggests that cost-based concerns would be the primary barrier to cessation treatment among low-income smokers. However, smokers enrolled in MHCP have access to free cessation medication and counseling,40 suggesting that other factors account for the low rates of cessation treatment utilization. Our analyses indicate that lower stigma smokers are less likely to receive cessation-related care from their health care provider, have social environments that are more conducive to smoking, and are less ready to quit; factors which may engender low rates of treatment utilization and quit attempts. These findings suggest that lower perceived stigma smokers are hard-to-reach population for cessation efforts, and the treatment effect observed provides strong evidence for the utility of proactive outreach strategies as means of both motivating lower stigma smokers to engage in treatment-facilitated quit attempts and a means of achieving abstinence.
Limitations and Future Research
With respect to the baseline analyses, the dependent and independent variables were measured at the same time so temporal relationships among these variables are unclear. Indeed, although higher stigma was associated with poorer mental health at baseline, it could be case that poor mental health leads to negative perceptions of how one is viewed, which could manifest as heightened perceptions of smoking stigma. Additionally, as these data were obtained from smokers who consented to participate in our trial, findings may not extend to the general population of smokers on state-funded insurance plans. It must be stressed that these analyses were conducted among low-income smokers, and these findings may not extend to smoking populations with a broader socioeconomic gradient. Indeed, future research should explore these associations within a more socioeconomically diverse population to gain a more complete understanding of the relationships between socioeconomic status, perceptions of smoking-related stigma, and the effects of proactive cessation strategies.
Conclusions
Given the high prevalence of smoking among socioeconomically disadvantaged populations, the development of effective cessation treatment strategies for these smokers is a public health priority. Our study demonstrates that proactive outreach interventions are an efficient cessation strategy for lower perceived stigma smokers, a group with unique health care provider, social environment, and psychosocial-related barriers to cessation. As such, we recommend that proactive outreach strategies be more widely implemented among populations with the highest prevalence of smoking, particularly low-income and ethnic minority groups, for whom smoking is more socially acceptable and less stigmatized. Proactive outreach may be particularly effective for motivating smokers enrolled in subsidized health insurance programs, who already have access to free cessation treatments, to take advantage of these resources.
Supplementary Material
Supplementary data are available at Nicotine & Tobacco Research online.
Funding
This work was supported by the National Cancer Institute (1R01CA141527-01), National Institutes of Health.
Declaration of Interests
The authors declare that they have no competing interests.
Supplementary Material
Acknowledgments
This material is the result of work supported with resources and the use of facilities at the Minneapolis VA Center for Chronic Disease Outcomes Research. The views expressed are those of the authors and do not necessarily reflect the position or policy of the Department of Veterans Affairs, or the United States government.
References
- 1. Jamal A, Homa DM, O’Connor E, et al. Current cigarette smoking among adults—United States, 2005-2014. MMWR Morb Mortal Wkly Rep. 2015;64(44):1233–1240. [DOI] [PubMed] [Google Scholar]
- 2. Hammond D, Fong GT, Zanna MP, Thrasher JF, Borland R. Tobacco denormalization and industry beliefs among smokers from four countries. Am J Prev Med. 2006;31(3):225–232. [DOI] [PubMed] [Google Scholar]
- 3. Bayer R, Stuber J. Tobacco control, stigma, and public health: rethinking the relations. Am J Public Health. 2006;96(1):47–50. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Chapman S, Freeman B. Markers of the denormalisation of smoking and the tobacco industry. Tob Control. 2008;17(1):25–31. [DOI] [PubMed] [Google Scholar]
- 5. Lavack AM. De‐normalization of tobacco in Canada. Soc Mar Q. 1999;5(3):82–5. [Google Scholar]
- 6. Fichtenberg CM, Glantz SA. Effect of smoke-free workplaces on smoking behaviour: systematic review. BMJ. 2002;325(7357):188. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Alamar B, Glantz SA. Effect of increased social unacceptability of cigarette smoking on reduction in cigarette consumption. Am J Public Health. 2006;96(8):1359–1363. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Wakefield MA, Loken B, Hornik RC. Use of mass media campaigns to change health behaviour. Lancet. 2010;376(9748):1261–1271. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Thompson L, Pearce J, Barnett JR. Moralising geographies: stigma, smoking islands and responsible subjects. Area. 2007. Dec;39(4):508–17. [Google Scholar]
- 10. Bayer R. Stigma and the ethics of public health: not can we but should we. Soc Sci Med. 2008;67(3):463–472. [DOI] [PubMed] [Google Scholar]
- 11. Stuber J, Galea S, Link BG. Smoking and the emergence of a stigmatized social status. Soc Sci Med. 2008;67(3):420–430. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Kim SH, Shanahan J. Stigmatizing smokers: public sentiment toward cigarette smoking and its relationship to smoking behaviors. J Health Commun. 2003;8(4):343–367. [DOI] [PubMed] [Google Scholar]
- 13. Link BG, Struening EL, Rahav M, Phelan JC, Nuttbrock L. On stigma and its consequences: evidence from a longitudinal study of men with dual diagnoses of mental illness and substance abuse. J Health Soc Behav. 1997;38(2):177–190. [PubMed] [Google Scholar]
- 14. Stokes JP, Peterson JL. Homophobia, self-esteem, and risk for HIV among African American men who have sex with men. AIDS Educ Prev. 1998;10(3):278–292. [PubMed] [Google Scholar]
- 15. Stuber J, Galea S. Who conceals their smoking status from their health care provider?Nicotine Tob Res. 2009;11(3):303–307. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Siahpush M, McNeill A, Borland R, Fong GT. Socioeconomic variations in nicotine dependence, self-efficacy, and intention to quit across four countries: findings from the International Tobacco Control (ITC) Four Country Survey. Tob Control. 2006. Jun;15Suppl 3:iii71–iii5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Gwaltney CJ, Metrik J, Kahler CW, Shiffman S. Self-efficacy and smoking cessation: a meta-analysis. Psychol Addict Behav. 2009;23(1):56–66. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Karasek D, Ahern J, Galea S. Social norms, collective efficacy, and smoking cessation in urban neighborhoods. Am J Public Health. 2012;102(2):343–351. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Cummings SR, Stein MJ, Hansen B, Richard RJ, Gerbert B, Coates TJ. Smoking counseling and preventive medicine. A survey of internists in private practices and a health maintenance organization. Arch Intern Med. 1989;149(2):345–349. [DOI] [PubMed] [Google Scholar]
- 20. Browning KK, Ferketich AK, Salsberry PJ, Wewers ME. Socioeconomic disparity in provider-delivered assistance to quit smoking. Nicotine Tob Res. 2008;10(1):55–61. [DOI] [PubMed] [Google Scholar]
- 21. Roter DL, Stewart M, Putnam SM, Lipkin M, Jr, Stiles W, Inui TS. Communication patterns of primary care physicians. JAMA. 1997;277(4):350–356. [PubMed] [Google Scholar]
- 22. Haas JS, Linder JA, Park ER, et al. Proactive tobacco cessation outreach to smokers of low socioeconomic status: a randomized clinical trial. JAMA Intern Med. 2015;175(2):218–226. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Fu SS, van Ryn M, Sherman SE, et al. Proactive tobacco treatment and population-level cessation: a pragmatic randomized clinical trial. JAMA Intern Med. 2014;174(5):671–677. [DOI] [PubMed] [Google Scholar]
- 24. Fu SS, van Ryn M, Burgess DJ, et al. Proactive tobacco treatment for low income smokers: study protocol of a randomized controlled trial. BMC Public Health. 2014;14:337. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Hughes JR, Keely JP, Niaura RS, Ossip-Klein DJ, Richmond RL, Swan GE. Measures of abstinence in clinical trials: issues and recommendations. Nicotine Tob Res. 2003;5(1):13–25. [PubMed] [Google Scholar]
- 26. Link BG, Cullen FT, Frank J, Wozniak JF. The social rejection of former mental patients: understanding why labels matter. Am J Sociol. University of Chicago Press; 1987. May;92(6):1461–500. [Google Scholar]
- 27. Link BG, Phelan JC. Conceptualizing stigma. Annu Rev Sociol. 2001;27(1):363–85. [Google Scholar]
- 28. Group PC. Unpublished Manual for the Patient-Reported Outcomes Measurement System (PROMIS) Version 1.1. [Internet]. 2009. [cited Jan 1, 2016]. Available from: http://www.nihpromis.org [Google Scholar]
- 29. California tobacco surveys. [Internet]. 2010 [cited Feb 11, 2009]. Available from: http://libraries.ucsd.edu/locations/sshl/data-gov-info-gis/ssds/guides/tobacco-surveys.html [Google Scholar]
- 30. Behavioral risk factor surveillance system questionnaire. 2010; [Google Scholar]
- 31. Davis RM. Healthcare report cards and tobacco measures. Tob Control. 1997;6Suppl 1:S70–S77. [PubMed] [Google Scholar]
- 32. Johnson RL, Saha S, Arbelaez JJ, Beach MC, Cooper LA. Racial and ethnic differences in patient perceptions of bias and cultural competence in health care. J Gen Intern Med. 2004;19(2):101–110. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Baldwin AS, Rothman AJ, Hertel AW, et al. Specifying the determinants of the initiation and maintenance of behavior change: an examination of self-efficacy, satisfaction, and smoking cessation. Health Psychol. 2006;25(5):626–634. [DOI] [PubMed] [Google Scholar]
- 34. Biener L, Abrams DB. The Contemplation Ladder: validation of a measure of readiness to consider smoking cessation. Health Psychol. 1991;10(5):360–365. [DOI] [PubMed] [Google Scholar]
- 35. Raghunathan T, Lepkowski J, Hoewyk J Van. A multivariate technique for multiply imputing missing values using a sequence of regression models. Surv Methodol. 2001;27:85–95. [Google Scholar]
- 36. Fu SS, van Ryn M, Nelson D, et al. Proactive tobacco treatment offering free nicotine replacement therapy and telephone counselling for socioeconomically disadvantaged smokers: a randomised clinical trial. Thorax. 2016;71(5):446–453. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Ibrahim JG, Lipsitz SR. Parameter estimation from incomplete data in binomial regression when the missing data mechanism is nonignorable. Biometrics. 1996;52(3):1071–1078. [PubMed] [Google Scholar]
- 38. Rubin DB. Multiple imputation for nonresponse in surveys. Wiley-Interscience; 2004. 287 p. [Google Scholar]
- 39. Park S, Lee JY, Song TM, Cho SI. Age-associated changes in nicotine dependence. Public Health. 2012;126(6):482–489. [DOI] [PubMed] [Google Scholar]
- 40. Minnesota Department of Human Services. MHCP Provider Manual—Pharmacy Services—Medicare [Internet]. 2016 [cited 2017 Jan 20]. Available from: www.dhs.state.mn.us/dhs16_140411 [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.