Smoking-Related Weight Concerns Among Underserved, Black Maternal Smokers

Abstract

Objective

To expand understanding of a smoking cessation barrier for women, weight concerns, in a medically underserved population.

Methods

Baseline weight concerns were examined among 235 low-income, black maternal smokers enrolled in a smoking trial. Logistic regression evaluated factors related to weight concerns.

Results

Higher BMI (OR 3.35, P<.001), intention to quit (OR 2.12, P=.02), more previous quit attempts (OR 1.14, P=.03), and less support for quitting (OR 0.81, P=.05) predicted weight concerns.

Conclusions

This is the first study to delineate factors predicting weight concerns in this population, thus expanding our understanding of a key cessation barrier and informing future cessation strategies in a population known to bear increased risk of tobacco-related disease.

Studying factors associated with smoking in groups with high risk for tobacco-related disease remains a public health priority. Weight concerns are an important factor contributing to smoking among women, but little is known about variability in weight concerns among high-risk groups, such as those living in medically underserved homes. Most smokers learn that nicotine suppresses appetite and weight; however, women are more likely than men to smoke to control weight and to express concern about postcessation weight gain.^1–3 Weight-concerned smokers are more likely to smoke to control weight,⁴ refrain from quitting for fear of postcessation weight gain,⁵ have poorer abstinence outcomes,¹ and have high attrition rates in smoking treatment.⁶ Improving our understanding of variability in weight concerns among medically underserved maternal smokers could inform future cessation strategies designed to reduce women and children’s tobacco morbidity and mortality risk. This paper examines weight concerns among mostly low-income, undereducated, black maternal smokers, a group known to experience unique and elevated tobacco use consequences compared to other subgroups of smokers.

A number of factors may predict weight concerns. Greater weight concern is often associated with higher levels of education and income,⁷ and women under 30 years are more likely to smoke to control weight⁸ and to cite weight gain as a reason for relapse.⁹ Social support and depressive symptoms, known factors associated with smoking in women,^10,11 may also relate to weight concerns. Because of potential weight concern-depressive symptoms associations,¹² together with the high prevalence of lifetime major depression among weight-concerned women,¹³ depressive symptoms may contribute to weight-concerned women’s challenges with quitting smoking and relapse. Depressive symptoms, social support, and weight concerns may be particularly important factors for maternal smokers with infants and toddlers. During the first couple of years postpartum, there is potential for depressive mood presentation, increased parenting stress, and reduced social support compared to that during pregnancy, and many women may experience increased motivation to return to prepregnancy weight.

In addition to these psychosocial factors, smoking-specific factors relate to weight concerns in females. For example, intention to quit has been associated with weight concerns¹² with greater concern relating to previous quit attempts and greater expectancies about weight-control benefits of smoking.¹⁴ Female smokers with greater concerns about postcessation weight gain tend to have less motivation and confidence about quitting,^15,16 and smoke more cigarettes per day than do unconcerned females;¹⁴ however, evidence of nicotine dependence-weight concern relations remains equivocal.^3,17–19

Some researchers have proposed weight concerns-body image associations,^20,21 predicting that lower BMI smokers would have greater weight concerns based on a broader literature supporting BMI-body esteem associations.²² One cessation trial suggested that females with weight concerns were more likely to have a lower BMI, be younger, and have white background.³ However, other studies have shown that weight-concerned women weighed more than their counterparts.²³ Perhaps a BMI-weight concern relation in smokers is represented as a U-shaped distribution, in which women with relatively low or high BMI express weight concerns.

Despite the growing understanding of factors influencing weight concerns, the literature focuses overwhelmingly on white, middle-to-upper socioeconomic populations, with a few exceptions.^24,27 This focus may have resulted from broader evidence that body image and weight concerns are more prevalent in affluent white women than other subgroups²⁸ and evidence that black women tend to have greater body satisfaction, are heavier, and have higher preferred weights than white women do.²⁸ However, this resulting focus may lead to a misinterpretation that weight concerns are inconsequential among lower-income black women who smoke. Recent research suggests that smoking-related weight concerns may be an issue among black women smokers who are contemplating or have previously attempted to quit smoking in treatment.^15,25

This study examined factors hypothesized to contribute to weight concerns in a sample of black, maternal, smokers enrolled at baseline in an ongoing second-hand smoke reduction trial with their children (<4 years old). This population is an important one to examine from a health disparities perspective because women, blacks, and maternal (eg, postpartum) women in particular have many unique tobacco-related health risks and often face greater challenges to adopting health promotion behaviors. In general, low-income, undereducated, black women experience myriad socioeconomic and psychosocial challenges that contribute to health risks and barriers to behavior change, including those challenges related to quitting smoking. For example, urban poverty confers increased risk for a wide range of health and social problems, and it tends to increase risk for specific outcomes such as smoking, obesity, secondhand smoke exposure, and secondary health consequences to those behaviors. These challenges are further reflected in higher smoking and relapse rates and higher disease risk compared to other groups of smokers. To reduce such disparities, researchers must recognize the challenges inherent in this community with respect to nicotine dependence. Women generally have greater difficulty quitting smoking than men, even in evidence-based treatments,²⁹ and postpartum smoking is a common problem for mothers who smoked before or during their pregnancy, particularly urban black women – with the vast majority of women relapsing within the first 2 years following delivery.^30,31

This study may improve our understanding of weight concerns in this high-risk group of maternal smokers and could present an important step to reducing health disparities in a population known to suffer greater tobacco-related risks.^30,32 Based on the previous weight-concern literature, we hypothesized that BMI, depressive symptoms, intention to quit, maternal age, and lack of social support would predict weight concerns in our study.

METHODS

Design and Participant Enrollment Procedures

This study examined prerandomized participant responses to assessment interviews administered as part of an ongoing, behavioral counseling trial, Philadelphia FRESH (Family Rules for Establishing Smoke-free Homes). FRESH was designed to reduce child exposure to secondhand smoke and maternal smoking rates (Collins: K07 CA93756 and R01 CA105183). Two hundred thirty-five participants in this study included black maternal smokers over 18 years old with children under 4 years old. In accordance with aims of the FRESH trial, investigators employed a purposive sampling strategy to access underserved, low-income, urban maternal smokers with exposed children. We recruited from pediatric primary care, Women, Infant, and Children’s (WIC) clinics and deployed targeted advertising (eg, local newspapers, posters on transit lines) in medically underserved Philadelphia neighborhoods. Recruitment in clinics included advertising on posters, brochures, and fliers. Some advertisement respondents called FRESH staff to complete the 10-minute telephone eligibility screening; in-clinic respondents completed prescreening forms and consented to have FRESH staff call them for screening. Inclusion criteria required that participants smoke at least 5 cigarettes per day. Exclusion criteria included diagnosis of current severe psychopathology (eg, psychotic disorder), pregnancy, and non-English speaking. Participants voluntarily enrolled in the trial following informed consent procedures and completion of a 70-minute, in-home baseline interview. Approximately 90% of enrolled participants were black, and only blacks were included in the sample for the present study based on the purpose of the study.

Measures

Measures were obtained via interviewer administration of the FRESH trial’s screening and baseline assessments. Screening interviews collected inclusion, exclusion, socio-demographic, clinic, and participant contact data. Baseline interviews included questions relevant to trial aims (eg, detailed smoking and exposure history, child health, and factors hypothesized to influence outcomes). In addition to standardized scales used to test primary trial aims, content valid interview-items below the seventh-grade Flesch-Kinkaid reading level were constructed by a panel of smoking intervention and public health experts, then tested for content validity via interview with a pilot sample of postpartum smokers. Limiting the number of standardized questionnaires in favor of short, content valid items decreased assessment duration, thereby minimizing participant burden. Items with common smoking-assessment terms (eg, relapse) included oral operational definitions with opportunity for interviewer clarification if queried or suspecting item misinterpretation. Training procedures for our licensed social worker and PhD student interview staff included (a) ongoing assessment of participant comprehension of items, (b) double data entry of participant responses to ensure accurate response coding, and (c) audiotaped quality assurance assessment by the principal investigator to ensure data collection integrity.

Outcome variable

Participant weight-concerns score was obtained by calculating the mean response to 4 items using a 6-point Likert scale (0=not at all, to 5=very much). Internal consistency was moderately high (α=.82, n=224). For multivariate analyses, weight-concerns score was dichotomized at the median (0–1.5=no; 1.5–5=yes). Items were derived from Borrelli and Mermelstein’s³³ scale. Specific items included (a) “In your decision to continue smoking in the past, how much was your concern about weight gain a reason for you to continue smoking?”; (b) “If there was a guarantee you would not gain too much weight after quitting smoking, how much do you believe you would be an ex-smoker within the next year?”; (c) “Since the birth of your baby, how much have you been concerned about your weight?”; and (d) “How much is any concern about weight gain keeping you from wanting to quit smoking now?”

Predictor variables

Predictor variables included demographic factors (education, mother’s age, children’s age, and marital status), psychosocial factors (depressive symptoms, social support), smoking factors (nicotine dependence, intention to quit smoking), and body mass index.

Demographic factors

Self-reported age and child age were continuous variables. Baby’s age was included to account for potential differential weight concerns based on time since birth (eg, postpartum). Mothers’ reported education was coded as a categorical response and converted into 2 dummy variables: (a) high school education versus less-than-a-high-school education, and (b) training beyond high school versus high school or less. (Table 1 reflects true distribution prior to dummy coding for analyses.) Similarly, self-reported marital status was dummy coded as 1 = married vs 0 = single. Married represented being married or living with a partner, whereas single included being single, widowed, divorced, or separated.

Table 1.

Sample Characteristics (N=235)

	Distributional Characteristics
Variable	Mean	SD	Frequency	Missing
Mean Weight Concern	1.81	1.52		0
Baseline CES-D Score	19.18	10.61		0
Baseline FTND Score	3.76	2.13		2 (0.9%)
Baseline ISEL Score	37.00	6.27		0
Body Mass Index (≥25 = overweight)	29.30	6.83		5 (2.1%)
Mean Number of Cigarettes Per Day	16.04	8.81		0
Mother’s Age in Years	29.60	7.95		0
Baby’s Age in Months	20.26	15.52		0
Number of Past Quit Attempts	2.17	2.55		0
Martial Status
Married/living with a partner			37 (15.7%)
Single			198 (84.3%)
Missing				0
Support for Quitting from Friends/family
None			14 (6.0%)
Not much			18 (7.7%)
A Little			16 (6.8%)
Some			26 (11.1%)
A Lot			70 (29.8%)
Very much			90 (38.5%)
Missing				1 (0.4%)
Intention to Quit Smoking in Next 30 Days
Yes			62 (26.4%)
No			173 (73.6%)
Missing				0
Education
Less than High School Equivalent			79 (33.6%)
High School Equivalent			101 (43.0%)
Postsecondary Training Beyond			54 (23.0%)
Missing				1 (0.4%)
Income
$15,000 or less			159 (67.2%)
More than $15,000			66 (28.1%)
Missing				11 (4.7%)
Employment Status
Employed			76 (32.3%)
Unemployed			159 (67.7%)

Note.

SD = standard deviation

Psychosocial factors

Depressive symptoms were obtained using the Center for Epidemiologic Studies Depression Scale (CES-D), a 20-item self-report instrument with adequate test-retest reliability (r=.57).³⁴ Its scores are related to clinical measures of depression severity and motives to smoke to reduce negative affect.³⁴ Scores greater than 16 can be interpreted as suggesting the potential for depressive disorder. The CES-D has been a reliable measure of depressive symptoms in a variety of homogeneous black subgroups.^35,36 To assess social support, we used the Interpersonal Support Evaluation List (ISEL) global support score to provide a meaningful index of social support.^37,38 Cohen and colleagues reported good overall test-retest reliability averaging around .87 and adequate concurrent and discriminant reliabilities (r = .46 and r =−.64).³⁷

Smoking factors

Nicotine dependence was measured by the Fagerström Test for Nicotine Dependence (FTND).³⁹ This 6-item self-report measure has adequate internal consistency (α=.64) and high test-retest reliability (r=.88).⁴⁰ Age of smoking initiation was included based on the response to the item, “How old were you when you started smoking at least one cigarette per day?” Self-reported cigarettes per day were assessed as a continuous variable during the past 14 days to correspond to questions assessing acute tobacco-related child illness presentation and severity. Participants also reported their intention to quit smoking within the next 30 days (0=No; 1=Yes) and the total number of previous quit attempts, defined as quitting for at least 3 consecutive days. Participants also reported the total number of smokers in the home. Although there is no existing literature suggesting an association with weight concerns, the number of smokers in participants’ homes could influence perceptions, social contingencies, and complex interrelationships between smoking, eating, and acceptable weight. Participant-reported perceived support specific to smoking cessation was assessed using a single item question with a 6-point Likert response scale to the question, “If you ever decided to quit smoking, how much support would you get from your friends and family?” (0=none, 5=very much). Body mass index (BMI = weight in kg/square of height in meters) was calculated from self-reported height and weight measures. For the multivariate analysis, BMI was dichotomized using the American College of Sport Medicine guidelines of normal weight and overweight (less than 25 = normal weight; greater than 25 = overweight).⁴¹

Analytic Procedures

Associations among predictors and criterion were analyzed using both univariate and multivariate procedures. Pearson bivariate correlations were used to evaluate simple associations between the criterion of mean weight concern and the various predictors. Because the criterion variable was dichotomous (high vs low weight concerns), data were analyzed using direct-entry (simultaneous) logistic regression with preselected hypothesized factors related to weight concerns. Factors that showed collinearity or those that did not show significant bivariate associations were excluded from the analysis. Multivariate associations are generally superior to univariate correlations because they better capture the full set of interrelationships among predictors and criteria.²⁵ Analyses were performed using SPSS 16.

RESULTS

Sample Characteristics and Weight Concerns

Table 1 shows the sample characteristics of the 235 black females in this study, which consisted of primarily single, unemployed, low-income and low-educated mothers who are moderate-to-heavy daily smokers. The mean weight-concern score was below “a little” on the 6-point response scale; however, only 17.9% reported no weight concerns. A substantial proportion of the sample reported some degree of weight concerns such that 42.1% reported “not too much” and 27.2% reported “a little” to “some” weight concerns, and 12.8% of the sample had “a lot” to “very much” concerns. Mean BMI and CES-D scores suggest that the majority of the sample was overweight or obese and above the cutoff considered to reflect potential for a depressive mood disorder.

Associations With Weight Concerns

Table 2 provides bivariate correlations between the predictors and the criterion of mean weight concerns. Five predictors showed bivariaate associations with the criterion: CES-D, ISEL, body mass index, intention to quit, and support for quitting smoking. Logistic regression analysis resulted in a statistically significant model, (χ² = 34.33, df = 10, P<.01) accounting for approximately 19% of the total variance in weight concerns. Converting the pseudo R² to Cohen’s f² statistic⁴² suggests a moderate-to-large effect size (f² = .23) for the model in discriminating participants with weight concerns versus those with no weight concerns.

Table 2.

Bivariate Correlations Between Predictors and Mean Weight Concern (N=235)

Predictor	Correlation to Mean Weight Concern
Baseline CES-D Score	.16*
Baseline ISEL score	−.19**
Body Mass Index	.29***
Support from Friends/Family on Quitting	−.15*
Intention to Quit Smoking in Nest 30 Days	.09
Mother’s Age in Years	.11
Baby’s Age in Months	−.02
Education Greater than High School Degree	−.09
Number of Smokers Living in the House	−.06
Number of Previous Quit Attempts	.10
Baseline FTND Score	.06
Average Number of Cigarettes Per Day	.04
Marital Status	.08
Age of Daily Smoking Initiation	−.01

Note.

^*P=.05,

^**P=.01,

^***P=.001.

Table 3 presents the regression model. Significant predictors in the model included BMI, intention to quit, number of quit attempts, and support for quitting smoking. The odds ratio for BMI indicated that overweight vs normal weight women in this sample had more than a 3-fold increase in smoking-related weight concerns. Women who intended to quit had more than double the risk of weight concerns. Women had moderate increased risk of weight concerns if they reported more previous quit attempts and less support for quitting smoking.

Table 3.

Logistic Regression Summary for Predicting Weight Concerns

	95% CI for Odds Ratio
Predictor	Odds Ratio	Lower	Upper	P
BMI (overweight)	3.35	1.71	6.54	<.001
Intention to Quit in Next 30 Days (yes)	2.12	1.10	4.07	.02
Number of Previous Quit Attempts	1.14	1.01	1.27	.03
Support for Quitting Smoking	0.81	0.65	1.00	.05
Baby’s Age (in months)	0.98	0.97	1.00	.07
Education >High School	1.66	0.84	3.26	.15
Average Cigarettes Per Day	1.02	0.99	1.06	.18
Number of Smokers in the House	0.82	0.57	1.17	.28
Baseline CES-D Score	1.02	0.99	1.06	.33
Baseline ISEL Score	0.99	0.94	1.05	.82
Constant	0.55			.65

DISCUSSION

This study adds to a growing awareness of weight concerns among black, female smokers. To our knowledge, this is the first study to demonstrate weight concerns in a medically underserved sample of predominantly low-income and single black maternal smokers with young children. Specifically, these results suggest that weight concerns are an issue for underserved black women smokers within the context of high BMI, particularly if they have previous experience quitting, are planning to quit smoking within the 30 days, and perceive less support for their upcoming quit attempt from family and friends. Considering that black women tend to report greater body satisfaction and higher preferred weights than do other groups of women, perhaps it is not surprising that almost 18% of the sample reported no weight concerns (0 on a 6-point scale) – an outcome similar to a study that examined weight concerns in black male and female smokers.²⁶ Nonetheless, the distribution of weight concerns in this study is informative, and results are consistent with others that suggest weight concerns may be an important issue for some female smokers.^17,18

Logistic regression analyses largely supported our hypotheses, demonstrating that 4 hypothesized factors contributed to variability in weight concerns. Most of the women in this sample were overweight, and BMI contributed the most unique variance to the model. This outcome is consistent with Davis and colleagues²³ but contrary to other research. For example, the study of Pollak and colleagues suggested obese black smokers were least likely to be concerned about weight;²⁶ however, their analyses combined males and females and their sample included fewer overweight and obese participants than were enrolled in the present study. Future research is needed to replicate our BMI results and to examine prospective relations among black women’s smoking-related weight concerns, current weight, and weight gain.

Intention to quit smoking in the next 30 days and previous number of quit attempts also were significant predictors of weight concerns in the regression model, supporting results in previous studies focusing on female smokers enrolled in treatment.^25,43,44 Blacks under 55 have higher risk of major weight gain following cessation attempts,⁴⁵ which may explain why higher BMI women with previous quit experience may report greater weight concerns in the context of a pending quit attempt. The increased prevalence of overweight and weight-related health conditions in black women⁴⁶ may affect weight concerns when preparing to quit, particularly if previous quit attempts resulted in substantial weight gain. Future studies could elucidate how interrelationships between previous quits, BMI, and current weight concerns may influence quit attempt outcomes. Previous quit attempts revealed a higher multivariate association with weight concern than the bivariate correlation. This situation is not uncommon, and most researchers generally accept the multivariate association as superior to the univariate correlations because they better capture the full network of relations among predictors and criteria. In the current case, previous attempts to quit showed a low redundancy with the other predictors, which resulted in its having a stronger multivariate effect.

Results also underscore the likelihood of greater concern about weight gain in the absence of support for quitting smoking⁴⁷ and the potential of social support to attenuate weight concerns among women intending to quit smoking. Treatment programs for smokers often encourage development of social support around behavior change goals (eg, abstinence), and smokers are likely to seek support while quitting.⁴⁸ An opportunity remains to learn more about characteristics of social support that relate to weight concerns and smoking. A better understanding of the covariation of BMI and support could lead to better treatment approaches.

Maternal age was not associated with weight concerns. This result is contrary to previous findings suggesting greater likelihood of weight concerns among younger smokers.⁸ Our contrary results could reflect restricted age range of smokers in our sample (women of childbearing age over 18 years old). Depressive symptoms (CES-D) and general social support (ISEL) were related to weight concerns in bivariate correlation analysis, but contrary to our hypotheses, were overwhelmed by the variance of other factors in the logistic model. Although greater depressive symptoms relate to fewer weight concerns among black women smokers,²⁷ our outcome was somewhat surprising given substantial evidence of overlapping associations between depressive symptoms, social support, and similar constructs reflecting concerns or anxiety/stress among female smokers with characteristics similar to those of our sample. One possibility with respect to CES-D is that the majority of our sample reported substantial depressive symptoms, thereby restricting its variability. Regarding social support, it is possible that alternative support-related constructs could relate better to weight concerns in this sample. For example, specific social contingencies rather than general social support may influence behaviors and perceptions about eating, smoking, weight, and plans to change health behaviors.

This latter point underscores a study limitation: our assessments did not capture other latent, potential confounding sociocultural factors around eating, food choices, and contingencies around weight management, smoking, and related health behaviors that could also influence smokers’ weight concerns. Thus, future research could use more refined measures of potential punitive and reinforcing social contingencies around behaviors that influence weight and smoking that could differentially predict weight concerns in different subgroups of smokers. Our cross-sectional, pretreatment data presented another limitation that restricted our ability to assess the causal influence of our predictors and did not allow for the assessment of potential mediating effects of weight concern on smoking outcomes. Future treatment outcome analyses from this study’s parent trial will enable tests of such effects.

Other limitations should be highlighted regarding our sampling and measurement strategies. Although purposive sampling is a standard, accepted recruitment strategy for behavioral intervention trials designed specifically to test hypotheses in population subgroups (eg, smokers who meet inclusion criteria and voluntarily enroll per institutional review board regulations), it does not represent an epidemiological sample. Such constraints present limitations to the study’s ability to generalize beyond our volunteer sample. Future research is necessary to replicate these findings and determine the extent to which results generalize to other subpopulations of smokers; however, our sample is appropriate for examining potential causal relationships or hypothesized associations that could affect larger populations of female smokers. Typically, associations based on hypotheses and theory generalize from selected subsamples; and studies examining systematic subsamples from an overarching population, such as the present study, are likely to contribute to our understanding of etiological processes. Therefore, our results inform the literature and can appropriately guide future prospective studies to expand our understanding of weight concerns in females and how to address these concerns in smoking treatment.

Our measures included content valid items, which could affect the extent to which the results generalize. However, this measurement strategy is accepted and not uncommon when there exist practical constraints to collecting a large battery of standardized assessments across multiple time points in populations known to demonstrate high attrition rates. Many intervention trial investigators must balance participant burden against their ability to implement a full standardized battery of assessments relevant to their project aims. Despite this limitation, the internal consistency reliability of our weight-concern measure (.82) was similar to the standardized Borrelli and Mermelstein scale (.87) referenced in the measures section. Nonetheless, an appropriate next step for research in this area would be to replicate our initial findings using a cross-sectional study design with a full set of relevant, standardized assessments outside the constraints of an intervention trial.

Despite these limitations, this study adds to a growing awareness of weight concerns among black women smokers that could inform future research and interventions designed to break the cessation–weight gain–smoking relapse– weight concern cycle that undermines smokers’ quit attempts. This study suggests that weight concerns exist in this population and that high BMI may be a key predictor increasing smoking-related weight concerns, particularly among those mothers who are intending to quit smoking within 30 days, who have a history of cessation attempts, and who perceive less support for quitting.

This study could guide future research designed to examine the potential mediating effect of weight concerns on smoking treatment outcomes among underserved black maternal smokers. In broader populations of female smokers, this evidence exists, along with strategies to enhance cessation outcomes through managing weight concerns.⁵ One alternative could include testing additional health education strategies to better inform black, overweight maternal smokers about the additive health risks of smoking and overweight in light of their increased risks associated with smoking and overweight-related disease.⁴⁹ This might include components that help these women and their families reframe potential postcessation weight gain as a manageable and healthier alternative to continuing to smoke. However, any strategy targeting underserved maternal smokers must be sensitive to the psychosocial and socioeconomic context within which behavior change is attempted. In recognizing contextual challenges, health education strategies and exclusively idiographic strategies may not be sufficient to promote health behavior change. Strategies to modify weight concerns among underserved maternal smokers may require additional components given the limited social support that many underserved maternal populations report together with the numerous barriers they face when attempting to adopt healthier behaviors. Therefore, future intervention studies that target this high-risk group of smokers could examine the efficacy of components designed to reinforce family support for quitting smoking in conjunction with weight-management strategies that include increased physical activity. In broader populations of female smokers, anxiety-management strategies reduced weight concerns and promoted smoking outcomes.⁵ Thus, in addition to building social support for health behavior change, future interventions for underserved maternal smokers could examine the utility of cognitive-behavioral counseling components to manage anxiety and weight concerns as well as consider testing the effectiveness of managing weight gain and weight concerns through pharmacotherapy.

Because concern about weight gain may relate to lower confidence in quitting,⁵⁰ future interventions may need to explicitly address weight concerns among black maternal smokers preparing to quit smoking. One study indicated that highly weight-concerned women are unlikely to attempt self-quitting or to seek smoking cessation treatment.¹⁸ Perhaps harm-reduction treatments such as FRESH (where women can enroll without pressure to committing to a quit attempt) can allow women to change their behavior gradually by first focusing on strategies such as children’s secondhand smoke exposure reduction. Such an approach could foster coping skills around weight concerns, develop self-efficacy, and shape smoking behavior change toward cessation. Future intervention studies are necessary to test such hypotheses.

In conclusion, it is important to expand our understanding of factors that may predict smoking-related weight concerns and how weight concerns may mediate smoking outcomes in groups of smokers that evidence increased risk of tobacco-related disease. Improved knowledge of these factors is an important step toward reducing tobacco-related disparities and addressing the unique challenges of underserved maternal smokers.