Abstract
Aims:
To estimate recent trends in cigarette use and health insurance coverage for United States adults with and without mental health and substance use disorders (MH/SUD).
Design:
Event study analysis of smoking and insurance coverage trends among US adults with and without MH/SUD using 2008–19 public use data from the National Survey on Drug Use and Health, an annual, cross-sectional survey.
Setting:
USA.
Participants:
A nationally representative sample of non-institutionalized respondents aged 18–64 years (n = 448 762).
Measurements:
Outcome variables were three measures of recent cigarette use and one measure of past-year health insurance coverage. We compared outcomes between people with and without MH/SUD (MH disorder: past-year mental illness, predicted from Kessler-6 and the World Health Organization–Disability Assessment Schedule impairment scale; SUD: met survey-based DSM-IV criteria for past-year alcohol, cannabis, cocaine or heroin use disorder) and over time.
Findings:
Comparing pooled data from 2008 to 2009 and from 2018 to 2019, current smoking rates of adults with MH/SUD decreased from 37.9 to 27.9% while current smoking rates of adults without MH/SUD decreased from 21.4 to 16.3%, a significant difference in decrease of 4.9 percentage points (pts) [95% confidence interval (CI) = 3.3–6.6 pts]. Daily smoking followed similar patterns (difference in decrease of 3.9 pts (95% CI = 2.3–5.4 pts). Recent smoking abstinence rates for adults with MH/SUD increased from 7.4 to 10.9%, while recent smoking abstinence rates for adults without MH/SUD increased from 9.6 to 12.0%, a difference in increase of 1.0 pts (95% CI = −3.0 to 0.9 pts). In 2018–19, 11% of net reductions in current smoking, 12% of net reductions in daily smoking and 12% of net increases in recent smoking abstinence coincided with greater gains in insurance coverage for adults with MH/SUD compared to those without MH/SUD.
Conclusions:
Improvements in smoking and abstinence outcomes for US adults with mental health and substance use disorders appear to be associated with increases in health insurance coverage.
Keywords: Health insurance, health reform, mental health, substance use disorder, survey data, tobacco
INTRODUCTION
The prevalence of smoking in the United States has fallen dramatically over the past 50 years, from 42 to 14% of the adult population [1]. Meanwhile, prevalence among individuals with mental health and substance use disorders (MH/SUD) has shown more limited reduction [2–4]. Risk ratios for smoking are 2.1 for any current psychiatric diagnosis; between 2 and 3 for many common individual diagnoses including depression, anxiety and alcohol use disorder; and greater than 3 for dual disorders and drug use disorders [5–7]. People with MH/SUD who smoke tend to smoke more heavily and report higher levels of nicotine dependence and withdrawal than those without MH/SUD [8–10]. They also bear a disproportionate share of the disease burden from tobacco use [9, 11].
Compared with people in the general population who smoke, people with MH/SUD who smoke report similar or greater desire to quit [9, 12]. They are also equally or more likely to attempt quitting [10, 13], more likely to have tried approved pharmacotherapy and more likely to be currently using medications for treatment than those without MH/SUD [14, 15]. Despite strong desire and efforts to quit, those with current or life-time MH/SUD who smoke have less success quitting than members of the general population [5, 8, 16]. In part, this may be the result of a lack of sustained support from health professionals and other service providers [17], as well as reduced availability and utilization of health services critical to providing information regarding the health harms of tobacco use and access to evidence-based treatments [18, 19]. For example, there may be reluctance among some providers to prescribe varenicline to people with MH/SUD due to concerns about psychiatric side effects [20], despite evidence of its effectiveness and safety among people with serious mental illness and with alcohol use disorder [21, 22].
In the United States, health insurance coverage can be provided from either public programs or private insurance plans, with most private insurance employer-based, while approximately 10% of the population remains uninsured [23, 24]. Those with MH/SUD are more likely to have reduced income or employment, limiting their access to insurance coverage in the United States and reducing affordability of health services, while also amplifying social and financial stressors that support ongoing tobacco use [9, 25, 26]. In addition, widely held beliefs that smoking may help individuals to manage MH/SUD symptoms and that smoking is a less immediate or lower-order treatment priority have historically been a barrier to promoting smoking cessation services in MH/SUD treatment settings, despite evidence that smoking cessation is associated with improved mental health [9, 25].
The Affordable Care Act (ACA) has led to major changes in the US health insurance market that may impact upon tobacco use among those with MH/SUD. ACA provisions implemented in 2014 include expanding eligibility for Medicaid, a government-funded program that helps with health-care costs for people with limited income and resources; launching market-places for non-group insurance, which assist individuals who are not covered by employers to find affordable coverage; requiring more payers to provide equal coverage of MH/SUD and medical conditions; and mandating benefits for MH/SUD and tobacco dependence (TDT) treatments. Additional ACA guidance, which was also issued in 2014, mandated comprehensive coverage for TDT for most private health plans and newly eligible Medicaid beneficiaries in states that expanded Medicaid coverage [27]. Together, these changes both expand the potential pool of individuals able to afford insurance coverage and improve evidence-based treatment options for those who have insurance. These changes are especially relevant as smoking in the United States has become increasingly concentrated among MH/SUD and other underserved populations (low socio-economic status, rural) who have historically had more limited access to care [28]. Access to TDT has been increasing in the United States since implementation of the ACA [29], although there is considerable variability between states and state-run programs in recent accessibility levels and how they have changed since ACA reforms took effect [30, 31].
Studies of early ACA implementation in a limited number of states provided preliminary evidence that insurance expansions were linked to increased cessation incidence [32] and reduced smoking prevalence among Medicaid-insured adults [29]. Individuals with psychological distress were more likely post-ACA to receive advice to quit smoking from their physicians [33]. In California, ACA Medicaid expansion led to a 50% drop in the proportion of smokers without insurance coverage from 27 to 13% [34]. These findings are consistent with the potential role of health insurance expansion in improving tobacco-related health outcomes among underserved populations, including lower income adults and those with MH/SUD. Although nationally representative trends in smoking for adults with MH/SUD have declined since the passage of the ACA [4], little has been reported about how insurance coverage changes have been associated with changes in smoking rates among adults with MH/SUD.
In this paper, we investigated whether differences by MH/SUD status (defined here as DSM-IV mental health, alcohol, cannabis, cocaine or heroin use disorders) in current and daily smoking prevalence and smoking abstinence have decreased in association with ACA insurance reforms, the majority of which first took effect in 2014. We also tested whether related trends in smoking outcomes were associated with gains in insurance coverage. We hypothesized that insurance expansion would have a larger effect on insurance coverage among those with MH/SUD compared to those without MH/SUD; and that increased insurance coverage would be associated with improved smoking outcomes among those with MH/SUD.
METHODS
Data and sample
We conducted a retrospective, observational study of pooled data from 12 annual cross-sections of the National Survey on Drug Use and Health (NSDUH), 2008–19. The NSDUH is a nationally representative, annual, computer- and self-administered survey overseen by the Substance Abuse and Mental Health Services Administration (SAMHSA). It is designed to collect data on substance use patterns and mental health status of the civilian, non-institutionalized US population aged 12 years and older [35].
We used 2008–19 data to assess smoking and insurance coverage trends among those with and without MH/SUD, examining outcomes throughout the first decade in which ACA reforms took effect (e.g., 2010 for expanded coverage of young adult family members under age 26, 2014 for state-based private coverage market-places and, in many states, income-based Medicaid expansion) in comparison to 2008–09, the 2-year period immediately preceding the ACA. To support a broad population health analysis, the study sample included all adults aged 18–64 years—an intentionally heterogeneous group with an array of subgroups with varying degrees of likelihood to be exposed to different combinations of most ACA reforms that took effect during the study period. We excluded only adults who were aged 65 years or older and most likely to be covered by Medicare (public health insurance covering elderly people) and therefore less subject to most ACA provisions. The NSDUH data collection protocol was approved by institutional review board at RTI International for all survey years included in this study. We did not pre-register the analytical plan for this study; the results should be considered exploratory.
Variables
Outcomes
There were two outcome measure categories: (1) health insurance coverage and (2) smoking status. For health insurance, we used a binary indicator of whether respondents had any insurance coverage for at least 10 of the prior 12 months when surveyed. For cigarette smoking status we used three binary measures: any current smoking, current smoking intensity and recent smoking abstinence. We defined current smoking as (a) smoked at least 100 life-time cigarettes and (b) smoked all or part of at least one cigarette in the 30 days prior to completing the survey. For intensity, we assessed daily smoking as (a) meeting current smoking criteria (above) and (b) smoked all or part of at least one cigarette during 25 or more of the 30 days prior to completing the survey. We defined recent abstinence as having smoked zero cigarettes in the last 30 days among the subsample of respondents who had (a) smoked at least 100 life-time cigarettes and (b) smoked all or part of at least one cigarette at any point in the 12 months prior to being surveyed.
Explanatory variables
To examine the degree to which these outcomes changed for people with MH/SUD between 2008 and 2019, we created an ordinal measure of time using a series of six 2-year intervals (2008–09, 2010–11, 2012–13, 2014–15, 2016–17 and 2018–19) and a binary measure of whether respondents met NSDUH-based criteria for past-year mental or substance use disorder or both (MH/SUD). The MH measure captures whether or not a respondent had mental illness in the past year, combining data from the Kessler-6 and the World Health Organization–Disability Assessment Schedule impairment scale in a prediction model that successfully predicted mental illness in follow-up diagnostic interviews [36]. The SUD measure was determined by a series of questions identifying illicit drug or alcohol abuse or dependence based on criteria specified in the Diagnostic and Statistical Manual of Mental Disorders, 4th edition (DSM-IV), and validated via clinical interview with NSDUH respondents [37]. For substance use disorder, we considered only alcohol, cannabis, cocaine and heroin [38]. In 2015, the NSDUH underwent changes to its questionnaire and definitions for all other substances; these changes produced a trend break making measures before and after 2015 incompatible [39]. However, the four substances we retained accounted for approximately 94% of all SUD indicated in the study pre-period and 90% in the post-period.
Covariates
In multivariable analyses, we adjusted for three self-reported demographic measures (age, sex and race/ethnicity). We did not include other commonly used socio-economic status measures such as marital status, education, employment, income and self-rated health status because these were time-varying, endogenous factors and may have been correlated outcomes or even influenced by the study outcome variables.
Statistical analyses
To test our hypotheses that, compared to those without MH/SUD, adults with MH/SUD would experience relative gains in insurance coverage and recent smoking abstinence and relative reductions in overall prevalence and frequency of current smoking after 2014 ACA implementation, we conducted two types of analyses, interrupted time–series analyses (ITSA) and adjusted logistic regression models.
First, we conducted ITSA, a quasi-experimental research design that accounts for differences in baseline levels, auto-correlation and trends in outcomes [40], to measure changes after 2014 ACA implementation in the level and trends (slopes) of outcomes in the MSUD group compared to the non-MSUD group.
Secondly, in adjusted models, we tested the significance of changes over 2-year intervals (2-year intervals were used to improve precision of comparisons) after adjustment for covariates, estimating logit models by MH/SUD group and time (2008–09 referent; 2010–11; 2012–13; 2014–15; 2016–17; 2018–19). We converted parameter estimates into predicted probabilities to aid with computation and interpretation of interaction terms [28, 41] and present differences by MH/SUD group and 2018–19 versus 2008–09. Exploratory analyses of trends by MH/SUD, time and sex and MH/SUD, time and race/ethnicity were conducted to assess if trends over time were disproportionately beneficial for historically marginalized gender and racial/ethnic groups.
In a follow-up set of analyses, we examined what proportion of differential smoking trends were explained by insurance. This proportion is calculated as the indirect association [comprised of (a) the association between msud*time and insurance and (b) the association between insurance and smoking] over the total association [comprised of the indirect association plus the direct adjusted association (c’) between msud*time and smoking]. This is similar to a mediation analysis comparing changes in a covariate of interest (msud*time) before and after adjustment for a mediator of interest (in this case, insurance) [42]. However, because the temporality of these relationships is unknown in non-experimental, cross-sectional NSDUH data, we use these mediation-like analyses only to suggest plausible mediation patterns for future studies that are able to establish causal links [43].
In all analyses, we accounted for the complex survey design to make results nationally representative, applying NSDUH sampling weights and accounting for the influence of stratified probability sampling on variance estimates [37]. The median weighted interview response rate for the 2008–19 study period was 71.4% (range = 64.9–75.6%). Missing data for select variables were imputed for approximately 2% of respondents by SAMHSA using a predictive mean neighborhoods imputation method modified specifically for the NSDUH [37]. We excluded the small number of remaining cases with non-imputed missing data of interest (< 1% of respondents), yielding a final unweighted sample size of 448 762 respondents (weighted n = 193 902 025, i.e. the average annual population of adults aged 18–64 years in the United States between 2008 and 2019). The unweighted sample size for the smoking abstinence subgroup analysis was 122 179 (weighted n = 49 216 863). We performed all statistical analyses in Stata/MP version 16.1 (StataCorp, College Station, TX, USA).
RESULTS
Sample characteristics
Table 1 describes sample characteristics showing differences between adults with and without MH/SUD for the full study period. Adults with MH/SUD tended to be younger, were more likely to be female, more likely to be non-Hispanic white and less likely to be Hispanic, non-Hispanic black or non-Hispanic Asian. They were also less likely to have had health insurance for at least 10 of the 12 months before survey completion (76.0 versus 80.4%).
TABLE 1.
Sample characteristics
| Total Weighted % (95% CI) | MH/SUD Weighted % (95% CI) | No MH/SUD Weighted % (95% CI) | |
|---|---|---|---|
| Unweighted, n | 448 762 | 130 071 | 318 691 |
| Weighted % (row) | 100.0 | 25.7 (25.4, 25.9) | 74.3 (74.1, 74.6) |
| Age, years | |||
| 18–21 | 9.0 (8.8, 9.1) | 10.8 (10.6, 11.1) | 8.3 (8.2, 8.4) |
| 22–25 | 8.7 (8.6, 8.8) | 11.1 (11.0, 11.3) | 7.9 (7.7, 8.0) |
| 26–29 | 8.9 (8.8, 9.0) | 10.8 (10.6, 11.0) | 8.3 (8.1, 8.4) |
| 30–34 | 10.6 (10.4, 10.7) | 12.2 (11.9, 12.5) | 10.0 (9.9, 10.2) |
| 35–49 | 31.8 (31.5, 32.0) | 30.6 (30.2, 31.0) | 32.1 (31.9, 32.4) |
| 50–64 | 31.1 (30.8, 31.4) | 24.4 (23.9, 24.9) | 33.4 (33.0, 33.8) |
| Sex | |||
| Male | 49.1 (48.9, 49.4) | 46.1 (45.6, 46.5) | 50.2 (49.9, 50.5) |
| Female | 50.9 (50.6, 51.1) | 53.9 (53.5, 54.4) | 49.8 (49.5, 50.1) |
| Race/ethnicity | |||
| White | 62.8 (62.4, 63.2) | 67.9 (67.5, 68.4) | 61.0 (60.6, 61.4) |
| Hispanic | 16.8 (16.5, 17.1) | 14.4 (14.0, 14.8) | 17.6 (17.3, 18.0) |
| Black | 12.4 (12.2, 12.6) | 10.8 (10.5, 11.2) | 12.9 (12.7, 13.2) |
| Asian | 5.6 (5.4, 5.7) | 3.7 (3.5, 3.9) | 6.2 (6.0, 6.4) |
| Other | 2.4 (2.4, 2.5) | 3.2 (3.0, 3.3) | 2.2 (2.1, 2.3) |
| Past-year MH/SUD | |||
| Mental disorder | – | 78.6 (78.3, 79.0) | – |
| SUD | – | 35.8 (35.4, 36.2) | – |
| Both | – | 14.4 (14.2, 14.7) | – |
| Health insurance | |||
| Yes | 79.3 (79.1, 79.5) | 76.0 (75.6, 76.3) | 80.4 (80.2, 80.7) |
| No | 20.7 (20.5, 20.9) | 24.0 (23.7, 24.4) | 19.6 (19.3, 19.8) |
| Currently smoking | |||
| Yes | 22.9 (22.7, 23.1) | 34.2 (33.8, 34.6) | 19.0 (18.8, 19.3) |
| No | 77.1 (76.9, 77.3) | 65.8 (65.4, 66.2) | 81.0 (80.7, 81.2) |
| Daily smoking | |||
| Yes | 16.3 (16.1, 16.4) | 24.2 (23.8, 24.6) | 13.5 (13.3, 13.7) |
| No | 83.7 (83.6, 83.9) | 75.8 (75.4, 76.2) | 86.5 (86.3, 86.7) |
| Recent smoking abstinencea | |||
| Yes | 9.7 (9.4, 9.9) | 8.9 (8.5, 9.3) | 10.1 (9.8, 10.5) |
| No | 90.3 (90.1, 90.6) | 91.1 (90.7, 91.5) | 89.9 (89.5, 90.2) |
Note: 2008–19 National Survey on Drug Use and Health (NSDUH) data (unweighted sample size: n = 448 762).
MH/SUD = mental health and substance use disorders; CI = confidence interval.
For recent smoking abstinence, analysis limited to subsample of adults who smoked at least 100 cigarettes in life-time and at any point in the past year (unweighted n = 122 179).
For smoking outcomes, adults with MH/SUD were far more likely to report current smoking (34.2 versus 19.0%) and daily smoking (24.2 versus 13.5%) and were less likely to report recent abstinence (8.9 versus 10.1%). Adults with MH/SUD category were composed of 64.2% mental illness only, 21.4% SUD only and 14.4% mental illness and SUD.
Trends in recent cigarette use and insurance coverage outcomes
Figure 1 displays 2008–19 trends in each outcome, comparing adults with and without MH/SUD. Interrupted time–series analyses identify that, in linear post-intervention trend analyses, the MH/SUD group experienced significantly greater declines in current smoking (P < 0.01) and daily smoking (P < 0.01) than the non-MH/SUD group. Significant increases in abstinence and rates of insurance were found both within the MH/SUD and the non-MH/SUD groups, but there were no significant between-group differences in increases on these outcomes. Full ITSA results are presented in Supporting information, Appendix S1.
FIGURE 1.
Trends in smoking-related outcomes and insurance coverage, 2008–19. Notes: Analysis of 2008–19 National Survey on Drug Use and Health (NSDUH) data (unweighted sample size: n = 448 762). aFor recent smoking abstinence, analysis limited to subsample of adults smoking at least 100 cigarettes in life-time and at any point in the past year (n = 122 179). Shaded areas indicated 95% confidence intervals computed from survey weighted logistic regressions. ACA, Affordable Care Act; MH/SUD, mental health and substance use disorders.
Table 2 presents adjusted estimates of within- and between-group differences comparing 2008–09 to 2018–19. In most cases, consistent patterns did not emerge until the 2014–19 post-ACA period (see Supporting information, Appendix S2 for full regression results for all years).
TABLE 2.
Adjusted estimates of smoking, abstinence and insurance outcomes, 2008–09 versus 2018–19
| Percentage Points (95% CI)’ | |||
|---|---|---|---|
|
| |||
| MH/SUD | No MH/SUD | Difference | |
| Current smoking | |||
| 2008–09 | 37.9% (36.8, 39.0) | 21.4% (20.7, 22.1) | 16.5 (15.2, 17.8) |
| 2018–19 | 27.9% (27.1, 28.7) | 16.3% (15.7, 16.8) | 11.6 (10.6, 12.5) |
| Difference | 10.1 (8.7, 11.4) | 5.2 (4.3, 6.0) | 4.9 (3.3, 6.6) |
| Daily smoking | |||
| 2008–09 | 27.3% (26.3, 28.2) | 15.7% (15.0, 16.3) | 11.6 (10.4, 12.8) |
| 2018–19 | 19.0% (18.1, 19.9) | 11.3% (10.8, 11.8) | 7.7 (6.8, 8.7) |
| Difference | 8.2 (6.9, 9.6) | 4.3 (3.5, 5.2) | 3.9 (2.3, 5.4) |
| Recent smoking abstinencea | |||
| 2008–09 | 7.4% (6.6, 8.3) | 9.6% (8.8, 10.4) | −2.2 (−3.4, −1.0) |
| 2018–19 | 10.9% (9.8, 11.9) | 12.0% (11.0, 13.0) | −1.2 (−2.7, 0.4) |
| Difference | −3.4 (−4.8, −2.1) | −2.4 (−3.7, −1.1) | −1.0 (−3.0, 0.9) |
| Any insurance | |||
| 2008–09 | 71.9% (70.7, 73.2) | 78.2% (77.5, 78.8) | −6.2 (−7.6, −4.8) |
| 2018–19 | 82.3% (81.6, 83.0) | 84.4% (83.8, 84.9) | −2.0 (−2.7, −1.3) |
| Difference | −10.4 (−11.8, −9.0) | −6.2 (−7.0, −5.4) | −4.2 (−5.7, −2.7) |
Note: 2008–19 National Survey on Drug Use and Health (NSDUH) data (unweighted sample size: n = 448 762). Linear probability models adjusted for age, sex, and race/ethnicity.
MH/SUD = mental health and substance use disorders; CI = confidence interval.
For recent smoking abstinence, analysis limited to subsample of adults smoking at least 100 cigarettes in life-time and at any point in the past year (n = 122 179).
Current smoking
In 2008–09, the adjusted prevalence of current smoking was 16.5 percentage points (pts) [95% confidence interval (CI) = 15.2–17.8] higher for adults with MH/SUD (37.9%) than for adults without MH/SUD (21.4%). In 2018–19, the adjusted prevalence of current smoking was 11.6 pts (95% CI = 10.6–12.5) higher for adults with MH/SUD (27.9%) than for adults without MH/SUD (16.3%). This represented a 4.9-pt (95% CI = 3.3–6.6) larger decrease for those with MH/SUD (10.1 pts, 95% CI = 8.7, 11.4) than those without MH/SUD (5.2 pts; 95% CI = 4.3, 6.0).
Daily smoking
In 2008–09, the adjusted prevalence of daily smoking was 11.6 pts (95% CI = 10.4–12.8) higher for adults with MH/SUD (27.3%) than for adults without MH/SUD (15.7%). In 2018–19, the adjusted prevalence of current smoking was 7.7 pts (95% CI = 6.8–8.7) higher for adults with MH/SUD (19.0%) than for adults without MH/SUD (11.3%). This corresponded to a 3.9-pt (95% CI = 2.3–5.4) larger decrease for those with MH/SUD (8.2 pts; 95% CI = 6.9–9.6) than those without MH/SUD (4.3 pts; 95% CI = 3.5–5.2).
Recent smoking abstinence
In 2008–09, the adjusted likelihood of recent smoking abstinence among those who had smoked in the past year was −2.2 pts (95% CI = −3.4 to 1.0) lower for adults with MH/SUD (7.4%) than for adults without MH/SUD (9.6%). By 2018–19 recent abstinence had increased for both groups (MH/SUD: 10.9%; no MH/SUD: 12.0%), and the difference between them was no longer statistically significant (−1.2 pts; 95% CI = −2.7 to –10.4). From 2008 to 2009 and from 2018 to 2019, the incidence of recent abstinence had risen significantly for both groups (MH/SUD: 3.4 pts; 95% CI = 2.1–4.8; no MH/SUD: 2.4 pts; 95% CI = 1.1–3.7). Pooling throughout 2014–19, the average annual increase in abstinence relative to 2008–09 was significantly larger for adults with MH/SUD than those without MH/SUD (1.7 pts; 95% CI = 0.2–3.2) (see Supporting information, Table S2).
Insurance coverage
In 2008–09, the adjusted prevalence of any insurance coverage was −6.2 pts (95% CI = −7.6 to –14.8) lower for adults with MH/SUD (71.9%) than for adults without MH/SUD (78.2%). In 2018–19, the MH/SUD–no MH/SUD difference had shrunk to −2.0 pts (95% CI = −2.7 to –1.3). This amounted to a 4.2-pt (95% CI = 2.7–5.7) larger increase in coverage for those with MH/SUD (10.4 pts; 95% CI = 9.0–11.8) than those without MH/SUD (6.2 pts; 95% CI = 5.4–7.0).
In exploratory triple difference analyses, we identified no significant differences in these differential smoking or insurance coverage trends by sex. We identified that black individuals had less favorable reductions in current smoking in 2014–15 (P = 0.044) and 2018–19 (P = 0.029), Hispanic individuals had less favorable reductions in current smoking in 2018–19 (P = 0.004), black and Hispanic individuals with MH/SUD had less favorable trends in daily smoking in 2014–15 (P = 0.05) and black (P = 0.011) and Hispanic individuals (P = 0.002) with MH/SUD had less favorable trends in daily smoking in 2018–19. However, these triple differences were not significant after Bonferroni correction for multiple comparisons (results available upon request).
The role of insurance status in explaining differential smoking rate trends by MH/SUD
Having any health insurance for at least 10 of the 12 months prior to being surveyed was strongly associated with a reduction in the likelihood of any current smoking (−14.2 pts; 95% CI = −14.7 to −13.7) and daily smoking (−12.3 pts; 95% CI = −12.8 to −11.8) and an increase in the likelihood of recent smoking abstinence (3.7 pts; 95% CI = 3.2–4.3).
Tables 3 and 4 present the results examining the proportion of differential smoking trends explained by insurance status. Given the strong associations of insurance coverage with smoking and abstinence outcomes, indirect associations tended to be detectable in follow-up periods when there were also detectable within- and between-group changes in rates of smoking and abstinence outcomes prior to accounting for insurance coverage (Table 2). For current and daily smoking, this included all three periods between 2014 and 2019. For smoking abstinence, this included all of 2016–17 (MH/SUD: 0.44 pts; 95% CI = 0.32–0.56; no MH/SUD: 0.28 pts; 95% CI = 0.20–0.35; difference: 0.16 pts; 95% CI = 0.04–0.28) and 2018–19 (MH/SUD: 0.42 pts; 95% CI = 0.31–0.52; no MH/SUD: 0.25 pts; 95% CI = 0.18–0.32; difference: 0.17 pts; 95% CI = 0.05–0.29).
TABLE 3.
Proportion of trends in current and daily smoking explained by insurance coverage gains, 2008–19
| Pts (95% CI) | ||||
|---|---|---|---|---|
|
|
||||
| Direct associations | Indirect associations | Total associations | Proportion mediated | |
| Current smoking | ||||
| 2010–11 | ||||
| MH/SUD | −2.54 (−4.25, −0.84) | 0.12 (−0.13, 0.36) | −2.43 (−4.16, −0.69) | −5% |
| No MH/SUD | −0.84 (−1.64, −0.03) | 0.24 (0.12, 0.36) | −0.60 (−1.42, 0.22) | −40% |
| Difference | −1.71 (−3.46, 0.05) | −0.12 (−0.40, 0.16) | −1.83 (−3.60, −0.06) | 7% |
| 2012–13 | ||||
| MH/SUD | −2.64 (−4.27, −1.01) | −0.05 (−0.30, 0.19) | −2.70 (−4.38, −1.01) | 2% |
| No MH/SUD | −1.49 (−2.35, −0.64) | 0.16 (0.03, 0.30) | −1.33 (−2.17, −0.49) | −12% |
| Difference | −1.15 (−2.81, 0.51) | −0.22 (−0.48, 0.05) | −1.37 (−3.06, 0.33) | 16% |
| 2014–15 | ||||
| MH/SUD | −4.48 (−5.90, −3.07) | −0.68 (−0.89, −0.47) | −5.16 (−6.60, −3.72) | 13% |
| No MH/SUD | −1.67 (−2.48, −0.87) | −0.36 (−0.47, −0.25) | −2.03 (−2.85, −1.20) | 18% |
| Difference | −2.81 (−4.46, −1.16) | −0.32 (−0.56, −0.08) | −3.13 (−4.81, −1.45) | 10% |
| 2016–17 | ||||
| MH/SUD | −6.76 (−8.28, −5.25) | −1.56 (−1.77, −1.36) | −8.33 (−9.88, −6.78) | 19% |
| No MH/SUD | −2.72 (−3.52, −1.91) | −0.96 (−1.07, −0.84) | −3.67 (−4.51, −2.84) | 26% |
| Difference | −4.05 (−5.72, −2.38) | −0.61 (−0.84, −0.38) | −4.66 (−6.37, −2.94) | 13% |
| 2018–19 | ||||
| MH/SUD | −8.89 (−10.24, −7.53) | −1.50 (−1.71, −1.30) | −10.39 (−11.78, −8.99) | 14% |
| No MH/SUD | −4.14 (−4.98, −3.29) | −0.89 (−1.00, −0.77) | −5.02 (−5.89, −4.16) | 18% |
| Difference | −4.75 (−6.34, −3.16) | −0.62 (−0.84, −0.40) | −5.36 (−7.01, −3.72) | 11% |
| Daily smoking | ||||
| 2010–11 | ||||
| MH/SUD | −1.68 (−3.11, −0.25) | 0.10 (−0.11, 0.31) | −1.58 (−3.04, −0.11) | −6% |
| No MH/SUD | −0.91 (−1.65, −0.18) | 0.21 (0.10, 0.31) | −0.71 (−1.44, 0.03) | −29% |
| Difference | −0.77 (−2.30, 0.76) | −0.11 (−0.35, 0.14) | −0.87 (−2.41, 0.67) | 12% |
| 2012–13 | ||||
| MH/SUD | −2.15 (−3.54, −0.76) | −0.05 (−0.26, 0.16) | −2.20 (−3.61, −0.79) | 2% |
| No MH/SUD | −1.48 (−2.33, −0.62) | 0.14 (0.02, 0.26) | −1.33 (−2.17, −0.50) | −11% |
| Difference | −0.68 (−2.33, 0.98) | −0.19 (−0.42, 0.04) | −0.86 (−2.50, 0.78) | 22% |
| 2014–15 | ||||
| MH/SUD | −4.01 (−5.21, −2.81) | −0.59 (−0.77, −0.40) | −4.60 (−5.82, −3.38) | 13% |
| No MH/SUD | −1.64 (−2.37, −0.92) | −0.31 (−0.41, −0.21) | −1.95 (−2.68, −1.22) | 16% |
| Difference | −2.37 (−3.79, −0.96) | −0.28 (−0.49, −0.07) | −2.65 (−4.08, −1.22) | 11% |
| 2016–17 | ||||
| MH/SUD | −5.68 (−7.07, −4.29) | −1.35 (−1.53, −1.18) | −7.03 (−8.45, −5.62) | 19% |
| No MH/SUD | −2.56 (−3.29, −1.83) | −0.83 (−0.93, −0.73) | −3.39 (−4.13, −2.64) | 24% |
| Difference | −3.12 (−4.69, −1.55) | −0.53 (−0.72, −0.33) | −3.65 (−5.25, −2.05) | 14% |
| 2018–19 | ||||
| MH/SUD | −7.28 (−8.62, −5.93) | −1.30 (−1.48, −1.12) | −8.57 (−9.94, −7.21) | 15% |
| No MH/SUD | −3.45 (−4.25, −2.65) | −0.77 (−0.87, −0.66) | −4.21 (−5.02, −3.40) | 18% |
| Difference | −3.83 (−5.35, −2.31) | −0.53 (−0.72, −0.34) | −4.36 (−5.91, −2.81) | 12% |
Note:2008–19 National Survey on Drug Use and Health (NSDUH) data (unweighted sample size: n = 448 762). Structural equation models adjusted for age, sex, and race/ethnicity.
MH/SUD = mental health and substance use disorders; CI = confidence interval.
TABLE 4.
Proportion of trends in recent smoking abstinence explained by insurance coverage gains, 2008–09
| Pts (95% CI) | ||||
|---|---|---|---|---|
|
|
||||
| Direct associations | Indirect associations | Total associations | Proportion mediated | |
| Recent smoking abstinence | ||||
| 2010–11 | ||||
| MH/SUD | −0.12 (−1.43, 1.18) | −0.02 (−0.12, 0.07) | −0.15 (−1.47, 1.17) | 16% |
| No MH/SUD | 0.18 (−0.99, 1.36) | −0.05 (−0.12, 0.03) | 0.14 (−1.05, 1.32) | −34% |
| Difference | −0.31 (−2.12, 1.50) | 0.02 (−0.10, 0.14) | −0.29 (−2.10, 1.53) | −8% |
| 2012–13 | ||||
| MH/SUD | 0.59 (−0.81, 1.99) | −0.01 (−0.11, 0.10) | 0.58 (−0.82, 1.98) | −1% |
| No MH/SUD | 0.41 (−0.88, 1.70) | −0.11 (−0.18, −0.03) | 0.30 (−0.97, 1.58) | −35% |
| Difference | 0.18 (−1.96, 2.31) | 0.10 (−0.04, 0.24) | 0.28 (−1.85, 2.40) | 36% |
| 2014–15 | ||||
| MH/SUD | 1.78 (0.53, 3.03) | 0.18 (0.09, 0.28) | 1.97 (0.72, 3.21) | 9% |
| No MH/SUD | 0.28 (−0.77, 1.34) | 0.09 (0.03, 0.15) | 0.37 (−0.68, 1.42) | 24% |
| Difference | 1.50 (−0.15, 3.15) | 0.10 (−0.02, 0.21) | 1.60 (−0.05, 3.25) | 6% |
| 2016–17 | ||||
| MH/SUD | 2.54 (1.16, 3.92) | 0.44 (0.32, 0.56) | 2.98 (1.58, 4.37) | 15% |
| No MH/SUD | 0.42 (−0.77, 1.61) | 0.28 (0.20, 0.35) | 0.70 (−0.49, 1.88) | 40% |
| Difference | 2.12 (0.32, 3.92) | 0.16 (0.04, 0.28) | 2.28 (0.48, 4.08) | 7% |
| 2018–19 | ||||
| MH/SUD | 3.09 (1.73, 4.44) | 0.42 (0.31, 0.52) | 3.50 (2.15, 4.86) | 12% |
| No MH/SUD | 2.11 (0.80, 3.42) | 0.25 (0.18, 0.32) | 2.36 (1.05, 3.66) | 10% |
| Difference | 0.98 (−0.97, 2.93) | 0.17 (0.05, 0.29) | 1.15 (−0.81, 3.11) | 15% |
Note: 2008–19 National Survey on Drug Use and Health (NSDUH) data (unweighted sample size: n = 122 179). Structural equation model adjusted for age, sex and race/ethnicity. Analysis limited to subsample of adults smoking at least 100 cigarettes in life-time and at any point in the past year.
MH/SUD = mental health and substance use disorders; CI = confidence interval.
The proportion of changes in rates of smoking behaviors explained by insurance coverage gains varied. For current smoking, the indirect path accounted for 10% of net MH/SUD–no MH/SUD reductions in current smoking in 2014–15, 13% in 2016–17 and 11% in 2018–19. Estimates of the indirect association for daily smoking were similar. For recent smoking abstinence, net insurance gains accounted for 7% of net increases in abstinence in 2016–17 and 15% in 2018–19.
DISCUSSION
During the first decade following passage of the ACA, which expanded both insurance eligibility and treatment coverage in the United States, adult NSDUH respondents with past-year MH/SUD (DSM-IV mental health, alcohol, cannabis, cocaine or heroin use disorders) experienced increases in prevalence of health insurance coverage and incidence of recent smoking abstinence. They also experienced decreases in prevalence of any current smoking and daily current smoking. Adults without MH/SUD demonstrated similar trends, but these changes tended to be significantly smaller than those found among adults with MH/SUD. Further, adults without MH/SUD did not experience significant changes in rates of abstinence. Statistically significant improvements in insurance coverage and recent smoking abstinence for adults with MH/SUD did not occur until after the 2014 ACA insurance expansions. For any current and daily smoking, however, within-group decreases also occurred in 2010–13 for those with MH/SUD. These findings, which hold after controlling for age, sex and race/ethnicity, are among the first to identify meaningful, population-level reductions in smoking and increases in abstinence among adults with MH/SUD, a group that has maintained significantly higher smoking rates in recent decades despite the implementation of numerous preventive public health measures and clinical interventions that have driven change in the general adult population [2].
Models exploring the percentage variance accounted for by insurance status indicate that absolute and relative gains in insurance coverage following ACA implementation for people with MH/SUD were strongly associated with observed reductions in smoking prevalence and intensity and increases in incidence of recent abstinence. Among the potential mechanisms that could account for the association between insurance coverage gains and these smoking outcomes are the reduced financial stress, alleviation of adverse social determinants and increased utilization of care (including greater interaction with health-care providers as well as availability and affordability of TDT) that are known to accompany insurance coverage [44]. The present study does not provide a means for comparing among these different mechanisms. Financial and social stressors may have been particularly relevant in the early years of the present study, which encompassed the Great Recession—when unemployment, foreclosures and other economic indicators were at concerning levels. The declines in both current and daily smoking among those both with and without MH/SUD observed prior to 2014 ACA insurance expansion may, in part, reflect gradual economic improvements during this period. Insurance expansion may also have increased access to and utilization of health-care within the MH/SUD population, including mental health treatment and in- and outpatient SUD treatment, which may in themselves support improved smoking outcomes. The ACA also increased mandates for TDT benefits [45], and increased availability of TDT may play an explanatory role in the increase in cessation observed after ACA implementation specifically among those with MH/SUD, given consistent evidence those with MH/SUD are more likely than those without MH/SUD to take advantage of medications and other available TDT both within and outside integrated health treatment settings [10, 15]. Finally, it is also likely that time-varying tobacco control policy measures, including cigarette price increases, smoke-free public places and mass media campaigns played a significant role in reducing smoking within the general population [46, 47]. We did not have access to NSDUH area-level variables for this study, and future research using data sets that incorporate variation in these tobacco control measures by area and time are needed.
Even after the reductions in current and daily smoking we identified in the present study, the prevalence of both current and daily smoking among adults with MH/SUD remained significantly higher than that for adults without MH/SUD. Continued efforts to address smoking within this population therefore remain critical. In contrast, however, we observed new evidence that, starting in 2014, adults with MH/SUD caught up to those without MH/SUD in terms of recent abstinence rates. Given the temporal association of these changes with 2014 ACA insurance expansions, further research is needed to assess any potential causal relationship between gaining coverage via Medicaid expansion and state insurance market-places and incidence of recent cessation among adults with MH/SUD. Research in other countries would also be important to confirm the potential role of insurance expansion in the context of different regulatory market-places. Further research is also needed to more clearly understand how specific insurance components (e.g. cost-sharing and benefits for specific types of TDT, including counseling and pharmacotherapy, as well as greater access to medical care more generally) may lead to reduced smoking. In the United States, Medicaid eligibility and TDT coverage policies have varied across states and time. For example, between 1991 and 2014, the number of prescriptions for nicotine replacement therapy (NRT) filled under Medicaid increased by more than 1900% [48]. Increasing enrollment in programs with more expansive coverage (e.g. Medicaid programs that paid for NRT) may thus be a mechanism underlying a portion of trends observed in this study. An updated understanding of the role of insurance coverage in increasing the use of evidence-based treatments among MH/SUD populations is needed. Other recent developments warranting further exploration include the increased use of e-cigarettes as a form of quitting aid or harm reduction [49].
While a strength of the NSDUH is its community-based assessment of mental health and substance use (as opposed to in clinical settings), the NSDUH is limited in its reliance upon self-reported and recall-based measures, which may be more vulnerable to bias than measures of observed behavior. A second limitation is that the recent smoking abstinence outcome variable available for this study focused upon behavior in the 30 days before the survey among respondents who smoked life-time and earlier in the last year. Unfortunately, richer measures identifying other recent periods of sustained smoking abstinence, which might be more closely indicative of cessation, were not available. Thirdly, the NSDUH does not survey individuals in criminal justice settings or long-term institutions such as nursing homes and psychiatric hospitals, settings where the prevalence of smoking and MH/SUD and need for cessation interventions may be disproportionately high. Fourthly, the response rate to the NSDUH declined during the study period; we conducted all analyses, however, with the recommended survey weighting procedures, which account in part for non-response. Last, because NSDUH is a point-in-time, cross-sectional survey and area-level variables were not available, we could not undertake a full causal examination of the effects of ACA insurance expansion on the outcomes we investigated.
CONCLUSIONS
During the decade following passage of expansions to both insurance eligibility and treatment coverage in the United States, adults with mental health and substance use disorders (MH/SUD) experienced significant increases in health insurance coverage. While they remained more likely than those without MH/SUD to smoke throughout this period, they also experienced significant reductions in multiple measures of smoking and significant increases in recent smoking abstinence. These changes in rates of smoking behaviors were significantly larger among those with MH/SUD than they were among adults without MH/SUD. A substantial proportion of the estimated improvements in each smoking and abstinence outcome for those with MH/SUD was explained by increases in health insurance coverage. It will be important to investigate other data to understand how diverse regulatory health reforms, changes in tobacco control policy, the growing use of e-cigarettes and other factors have contributed to these trends. Overall, however, this study presents new evidence that smoking among adults living with MH/SUD has begun to decline substantially and that health insurance coverage expansions may play a supporting role in these important reductions. Mechanisms for these changes may include increased access to and use of tobacco dependence treatment, improvements in behavioral health outcomes due to behavioral health treatment, or indirect financial or environmental improvements. Future policy reforms continuing to expand the availability and adoption of insurance coverage may hold promise for building on these gains.
Supplementary Material
ACKNOWLEDGMENTS
Dr. Creedon is an employee of the Office of the Assistant Secretary for Planning and Evaluation (ASPE). The views and opinions expressed in this article are those of the authors, and no official endorsement by ASPE or the US Department of Health and Human Services is intended or should be inferred.
Funding information
This project was supported by the National Cancer Institute (R01CA229355-03).
Footnotes
DECLARATION OF INTERESTS
None.
SUPPORTING INFORMATION
Additional supporting information can be found online in the Supporting Information section at the end of this article.
REFERENCES
- 1.Wang TW, Asman K, Gentzke AS, Cullen KA, Holder-Hayes E, Reyes-Guzman C, et al. Tobacco product use among adults—United States, 2017. Morb Mort Wkly Rep. 2018;67:1225–32. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Cook BL, Wayne GF, Kafali EN, Liu Z, Shu C, Flores M. Trends in smoking among adults with mental illness and association between mental health treatment and smoking cessation. JAMA. 2014;311: 172–82. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Lawrence D, Williams JM. Trends in smoking rates by level of psychological distress—time series analysis of US National Health Interview Survey Data 1997–2014. Nicotine Tob Res. 2016;18: 1463–70. [DOI] [PubMed] [Google Scholar]
- 4.Han B, Volkow ND, Blanco C, Tipperman D, Einstein EB, Compton WM. Trends in prevalence of cigarette smoking among US adults with major depression or substance use disorders, 2006–2019. JAMA. 2022;327:1566–76. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Smith PH, Chhipa M, Bystrik J, Roy J, Goodwin RD, McKee SA. Cigarette smoking among those with mental disorders in the US population: 2012–2013 update. Tob Control. 2020;29:29–35. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Conway KP, Green VR, Kasza KA, Silveira ML, Borek N, Kimmel HL, et al. Co-occurrence of tobacco product use, substance use, and mental health problems among adults: findings from wave 1 (2013–2014) of the population assessment of tobacco and health (PATH) study. Drug Alcohol Depend. 2017;(177):104–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Parker MA, Sigmon SC, Villanti AC. Higher smoking prevalence among United States adults with co-occurring affective and drug use diagnoses. Addict Behav. 2019;99:106112. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Forman-Hoffman VL, Hedden SL, Glasheen C, Davies C, Colpe LJ. The role of mental illness on cigarette dependence and successful quitting in a nationally representative, household based sample of U.S. adults. Ann Epidemiol. 2016;26:447–54. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Prochaska JJ, Das S, Young-Wolff KC. Smoking, mental illness, and public health. Annu Rev Public Health. 2017;(38):165–85. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Brose LS, Brown J, Robson D, McNeill A. Mental health, smoking, harm reduction and quit attempts—a population survey in England. BMC Public Health. 2020;20:1237. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Callaghan RC, Veldhuizen S, Jeysingh T, Orlan C, Graham C, Kakouris G, et al. Patterns of tobacco-related mortality among individuals diagnosed with schizophrenia, bipolar disorder, or depression. J Psychiatr Res. 2014;48:102–10. [DOI] [PubMed] [Google Scholar]
- 12.Siru R, Hulse GK, Tait RJ. Assessing motivation to quit smoking in people with mental illness: a review. Addiction. 2009; 104:719–33. [DOI] [PubMed] [Google Scholar]
- 13.Kalkhoran S, Thorndike AN, Rigotti NA, Fung V, Baggett TP. Cigarette smoking and quitting-related factors among US adult health center patients with serious mental illness. J Gen Intern Med. 2019;34:986–91. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Cummins SE, Zhu SH, Tedeschi GJ, Gamst AC, Myers MG. Use of e-cigarettes by individuals with mental health conditions. Tob Control. 2014;23:iii48–53. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Young-Wolff KC, Kline-Simon AH, Das S, Mordecai DJ, Miller-Rosales C, Weisner C. Smoking trends among adults with behavioral health conditions in integrated health care: a retrospective cohort study. Psychiatr Serv. 2016;67:996–1003. [DOI] [PubMed] [Google Scholar]
- 16.Streck JM, Weinberger AH, Pacek LR, Gbedemah M, Goodwin RD. Cigarette smoking quit rates among persons with serious psychological distress in the United States from 2008 to 2016: are mental health disparities in cigarette use increasing? Nicotine Tob Res. 2020;22:130–4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Twyman L, Bonevski B, Paul C, Bryant J. Perceived barriers to smoking cessation in selected vulnerable groups: a systematic review of the qualitative and quantitative literature. BMJ Open. 2014;4: e006414. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Tobacco Use and Dependence Guideline Panel. Treating Tobacco Use and Dependence: 2008 Update Rockville, MD: US Department of Health and Human Services; 2008. [cited 2021 Oct 22]. Available at: https://www.ncbi.nlm.nih.gov/books/NBK63952/ [Google Scholar]
- 19.Centers for Disease Control and Prevention. Best Practices for Comprehensive Tobacco Control Programs—2014 [internet] Atlanta, GA: Office on Smoking and Health, National Center for Chronic Disease Prevention and Health Promotion, Centers for Disease Control and Prevention, U.S. Department of Health and Human Services; 2014. [cited 2021 Oct 22]. Available at: https://www.cdc.gov/tobacco/stateandcommunity/best_practices/pdfs/2014/comprehensive.pdf [Google Scholar]
- 20.Taylor GMJ, Itani T, Thomas KH, Rai D, Jones T, Windmeijer F, et al. Prescribing prevalence, effectiveness, and mental health safety of smoking cessation medicines in patients with mental disorders. Nicotine Tob Res. 2020;22:48–57. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Smith RC, Amiaz R, Si TM, Maayan L, Jin H, Boules S, et al. Varenicline effects on smoking, cognition, and psychiatric symptoms in schizophrenia: a double-blind randomized trial. PLOS ONE. 2016; 11:e0143490. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Ray LA, Green R, Enders C, Leventhal AM, Grodin EN, Li G, et al. Efficacy of combining varenicline and naltrexone for smoking cessation and drinking reduction: a randomized clinical trial. Am J Psychiatry. 2021;178:818–28. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Yabroff KR, Valdez S, Jacobson M, Han X, Fendrick AM. The changing health insurance coverage landscape in the United States. Am Soc Clin Oncol Educ Book. 2020;e264–74. [DOI] [PubMed] [Google Scholar]
- 24.Kaiser Family Foundation. Health Insurance Coverage of the Total Population (CPS) [internet] Kaiser Family Foundation; 2022. [cited 2021 Oct 22]. Available at: https://www.kff.org/other/state-indicator/health-insurance-coverage-of-the-total-population-cps/?currentTimeframe=0&sortModel=%7B%22colId%22:%22Location%22,%22sort%22:%22asc%22%7D
- 25.Schroeder SA, Morris CD. Confronting a neglected epidemic: tobacco cessation for persons with mental illnesses and substance abuse problems. Annu Rev Public Health. 2010;31:297–314. [DOI] [PubMed] [Google Scholar]
- 26.Centers for Disease Control and Prevention (CDC). Vital signs: current cigarette smoking among adults aged ≥18 years with mental illness—United States, 2009–2011. Morb Mortal Wkly Rep. 2013;62: 81–7. [PMC free article] [PubMed] [Google Scholar]
- 27.McAfee T, Babb S, McNabb S, Fiore MC. Helping smokers quit: opportunities created by the affordable care act. N Engl J Med. 2015;372:5–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Creedon TB, Cook BL. Access to mental health care increased but not for substance use, while disparities remain. Health Aff (Millw). 2016;35:1017–21. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Bailey SR, Marino M, Ezekiel-Herrera D, Schmidt T, Angier H, Hoopes MJ, et al. Tobacco cessation in affordable care act Medicaid expansion states versus non-expansion states. Nicotine Tob Res. 2019;22:1016–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Ku L, Bruen BK, Steinmetz E, Bysshe T. Medicaid tobacco cessation: big gaps remain in efforts to get smokers to quit. Health Aff Proj Hope. 2016;35:62–70. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.DiGiulio A, Haddix M, Jump Z, Babb S, Schecter A, Williams KAS, et al. State Medicaid expansion tobacco cessation coverage and number of adult smokers enrolled in expansion coverage—United States, 2016. Morb Mortal Wkly Rep. 2016;65:1364–9. [DOI] [PubMed] [Google Scholar]
- 32.Koma JW, Donohue JM, Barry CL, Huskamp HA, Jarlenski M. Medicaid coverage expansions and cigarette smoking cessation among low-income adults. Med Care. 2017;55:1023–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Tan ASL, Young-Wolff KC, Carter-Harris L, Salloum RG, Banerjee SC. Disparities in the receipt of tobacco treatment counseling within the US context of the affordable care act and meaningful use implementation. Nicotine Tob Res. 2018;20:1474–80. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Zhu SH, Anderson CM, Wong S, Kohatsu ND. The growing proportion of smokers in Medicaid and implications for public policy. Am J Prev Med. 2018;55:S130–7. [DOI] [PubMed] [Google Scholar]
- 35.Center for Behavioral Health Statistics and Quality. 2018 National Survey on Drug Use and Health Public Use File Codebook [internet] Rockville, MD: Substance Abuse and Mental Health Services Administration; 2019. [cited 2021 Oct 22]. Available at: https://www.datafiles.samhsa.gov/study-dataset/national-survey-drug-use-and-health-2018-nsduh-2018-ds0001-nid18758 [Google Scholar]
- 36.Aldworth J, Colpe LJ, Gfroerer JC, Novak SP, Chromy JR, Barker PR, et al. The National Survey on drug use and health mental health surveillance study: calibration analysis. Int J Methods Psychiatr Res. 2010;19:61–87. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Center for Behavioral Health Statistics, and Quality. 2016 National Survey on drug use and health: methodological summary and definitions [internet] Rockville, MD: Substance Abuse and Mental Health Services Administration; 2017. [cited 2021 Oct 22]. Available at: https://www.samhsa.gov/data/sites/default/files/NSDUH-MethodSummDefs-2016/NSDUH-MethodSummDefs-2016.htm [Google Scholar]
- 38.Olfson M, Wall MM, Barry CL, Mauro C, Choi CJ, Mojtabai R. Effects of the ACA on health care coverage for adults with substance use disorders. Psychiatr Serv. 2021;72:905–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Substance Abuse and Mental Health Services Administration. Summary of the Effects of the 2015 NSDUH Questionnaire Redesign: Implications for Data Users [internet] Rockville, MD: Substance Abuse and Mental Health Services Administration; 2016; (CBHSQ Methodology Report). [cited 2021 Oct 22]. Available at: https://www.ncbi.nlm.nih.gov/books/NBK524967/ [PubMed] [Google Scholar]
- 40.Wagner AK, Soumerai SB, Zhang F, Ross-Degnan D. Segmented regression analysis of interrupted time series studies in medication use research. J Clin Pharm Ther. 2002;27:299–309. [DOI] [PubMed] [Google Scholar]
- 41.Ai C, Norton EC. Interaction terms in logit and probit models. Econ Lett. 2003;80:123–9. [Google Scholar]
- 42.Muller D, Judd CM, Yzerbyt VY. When moderation is mediated and mediation is moderated. J Pers Soc Psychol. 2005;89:852–63. [DOI] [PubMed] [Google Scholar]
- 43.Shrout PE, Bolger N. Mediation in experimental and nonexperimental studies: new procedures and recommendations. Psychol Methods. 2002;7:422–45. [PubMed] [Google Scholar]
- 44.Baicker K, Taubman SL, Allen HL, Bernstein M, Gruber JH, Newhouse JP, et al. The Oregon experiment—effects of Medicaid on clinical outcomes. N Engl J Med. 2013;368:1713–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45.Beronio K, Glied S, Frank R. How the affordable care act and mental health parity and addiction equity act greatly expand coverage of behavioral health care. J Behav Health Serv Res. 2014;41:410–28. [DOI] [PubMed] [Google Scholar]
- 46.Warner KE, Sexton DW, Gillespie BW, Levy DT, Chaloupka FJ. Impact of tobacco control on adult per capita cigarette consumption in the United States. Am J Public Health. 2014;104:83–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47.Gravely S, Giovino GA, Craig L, Commar A, D’Espaignet ET, Schotte K, et al. Implementation of key demand-reduction measures of the WHO framework convention on tobacco control and change in smoking prevalence in 126 countries: an association study. Lancet Public Health. 2017;2:e166–74. [DOI] [PubMed] [Google Scholar]
- 48.Yue X, Guo JJ, Wigle PR. Trends in utilization, spending, and prices of smoking-cessation medications in Medicaid programs: 25 years empirical data analysis, 1991–2015. Am Health Drug Benefits. 2018; 11:275–85. [PMC free article] [PubMed] [Google Scholar]
- 49.Weinberger AH, Zhu J, Barrington-Trimis JL, Wyka K, Goodwin RD. Cigarette use, E-cigarette use, and dual product use are higher among adults with serious psychological distress in the United States: 2014–2017. Nicotine Tob Res. 2020;22:1875–82. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.