Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2020 May 1.
Published in final edited form as: Drug Alcohol Depend. 2019 Mar 26;198:168–175. doi: 10.1016/j.drugalcdep.2019.02.014

The interaction of sociodemographic risk factors and measures of nicotine dependence in predicting maternal smoking during pregnancy *

Alexandra N Houston-Ludlam a,b, Kathleen K Bucholz c, Julia D Grant d, Mary Waldron e, Pamela A F Madden c, Andrew C Heath c
PMCID: PMC6467711  NIHMSID: NIHMS1525476  PMID: 30939374

Abstract

Background:

Understanding differences in nicotine dependence assessments’ ability to predict smoking cessation is complicated by variation in quit attempt contexts. Pregnancy reduces this variation, as each pregnant smoker receives the same strong cessation incentive. Cigarette smoking during pregnancy (SDP) provides a powerful paradigm for analyzing the interplay between nicotine dependence measures and sociodemographics in predicting cessation failure.

Methods:

Data from a female twin cohort (median birth year 1980), assessed in teens and early twenties, were merged with birth records to identify those with smoking history who experienced childbirth (N=1657 births, N=763 mothers). Logistic regression predicting SDP, as a function of birth record sociodemographic variables, generated a sociodemographic risk-score. Further analysis incorporated the risk-score with data from research interviews on DSM-IV-Nicotine Dependence symptom count, Heaviness of Smoking Index.

Results:

Low maternal educational level, younger age at childbirth, and being unmarried all contributed risk for SDP. In addition to sociodemographic risk-score, the best predictors of SDP included HSI-score (OR:1.51), their two-way interaction (OR:0.39; reduced impact of dependence at intermediate-high sociodemographic risk), history of ≥ two failed quit attempts (OR:1.37), and a dummy variable for prior pregnancy at time of assessment (OR:1.82). DSMIV-Nicotine Dependence symptoms underperformed the Heaviness of Smoking Index and did not improve prediction when added to the best model.

Conclusions:

The 2-item Heaviness of Smoking Index measure and report of ≥ two failed quit attempts performed best for predicting SDP. The contribution of either nicotine dependence measure to SDP risk was diminished at increased levels of sociodemographic risk.

Keywords: Classification, Tobacco Dependence, Prediction, Pregnant Smokers, Heaviness of Smoking Index, DSM Nicotine Dependence

1. Introduction

While risk and protective factors for onset of substance use and problem use are well-described, determinants of successful recovery and cessation are less clear, particularly given variability in the contexts of quit attempts (e.g., proximal stressors, major life-changing events, self-efficacy for cessation; (Bolman et al., 2018; Koslovsky et al., 2018; Volz et al., 2014). Utilizing outcomes of a large number of users under similar conditions enable inference about risk- and protective-mechanisms in addiction persistence and desistance. Cigarette smoking during pregnancy is a useful model for identifying predictors of cessation failure because many smokers experience a common, strong incentive for cessation: pregnancy, or the anticipation of pregnancy (Practice Committee of the American Society for Reproductive Medicine, 2018).

Smoking during pregnancy prevalence ranges from 8.4%−16.5%, depending whether it is reported from surveillance data or a U.S. nationally-representative sample (Brown et al., 2016; Curtin and Matthews, 2016). These rates exceed the Healthy People 2020 goal of reduction to 1.4% (U.S. Department of Health and Human Services). Smoking during pregnancy has been associated with newborn outcomes such as low birth weight and premature birth (Cnattingius, 2004; Dietz et al., 2010) and more distal outcomes such as chronic respiratory issues (McEvoy and Spindel, 2017; Tager et al., 1983), structural brain changes (El Marroun et al., 2016), and psychiatric phenotypes, especially hyperactive/impulsive symptoms (Knopik et al., 2016; for review see Scherman et al., 2018). Smoking during pregnancy provides a powerful model system for examining the interplay of biological factors, such as inherited differences in nicotine dependence vulnerability (for review see Sharp and Chen, 2019), and sociodemographic factors, (for review see Chamberlain et al., 2017; Riaz et al., 2018). Given low rates of smoking cessation in pregnancy and high rates of post-partum relapse (Cnattingius, 2004), identifying early predictors of future smoking during pregnancy risk will improve early screening and intervention.

A measure’s predictive utility for smoking during pregnancy is influenced by how nicotine dependence is assessed. The nicotine dependence construct aims to assess problems associated with nicotine use that have individual and societal impact (IARC Handbooks of Cancer Prevention, 2008; Piper et al., 2006). Nicotine dependence measures commonly assess themes of difficulty quitting, experiencing problems related to use, and heaviness of use. There are global assessments, such as the Fagerström Test for Nicotine Dependence (FTND; Heatherton et al., 1991) and the Diagnostic and Statistical Manual of Mental Disorders Nicotine Dependence criteria (DSM; APA, 1994), now assessed as Tobacco Use Disorder in the most recent DSM-5 (APA, 2013), and more focused measures for identifying and defining subgroups of smokers (Piper et al., 2008), including the Nicotine Dependence Syndrome Scale (Shiffman et al., 2004) and Wisconsin Inventory of Smoking Dependence Motives (Piper et al., 2004). The six-item FTND and related Heaviness of Smoking Index (HSI; Heatherton et al., 1989), containing two FTND items, measure physical nicotine dependence. Both have been shown to predict smoking cessation outcomes (Baker et al., 2012). The DSM-IV assessment of nicotine dependence uses a medical model framework to identify “a maladaptive pattern of nicotine use, leading to clinically significant impairment or distress” (APA, 2000) and has modest ability to predict smoking cessation outcomes, yet generally underperforms other measures (Baker et al., 2012).

Understanding the relationship between nicotine dependence, sociodemographic factors, and risk of smoking during pregnancy is arguably best achieved prospectively, since failure of smoking cessation during pregnancy may impact a participant’s evaluation of own nicotine dependence (e.g., greater likelihood of endorsement of nicotine dependence). In this report, we utilize data from a large female twin cohort, with many participants assessed prior to first pregnancy, in tandem with data from state birth records, to analyze the interplay of sociodemographic characteristics at the time of childbirth, and research assessments of nicotine dependence, in predicting smoking during pregnancy as recorded on the birth record.

2. Methods and Materials

2.1. Sample

Analyses used Missouri Adolescent Female Twin Study (MOAFTS) data, described previously (Heath et al., 2002). Briefly, a like-sex female twin pair cohort (monozygotic and dizygotic), born July 1st 1975-June 30th 1985, was ascertained from Missouri birth records. Parents (“Generation 1” – G1) and their twin daughters (“Generation 2” – G2) were recruited using a cohort-sequential design for a prospective study initially focused on the predictors and long-term outcomes of alcohol misuse in adolescent girls. Data for this report comes from the first major follow-up telephone diagnostic interview of the original target sample of twins, after G2 age 18 (G2 median age 22, range 18–28, N=3787 twins, 81.7% of the total cohort identified from birth records, 14% African American (Sartor et al., 2013). Informed consent was obtained. Study procedures were approved by the Washington University Institutional Review Board and comply with the Declaration of Helsinki. State birth record data from 1993–2016 were used to identify children (“Generation 3”- G3) that were born to MOAFTS twins (G2). The final sample included G2 participants who were regular smokers (defined in 2.2.1), had given birth and participated in the follow-up interview: a total of 1,657 births to 763 White participants and 177 births to 65 African American participants were identified. Rates of smoking during pregnancy in African American mothers are lower (Mumford et al., 2014), and smoking trajectories for smoking during pregnancy differ between White and African American mothers (De Genna et al., 2017). Due to these differences, we had insufficient power to conduct separate analyses in African American mothers. We find it inappropriate to collapse these samples by race, given the limited sociodemographic overlap of White and African American twins, which would lead to predictions for cells with no data (Imbens and Rubin, 2015). Thus, we present analyses only for White mothers.

2.2. Measures

Follow-up interview assessments of the G2 twins were based on the Semi-Structured Assessment of the Genetics of Alcoholism (Bucholz et al., 1994) adapted for telephone interview. This measure is an extensively-validated semi-structured interview initially developed for the Collaborative Study on the Genetics of Alcoholism (Begleiter et al., 1995; Bucholz et al., 1994; Hesselbrock et al., 1999). The smoking section of the interview is based on items from the Composite International Diagnostic Interview (WHO, 1997).

2.2.1. Nicotine Dependence Measures from the Research Data.

We adopted the commonly used definition of regular smoking recommended by the CDC: ever smoking ≥ 100 lifetime cigarettes. This definition includes ‘current’ and ‘former’ smokers at time of research interview, who is the pool of people at risk for smoking during pregnancy. The tobacco use module included assessment of the Heaviness of Smoking Index (HSI) and criteria for DSM-IV Nicotine dependence diagnosis. The HSI (Heatherton et al., 1991; Heatherton et al., 1989) is a widely-used two-item measure to assess nicotine dependence containing two-items: Time-to-First-Cigarette and Cigarettes-(smoked)-Per-Day. The variables included in the present analyses include: 1) Total HSI score, as a sum of the points for each item ranging from 0–7, with higher numbers indicating more severe nicotine dependence, coded into a six-level ordinal variable corresponding to HSI scores of: 0, 1, 2, 3, 4–5, 6–7; 2) individual Time-to-First-Cigarette item and 3) individual Cigarettes-Per-Day item, coded as above.

DSM-IV-Nicotine Dependence was assessed via structured questions to capture the following items: tolerance, great deal of time using, important activities given up, smoking more than intended, difficulty quitting (experienced ≥ 2 lifetime failed quit attempts) or persistent desire to quit, withdrawal syndrome (experienced ≥ 4 symptoms within 24 hours of last cigarette use), and use despite experiencing physical or emotional health problems related to smoking. The DSM-based variables used in the present analyses include: 1) A symptom count variable characterized as a six-level ordinal variable corresponding to 0, 1, 2, 3, 4, ≥ 5 symptoms; ND-sx, 2) a dichotomous variable indicating a DSM-IV-Nicotine Dependence diagnosis, by endorsing three or more symptoms; ND-dx and 3) individual symptom level data. To reflect the DSM criterion, the two or more failed quit attempts item was combined with persistent desire to quit for the total DSM symptom score, but these two symptoms were analyzed separately for the individual symptom level analyses. For more detail on coding of Nicotine Dependence measures, see Supplemental Methods1.

2.2.2. Pregnancy Data.

Given that the experience of difficulty quitting during pregnancy could influence a participant’s endorsement of nicotine dependence symptoms, we compared prospective with retrospective prediction. Prospective prediction occurred when nicotine dependence items were assessed prior to the participant ever experiencing pregnancy. Retrospective prediction occurred when nicotine dependence assessment occurred after a participant experienced their first pregnancy. There was 98.6% concordance between self-report pregnancy history and state birth record data on childbirth. In cases of discordance, birth record data superseded self-report.

2.2.3. Birth Record Data.

Missouri birth records from 1993—2016 including identifiers were obtained with permission from the state Department of Health and Senior Services and linked at the individual level for all in-state births of G3 children to G2 MOAFTS participants. Sociodemographic variables from the birth record were recoded when necessary for consistency across years and used to create the sociodemographic risk score (SRS), see 2.3.1. Variables used for the SRS: maternal age, maternal highest educational attainment, marital status, parity (yes/no indicating first born child, due to lower rates of smoking during pregnancy for first-pregnancy compared with subsequent-pregnancy; Meernik and Goldstein, 2015), and a dummy variable to represent a coding change initiated by the CDC, implemented in 2010 (2010-onwards: smoking during pregnancy by trimester; 1993–2009: dichotomous smoking at any time during pregnancy). The primary outcome is thus smoking during pregnancy as recorded on the birth record, with cases for 2010-onwards reflecting smoking during pregnancy beyond the first trimester. For more details about birth record variables and our coding of these variables, see Supplementary Methods1.

2.3. Data Analysis

Descriptive statistics were calculated with SAS Version 9.2 (SAS Institute Inc., Cary, NC). Descriptive statistics for sample characteristics were calculated based on smoking status at time of research interview: never smokers (reported never smoking or smoking < 100 lifetime cigarettes), regular smokers (≥ 100 lifetime cigarettes) with no reports of smoking during pregnancy on the birth record, and regular smokers with ≥1 report of smoking during pregnancy on the birth record. Descriptive statistics for smoking variables were calculated on both groups of smokers. Generalizability of findings from twin cohort research data was assessed by comparing MOAFTS participants versus other non-twin mothers, who were born in state during the same birth years as the MOAFTS participants (1975–1985), on key sociodemographic variables.

2.3.1. Generation of Sociodemographic Risk Score (SRS).

Logistic regression analyses predicting smoking during pregnancy from the birth record as the primary outcome were calculated using STATA Version 15.1 (StataCorp, College Station, TX). All births were included; thus, the same mother could be represented on multiple records. To generate the SRS, predicted probabilities of smoking during pregnancy were estimated using logistic regression with the birth record variables described in 2.2.3., adjusted for clustered samples using the STATA robust cluster command. SRS was then sextiled for use in subsequent analyses to protect confidentiality, a requirement for working with state birth record data.

2.3.2. Analyses using SRS, HSI, DSM-ND.

To identify which smoking variables had best utility for predicting smoking during pregnancy, individual adjusted logistic regressions were performed predicting smoking during pregnancy from smoking characteristic separately: HSI score, HSI individual items; DSM-based ND-sx (symptom count), ND-dx (endorsement of ≥ 3 symptoms), and individual DSM-IV-Nicotine Dependence items. The HSI analyses were adjusted for SRS. The DSM-IV-Nicotine Dependence analyses were adjusted for SRS and individual HSI items to evaluate their ability to improve prediction of smoking during pregnancy beyond HSI and sociodemographic risk. Additional logistic regression models, stratified by prospective (no pregnancy by research interview) versus retrospective (experienced at least one pregnancy by time of research interview) assessment of nicotine dependence relative to first pregnancy, investigated the interaction between SRS and smoking characteristic (HSI, ND-sx) on predicting smoking during pregnancy risk.

3. Results

3.1. Sample Characteristics

Sociodemographic characteristics of never-smokers, regular smokers with no report of smoking during pregnancy, and regular smokers with report of smoking during at least one pregnancy (“pregnancy smokers”) are summarized in Table 1. These show the expected differences: Compared to never smokers, both groups of smokers were younger at childbirth, had completed fewer years of education, and were more likely to be unmarried, with these differences more pronounced for pregnancy smokers. For both groups of regular smokers, the endorsement rate for nicotine dependence symptoms at interview is shown in Table 2. Across all variables except difficulty quitting/persistent desire to quit, higher endorsement rates were seen among pregnancy smokers.

Table 1.

Maternal sociodemographic characteristics at childbirth recorded on birth record

Smoking Status at Follow Up Interview Never Smokers Regular Smokers, Never SDP Regular Smokers, SDP for ≥ 1 birtha
 Maternal Age N % N % N %
  Under 18 28 1.41 25 3.26 61 6.01
  18–19 90 4.54 44 5.74 111 12.35
  20–23 291 14.69 177 23.11 276 30.70
  24–26 426 21.50 130 16.97 172 19.13
  27–30 (reference) 667 33.67 209 27.28 173 19.24
  31 or greater 479 24.18 181 23.63 106 11.79
 Maternal Marital Status
  Unmarried, Partner Unnamed 107 5.40 59 7.70 185 20.58
  Unmarried, Partner Named 281 14.18 173 22.58 342 38.04
  Married, Partner Named (reference)b 1590 80.26 532 69.45 349 38.82
  Married, Partner Unnamedb 2 0.15 2 0.26 23 2.56
 Years Education Completed
  Less than 8 years 8 0.40 12 1.57 42 4.75
  9–12 years (reference) 495 25.03 313 40.86 617 69.80
  13–15 years 501 25.33 210 27.42 193 21.83
  16 years 651 32.91 143 18.67 23 2.60
  17 years 323 16.33 88 11.49 9 1.02
Open in a new tab
a

“Regular Smoker” status (100+ cigarettes lifetime) determined at time of research follow-up interview, median age 22. SDP status determined across all birth records to a given mother ascertained from birth record data.

b

Since under state law a woman married at any stage during pregnancy is recorded as married, this group could include women who had already divorced their partner.

Sample size, N, refers to number of births, since maternal characteristics could change between births.

Table 2.

Maternal smoking characteristics at research interview.

Smoking Status at Follow Up Interview Regular Smokers, Never SDPa Regular Smokers, SDP for > 1 birtha
 Heaviness of Smoking Index (HSI) Score N = 391 % N = 369 %
   0 86 21.99 11 2.98
   1 71 18.16 32 8.67
   2 65 16.62 39 10.57
   3 53 13.55 46 12.47
   4 37 9.46 54 14.63
   5 33 8.44 58 15.72
   6 32 8.18 48 13.01
   7 14 3.58 81 21.95
 HSI: Cigarettes Per Day N = 391 % N = 369 %
   1–5 108 27.62 19 5.15
   6–10 109 27.88 85 23.04
   10–15 70 17.90 82 22.22
   16–19 48 12.28 51 13.82
   20+ 56 14.32 132 35.77
 HSI: Time to First Cigarette N = 394 % N = 369 %
   >60 minutes A 203 51.52 69 18.70
   31–60 minutes 70 17.77 56 15.18
   6–30 minutes 87 22.08 109 29.54
   ≤ 5 minutes 34 8.63 135 36.59
 DSM IV Nicotine Dependence N = 394 % N = 369 %
  Symptom Count
   0 38 9.64 21 5.69
   1 95 24.11 41 11.11
   2 94 23.86 87 23.58
   3 62 15.74 66 17.89
   4 61 15.48 65 17.62
   ≥ 5 44 11.17 89 24.12
  Diagnosis (≥ 3 Symptoms) 158 40.10 207 56.10
 DSM IV Nicotine Dependence Symptoms N = 391 % N = 369 %
   Tolerance 150 38.07 224 60.70
   Great Deal of Time Using 91 23.10 105 28.46
   Important Activities Given Up 19 4.82 32 8.67
   Smoke More Than Intended 168 42.64 211 57.18
   Difficulty Quitting/Persistent Desire To Quit 321 81.47 304 82.38
    ≥ Two Failed Quit Attempts 171 43.40 222 60.16
   Withdrawal Syndrome (≥ 4 Symptoms) 134 34.01 166 44.99
   Use Despite Physical/Emotional Problems 67 17.01 14 30.89
Open in a new tab
a

“Regular Smoker” status (100+ cigarettes lifetime) determined at time of research follow-up interview, median age 22. SDP status determined across all birth records to a given mother ascertained from birth record data.

Sample size N refers to number of mothers with history of regular smoking who have given birth in state.

Supplementary Table 11 presents additional birth record sample characteristics for mother and child at time of childbirth: Pregnancy smokers were, on average, younger at first childbirth. Sociodemographic variables at childbirth between MOAFTS participants and all other mothers, themselves born in state 1975–1985, show minimal differences (Spearman rhos <|0.02|; Supplementary Table 2)2. For participants who experienced first pregnancy after research interview (prospective nicotine dependence assessment), children were born 7.0 years after research interview (N=593 births). For participants who had at least one pregnancy by time of research interview (retrospective nicotine dependence assessment), children were born on average 2.9 years before research interview (N=577 births) and 5.3 years after research interview (N=490 births).

3.2. Sociodemographic Predictors of Smoking During Pregnancy Risk

The results of fitting a multiple logistic regression predicting smoking during pregnancy are summarized in Table 3. Smoking during pregnancy risk is reduced in those younger than age 18 at childbirth and those who completed more years of education, with the protective effect of education particularly strong. Compared to the married, partner named reference group, smoking during pregnancy risk is elevated both by being unmarried during pregnancy and having an unnamed reproductive partner. The control variable correcting for the effects of the CDC coding change was non-significant. In a separate analysis including a dummy variable for whether first childbirth occurred before or after research interview, an increased risk in those interviewed after their first birth was noted (OR:1.82, 95% CI: [1.33, 2.49]). When comparing rates of smoking during pregnancy by birth year between the retrospective and prospective analyses, we observe higher rates for each birth year in those who had first pregnancy by research assessment (unpublished data, available upon request).

Table 3.

Prediction of smoking during pregnancy using sociodemographic variables at time of childbirth analyzed with logistic regression

 Maternal Age (years) Odds Ratio 95% CI
  Under 18 0.46 0.23 0.90
  18–19 0.80 0.49 1.30
  20–23 0.82 0.56 1.22
  24–26 0.94 0.66 1.34
  27–30 (reference) - - -
  31 or greater 0.85 0.56 1.28
 Maternal Marital Status
  Married, Partner Named (reference) - - -
  Married, Partner Unnamed 6.74 2.80 16.25
  Unmarried, Partner Named 2.14 1.58 2.90
  Unmarried, Partner Unnamed 2.26 1.56 3.27
 Maternal Educational Attainment (years)
  Less than 8 1.74 0.66 4.56
  9–12 (reference) - - -
  13–15 0.58 0.42 0.79
  16 0.15 0.08 0.27
  17 0.11 0.04 0.26
 CDC Coding Change (dummy variable) 1.09 0.71 1.68
 First Born Child (yes/no) 0.79 0.62 1.00
Open in a new tab

3.3. Joint Effects of Sociodemographic Risk and Individual Nicotine Dependence Measures

Spearman correlations among the nicotine dependence variables were low to modest, ranging from 0.05 to 0.34 (Supplementary Table 3)1. Table 4 summarizes the association between HSI measures obtained at research interview and smoking during pregnancy, controlling for sociodemographic variables (Model 1a: HSI-total score; Model 1b: HSI time-to-first-cigarette; Model 1c: HSI cigarettes-per-day); and then between DSM-IV-Nicotine Dependence measures and smoking during pregnancy, controlling for both SRS and HSI items (Model 2a: ND-sx count; Model 2b: ND-dx; Model 3, individual nicotine dependence symptoms, entered separately). The HSI items improve prediction of smoking during pregnancy risk, but the DSM symptom count (OR:1.01, 95% CI [0.92–1.13]) and binary diagnostic measures (OR:1.00, 95% CI [0.74–1.35]) show a striking lack of further improvement in prediction. Only a single DSM-based measure, reporting two or more unsuccessful efforts to quit smoking, improves prediction.2 Spearman correlation between HSI items was moderate (0.54), while those between HSI items and the ≥2 failed quit attempts item were low: time-to-first-cigarette and ≥2 failed quit attempts (0.21) and cigarettes-per-day and ≥2 failed quit attempts (0.15).

Table 4.

Comparison of the ability of smoking characteristics from research data to predict smoking during pregnancy beyond sociodemographic variables at time of childbirth analyzed with logistic regression.

Model Odds Ratio 95% CI
1  Heaviness of Smoking Indexa
1a  Heaviness of Smoking Index Score 1.51 1.36 1.66
1b  Time To First Cigarette
  ≥ 60 minutes (reference) - - -
  31–60 minutes 2.39 1.51 3.78
  6–30 minutes 2.72 1.84 4.03
  ≤ 5 minutes 6.05 3.87 9.46
1c  Cigarettes Per Day
  1–5 (reference) - - -
  6–10 3.27 1.83 5.85
  11–15 4.18 2.31 7.57
  16–20 4.83 2.51 9.30
  ≥ 21 7.73 4.29 13.92
2  DSM IV Nicotine Dependenceb
2a   Symptom Count 1.01 0.92 1.13
2b   Diagnosis (≥ 3 Symptoms) 1.00 0.74 1.35
3  DSM IV Nicotine Dependence Symptoms
  Tolerance 1.01 0.72 1.43
  Great Deal of Time Using 0.86 0.61 1.23
  Important Activities Given Up 0.92 0.54 1.54
  Smoke More Than Intended 1.36 1.00 1.84
  Difficulty Quitting/Persistent Desire To Quit 1.04 0.72 1.50
  ≥ Two Failed Quit Attempts 1.37 1.02 1.86
   Withdrawal Syndrome (≥ 4 Symptoms) 0.80 0.59 1.11
  Use Despite Physical/Emotional Problems 1.03 0.75 1.42
Open in a new tab
a

Heaviness of Smoking Index analyses were run as three separate logistic regression models adjusted for sociodemographic risk score at time of childbirth ascertained from birth record data.

b

DSM-IV-Nicotine Dependence analyses were run as separate logistic regression models per variable listed, adjusted for sociodemographic risk score and the individual Heaviness of Smoking Index Items (Time to First Cigarette and Cigarettes Per Day).

Figure 1 summarizes predicted probability of smoking during pregnancy as a joint function of SRS (continuous) and nicotine dependence score (HSI, panels a, c; ND-sx, panels b, d), stratified by prospective (panels a, b) versus retrospective (panels c, d) prediction. The interaction between HSI and SRS was statistically significant for prospective prediction (HSI-prospective: OR: 0.39, 95% CI [0.170, 0.876]; HSI-retrospective: OR: 0.54, 95% CI [0.256, 1.15]). The interaction between ND-sx count and SRS was trend-level significant (ND-sx prospective: OR:0.40, 95% CI [0.168, 0.961], ND-sx retrospective: OR:0.64, 95% CI [0.323, 1.28]). Four major conclusions may be drawn from the figure. First, confirmation of HSI-score as a superior predictor of smoking during pregnancy risk, shown by greater separation of the smoking during pregnancy predicted probability curves as a function of level of nicotine dependence at every level of sociodemographic risk. Second, the probability curves show greatly diminished separation at high levels of sociodemographic risk, indicating that at high levels of sociodemographic risk, degree of nicotine dependence at research interview is a much weaker predictor of outcome. Third, there is broad consistency of findings regardless of whether reporting of nicotine dependence occurred after or before first childbirth (i.e., retrospectively or prospectively). Lastly, the predictive utility of nicotine dependence shows a trend towards strengthened prospective versus retrospective prediction.

Figure 1.

Figure 1.

Open in a new tab

Predicted probability of smoking during pregnancy as a function of sociodemographic risk score at time of pregnancy and of nicotine dependence measured either prior to or after first pregnancy.

Interaction between Nicotine Dependence measure (Heaviness of Smoking Index score or DSMIV-Nicotine Dependence symptom count), assessed during late teens/early twenties from research data and sociodemographic risk score, estimated using birth record data at time of childbirth, in predicting SDP. Analyses are separated by whether assessment of Nicotine Dependence occurred before or after first pregnancy. In panels A) and C), the interaction between HSI score and sociodemographic risk score (continuous) is presented, separated by timing of first pregnancy relative to follow-up interview. In panels B) and D), the interaction between sextiled DSM ND-sx count and sociodemographic risk score (continuous) is presented, separated by timing of first pregnancy relative to follow-up interview. As sociodemographic risk increases, the importance of HSI as a predictor of future SDP is diminished. The strength of this relationship is weakened for ND-sx relative to HSI. Smoking assessment prior to first pregnancy: N = 589 births; smoking assessment after first pregnancy: N = 1048 births. (ND-sx: DSM-IV Nicotine Dependence Symptom Count; HSI: Heaviness of Smoking Index).

4. Discussion

4.1. Conclusions

Characterizing influential sociodemographic, individual, and environmental risk-factors for failure of smoking cessation during pregnancy is important for improving intervention efforts. We used nicotine dependence data from research cohort of female twins, combined with the birth records of children born to these twins, to investigate the predictive utility for smoking during pregnancy by proximal sociodemographic variables recorded on the birth record, nicotine dependence measured in a research setting, and their interaction. Consistent with previous research, we found that birth record data on education, marital status, age, and parity, were significant predictors of smoking during pregnancy. The best predictors of smoking during pregnancy were SRS, HSI, their two-way interaction, and a single item assessing lifetime report of failing two or more smoking cessation attempts. The DSM-IV symptom count, DSM-IV-Nicotine Dependence diagnosis, and individual DSM-IV-Nicotine Dependence items all failed to improve prediction of smoking during pregnancy when accounting for SRS and HSI items. Our ability to prospectively predict smoking during pregnancy risk from smoking characteristics was stronger than retrospective prediction.

4.2. Assessment of Nicotine Dependence in Pregnancy

While both the DSM-IV-Nicotine Dependence and HSI aim to measure physiological nicotine dependence, the HSI is a shorter measure, requiring minimal resources to administer, containing the majority of the information important for predicting smoking during pregnancy. Our work is consistent with a previous report showing that prediction of late-pregnancy smoking status was best achieved by using both HSI items and failed pre-pregnancy quit attempts were associated with higher risk of late-pregnancy continued smoking (Kurti et al., 2016). This conflicts with reports from the general population, where the time-to-first-cigarette item is more important for predicting smoking cessation (Baker et al., 2007). This present report highlights aspects of smoking behavior associated with smoking during pregnancy risk, setting the stage for improved assessment of risk and identifying at-risk individuals to target for increased cessation support. We identified three aspects of smoking behavior most important for assessing smoking during pregnancy risk—how soon one smokes their first cigarette after waking, number of cigarettes smoked per day, and history of cessation attempts/number of failed attempts. Using this information about an individual’s smoking behavior in the context of their sociodemographic status—age, years of education completed, marital status/reproductive partner support, and history of prior pregnancy and childbirth, yields important information for providers in determining smoking during pregnancy risk and indicates need for smoking cessation intervention. Importantly, given our observed strengthened prospective prediction, our results support the importance of beginning screening years in advance of first pregnancy, thus maximizing opportunities to achieve successful smoking cessation prior to pregnancy.

The predictive utility of nicotine dependence measures was substantially diminished in individuals with intermediate-high sociodemographic risk, and this result was consistent in both prospective and retrospective analyses. Since more severe dependence is associated with greater difficulty quitting, including in pregnancy (Riaz et al., 2018), this implies an important health disparities challenge: the failure to accomplish cessation for sociodemographically disadvantaged pregnant individuals, despite lower levels of nicotine dependence, which should otherwise facilitate successful quitting. Our results suggest that nicotine dependence is not the sole indicator of smoking during pregnancy risk, and for some patients, sociodemographic risk proximal to time of childbirth may explain more risk for smoking during pregnancy. In their review, Boucher and Konkle (2016) emphasize the role of socioeconomic status, nicotine dependence, social support, culture, mental health, and access to health services as important contributors to smoking during pregnancy risk and emphasize pregnancy as an important point of contact for smoking cessation services to be delivered to patients. Less than half of patients report that their healthcare providers discourage their continued smoking during pregnancy (Hoekzema et al., 2014), despite evidence that two or more healthcare providers asking about smoking status increases odds of a past-year quit attempt and increased intention to quit in the next six months in the general population (An et al., 2008). Our results corroborate the need to address factors other than nicotine dependence as an important focus of patient-centered care and intervention efforts. In a recent analysis of the same data, we demonstrated unique risk of smoking during pregnancy associated with relationship/marital status at time of birth (Waldron et al., 2017).

4.3. Combining Research and Administrative Data in Health Research

The analytic framework presented in this report is applicable to a broad array of health problems, where research data and administrative (e.g., birth record, driver’s license) datasets can be combined to yield important insights not achieved in analyses considering only one data type. Our analyses of phenotype-environment interaction, predicting smoking during pregnancy as a joint function of nicotine dependence and environmental risk-factors at time of pregnancy, illustrate a framework that can be extended to other questions of interaction between individual and environmental risk: identify key environmental domains to test for interaction effects on cessation failure in addiction; on continued weight gain or failure of weight-loss efforts in obesity; or on relapse/recurrence risk across a broader range of psychiatric phenotypes. Such an approach can then be used to guide genotype-interaction analyses in genetic association data, or twin/family study data, supplementing preexisting research data with administrative data at modest cost.

4.4. Limitations

First, our analyses include only White non-Hispanic mothers due to limited numbers of African American twin pairs. It is important that future research identifies both risk and protective factors for groups experiencing higher rates of smoking during pregnancy and associated risk factors (e.g., Native Americans; CDC, 2018; Curtin and Matthews, 2016). Second, while aggregate birth record data are useful for understanding predictors of smoking during pregnancy, individual cases of both false positive and, more frequently, false negative reports can be identified. While some under-reporting of smoking during pregnancy is anticipated, this does not appear to be a major problem in our data. Some age × education combinations report rates as high as 60% (e.g., White non-Hispanic mothers, less than high school education, giving birth in their early 20s; data available on request), compared to the 8.4% national average. Moreover, underestimation of smoking during pregnancy rates does not change observed effect estimates between smoking during pregnancy and offspring outcomes (Bakker et al., 2011). Third, while identifying individual factors associated with smoking during pregnancy risk is important for intervention efforts, this risk undoubtedly reflects an interplay of sociodemographic risks, history of exposure to neighborhoods with high rates of smoking during pregnancy, and other individual (including genetic, psychiatric history) and family-of-origin effects which may be only partially mediated through nicotine dependence. The present report does not address important aspects of social influence, either by reproductive partner, cohabitants during pregnancy, or neighborhood rates of smoking during pregnancy. We include reproductive partner influence obliquely, by analyzing the effect of marital status, but future research would benefit from more thorough analysis of these influences and their effect on achieving successful smoking cessation by time of pregnancy. Fourth, there is relatively high rates of co-use of tobacco and other drugs, especially cannabis and alcohol, during pregnancy (Dukes et al., 2017; Oga et al., 2018). Future research should examine the roles of substance use and dependence, including, alcohol, marijuana, illicit substances, and electronic nicotine delivery systems, on risk for cigarette use during pregnancy. This would further inform how healthcare providers should use assessment of alcohol and other substance use during pregnancy to tailor risk assessment and treatment. Finally, lower levels of nicotine dependence may become more severe over time. Future work could address this by examining changes in sociodemographic risk and nicotine dependence between pregnancies and evaluate the roles of sociodemographic risk and nicotine dependence in sister-sister discordance for smoking during pregnancy.

Supplementary Material

1

Highlights.

  • Heaviness of Smoking Index outperforms DSM in predicting smoking during pregnancy

  • Sociodemographics and nicotine dependence predict SDP risk in nonlinear interaction

  • Nicotine dependence is weak SDP predictor at high sociodemographic risk levels

  • Research and administrative data can be combined to predict long-term outcome

Acknowledgements

The authors thank the subjects whose participation made this work possible. The authors are grateful to Stacey Marion and Denise Schmitz for management and collection of the research data and Dejan Jovanovic and Radivoje Todorovic for their technical assistance.

Role of Funding Source

This work was supported by the National Institutes of Health (ANHL, Award Number TL1TR002344; ACH, Award Number AA017688; ACH, PAFM, Award Number AA021492; ACH, KKB, PAFM, Award Number AA023487). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. The NIH had no further role in study design, analysis, interpretation, writing of the report, or decision to submit the report for publication.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

*

Supplementary material can be found by accessing the online version of this paper at http://dx.doi.org and by entering doi.

Conflict of Interest

No conflict declared.

1

Supplementary material can be found by accessing the online version of this paper at http://dx.doi.org and by entering doi:…

2

Given that past experience of failed quit attempts during pregnancy might influence responding to this item, we ran an additional model predicting smoking during pregnancy including a dichotomous variable indicating whether a participant experienced first pregnancy by time of follow-up interview, the symptom of experiencing two or more failed quit attempts, and their two-way interaction, adjusted for SRS and two HSI items, and observed a significant interaction term (OR: 1.97, 95% CI: [1.07, 3.60]), confirming our concern.

References

  1. American Psychiatric Association (APA), 1994. Diagnostic and Statistical Manual of Mental Disorders (fourth edition). Washington, DC. [Google Scholar]
  2. American Psychiatric Association (APA), 2000. Diagnostic and Statistical Manual of Mental Disorders (fourth edition, text revision). Washington, DC. [Google Scholar]
  3. American Psychiatric Association (APA), 2013. Diagnostic and Statistical Manual of Mental Disorders (fifth edition), Washington, DC. [Google Scholar]
  4. An LC, Foldes SS, Alesci NL, Bluhm JH, Bland PC, Davern ME, Schillo BA, Ahluwalia JS, Manley MW, 2008. The impact of smoking-cessation intervention by multiple health professionals. Am. J. Prev. Med 34, 54–60. 10.1016/j.amepre.2007.09.019 [DOI] [PubMed] [Google Scholar]
  5. Baker TB, Breslau N, Covey L, Shiffman S, 2012. DSM criteria for tobacco use disorder and tobacco withdrawal: a critique and proposed revisions for DSM-5. Addiction. 107, 263–275. 10.1111/j.1360-0443.2011.03657.x [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Baker TB, Piper ME, McCarthy DE, Bolt DM, Smith SS, Kim S-Y, Colby S, Conti D, Giovino GA, Hatsukami D, Hyland A, Krishnan-Sarin S, Niaura R, Perkins KA, Toll BA, Transdisciplinary Tobacco Use Research Center (TTURC) Tobacco Dependence Phenotype Workgroup, 2007. Time to first cigarette in the morning as an index of ability to quit smoking: Implications for nicotine dependence. Nicotine Tob. Res 9, S555–S570. 10.1080/14622200701673480 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Bakker R, Kruithof C, Steegers EA, Tiemeier H, Mackenbach JP, Hofman A, Jaddoe VW, 2011. Assessment of maternal smoking status during pregnancy and the associations with neonatal outcomes. Nicotine Tob. Res 13, 1250–1256. 10.1093/ntr/ntr117 [DOI] [PubMed] [Google Scholar]
  8. Begleiter H, Reich T, Hesselbrock V, Porjesz B, Begleiter H, Li TK, Schuckit MA, Edenberg HJ, Rice JP, 1995. The collaborative study on the genetics of alcoholism. Alcohol Health Res. World 19, 228–236. [PMC free article] [PubMed] [Google Scholar]
  9. Bolman C, Verboon P, Thewissen V, Boonen V, Soons K, Jacobs N, 2018. Predicting smoking lapses in the first week of quitting: an ecological momentary assessment study. J. Addict. Med 12, 65–71. 10.1097/adm.0000000000000365 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Boucher J, Konkle AT, 2016. Understanding inequalities of maternal smoking--bridging the gap with adapted intervention strategies. Int. J. Environ. Res. Public Health 13, 282 10.3390/ijerph13030282 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Brown QL, Hasin DS, Keyes KM, Fink DS, Ravenell O, Martins SS, 2016. Health insurance, alcohol and tobacco use among pregnant and non-pregnant women of reproductive age. Drug Alcohol Depend. 166, 116–124. 10.1016/j.drugalcdep.2016.07.001 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Bucholz KK, Cadoret R, Cloninger CR, Dinwiddie SH, Hesselbrock VM, Nurnberger JI Jr., Reich T, Schmidt I, Schuckit MA, 1994. A new, semi-structured psychiatric interview for use in genetic linkage studies: a report on the reliability of the SSAGA. J. Stud. Alcohol 55, 149–158. [DOI] [PubMed] [Google Scholar]
  13. Centers for Disease Control and Prevention (CDC) Office on Smoking and Health, American Indians/Alaska Natives and Tobacco Use. cdc.gov/tobacco/disparities/american-indians. (Accessed on 03/26/2018) [Google Scholar]
  14. Chamberlain C, O’Mara-Eves A, Porter J, Coleman T, Perlen SM, Thomas J, McKenzie JE, 2017. Psychosocial interventions for supporting women to stop smoking in pregnancy. Cochrane Database Syst. Rev 2, CD001055 10.1002/14651858.CD001055.pub5 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Cnattingius S, 2004. The epidemiology of smoking during pregnancy: smoking prevalence, maternal characteristics, and pregnancy outcomes. Nicotine Tob. Res 6, S125–S140. 10.1080/14622200410001669187 [DOI] [PubMed] [Google Scholar]
  16. Curtin SC, Matthews TJ, 2016. Smoking prevalence and cessation before and during pregnancy: data from the birth certificate, 2014. Natl. Vital. Stat. Rep 65, 1–14. [PubMed] [Google Scholar]
  17. De Genna NM, Goldschmidt L, Day NL, Cornelius MD, 2017. Maternal trajectories of cigarette use as a function of maternal age and race. Addict. Behav 65, 33–39. 10.1016/j.addbeh.2016.09.011 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Dietz PM, England LJ, Shapiro-Mendoza CK, Tong VT, Farr SL, Callaghan WM, 2010. Infant morbidity and mortality attributable to prenatal smoking in the U.S. Am. J. Prev. Med 39, 45–52. 10.1016/j.amepre.2010.03.009 [DOI] [PubMed] [Google Scholar]
  19. Dukes K, Tripp T, Willinger M, Odendaal H, Elliott AJ, Kinney HC, Robinson F, Petersen JM, Raffo C, Hereld D, Groenewald C, Angal J, Hankins G, Burd L, Fifer WP, Myers MM, Hoffman HJ, Sullivan L, 2017. Drinking and smoking patterns during pregnancy: development of group-based trajectories in the safe passage study. Alcohol, 62, 49–60. 10.1016/j.alcohol.2017.03.001 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. El Marroun H, Tiemeier H, Franken IHA, Jaddoe VW, van der Lugt A, Verhulst FC, Lahey BB, White T, 2016. Prenatal cannabis and tobacco exposure in relation to brain morphology: a prospective neuroimaging study in young children. Biol. Psychiatry 79, 971–9. 10.1016/j.biopsych.2015.08.024 [DOI] [PubMed] [Google Scholar]
  21. Heath AC, Martin NG, Lynskey MT, Todorov AA, Madden PA, 2002. Estimating two-stage models for genetic influences on alcohol, tobacco or drug use initiation and dependence vulnerability in twin and family data. Twin Res. 5, 113–124. 10.1375/1369052022983 [DOI] [PubMed] [Google Scholar]
  22. Heatherton TF, Kozlowski LT, Frecker RC, Fagerstrom KO, 1991. The fagerström test for nicotine dependence: a revision of the fagerström tolerance questionnaire. Br. J. Addict 86, 1119–1127. [DOI] [PubMed] [Google Scholar]
  23. Heatherton TF, Kozlowski LT, Frecker RC, Rickert W, Robinson J, 1989. Measuring the heaviness of smoking: using self-reported time to the first cigarette of the day and number of cigarettes smoked per day. Br. J. Addict 84, 791–799. [DOI] [PubMed] [Google Scholar]
  24. Hesselbrock M, Easton C, Bucholz KK, Schuckit M, Hesselbrock V, 1999. A validity study of the SSAGA--a comparison with the SCAN. Addiction 94, 1361–1370. [DOI] [PubMed] [Google Scholar]
  25. Hoekzema L, Werumeus Buning A, Bonevski B, Wolke L, Wong S, Drinkwater P, Stewart K, George J, 2014. Smoking rates and smoking cessation preferences of pregnant women attending antenatal clinics of two large australian maternity hospitals. Aust. N. Z. J. Obstet. Gynaecol 54, 53–58. 10.1111/ajo.12148 [DOI] [PubMed] [Google Scholar]
  26. IARC Handbooks of Cancer Prevention, 2008. Tobacco Control, Vol. 12: Methods for Evaluating Tobacco Control Policies, Lyon, France. [Google Scholar]
  27. Imbens G, Rubin D, 2015. Causal Inference for Statistics, Social, and Biomedical Sciences: An Introduction. Cambridge University Press, New York. [Google Scholar]
  28. Knopik VS, Marceau K, Bidwell LC, Palmer RH, Smith TF, Todorov A, Evans AS, Heath AC, 2016. Smoking during pregnancy and ADHD risk: A genetically informed, multiple-rater approach. Am. J. Med. Genet. B Neuropsychiatr. Genet 171, 971–981. 10.1002/ajmg.b.32421 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Koslovsky MD, Hébert ET, Swartz MD, Chan W, Leon-Novelo L, Wilkinson AV, Kendzor DE, Businelle MS, 2018. The time-varying relations between risk factors and smoking before and after a quit attempt. Nicotine Tob. Res 20, 1231–1236. 10.1093/ntr/ntx225 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Kurti AN, Davis DR, Skelly JM, Redner R, Higgins ST, 2016. Comparison of nicotine dependence indicators in predicting quitting among pregnant smokers. Exp. Clin. Psychopharmacol 24, 12–17. 10.1037/pha0000056 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. McEvoy CT, Spindel ER, 2017. Pulmonary effects of maternal smoking on the fetus and child: effects on lung development, respiratory morbidities, and life long lung health. Paediatr. Respir. Rev 21, 27–33. 10.1016/j.prrv.2016.08.005 [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Meernik C, Goldstein AO, 2015. A critical review of smoking, cessation, relapse and emerging research in pregnancy and post-partum. Br. Med. Bull 114, 135–146. 10.1093/bmb/ldv016 [DOI] [PubMed] [Google Scholar]
  33. Mumford EA, Hair EC, Yu T-C, Liu W, 2014. Women’s longitudinal smoking patterns from preconception through child’s kindergarten entry: profiles of biological mothers of a 2001 US birth cohort. Matern. Child Health J. 18, 810–820. 10.1007/s10995-013-1305-y [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Oga EA, Mark K, Coleman-Cowger VH, 2018. Cigarette smoking status and substance use in pregnancy. Matern. Child Health J. 22, 1477–1483. 10.1007/s10995-018-2543-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Piper ME, McCarthy DE, Baker TB, 2006. Assessing tobacco dependence: a guide to measure evaluation and selection. Nicotine Tob. Res 8, 339–351. 10.1080/14622200600672765 [DOI] [PubMed] [Google Scholar]
  36. Piper ME, McCarthy DE, Bolt DM, Smith SS, Lerman C, Benowitz N, Fiore MC, Baker TB, 2008. Assessing dimensions of nicotine dependence: An evaluation of the nicotine dependence syndrome scale (NDSS) and the wisconsin inventory of smoking dependence motives (WISDM). Nicotine Tob. Res 10, 1009–1020. 10.1080/14622200802097563 [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Piper ME, Piasecki TM, Federman EB, Bolt DM, Smith SS, Fiore MC, Baker TB, 2004. A multiple motives approach to tobacco dependence: the wisconsin inventory of smoking dependence motives (WISDM-68). J. Consult. Clin. Psychol 72, 139–154. 10.1037/0022-006X.72.2.139 [DOI] [PubMed] [Google Scholar]
  38. Practice Committee of the American Society for Reproductive Medicine, 2018. Smoking and infertility: a committee opinion. Fertil. Steril 110, 611–618. 10.1016/j.fertnstert.2018.06.016 [DOI] [PubMed] [Google Scholar]
  39. Riaz M, Lewis S, Naughton F, Ussher M, 2018. Predictors of smoking cessation during pregnancy: a systematic review and meta-analysis. Addiction 113, 610–622. 10.1111/add.14135 [DOI] [PubMed] [Google Scholar]
  40. Sartor CE, Waldron M, Duncan AE, Grant JD, McCutcheon VV, Nelson EC, Madden PAF, Bucholz KK, Heath AC, 2013. Childhood sexual abuse and early substance use in adolescent girls: the role of familial influences. Addiction 108, 993–1000. 10.1111/add.12115 [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Scherman A, Tolosa JE, McEvoy C, 2018. Smoking cessation in pregnancy: a continuing challenge in the United States. Ther. Adv. Drug Saf. 9, 457–474. 10.1177/2042098618775366 [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Sharp BM, Chen H, 2018. Neurogenetic determinants and mechanisms of addiction to nicotine and smoked tobacco. Eur. J. Neurosci [ePub ahead of print]. 10.1111/ejn.14171 [DOI] [PubMed] [Google Scholar]
  43. Shiffman S, Waters A, Hickcox M, 2004. The nicotine dependence syndrome scale: a multidimensional measure of nicotine dependence. Nicotine Tob. Res 6, 327–348. 10.1080/1462220042000202481 [DOI] [PubMed] [Google Scholar]
  44. Tager IB, Weiss ST, Muñoz A, Rosner B, Speizer FE, 1983. Longitudinal study of the effects of maternal smoking on pulmonary function in children. N. Engl. J. Med 309, 699–703. 10.1056/NEJM198309223091204 [DOI] [PubMed] [Google Scholar]
  45. U.S. Department of Health and Human Services, Healthy People 2020. https://www.healthypeople.gov/2020/topics-objectives/topic/maternal-infant-and-child-health/objectives. (Accessed on 03/26/2018)
  46. Volz AR, Dennis PA, Dennis MF, Calhoun PS, Wilson SM, Beckham JC, 2014. The role of daily hassles and distress tolerance in predicting cigarette craving during a quit attempt. Nicotine Tob. Res 16, 872–875. 10.1093/ntr/ntt286 [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Waldron M, Bucholz KK, Lian M, Lessov-Schlaggar CN, Miller RH, Lynskey MT, Knopik VS, Madden PAF, Heath AC, 2017. Single motherhood, alcohol dependence, and smoking during pregnancy: a propensity score analysis. J. Stud. Alcohol Drugs 78, 745–753. 10.15288/jsad.2013.74.337 [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. World Health Organization (WHO), 1997. Composite International Diagnostic Interview Schedule-Core Version 2.1 Geneva, Switzerland. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS1525476-supplement-1.docx (28.8KB, docx)

RESOURCES