Smoking and Behavioral Health of Women

Abstract

Background:

Using data from a study of reliability and validity of a screening tool for co-occurring substance abuse and mental health problems, our objective was to compare behavioral health issues of female smokers and nonsmokers and explore correlates of smoking.

Methods:

Using a convenience sample (n=1021), we recruited participants to complete an online survey conducted in substance abuse treatment, primary care, mental health services, senior, and public settings. The survey included demographic questions, smoking status, the co-occurring disorders screening tool, the Global Appraisal of Individual Needs-Short Screener (GAIN-SS) and the Postraumatic Stress Disorder Checklist (PCL)-Civilian.

Results:

One third of participants self-identified as smokers, and African American, American Indian, and bisexual women reported the highest rates of smoking. Seventy-two percent of women reported at least one mental health problem in the past year; 29% had a past year substance abuse problem, and 26% reported a past year co-occurring disorder of both. Smokers had significantly higher rates of posttraumatic stress disorder (PTSD), past year depression and anxiety, suicidality, past year substance abuse, and co-occurring disorders. Smokers also had significantly higher rates of lifetime intimate partner violence (IPV) and childhood abuse.

Conclusions:

Smoking in women was associated with significantly higher rates of mental health and substance abuse problems. Substance abuse, being in a treatment setting, IPV, African American and mixed ethnicity, Medicaid insurance status, reduced income, and no home ownership were identified as predictors of smoking. Screening and evaluation of smoking status, mental health, substance use disorders, and the presence and impact of violence are essential for women's health.

Introduction

Tobacco kills 443,000 people in the United States every year, and 40% of those deaths occur in women smokers.¹ Among American women, 18.3% smoke; 22.4% of American Indian/Alaska Native, 20.6% of white, 17.8% of African American, 10.7% of Hispanic, and 4.7% of Asian women smoke.² Lung cancer is now the leading cause of cancer-related deaths among American women, with almost 80% of all lung cancer deaths in women due to smoking.^3,4 The close relationship between tobacco addiction and co-occurring substance use disorders (SUD)^5,6 and mental illness (MI)^7

–12 has been well documented, yet health and human services systems, including women-specific programs, lag in their tobacco addiction screening and smoking cessation services.^13
–15 The purpose of this research was to examine data on female participants in a study that explored the reliability and validity of an instrument to screen for co-occurring SUD and MI in order to explore smoking and behavioral health.

Smoking-related health problems for women include coronary artery disease (CAD), chronic obstructive pulmonary disease (COPD), stroke, respiratory diseases and lung cancer, osteoporosis, miscarriage, stillbirth, low birth weight infants, and sudden infant death syndrome (SIDS).^1,16 Women with SUD experience more severe health problems because of the synergistic effects of alcohol, drugs, and tobacco^17,18 and are more likely to die from tobacco-related causes than from alcohol-related disease.¹⁹ Recent evidence also suggests that women may have a heightened susceptibility to COPD²⁰ and tobacco smoke and may be more vulnerable to development of lung cancer.²¹ Smoking rates are particularly high for persons with mental health problems¹⁴; combined with drug-dependent people, they consume 44% of all cigarettes sold in the United States.¹¹

SUD and MI are associated with an increased vulnerability to the reinforcing and addictive properties of nicotine,^11,22 with mood enhancement as a central reason for smoking.^22,23 Women smokers are more likely than men to report past or current depression,^9,24 and smoking is associated with an increased risk for developing depression. In a 26-year population-based follow-up study of 6999 women, those who smoked one-half to one pack of cigarettes per day had an adjusted risk for depression of 1.74 (confidence interval [CI] 1.33-2.27); for women smoking more than one pack per day (>22 100 mm filtered cigarettes), the risk was 2.17 (CI 1.45-3.26).¹² Research literature on posttraumatic stress disorder (PTSD) indicates a strong positive relationship between PTSD and nicotine dependence, with odds ratios (OR) ranging from 2.04 to 4.52.¹⁰ Evidence further suggests that PTSD rather than exposure to trauma alone is more likely to influence the risk for smoking. When culture, ethnicity, gendered roles, and sexual orientation are considered, distinct variations exist in smoking trajectories, experiences, and outcomes for women smokers.^7,25

–29

In an effort to develop a tool designed to screen for co-occurring substance use and mental health disorders in adults, the Co-Occurring Joint Action Council (COJAC) Workgroup of the California Department of Alcohol and Drug Programs and the California Department of Mental Health created the COJAC Screening Tool (CST). The purpose of the primary study was to explore the reliability and validity of the CST, and the results of that work are reported elsewhere.³⁰ Here, we report on findings from our analyses exploring smoking and behavioral health among the female participants in the COJAC study.

Materials and Methods

Design

Multiple treatment sites and public settings were used to collect one-time survey data from participants. Administration options included self-administration on paper or on a laptop via Survey Monkey^™,³¹ an interview by a research assistant (RA) who recorded their responses via Survey Monkey, or survey responses via an e-mail-delivered link to the survey on Survey Monkey. Participants who elected to have the RA ask them the survey questions did so because of their limited computer knowledge or visual ability, and in these cases, the participant and RA adjourned to a confidential space within the site for the interview.

Sample, settings, and recruitment

Inclusion criteria required that participants be ≥18 years of age, a California resident, at a fifth grade level in English reading ability, and willing and able to participate in the study. We collected data in waiting rooms of two public sector mental health centers, five substance abuse treatment programs, three primary care clinics, and numerous public settings (e.g., cafes, libraries, senior centers) in urban, suburban, and rural areas in northern California. Participants were also recruited via the internet using snowball recruitment procedures with a recruitment message sent by the investigators. A total of 2262 potential participants were approached at the sites to participate in the study, and 1951 agreed, resulting in a response rate of 86%. Of the 1021 female participants, most (70.4%) completed the survey via Survey Monkey, some with paper surveys (20.4%) and some (9.2%) via online invitations.

The study survey

The study survey questionnaire containing 59 items took 10–20 minutes to complete and consisted of (1) 18 demographic questions, (2) the 9-item CST, (3) a revision of 6 items of the CST to include a specific time frame, (4) the 20-item Global Appraisal of Individual Needs–Short Screener (GAIN-SS),³² and (5) the 6-item PTSD Checklist (PCL-Civilian).³³ Demographic questions included queries on age, gender, ethnicity, education, employment, sexual orientation, marital status, insurance, substance abuse and mental health treatment history, caregiver role, veteran status, and smoking status.

The CST was evaluated in the original study, and a detailed description of the CST is provided in the report of the study results.³⁰ For women in this study, the 3-item substance abuse scale had an alpha reliability of 0.83 and was correlated with the GAIN-SS substance abuse scale at 0.75. The CST mental health and trauma items did not perform well as scales with Cronbach's alphas of 0.64 and 0.56, respectively; therefore, we used the individual items as demographic information.

The GAIN-SS (www.chestnut.org/LI/gain/GAIN_SS/index.html) is a 3–5-minute 20-item screening tool for use in general populations for identification of persons with a substance abuse or mental health problem. The GAIN-SS has four 5-item subscales (internalizing disorders, externalizing disorders, substance disorders, and crime/violence) and is a periodic measure of change over time. Each item is scored 0–3 (0=never, 1=1+year ago, 2=2–12 months ago, 3=within past month), and the number of 0s, 1s, 2s, and 3s are then summed. Dennis et al.³² have indicated that for adults and adolescents, the 20-item screener and the four 5-item subscales had good internal consistency (alpha of 0.96 on total screener), had excellent sensitivity (90% or more) for identifying individuals with a disorder, were highly correlated (r=0.84–0.94) with the 123-item scales in the full GAIN-I, and had excellent specificity (92% or>) for accurately ruling out people who do not have a disorder. Research suggests that the subscreeners have good discriminant validity and that the total structure is consistent with the model used with the full GAIN.³²

The PCL-Civilian^33,34 is a 3-minute 6-item screener available for use in the public domain for identification of civilians who have experienced trauma. Each item is scored 1–5, indicating the severity of the complaint as experienced in the last month (1=not at all to 5=extremely). A score of 14 is the suggested tool cutoff point for referral for further evaluation. The PCL is reported to be the second most frequently used self-report tool among internationally based traumatic stress treatment clinicians.³⁵ In this sample of women, the reliability was 0.90.

Procedures

Study procedures were approved by the Committee for the Protection of Human Subjects, Department of Health and Human Services, State of California, and the Committee on Human Research at the University of California, San Francisco. Three study sites required approval of the study protocol by their site institutional review boards (IRB), and those were submitted and site IRB approval was granted before study initiation at those sites.

A flier inviting participation in the study was either distributed by the site staff or posted in the treatment service setting. Research staff approached participants as they awaited appointments in the treatment services waiting rooms and in the public settings. In some cases, the program setting staff supported recruitment during breaks in group treatment processes. When contact was made between study staff and a participant, the study was explained and consent procedures were conducted. There were no signed consent forms, as they would be the only link to the participants in this sensitive study.

Data collection was conducted from December 2008 through June 2009. The demographic data were monitored weekly by the investigators to ensure sample diversity (e.g., age, gender, race/ethnicity, sexual orientation). Consequently, in order to acquire a fully diverse and representative sample, two additional study sites were added approximately 2 months after data collection began.

Data management and analyses

The study research assistants entered data from all paper surveys into the secured Survey Monkey website. Data from all participant surveys were then downloaded from the site into Excel 2003, where the data were cleaned and then uploaded into SPSS (version 16). The data were then processed, checked, and rechecked. Descriptive statistics (means, standard deviations [SD], and frequency distributions) were generated. Statistical inferences were gathered by generating chi-square tests, t tests, analyses of variance (ANOVA), and logistic regression.

Results

Demographics

Data were collected from multiple sites representing five categories: public settings (30%), primary care (27%), substance abuse treatment (24%), senior centers (10%), and mental health clinics (9%). In our sample, 67% of the women were from community settings, which included public settings (e.g., coffee houses, library), primary healthcare settings, and senior centers. The rest of the sample was from treatment settings, that is, from substance abuse and mental health services settings. Female study participants (n=1021) ranged in age from 18 to 93 (mean 41.2, SD 17.6); 50% identified as Caucasian, 15% as mixed race/ethnicity, 14% as African American, 12% as Latina, 8% as Asian/Pacific Islander, and 2% as American Indian. Table 1 shows the distribution of racial/ethnic groups by recruitment site. Single participants comprised 63% of the sample, 24% were married, and 14% reported being unmarried and living with a partner. Eighty-three percent (n=832) identified as heterosexual, 9% (n=87) were bisexual, and 8% (n=85) were lesbians. Over half (53%) of the women had more education than a general equivalency diploma (GED) or high school diploma. About a quarter (28%) of women reported they had absolutely no health insurance, and half (50%) were unemployed. On a scale from 0 to 10 (0=not at all adequate), the average adequacy of income was 4.90 (SD 3.4); this translates into a third (34%) reporting that their income was inadequate to meet their needs (0–3 on the 0–10 scale). Thirty-seven percent of participants were caregivers of dependent adults or children, 27% were homeowners, and 23% of participants reported they lived alone. Eighty-five percent of participants were U.S.-born citizens, and 8 women (1%) had served in active military duty in a war zone.

Table 1.

Ethnicity by Community and Treatment Setting (n=969)

Ethnicity	Community, n (%)	Treatment, n (%)
African American	46 (35)	86 (65)
American Indian	8 (44)	10 (56)
Asian Pacific Islander	59 (79)	16 (21)
Latina	83 (73)	31 (27)
White	352 (73)	133 (27)
Mixed ethnicity	102 (70)	43 (30)

Survey item on ethnicity declined by 52 participants.

Smoking

Approximately a third (32%, n=323) of women participants identified themselves as smokers in response to the question: Do you currently smoke cigarettes? (yes/no option only). The smoking rate varied significantly by ethnicity, with African American women reporting the highest rate of smoking (58%, 95% CI 50-66), followed by American Indians (56%, 95% CI 33-76), mixed ethnicity (37%, 95% CI 30-45), Caucasians (28%, 95% CI 24-32), Asian Pacific Islanders (23%, 95% CI 15-33), and Latinas (21%, 95% CI 14-29; chi-square-60.5, p<0.001). Age was related to smoking rates, with those >65 smoking less (8% vs. 25% for those aged <26, 47% for those aged 26–45, and 29% for those aged 46–65; chi-square=72.6, p<0.001). Smoking rates also varied significantly by sexual orientation, with bisexual women smoking most (53%, 95% CI 42-63), followed by heterosexuals (30%, 95% CI 27-34), and by lesbians (26%, 95% CI 18-36; chi-square=19.79, p<0.001). Smoking also varied significantly by educational level, with high school graduates and GED or less schooling having a smoking rate of 40%, whereas those with more education had a smoking rate of 27% (chi-square=17.2, p<0.001). The lowest smoking rates (14%) were those with Master's degrees or higher. Rates of smoking by recruitment site are noted in Table 2.

Table 2.

Rates of Smoking in Women by Site (n=1,005)

Site	% Smokers
Substance abuse treatment	68.8
Mental health services	65.2
Primary care	21.3
Public settings	12.1
Senior recreation centers	6.9

Missing data for smoking status of 16 participants.

Those who were U.S.-born citizens were significantly more likely to smoke (34% vs. 19%, p<0.001), as were those who did not have a job (41% vs. 23%, p<0.001) and those who did not own their home (40% vs. 10%, p<0.001). There were no significant differences in smoking by caregiving status or if someone lived alone. There were significant differences by marital status, with those who were married having the lowest smoking rate of 24% compared with singles at 33% and those who were living with a partner at 39% (chi-square=11.56, p=0.003). Smoking rates varied significantly by insurance status (p<0.001); 35% of study participants with no insurance, 59% of Medicaid recipients, 28% of Medicare recipients, and 18% of private payers smoked.

Mental health

PTSD

There were significant differences by age group (F=13.5, p<0.001), with the >65 age group having lower scores for PTSD. PTSD scores also varied significantly by sexual orientation, with bisexual women having the highest score of 15.7 (F=10.9 p<0.001). There were no significant differences between lesbian and heterosexual women. There were significant differences by ethnicity, with American Indians having the highest score of 16.0 (F=4.91, p<0.001). On average, 38% of our sample had PCL scores ≥14, indicative of needing further evaluation. There were significant differences in PTSD scores by smoking status (t=7.63, p<0.001), with smokers having a mean score of 15.1 (SD 6.1) and nonsmokers having a mean score of 12.1 (SD 5.3) (Table 3).

Table 3.

Indicators of Behavioral Health Status in Smoking and Non-smoking Women in Previous Year

	Smokers (n = 322)	Non-Smokers (n = 675)	Statistic	Significance
Trapped, lonely, depressed etc^a	74.6%	56.0%	χ² = 38.7	p<.001
Anxiety^a	74.0%	53.5%	χ² = 38.3	p<.001
Suicidality^a	24.1%	12.4%	χ² = 22.2	p<.001
Any mental health issue^a	86.0%	66.1%	χ² = 43.4	p<.001
Substance abuse^b	59.1%	15.5%	χ² = 200.1	p<.001
Co-occurring mental health & substance abuse^c	55.0%	12.3%	χ² = 206.7	p<.001
IPV^d	65.1%	31.3%	χ² = 102.1	p<.001
Childhood physical abuse^e	39.8%	21.1%	χ² = 38.5	p<.001
Childhood sexual abuse^e	42.1%	27.6%	χ² = 21.0	p<.001
PCL ≥14^f	52.5%	31.0%	χ² = 42.9	p<.001

Per the Global Assessment of Individual Needs-Short Screener (GAIN-SS).

Calculated from the three-item substance abuse subscale of the COJAC Screening Tool (CST).

Calculated from the combination of the GAIN and the COJAC CST.

Intimate partner violence.

Calculated from the CST.

Score ≥14 indicating probable PTSD and need for further evaluation.

Depression

On average, 62% of our sample reported they had symptoms of depression during the prior year. There were significant differences by age group (chi-square=43.2, p<0.001), with the >65 age group having the lowest rate at 36%. Depression rates during the prior year also varied significantly by sexual orientation, with bisexual women having more depression at 72% (chi-square=5.92, p=0.05). There were no significant differences between lesbian and heterosexual women. There were significant differences by ethnicity, with American Indians having the highest rates of depression at 82% (chi-square=19.1, p=0.002). There were significant differences in depression during the past year by smoking status (chi-square=37.9, p<0.001), with 76% of smokers and 56% of nonsmokers reporting depression.

Anxiety

On average, 60% of our sample reported they had symptoms of anxiety during the prior year. There were significant differences by age group (chi-square=41.9, p<0.001), with the >65 age group having the lowest rate at 35% and the 26–45 age group having the highest at 68% (chi-square=7.6, p=0.022). Anxiety rates during the prior year also varied significantly by sexual orientation, with bisexual women having more anxiety at 78%, followed by heterosexual women at 60%, and lesbians at 45% (chi-square=20.3, p<0.001). There were significant differences by ethnicity, with American Indians (78%) and African Americans (77%) having the highest rates of anxiety (chi-square=27.4, p<0.001). There were significant differences in anxiety during the past year by smoking status (chi-square=37.2, p<0.001), with 74% of smokers and 54% of nonsmokers reporting anxiety.

Suicidality

On average, 16% of our sample reported having had significant thoughts about committing suicide over the prior year. There were significant differences by age group (chi-square=43.2, p<0.001), with the >65 age group having the lowest rate of suicidal thoughts at 7% and those between 46 and 65 having the highest rate at 20%. Thoughts about committing suicide varied significantly by sexual orientation, with bisexual women having the highest rates at 33% (chi-square=20.0, p<0.001). There were no significant differences between lesbian and heterosexual women. There were no significant differences in suicidal thoughts by ethnicity. There were significant differences in thoughts of suicide during the past year by smoking status (chi-square=21.3, p<0.001), with 24% of smokers and 12% of nonsmokers reporting suicidal thoughts.

In summary, 72% of our sample of women reported at least one mental health issue (depression, anxiety, suicidality). There were significant age differences, with oldest participants (>65 group) reporting significantly less problems at 47% (chi-square=8.50, p<0.001). There were significant differences by sexual orientation, with bisexuals reporting significantly more mental health issues at 85% (chi-square=13.2, p=0.002). There were no significant differences between lesbian and heterosexual women. There were significant differences by ethnicity, with American Indians (82%) and African Americans (86%) having the highest rates of mental health issues (chi-square=21.8, p=0.001). There were significant differences in mental health issues during the past year by smoking status (chi-square=42.4, p<0.001), with 86% of smokers and 66% of nonsmokers reporting at least one mental health issue.

Substance abuse

Twenty nine percent of the women (n=301) reported a substance abuse issue during the prior year. There were significant age differences, with the oldest participants (>65 group) reporting significantly less problems at 2%, and participants aged 26–45 reporting the most problems at 43% (chi-square=74.1, p<0.001). There were significant differences by sexual orientation, with bisexuals reporting significantly more substance abuse at 47% (chi-square=15.2, p=0.002). There were no significant differences between lesbian and heterosexual women. There were significant differences by ethnicity, with Asian/Pacific Islanders reporting the least problems (15%) and African Americans reporting the most problems at 54% (chi-square=53.5, p<0.001). There were significant differences in substance abuse during the past year by smoking status (chi-square=198.0, p<0.001), with 59% of smokers and 16% of nonsmokers reporting a problem with substance abuse.

Co-occurring disorders

Some women (26%, n=262) reported both a substance abuse and mental health issue during the prior year. There were significant age differences, with oldest participants (>65 group) reporting no problems and participants aged 26–45 reporting the most problems at 38% (chi-square=67.1, p<0.001). There were significant differences by sexual orientation, with bisexuals reporting significantly more co-occurring disorders at 44% (chi-square=17.9, p<0.001). There were no significant differences between lesbian and heterosexual women. There were significant differences by ethnicity, with African Americans reporting the most problems at 50% (chi-square=48.1, p<0.001). There were significant differences in co-occurring disorders during the past year by smoking status (chi-square=206.7, p<0.001), with 55% of smokers and 12% of nonsmokers reporting a problem with both substance abuse and mental health issues.

Violence

Among women study participants, 42% reported a history of lifetime intimate partner violence (IPV); 27% percent of women reported a history of childhood physical abuse before the age of 13, and 32% reported childhood sexual abuse before the age of 13 (Table 3 shows the demographics of violence items). There were significant differences in rates of IPV by smoking status (chi-square=102.1, p<0.001), with 65% of smokers and 31% of nonsmokers reporting IPV. There were significant differences in rates of childhood physical abuse by smoking status (chi-square=37.6, p<0.001) with 40% of smokers and 21% of nonsmokers reporting physical abuse as a child. There were significant differences in rates of childhood sexual abuse by smoking status (chi-square=37.6, p<0.001), with 42% of smokers and 28% of nonsmokers reporting sexual abuse as a child (Table 4).

Table 4.

Demographics for Violence Items

Item	IPV, % (n)	Childhood physical abuse, % (n)	Childhood sexual abuse, % (n)
Age group
≤25	28.9 (73)	22 (56)	25.3 (64)
26–45	52.9 (193)	34.2 (125)	37.4 (136)
46–65	48.5 (126)	29.3 (76)	37.5 (97)
>66	21.2 (22)	7.7 (8)	16.2 (17)
Sexual orientation
Bisexual	60.9 (53)	41.4 (36)	46.0 (40)
Heterosexual	40.0 (328)	24.5 (201)	28.9 (237)
Lesbian	44.7 (38)	36.5 (31)	48.2 (41)
Ethnicity
African American	59.5 (78)	35.4 (46)	43.1 (56)
Asian/Pacific Islander	32.0 (24)	17.3 (13)	20.0 (15)
American Indian	66.7 (12)	50.0 (9)	55.6 (10)
Latina	35.1 (39)	30.7 (35)	29.8 (34)
Mixed	46.2 (66)	33.8 (48)	32.6 (47)
White	39.4 (189)	21.9 (105)	30.3 (145)
U.S.-born citizen
Yes	44.8 (383)	27.6 (236)	32.9 (281)
No	26.0 (38)	24.7 (36)	25.9 (38)

Behavioral health treatments

For all the behavioral health treatment items, smokers have received significantly more counseling and medication than have nonsmokers. Table 5 contains the specific details.

Table 5.

Participant Treatment Experiences

Treatment type	Smokers % (n)	Nonsmokers % (n)	Chi-square	p value
Seen a counselor/therapist/psychologist
Never	14.2 (45)	31.3 (211)
In the past	39.7 (126)	48.7 (328)	79.7	<0.001
Currently	46.1 (146)	20.0 (135)
Seen a psychiatrist
Never	34.8 (108)	66.5 (442)
In the past	29.4 (91)	23.5 (156)	117.1	<0.001
Currently	35.8 (111)	10.1 (67)
Been in hospital for mental health issue
Never	58.7 (186)	87.8 (585)
In the past	36.3 (115)	11.0 (73)	108.3	<0.001
Currently	5.0 (16)	1.2 (8)
Taken medicine for depression for longer than a month
Never	43.4 (138)	69.9 (467)
In the past	22.0 (70)	16.6 (111)	75.5	<0.001
Currently	34.6 (110)	13.5 (90)
Taken medicine for mental health issue
Never	50.2 (158)	78.4 (522)
In the past	16.2 (51)	10.4 (69)	88.6	<0.001
Currently	33.7 (106)	11.3 (75)
Taken anxiety medicine for longer than a month
Never	55.1 (173)	80.5 (535)
In the past	19.4 (61)	10.7 (71)	72.5	<0.001
Currently	25.5 (80)	8.9 (59)
Taken sleeping pills for longer than a month
Never	60.3 (190)	83.9 (561)
In the past	21.0 (66)	8.8 (59)	65.8	<0.001
Currently	18.7 (59)	7.3 (49)

Predictors of smoking

In logistic regression analyses intended to find candidate variables for a final, combined model, all demographic variables were explored, and ethnicity, home ownership, income, insurance, and setting were significantly related to smoking when combined in the same model. We then explored the associations of smoking status with mental health, substance abuse, and violence variables separately in sets to identify candidate variables for a multivariable model containing predictors across sets. Of these, only substance abuse and the presence of intimate partner violence were significant. Table 6 describes the ORs and 95% CIs for the final model. The McFadden pseudo-R-squared^36,37 for the model was 0.32, indicating a 32% improvement in fit for the final model compared to a null model. The point biserial correlation of smoking with the predicted probability of smoking³⁸ was 0.62 (p<0.00005), indicating that 38% of the variance in smoking status was predicted by the model.

Table 6.

Logistic Regression of Smoking Status on Demographic Variables, Setting, Substance Abuse Status, and Intimate Partner Violence Status

							95% CI for OR
Covariate	B	S.E.	Z (Chi-square)	df	p value	Odds ratio	Lower	Upper
Own home	−0.85	0.27	−3.11	1	0.002	0.44	0.26	0.74
Ethnicity (Referent Latina)			(9.47)	4	0.05
Black/African American	1.01	0.37	2.75	1	0.006	2.74	1.34	5.62
Asian/Pacific Islander	0.57	0.44	1.30	1	0.195	1.76	0.75	4.15
Caucasian	0.74	0.31	3.99	1	0.019	2.09	1.13	3.88
Mixed/other	0.96	0.35	2.35	1	0.007	2.60	1.30	5.19
Income	−0.11	0.03	−3.84	1	0.000	0.89	0.84	0.95
Insurance (Referent Private)			(11.21)	3	0.01
None	0.05	0.24	0.23	1	0.82	1.06	0.66	1.68
MediCal	0.76	0.26	2.96	1	0.003	2.14	1.29	3.56
Medicare	0.29	0.31	0.94	1	0.35	1.33	0.73	2.44
Community setting (Referent Clinical)	−1.47	0.21	−6.92	1	0.000	0.23	0.15	0.35
Substance abuse	0.79	0.21	3.72	1	0.000	2.20	1.45	3.34
Intimate partner violence	0.38	0.19	1.98	1	0.048	1.47	1.003	2.15
Intercept	−0.21	0.39	−0.54	1	0.59

Discussion

In this study, we examined smoking and behavioral health of women participants in the COJAC study, and we found higher rates of substance abuse, depression, and anxiety among smoking than among nonsmoking women. Significantly more smokers than nonsmokers had higher rates of substance abuse, depression, anxiety, childhood physical and sexual abuse, and violence and also reported a co-occurring substance abuse and mental health problem during the past year. Twice as many smokers experienced suicidal thoughts compared to nonsmokers. Substance abuse, being in a treatment setting, IPV, African American and mixed ethnicity, Medicaid insurance status, reduced income, and lack of home ownership were identified as predictors of smoking in this sample of women. Our findings suggest that women who smoke may have chronic yet urgent co-occurring behavioral health problems and may have experienced IPV during their lifetimes as well.

That substance abuse was identified as a predictor of smoking for women in this sample contributes further data on the gendered nature of smoking and supports established research on adult smokers with substance use disorders and mental health problems. Our finding that IPV was a predictor of smoking also expands research on the significant association between violence (IPV) and smoking by women, as described in a growing number of population-based studies^39

–42 and in correlational studies of IPV and women's health risk behaviors.^43,44 This study did not examine specific mechanisms underlying the relationship between IPV and smoking. However, that association has been described as fundamentally rooted in stress and coping models,⁴⁵ where women's smoking is in response to “mental distress,” including interpersonal violence.^46,47 Relational perspectives on the mechanisms of IPV and women's smoking have described functional aspects of smoking, such as helping women with social relationships, creating a positive self-image, and forming identity in female survivors of abuse.⁴⁸ Within this relational framework, patterns of power dynamics and abuse are known to persist when partners of perinatal smokers use verbal abuse, isolation, economic control, and intimidation to influence smoking cessation during pregnancy.⁴⁹

Our findings are similar to those of Gerber et al.,⁵⁰ who examined adverse health behaviors and IPV among female patients in healthcare settings and found that smoking significantly increased the probability of lifetime IPV. Although data indicate that about 1 in 4 women will experience IPV⁵¹ and those aged 20–24 years are at greatest risk for nonfatal IPV,⁵² in this sample of women, IPV was at higher than expected rates across all age groups, with about half of women aged 26–65 reporting a history of lifetime IPV. The positive association between smoking and childhood abuse has been described among perinatal women,⁵³ college-age women,⁵⁴ and adult women with childhood adversities.⁵⁵ Similar to the work of Dube et al.,⁵⁵ we found that women smokers in this sample were more likely than nonsmokers to have experienced childhood physical and sexual abuse, but we did not examine the association of smoking to levels of other childhood psychologic burden, as described in their population-based study. Bisexual women also had the highest rates of smoking and were at significantly higher risk for a COD, PTSD, depression, anxiety, suicidality, and substance abuse; however, they were overrepresented in the treatment settings. Our findings on the mental health problems of bisexual women reinforce other research indicating higher rates of current smoking and suicidality.^56,57 From these data, it is clear that groups of sexual minority women should not be combined together as one group in processes of data analyses or interpretation.

American Indian and African American women, with the highest rates of smoking, were also found to be at significantly higher risk for mental health problems. Of interest in our sample is that when examined by ethnicity, the smoking rates for all groups greatly exceed smoking rates of women in published surveillance data reports during the same time period.⁵⁸ Again in our sample, African American women are likely overrepresented in the treatment settings. Together, these findings suggest the need for group-specific prevention and treatment. The smoking rates of women per site (i.e., primary care, substance abuse treatment, mental health services) were consistent with other published reports on patient smoking rates per treatment settings.^14,59,60

Universal and routine screening, assessment, and treatment of women for nicotine dependence and its associated cluster of other health problems, including substance use disorders, mental health conditions, and IPV, are recommended. When any one problem in the cluster is detected, the potential for others may be assumed by the clinician. Known benefits of smoking cessation include immediate improvements in health,⁶¹ enhanced substance abstinence and recovery for people with SUD,^62,63 and benefits from available smoking cessation treatments for individuals with mental health disorders.⁶⁴ Opportunities for smoking cessation should also be provided to women in domestic violence programs and shelters and in trauma-responsive gender-specific substance abuse treatment models, as evidence in support of this approach to treatment services integration has clearly accumulated.

Limitations of this study include a bias in sampling as a result of conducting the survey in settings known to have higher rates of smoking and behavioral health problems; therefore, the results may not be generalizable to women in other treatment settings. This study did not use population-based sampling, although the sample size was large enough to detect statistically significant differences between smokers and nonsmokers. Additionally, although higher scores on the PCL were detected in select groups of women including smokers, the PCL is a screening tool and is not diagnostic of PTSD. Thus, the findings suggest only a possible association to PTSD. Because of funding and time constraints, the study was conducted in English only. Finally, because of the primary objective of the study and the length of the survey, smoking status was determined by a single survey item. Measurements of nicotine dependence, including daily cigarettes consumed, biochemical measures of heaviness of nicotine dependence, and past smoking status, were not conducted. In particular, further examination of ex-smokers might well yield a distinctly different psychologic profile from the current smokers we describe.

Conclusions

Our study shows that smoking in women is associated with significantly higher rates of both mental health and substance abuse problems. In this sample of adult women, substance abuse and IPV were predictive of smoking. This close association of smoking with behavioral health problems and, in our findings, with IPV underscores the rationale for routine screening and assessment of women for both smoking and the other sentinel problems in this cluster of potential markers for smoking. Health practitioners and behavioral health treatment staff are especially well placed to ask and intervene for treatment and recovery from these co-occurring and significant barriers to women's health and recovery. Future research should include model testing of smoking outcomes as well as qualitative research to inform the models for testing.

Footnotes

Acknowledgments

This research was supported by the California Department of Alcohol and Drug Programs (07-00168), with funds from the California Mental Health Services Act of 2004. We thank Renée Zito, past Director of the California Department of Alcohol and Drug Programs (ADP), and staff of ADP and Stephen W. Mayberg, past Director of the California Department of Mental Health (DMH), and staff of DMH for their assistance and support for this project. We also extend our thanks to each member of COJAC, to the COJAC Chairs, Cheryl Trenwith, M.F.T., and Marvin Southard, D.S.W., and to members of the COJAC Screening Committee, and we acknowledge them for their work. This study could not have been conducted without the collaboration of the staff and managers at the study sites who gave generously of their time and energy to support the study, and we thank them for their contribution to the work.

Disclosure Statement

The authors have no conflicts of interest to report.

References

Centers for Disease Control and Prevention. Annual smoking-attributable mortality, years of potential life lost, and productivity losses—United States, 2000–2004. MMWR, 2008; 57:1226–1228. www.cdc.gov/mmwr/preview/mmwrhtml/mm5745a3.htm. 2010 January 21.

Centers for Disease Control and Prevention. Cigarette smoking among adults and trends in smoking cessation—United States, 2008. National Health Interview Survey. MMWR, 2009; 58:1227–1232. www.cdc.gov/mmwr/preview/mmwrhtml/mm5844a2.htm#tab. 2010 January 21.

Centers for Disease Control and Prevention. Rates of new lung cancer cases. www.cdc.gov/Features/dsLungCancer/. 2011 February 10.

U.S. Department of Health and Human Services, Centers for Disease Control and Prevention. 1999–2007 Cancer incidence and mortality data. www.cdc.gov/uscs. 2011 February 10.

Kalman

, Morissette

, George

. Co-morbidity of smoking in patients with psychiatric and substance use disorders. Am J Addict, 2005; 142:106–123.

Degenhardt

, Hall

. The relationship between tobacco use, substance-use disorders and mental health: Results from the National Survey of Mental Health and Well-being. Nicotine Tob Res, 2001; 3:225–234.

Greaves

, Hemsing

. Women and tobacco control policies: Social-structural and psychosocial contributions to vulnerability to tobacco use and exposure. Drug Alcohol Depend, 2009; 104S:S121–S130.

Epstein

, Induni

, Wilson

. Patterns of clinically significant symptoms of depression among heavy users of alcohol and cigarettes. Prev Chronic Dis Public Health Res Pract Policy, 2009; 6:1–10.

Husky

, Mazure

, Paliwal

, McKee

. Gender differences in the comorbidity of smoking behavior and major depression. Drug Alcohol Depend, 2008; 93:176–179.

10.

, McFall

, Saxon

et al. Post-traumatic stress disorder and smoking: A systematic review. Nicotine Tob Res, 2007; 9:1071–1084.

11.

Lasser

, Boyd

, Woodhandler

, Himmelstein

, McCormick

, Bor

. Smoking and mental illness: A population-based prevalence study. JAMA, 2000; 284:2606–2610.

12.

Flensborg-Madsen

, von Scholten

, Flachs

, Mortensen

, Prescott

, Tolstrup

. Tobacco smoking as a risk factor for depression: A 26-year population-based follow-up study. J Psychiatr Res, 2011; 45:143–149.

13.

Substance Abuse and Mental Health Services Administration U.S. Department of Health and Human Services. 2007 National Survey of Substance Abuse Treatment Services (N-SSATS) www.oas.samhsa.gov/nssats2k7/NSSATS2k7Hi.htm. 2009 March 20.

14.

Schroeder

, Morris

. Confronting a neglected epidemic: Tobacco cessation for persons with mental illness and substance abuse problems. Annu Rev Public Health, 2010; 31:16.1–16.18.

15.

Reif

, Horgan

, Garnick

, McLellan

. Systems-level smoking cessation activities by private health plans. Prev Chronic Dis, 2011; 8:A14. www.cdc.gov/pcd/issues/2011/jan/09_0236.htm. 2010 January 21.

16.

U.S. Department of Health and Human Services, Centers for Disease Control and Prevention. 2004 Surgeon General's report—The health consequences of smoking. www.cdc.gov/tobacco/data_statistics/sgr/2004/index.htm. 2011 February 10.

17.

Cherubin

, Sapira

. The medical complications of drug addiction and the medical assessment of intravenous drug use 25 years later. Ann Intern Med, 1993; 119:1017–1028.

18.

Salaspuro

. Interrelationship between alcohol, smoking, acetaldehyde and cancer. Chadwick

, Goode

. Acetaldehyde-related pathology: Bridging the trans-disciplinary divide: Novartis Foundation Symposium 285. Chichester, UK: John Wiley & Sons, 2007; 285:80–89.

19.

Hurt

, Offord

, Croghan

et al. Mortality following inpatient addictions treatment. JAMA, 1996; 275:1097–1103.

20.

Sin

, Cohen

, Day

, Coxson

, Paré

. Understanding the biological differences in susceptibility to chronic obstructive pulmonary disease between men and women. Proc Am Thorac Soc, 2007; 4:671–674.

21.

Kiyohara

, Ohno

. Sex differences in lung cancer susceptibility: A review. Gender Med, 2010; 7:381–401.

22.

Benowitz

. Nicotine addiction. N Engl J Med, 2010; 362:2295–2303.

23.

Heinz

, Kassel

, Berbaum

, Mermelstein . Adolescents' expectancies for smoking to regulate affect predict smoking behavior and nicotine dependence over time. Drug Alcohol Depend, 2010; 111:128–135.

24.

Currie

, Hodgins

, el-Guebaly

, Campbell

. Influence of depression and gender on smoking expectancies and temptations in alcoholics in early recovery. J Subst Abuse, 2001; 13:443–458.

25.

Gruskin

, Greenwood

, Matevia

, Pollack

, Bye

. Disparities in smoking between the lesbian, gay, and bisexual population and the general population in California. Am J Public Health, 2007; 97:1496–1502.

26.

Rooney

, Choudhary

, Bliss

. Social determinants of smoking among Hmong Americans residing in Wisconsin. Wis Med J, 2009; 108:439–446.

27.

Mickens

, Ameringer

, Brightman

, Leventhal

. Epidemiology, determinants, and consequences of cigarette smoking in African American women: An integrative review. Addict Behav, 2010; 35:383–391.

28.

Afifi

, Nakkash

, Khawaja

. Social capital, women's autonomy and smoking among married women in low-income urban neighborhoods of Beirut, Lebanon. Womens Health Issues, 2010; 20:156–176.

29.

, Heck

, Dresler

et al. Determinants of smoking initiation among women in five European countries: A cross-sectional survey. BMC Public Health, 2010; 10:74. www.biomedcentral.com/1471-2458/10/74. 2011 February 10.

30.

Jessup

, Dibble

. Validity and reliability of the COJAC screening tool for co-occurring disorders. Am J Addict, 2011; 20:264–270.

31.

Finley

Survey Monkey^™ online software application. Portland, OR, 1999. www.surveymonkey.com. December 2008–June 2009 .

32.

Dennis

, Chan

, Funk

RR.

Development and validation of the GAIN Short Screener (GAIN-SS) for psychopathology and crime/violence among adolescents and adults. Am J Addict, 2006; 15,Suppl 1:80–91.

33.

Lang

, Stein

. An abbreviated PTSD checklist for use as a screening instrument in primary care. Behav Res Ther, 2005; 43:585–594.

34.

U.S. Department of Veterans Affairs, National Center for Post-traumatic Stress Disorder. Short form of the PTSD checklist—Civilian version. www.ptsd.va.gov/professional/pages/assessments/civilian-ptsd-checklist.asp. 2008 August 20.

35.

Elhai

, Gray

, Kashdan

, Franklin

. Which instruments are most commonly used to assess traumatic event exposure and posttraumatic effects? A survey of traumatic stress professionals. J Trauma Stress, 2005; 18:541–545.

36.

McFadden

. The measurement of urban travel demand. J Public Econ, 1974; 3:303–328.

37.

Menard

. Coefficients of determination for multiple logistic regression analysis. Am Stat, 2000; 54:17–24.

38.

Mittlbock

, Schemper

. Explained variation for logistic regression. Stat Med, 1996; 15:1987–1997.

39.

Yoshihama

, Horrocks

, Bybee

. Intimate partner violence and initiation of smoking and drinking: A population-based study of women in Yokohama, Japan. Soc Sci Med, 2010; 71:1190–1207.

40.

Hathaway

, Mucci

, Silverman

, Brooks

, Mathews

, Pavlos

. Health status and health care use of Massachusetts women reporting partner abuse. Am J Prev Med, 2000; 19:302–307.

41.

Breiding

, Black

, Ryan

. Chronic disease and health risk behaviors associated with intimate partner violence—18 U.S. states/territories, 2005. Ann Epidemiol, 2008; 18:538–544.

42.

Lemon

, Verhoek-Oftedahl

, Donnelly

. Preventive healthcare use, smoking, and alcohol use among Rhode Island women experiencing intimate partner violence. J Womens Health Gend Based Med, 2002; 11:555–562.

43.

Gass

, Stein

, Williams

, Seedat

. Intimate partner violence, health behaviors, and chronic physical illness among South African women. S Afr Med J, 2010; 100:582–585.

44.

McFarlane

, Parker

, Soeken

. Physical abuse, smoking, and substance use during pregnancy: Prevalence, interrelationships, and effects on birth weight. J Obstet Gynecol Neonatal Nurs, 1996; 25:313–320.

45.

Jun

H-J

, Rich-Edwards

, Boynton-Jarrett

, Wright

. Intimate partner violence and cigarette smoking: Association between smoking risk and psychological abuse with and without co-occurrence of physical and sexual abuse. Am J Public Health, 2008; 98:527–535.

46.

Vest

, Catlin

, Chen

, Brownson

. Multi-state analysis of factors associated with intimate partner violence. Am J Prev Med, 2002; 22:156–164.

47.

McNutt

, Carlson

, Persaud

, Postmus

. Cumulative abuse experiences, physical health and health behaviors. Ann Epidemiol, 2002; 12:123–130.

48.

Greaves

. Smoke screen: Women's smoking and social control. Winnipeg: Fernwood Publishing, 1996.

49.

Greaves

, Kalaw

, Bottorff

. Case studies of power and control related to tobacco use during pregnancy. Womens Health Issues, 2007; 17:325–332.

50.

Gerber

, Ganz

, Lichter

, Williams

, McCloskey

. Adverse health behaviors and the detection of partner violence by clinicians. Arch Intern Med, 2005; 165:1016–1021.

51.

Center for Disease Control and Prevention. Adverse health conditions and health risk behaviors associated with intimate partner violence. MMWR, 2008; 57:113–117. www.cdc.gov/mmwr/preview/mmwrhtml/mm5705a1.htm. 2011 February 10.

52.

Catalano

. Intimate partner violence in the United States. US Department of Justice, Bureau of Justice Statistics. bjs.ojp.usdoj.gov/content/intimate/ipv.cfm. 2011 February 10.

53.

Chung

, Nurmohamed

, Mathews

, Elo

, Coyne

, Culhane

. Risky health behaviors among mothers-to-be: The impact of adverse childhood experiences. Acad Pediatr, 2010; 10:245–251.

54.

De Von Figueroa-Moseley

, Abramson

, Williams

. College women: History of childhood abuse and its relationship to smoking. Violence Against Women, 2010; 16:1242–1251.

55.

Dube

, Cook

, Edwards

. Health-related outcomes of adverse childhood experiences in Texas. 2002. www.cdc.gov/pcd/issues/2010/may/09_0158.htm. 2010 January 31.

56.

Steele

, Ross

, Dobinson

, Veldhuizen

, Tinmouth

. Women's sexual orientation and health: Results from a Canadian population-based survey. Women Health, 2009; 49:353–367.

57.

Conron

, Mimiaga

, Landers

. A population-based study of sexual orientation identity and gender differences in adult health. Am J Public Health, 2010; 100:1953–1960.

58.

California Department of Public Health, California Tobacco Control Program. California tobacco control update 2009: 20 years of tobacco control in California. 2009. www.cdph.ca.gov/programs/tobacco/Documents/CTCPUpdate2009.pdf. 2010 February 14.

59.

Jones

, Knutson

, Haines

. Common problems in patients recovering from chemical dependency. Am Fam Physician, 2003; 68:1971–1978.

60.

Mertens

, Lu

, Parthasarathy

, Moore

, Weisner

. Medical and psychiatric conditions of alcohol and drug treatment patients in an HMO: Comparison with matched controls. Arch Intern Med, 2003; 163:2511–2517.

61.

Fagerström

. The epidemiology of smoking: Health consequences and benefits of cessation. Drugs, 2002; 62,Suppl 2:1–9.

62.

Tsoh

, Chi

, Mertens

, Weisner

. Stopping smoking during first year of substance use treatment predicted 9-year alcohol and drug treatment outcomes. Drug Alcohol Depend, 2011; 114:110–118.

63.

Prochaska

, Delucchi

, Hall

. A meta-analysis of smoking cessation interventions with individuals in substance abuse treatment or recovery. Postprints, Multi-Campus. escholarship.org/uc/item/0r8673wv. 2011 February 10.

64.

Hall

, Prochaska

. Treatment of smokers with co-occurring disorders: Emphasis on integration in mental health and addiction treatment settings. Annu Rev Clin Psychol, 2009; 5:409–431.