Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2021 Aug 1.
Published in final edited form as: Alcohol Clin Exp Res. 2020 Jul 23;44(8):1675–1685. doi: 10.1111/acer.14394

Alcohol Use Trajectories Before and After Pregnancy Among Adolescent and Young Adult Mothers

Irene Tung 1, Tammy Chung 1, Robert T Krafty 2, Kate Keenan 3, Alison E Hipwell 1
PMCID: PMC7484369  NIHMSID: NIHMS1615925  PMID: 32701186

Abstract

Background:

Moderate to heavy alcohol use during pregnancy (≥3 drinks/occasion) is linked to a range of adverse offspring outcomes. Prior studies suggest that adolescent and young adult mothers may be particularly vulnerable to these risky drinking behaviors during and after pregnancy. This study used latent class growth analysis (LCGA) to examine changes in risky drinking from pre- to post-pregnancy and identified pre-pregnancy predictors representing cognitive and social factors to inform prevention.

Methods:

Participants included 432 adolescent and young adult mothers (aged 13-24; 78% Black) assessed annually since childhood in the Pittsburgh Girls Study. Participants prospectively reported frequency of risky drinking (≥3 drinks/occasion) in the year of pregnancy and two years before and after pregnancy, as well as social-cognitive risk factors for drinking (alcohol expectancies, drinking motives, caregiver and peer alcohol use).

Results:

Approximately 11% of young mothers in this community-based sample engaged in risky drinking at least once a month during the year of pregnancy, with greater frequency observed for young adult (vs. adolescent) mothers. LCGA revealed four risky drinking trajectories across the perinatal period, including three relatively stable trajectories (low-, moderate-, and high-frequency) and one distinct trajectory (‘postpartum initiators’) characterized by increased risky drinking post-childbirth. Peer alcohol use during and after pregnancy distinguished higher-risk from lower-risk trajectories, controlling for age, race, and age of drinking onset. Exposure to caregiver alcohol use and pre-pregnancy coping motives for drinking also elevated risk, but only for mothers who became pregnant in early adulthood (vs. adolescence).

Conclusions:

Adolescent and young adult mothers show heterogeneous risky drinking trajectories from pre- to post-pregnancy. Pregnant women in the transitional age of early adulthood who report exposure to caregiver alcohol use, more friends who drink, and coping motives for drinking may have elevated risk for moderate-heavy drinking during the perinatal period and may benefit from targeted intervention.

Keywords: alcohol, pregnancy, young mothers, adolescence, early adulthood, growth trajectories

Introduction

Adolescence and early adulthood are transitional periods of development characterized by increased alcohol use and risky sexual behaviors that can lead to pregnancy (Patrick and Schulenberg, 2014; Salas-Wright et al., 2015). Despite the frequent co-occurrence of drinking and sexual behaviors during adolescence and early adulthood (Cavazos-Rehg et al., 2011), few studies have investigated longitudinal changes in alcohol use from pre-pregnancy to the postpartum period among young mothers, a unique group of women who likely show different trajectories of drinking than same-aged peers.

The U.S. Department of Health and Human Services currently recommends that pregnant women should not drink alcohol (USDHHS, 2015), emphasizing that “no safe level of alcohol use during pregnancy has been established” (CDC, 2018). Indeed, alcohol use during pregnancy and early motherhood can have long-term implications for offspring outcomes, including behavioral and neurodevelopmental impairments, though effects likely differ based on the frequency, quantity, and timing of alcohol use (Gupta et al., 2016). Although there is active ongoing research on the effects of infrequent and low-level drinking on offspring outcomes, a substantial body of evidence shows that ‘moderate to heavy alcohol use’ during pregnancy (typically quantified as ≥ 3 drinks/occasion) significantly heightens risk for fetal alcohol spectrum disorders and adverse birth outcomes (Meyer-Leu et al., 2011). Most women decrease drinking during pregnancy, but at least 1 in 9 pregnant women continue to consume alcohol (Denny, 2019). Furthermore, longitudinal studies suggest that adolescents may be more likely than women of other age groups to rapidly resume drinking during the first year postpartum (Bottorff et al., 2014; De Genna et al., 2017). Identifying subgroups of young mothers who engage in risky drinking during and after pregnancy and understanding what psychosocial factors influence these trajectories is an important public health priority.

Trajectories of Alcohol Use Before and After Pregnancy

Existing studies suggest that on average, adolescent mothers reduce or stop drinking during pregnancy, and then resume drinking in the first year postpartum (Chapman and Wu, 2013; Tung et al., 2019). However, adolescent mothers also vary widely in pre-pregnancy levels of drinking and the extent to which they change their use during and after pregnancy (Chang et al., 2011; De Genna et al., 2007; Skagerstróm et al., 2011). For example, one study of 305 pregnant adolescents examined changes in drinking frequency and found significant between-person differences in its slope after pregnancy, which was predicted by greater retrospectively-reported history of pre-pregnancy drinking (Spears et al., 2010). Similarly, in a study of 240 adolescents, latent class growth models from 1-10 years postpartum yielded four classes distinguished by frequency of postpartum drinking (low-frequency drinkers, early decliners, late decliners, and increasers); younger age at first drink and older age at first birth were associated with higher frequency drinking trajectories after pregnancy (Oxford et al., 2003). Together, these results highlight the importance of elucidating individual differences in drinking patterns for adolescent and young adult mothers. However, because these studies did not differentiate frequency of low-level drinking from moderate-heavy drinking, the implications of these perinatal trajectories are unclear. Prior blood alcohol concentration studies suggest using a 3-drink threshold to represent ‘binge drinking’ for adolescent females (ages 13-17; Donovan, 2009), and pregnancy studies similarly have linked ‘moderate to heavy drinking’ during pregnancy (defined as ‘3+ drinks/occasion’) to serious adverse fetal outcomes (Meyer-Leu et al., 2011). Building on these studies, the present study will focus on characterizing changes in the frequency of ‘risky drinking’ (3+ drinks/occasion) from pre- to post-pregnancy in a sample of pregnant adolescent and young adult women.

Predictors of Risky Drinking During and After Pregnancy

To date, almost all studies investigating predictors of perinatal alcohol use have recruited participants when they are already pregnant and, thus, have relied on retrospective reporting of pre-pregnancy risk factors. In addition to increasing vulnerability to recall bias, this approach limits the types of risk factors that can be investigated. Predictors of drinking during adolescence can differ from later adulthood (Windle, 2016). For example, among adult women, predictors of prenatal drinking include factors such as being unmarried, having only one child, and a history of alcohol use problems (Skagerstróm et al., 2011; Tran et al., 2015). These risk factors may not generalize to pregnant women in adolescence/early adulthood, a transitional developmental period characterized by normative patterns of alcohol use initiation and increased binge drinking. Importantly, risk factors traditionally linked to alcohol use problems in adolescence, such as social context, alcohol expectancies, and drinking motives (Leung et al., 2014; Patrick and Schulenberg, 2014), are ideally assessed using prospective studies, beginning prior to pregnancy.

In this study, we draw from social learning models of alcohol use (Bandura, 1969; Cooper et al., 1988; Maisto et al., 1999) to examine several pre-pregnancy predictors of risky drinking during and after pregnancy: exposure to drinking in the social environment (caregivers, peers), early alcohol expectancies, and motives for drinking. Although a large body of evidence supports the predictive influence of these factors on normative trajectories of alcohol use during adolescence (Alati et al., 2014; Patrick and Schulenberg, 2014; Patton et al., 2018), almost no studies have characterized these factors specifically for pregnant adolescents transitioning to motherhood. Frequent exposure to caregiver alcohol use, particularly during early adolescence (Chartier et al., 2010), is associated with early initiation of drinking and binge drinking (Alati et al., 2014). These effects may be particularly salient during the transition to early motherhood, when a history of growing up with frequent caregiver alcohol use may influence a new mother’s decisions to engage in risky drinking through social learning (modeling) and positive expectancies about alcohol effects (Brown et al., 1999; Cranford et al., 2010). Thus, prior exposure to frequent caregiver alcohol use may increase vulnerability to risky postpartum drinking, a hypothesis that will be tested.

Beyond these early risk factors, peer relationships also represent a primary social context for risky drinking during adolescence (Leung et al., 2014; McCann et al., 2019). However, few studies have investigated how peer drinking before, during, and after pregnancy influences trajectories of perinatal alcohol use for adolescent mothers. Existing studies support the important role of peer socialization on perinatal drinking: substance use patterns of best friends significantly predicted drinking during pregnancy for adolescent mothers, even after controlling for (retrospectively reported) pre-pregnancy drinking (Lohr et al., 1992). Similarly, pregnant adolescents with substance-using boyfriends were more likely to resume drinking postpartum (Spears et al., 2010). These studies suggest that exposure to peer drinking may play a critical role in shaping risky drinking patterns during and after pregnancy. However, because no studies have examined these factors prospectively starting before pregnancy, it is unclear how the timing of peer drinking exposure influences changes in risky drinking from pre- to post-pregnancy.

Finally, among women who are already drinking before pregnancy, drinking motives may play an important role in drinking trajectories across time (Cooper et al., 2016). The two drinking motives most strongly linked with risky drinking patterns are drinking to enhance positive emotions (‘enhancement motives’) and dampen negative emotions (‘coping motives’) (Cooper et al., 2016; Kuntsche et al., 2005), with enhancement motives endorsed more frequently during adolescence (Kuntsche and Cooper, 2010). However, adolescent and young adult mothers may experience unique stressors compared to same-aged peers. Indeed, for some women, the transition to motherhood is characterized by increased vulnerability to negative emotions and increased responsibilities as part of the social role transition into parenthood (Rutherford and Mayes, 2019). These experiences may be particularly challenging for young mothers who often have more limited resources (Torres et al., 2017). Thus, young women who drink to cope with negative emotions prior to pregnancy may be particularly vulnerable to risky drinking during and after pregnancy, although no study to our knowledge has examined pre-pregnancy drinking motives as a predictor of risky drinking during and after pregnancy.

The Present Study

Understanding individual differences in trajectories of risky drinking during and after pregnancy and identifying factors that predict these trajectories is important for informing prevention of alcohol-related problems for adolescent and young adult mothers. Although many studies have investigated predictors of alcohol use initiation and risky drinking in adolescence and early adulthood more generally, few studies have examined whether these factors predict risky drinking trajectories for adolescent/young adult mothers, who may show unique changes in drinking patterns during and after pregnancy compared to their peers across this age range. Drawing from a longitudinal sample of adolescent/young adult mothers (aged 13-24 years old) assessed annually since childhood, we examined latent trajectories of risky alcohol use across pre-pregnancy, prenatal, and postpartum periods, and then characterized these trajectory groups with pre-pregnancy factors previously linked to risky drinking. Whereas prior studies of pregnant adolescents have largely relied on retrospective data, we used a prospective study design to examine predictors collected prior to pregnancy, such as alcohol-related cognitions (e.g., alcohol expectancies) and social environment factors (e.g., caretaker and peer alcohol use).

Based on prior studies (Chapman and Wu, 2013; De Genna et al., 2017), we hypothesized that several distinct patterns of perinatal risky drinking would emerge, including ‘stable low’ and ‘stable high’ trajectories, a ‘decreasing’ trajectory, and a ‘rebound’ trajectory (decrease during pregnancy that then increases after birth). Given that this is the first study of adolescent/young adult mothers to prospectively model trajectories of risky drinking from pre-pregnancy to post-pregnancy, we note the exploratory nature of these hypotheses. We hypothesized that women who engaged in frequent risky drinking during and after pregnancy would be characterized by greater exposure to caretaker alcohol use and more positive expectancies about alcohol effects in early adolescence; report more friends who drink alcohol before, during, and after pregnancy; and report a greater pre-pregnancy tendency to use alcohol to cope with negative emotions. Given the unique age-expected changes in alcohol use during adolescence and early adulthood (Chassin et al., 2002), we examined these predictive associations while adjusting for the linear and quadratic effects of age on drinking and explored whether risk factors for high-risk trajectories differ based on age at pregnancy.

Materials and Methods

Participants and Procedures

The present study included a subsample of 432 adolescent and young adult pregnant women (age 13-23 at pregnancy) drawn from the Pittsburgh Girls Study (PGS; Hipwell et al., 2002; Keenan et al., 2010). The PGS is a longitudinal study of 2,450 girls (52% Black, 41% White, 7% multiracial/other) initially recruited in childhood (age 5-8) in 1999-2000 based on a stratified, random sampling of 103,238 households that oversampled from low-income neighborhoods in Pittsburgh, PA (USA). Over the past 18 years, participants have been assessed annually with high retention (range: 97.2% in wave 2 to 85.6% in wave 18), and 791 total participants have delivered a live birth. To assess alcohol use before, during, and after pregnancy, data from the two years before pregnancy (T-2 and T-1), year of pregnancy (T0), and the two years following childbirth (T+1 and T+2) were identified. Participants were included in analyses if they completed an assessment while pregnant (N=109 excluded) and at least one assessment before and after pregnancy (N=250 excluded), resulting in a final sample of N=432. We included data from first births only. Compared to participants who delivered a live birth but had insufficient data, the 432 participants included were slightly younger (mean childbirth age = 19.33 vs. 20.83; t=−8.91, p<.01) but did not differ on age of alcohol initiation or any other demographics. Compared to PGS participants who did not deliver a child, the 432 participants included in this study were more likely to identify as minority race (84.0% vs. 54.2%; χ2=82.25, p<.001) and have a history of using public assistance (24.2% vs. 18.2% at PGS wave 1; χ2=6.94, p=.008); groups did not differ on age of alcohol initiation (t=.43, p=.67).

All study procedures were approved by the university’s Institutional Review Board. Caregivers (94% female) provided written consent and participants provided verbal assent prior to age 18, after which participants provided their own written informed consent.

Measures

Risky alcohol use.

Frequency and quantity of alcohol use in the past year were self-reported in the Nicotine, Alcohol and Drug Use questionnaire (NADU; adapted from Pandina et al., 1984). To generate a ‘frequency of risky drinking’ variable, participants reported their frequency of drinking in the past year on a 6-point scale: 0 (none), 1 (less than 5 times), 2 (more than 5 times but less than 1 month), 3 (about once a month), 4 (about once a week), 5 (a couple times a week or more). Next, participants reported their typical quantity of drinks on a 5-pt scale: 0 (none), 1 (less than 1 drink), 2 (1-2 drinks), 3 (3-5 drinks) and 4 (6 or more drinks). Prior blood alcohol concentration studies suggest using a 3-drink threshold to represent ‘binge drinking’ for adolescents (Donovan, 2009), consistent with the 3-drink threshold of ‘moderate to heavy drinking’ linked to adverse fetal outcomes in pregnancy studies (Meyer-Leu et al., 2011). Thus, we used a ‘≥3 drinks’ threshold by dichotomizing the quantity variable to indicate presence or absence of ‘risky drinking’. We combined the frequency and dichotomous quantity data to generate the final ‘frequency of risky drinking’ variable for each time point. Thus, frequency of risky drinking (≥3 drinks/occasion) in the past year ranged from 0 (none) to 5 (a couple times a week or more) and was analyzed as an ordinal variable at each time point.

Alcohol expectancies.

The Drug, Alcohol, and Tobacco Expectancy Questionnaire (adapted from Hornik et al., 2002) assessed positive and negative expectancies about the effects of alcohol in early adolescence (age 11-12). Four items assess positive expectancies (e.g., “I would have a good time with friends,” “I would be more relaxed”), and four items assess negative expectancies (e.g., “It would upset my parents,” “I would lose control of myself”). Items were rated on a 5-point scale ranging from 0 (very unlikely) to 4 (very likely); positive and negative items were summed separately, with higher scores indicating either stronger positive or negative expectancies about alcohol. We used average scores across ages 11-12 for positive (McDonald’s ω=.90-91, GLB=.89-.93) and negative expectancies (ω=.50-.60, GLB=.58-65) because these were the years of available data prior to the earliest pregnancy (age 13).

Drinking motives.

The Drinking Motives Questionnaire (Cooper, 1994) assessed pre-pregnancy reasons for drinking for participants who endorsed recurring alcohol use (≥5x/year) during the two years before pregnancy (T-1 and T-2) (N=145). Four items assessed coping motives (ω=.89-.90, GLB=.91-93; e.g., “drinking to forget problems,” “drinking when depressed/anxious”) and five items assessed enhancement motives (ω=.72-79, GLB=.77-.83; e.g., “drinking because it is fun”; “drinking because I like the feeling”). Items were rated on a 3-point scale: 0 (never), 1 (some of the time), 2 (most of the time). Items were summed to form coping and enhancement subscales, and T-1 and T-2 scores were averaged to represent pre-pregnancy coping and enhancement motivations for drinking. Coping motives were stable from T-2 to T-1 (2.35 vs. 2.56; t=−0.61, p=.54), whereas enhancement motives were slightly higher at T-2 (4.81 vs. 4.23; t=2.11, p=.04).

Caregiver alcohol use.

The Alcohol Use Disorders Identification Test (AUDIT; Saunders et al., 1993) assessed caregiver’s self-reported alcohol use when the girls were in early adolescence (ages 11-12). The AUDIT is a widely used measure with 10 self-report items assessing the frequency/intensity of drinking, alcohol dependence, and alcohol-related impairments. Higher scores indicate greater problems with alcohol, with scores of 8 or more indicating harmful levels of drinking. Because the youngest pregnant participant in this sample was 13 years old, we used average caregiver AUDIT scores when the girls were aged 11 (ω=.85, GLB=.91) and 12 (ω=.81, GLB=.86) to represent caregiver alcohol use in early adolescence.

Peer alcohol use.

Participants reported the number of friends in the past year who used alcohol as part of the Peer Delinquency Scale (Loeber et al., 1998) using a 4-point scale ranging from 0 (none) to 3 (all of them). We examined peer alcohol use at three time points: before pregnancy (T-1 and T-2), year of pregnancy (T0), and after childbirth (T+1 and T+2).

Covariates.

Demographic variables and potential confounding factors were evaluated as covariates by examining their correlations with risky drinking. Race data were collected from caregivers as part of the original longitudinal study. Age at pregnancy assessment (T0) was calculated as the number of years between the participant’s date of birth and T0 date of assessment. Due to the annual assessment schedule, some participants completed T0 assessments earlier in pregnancy than others; thus, we covaried month of pregnancy at T0 by using an estimated date of conception (based on a 40-week pregnancy). Age of alcohol initiation was the age in which respondents first reported any alcohol consumption on the NADU (described above). History of conduct problems was measured by the Child Symptom Inventory (CSI), a standardized behavioral rating scale of DSM-IV oppositional defiant disorder (ODD) and conduct disorder (CD) symptoms that has shown concurrent validity and sensitivity/specificity of symptom scores to clinician diagnoses (Gadow and Sprafkin, 1994). From ages 8-13, parents and teacher separately rated symptoms of ODD and CD on a 4-pt scale (0=never, 3=all the time). We combined parent- and teacher reports by selecting the higher reported severity for each item; items were summed into a total ODD/CD severity score. We used average ODD/CD severity from ages 8-13 to represent history of childhood conduct problems.

Data Analysis Plan

To model trajectories of risky drinking from pre- to post-pregnancy, latent class growth analysis (LCGA) was conducted in Mplus 6.12. Frequency of risky drinking was modeled as an ordered categorical (ordinal) variable across the five time points while allowing for linear and nonlinear (quadratic, cubic) patterns of change. To identify the best fitting number of classes (from 1 to 6), we used goodness of fit indices (lower values indicate better fit) including the Akaike information criterion (AIC) and sample-adjusted Bayesian information criterion (BIC), as well as the Lo-Mendell-Rubin (LMR) likelihood ratio test (comparing k class with k–1 class models; Lo, Mendell, & Rubin, 2001). Posterior probabilities assign each individual fractionally to all classes, rather than forcing a 0/1 classification (Muthén, 2004). This is important because it is unlikely that participants have 100% probability of membership in a single class. After selecting the fitted latent class model, we conducted post-hoc probing of each class to evaluate whether the probability (odds ratio; OR) of engaging in risky drinking ‘at least once a month or more’ significantly changed year-to-year from 2-years pre-pregnancy to 2-years post-pregnancy.

Predictive models.

Next, to examine predictors of trajectory class, we used two sets of fractional multinomial logit models (fmlogit; Buis, 2017) in Stata 15.1. Fractional multinomial logit models are an extension of the multinomial logit model and allow for joint estimation of the probability of each class while accounting for their fractional nature. Thus, we modeled predictions of the probability of membership in each class while simultaneously accounting for the fractional probability of membership in other classes. In Model 1 (N=432), early adolescent expectancies about alcohol effects, caregiver alcohol use, and peer alcohol use (before, during, and after pregnancy) were entered as predictors of the probability of membership in each class. Next, in the subset of mothers (N=145) who drank alcohol (5+ times/year) before pregnancy (Model 2), pre-pregnancy coping and enhancement motivations for drinking were modeled as predictors of each trajectory class. All models adjusted for linear and quadratic effects of (T0) age on drinking trajectory. Finally, to explore whether predictive effects differed by age, secondary analyses tested interactions between age and each independent variable. To probe significant interactions, we used regions of significance analyses to identify the specific ages in which predictor variables differentiated trajectory class membership (Preacher et al., 2006).

Rates of missing data ranged from 0% to 3.9% for each covariate and predictor variable, with 88% of the sample having complete data on all variables. To minimize parameter biases associated with missing data and increase power, we employed multiple imputation by chained equation (MICE) using 50 iterations (White et al., 2011). MICE employs an iterative regression technique that treats each variable as an outcome variable and then imputes any missing values based on the remaining variables in the model. MICE has several advantages to other missing data methods (e.g., mean imputation, listwise deletion) including generating more accurate standard errors by incorporating a random element to account for uncertainty in the imputations.

Results

Descriptive statistics

During the year before pregnancy (T-1), 16% of the sample (N=61) reported drinking 3 or more drinks/occasion at least once a month (10% reporting once a week or more). Approximately 11% (N=45) engaged in risky drinking (≥3 drinks/occasion) at least once a month during the year of pregnancy (T0), which increased to 18% the year after childbirth (T+1). Correlations among study variables appear in Table 1. Frequency of risky drinking during pregnancy (T0) showed low to moderate correlations with risky drinking before (T-1, T-2) and after pregnancy (T+1, T+2), supporting the need to examine patterns of change over time. For potential covariates, minority race, age at T0, and age of alcohol initiation (but not childhood conduct problems) were correlated with risky drinking before, during, or after pregnancy and thus were included in subsequent predictive models.

Table 1.

Spearman’s correlations (rho) between observed study variables

M (SD) or % 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
1. Minority race 84.0% --
2. Age at T0 (preg) 18.96 (2.03) −.11* --
3. Age of first drink 15.81 (3.55) .11* <.01 --
4. Child conduct problems 10.77 (5.46) .07 −.18** −.24** --
5. Positive expectancies 2.87 (3.07) .05 −.07 −.08 .04 --
6. Negative expectancies 5.19 (2.46) −.02 <.01 −.17** .06 −.05 --
7. Coping motives 2.20 (2.21) −.02 −.06 −.08 −.08 .03 −.07 --
8. Enhancement motives 4.17 (1.92) −.07 −.12 −.24** .17* .12 .08 .27** --
9. Caregiver AUDIT 2.45 (2.91) .02 −.10* −.07 .08 <.01 <.01 −.03 .17* --
10. Peer alcohol before preg 1.54 (0.94) −.04 .27** −.33** .04 −.03 .05 −.02 .24** .02 --
11. Peer alcohol during preg 1.70 (1.06) −.03 .25** −.23** .04 −.04 .09 .20* .25** <.01 .52** --
12. Peer alcohol after preg 1.63 (0.95) −.01 −.17** −.19** >.01 .16** <.01 −.01 .13 −.01 .17** .19** --
13. T-2 risky drinking 0.55 (1.27) −.11* .29** −.25** .02 .01 .04 .10 .22* .04 .37** .30** .05 --
14. T-1 risky drinking 0.79 (1.53) −.08 .24** −.22** .03 .02 .09 .25** .41** <.01 .48** .37** .04 .50** --
15. T0 risky drinking 0.57 (1.31) −.07 .24** −.22** .01 .08 .04 .19* .02 .02 .28** .33** .12* .36** .43** --
16. T+1 risky drinking 0.87 (1.51) .02 .26** −.23** .04 .03 −.01 .04 −.05 −.05 .19** .28** .19** .25** .33** .37** --
17. T+2 risky drinking 0.99 (1.55) −.07 .24** −.17** .10 .07 −.01 .05 .03 .04 .33** .27** .20** .26** .30** .36** .50** --
Open in a new tab

Note.

*

p < .05

**

p < .01. T-2 = two years before pregnancy, T-1 = one year before pregnancy, T0 = during pregnancy, T+1 = one year after childbirth. T+2 = two years after childbirth. Risky drinking variables were measured on an ordinal frequency scale ranging from 0-5: 0=never, 1=less than 5x/year, 2=more than 5x/year but less than 1x/month, 3=about once a month, 4=about once a week, 5=a couple times a week or more

Trajectories of alcohol use from pre- to post-pregnancy

Goodness-of-fit comparisons for unconditional LCGA models (classes ranging from 1-6) are shown in Table 2. Overall, fit indices supported the 4-class model as the best fit. Because the LMR test comparing the 4-class to 3-class solution was not significant, we examined the distribution of classes for these solutions. Both the 4-class and 3-class models produced a ‘low-risk’ and ‘postpartum initiating’ class; however, whereas the 3-class model grouped the rest of participants into a large ‘moderate risky drinking’ class, the 4-class model distinguished between a ‘moderate risk’ and a smaller ‘high-risk drinking’ class. Given the conceptual significance of differentiating between these two groups, we followed the fit indices and used the 4-class solution in subsequent analyses. In general, entropy was modest for all solutions, indicating some imprecision in the classification of individuals into their ‘most likely class.’ Thus, primary models focused on predicting participants’ posterior probability of membership in all classes rather than forcing each individual into a categorical ‘most likely class’ variable (Muthén, 2004).

Table 2.

Fit indices for latent class growth analysis (LCGA) models with 1-6 classes

Number
of Classes		 AIC Sample-adjusted BIC LMR Test Entropy
1 3870.428 3877.588 -- --
2 3607.635 3619.269 <.0001 0.747
3 3587.125 3603.235 .0381 0.623
4 3573.993 3594.577 .1914 0.659
5 3571.277 a 3596.337 .2717 0.700
6 3572.756 3602.291 n/a b 0.732
Open in a new tab

Note. Bold indicates best fitting model. AIC = Akaike Information Criterion; BIC = Bayesian Information Criterion; LMR = Lo-Mendell-Rubin likelihood ratio test.

a

Although the AIC value for the 5-class model was slightly lower than the 4-class model, this change in AIC was nominal and does not suggest a significant improvement in fit.

b

LMR test results were unavailable for the 6-class solution due to a small N=3 in one category preventing model convergence.

Table 3 summarizes frequency of risky drinking from two years pre-pregnancy to two years post-pregnancy for each trajectory class. Approximately half of the sample were members of class 1 (49%; ‘low-risk’), characterized by a near-zero probability of risky drinking that remained stable in pregnancy and postpartum (post-hoc comparisons between adjacent time points: ORs=.86-.87, p’s>.10) (Figure 1). Members of class 2 (21%; ‘postpartum initiation’) had similarly near-zero rates of risky drinking pre-pregnancy, but some increased probability of initiating risky drinking during pregnancy (OR=7.19, p=.011) that then substantially increased during first year postpartum (OR=8.30, p=.007). In contrast, class 3 (24%; ‘moderate-risk’) was characterized by a marginal increase in risky drinking during the two years leading up to pregnancy (OR=1.66, p=.062) that then stabilized during pregnancy and postpartum (ORs=.56-.64, p’s>.10). Finally, a minority (6%) of participants followed a ‘high-risk’ drinking trajectory (class 4), with 75% of this class engaging in risky drinking at least once a month and 58% once a week or more during pregnancy (Table 3). Post-hoc comparisons indicated no changes in risky drinking from pre-to post-pregnancy for this high-risk class (ORs=.86-1.21, p’s>.10).

Table 3.

Frequency of risky drinking (3+ drinks) and demographic characteristics by trajectory class

Risky drinking frequency C1: Low risk
N=214 (49%)		 C2: Postpartum initiation
N=78 (21%)		 C3: Moderate risk
N=99 (24%)		 C4: High risk
N=24 (6%)
% of C1 % of C2 % of C3 % of C4
T-2 (before pregnancy)
 ‘1x/month or more’ 1.9% 0% 17.4% 78.3%
 ‘1x/week or more’ 0.9% 0% 10.9% 69.6%
T-1 (before pregnancy)
 ‘1x/month or more’ 1.4% 0% 43.4% 90.5%
 ‘1x/week or more’ 1.0% 0% 25.6% 71.4%
T0 (during pregnancy)
 ‘1x/month or more’ 0% 1.3% 27.3% 75.0%
 ‘1x/week or more’ 0% 0% 16.2% 58.3%
T+1 (after pregnancy)
 ‘1x/month or more’ 0% 34.3% 32.9% 83.3%
 ‘1x/week or more’ 0% 17.1% 17.6% 55.6%
T+2 (after pregnancy)
 ‘1x/month or more’ 0% 46.2% 22.5% 90.9%
 ‘1x/week or more’ 0% 21.5% 10.0% 68.2%
Demographic variables M (SD) or % M (SD) or % M (SD) or % M (SD) or %
Age at first drink 16.81 (3.65) 15.46 (3.60) 14.39 (2.74) 13.96 (2.56)
Age at childbirth 18.78 (2.03) 19.50 (1.53) 20.04 (1.93) 21.27 (1.79)
Age group at pregnancy (T0)
 Teenager (13-17) 34.1% 14.8% 14.9% 5.0%
 Pre-legal adult (18-20) 54.3% 72.2% 50.7% 25.0%
 Legal drinking age (≥ 21) 11.6% 13.0% 34.3% 60.0%
Living with caregiver
 T0 (pregnancy) 63.3% 66.1% 56.6% 43.5%
 T+1 (after pregnancy) 47.9% 34.3% 40.5% 33.3%
Living with intimate partner
 T0 (pregnancy) 15.8% 19.3% 18.4% 21.7%
 T+1 (after pregnancy) 16.1% 14.9% 10.7% 11.1%
Employed full- or part-time
 T0 (pregnancy) 38.1% 40.0% 41.7% 65.2%
 T+1 (after pregnancy) 41.2% 67.1% 58.8% 72.2%
Receiving WIC assistance
 T0 (pregnancy) 32.7% 40.4% 40.3% 17.6%
 T+1 (after pregnancy) 69.8% 67.9% 73.2% 66.7%
Open in a new tab

Note. T-2 = two years before pregnancy, T-1 = one year before pregnancy, T0 = during pregnancy, T+1 = one year after childbirth. T+2 = two years after childbirth. WIC = Women, Infants, and Children federal assistance supplemental nutrition program for low-income families

Figure 1.

Figure 1.

Open in a new tab

Probability of engaging in risky drinking (3 or more drinks per occasion) before, during, and after pregnancy for each trajectory class. Participants ranged from ages 13-23 during pregnancy (T0) and 13-24 at childbirth: T-2 = two years before pregnancy, T-1 = one year before pregnancy, T0 = during pregnancy (age 13-23), T+1 = one year after childbirth, T+2 = two years after childbirth. Post-hoc analyses compared the probability of risky drinking between adjacent time points; significant (*p<.05; **p<.01) and marginally significant changes († p<.10) are marked.

Predicting alcohol trajectory class membership

Demographic differences.

Table 3 summarizes the demographic characteristics of each trajectory class. One-way ANOVAs revealed that the four trajectory classes differed significantly on age of alcohol initiation (F=14.60, p<.001) and age at childbirth (F=19.33, p<.001): the moderate and higher risk trajectories were characterized by earlier age of alcohol initiation combined with later age at childbirth. Relative to other groups, the high-risk group had the greatest proportion of women who were at or above the legal drinking age (21 years) during pregnancy (60% vs. 11-34%), and most of this class was employed during and after pregnancy. In contrast, the postpartum initiation group had the greatest proportion of women in the ‘pre-legal drinking age range’ of 18-20 years-old during pregnancy (72%). Regarding contextual factors, similar proportions of women in the low-risk (63%) and postpartum initiation (66%) groups lived with their caregivers during pregnancy; however, by first year postpartum, only half of these women in the postpartum initiation group lived with caregivers, whereas most of the low-risk group living with caregivers continued to do so postpartum.

Alcohol expectancies and social factors.

Results from the first fractional logistic regression (Model 1; N=432) appear in Table 4, in which the probability of membership in each trajectory class was compared to the low-risk class. Positive and negative alcohol expectancies and caregiver alcohol use during early adolescence did not have a main effect on risky drinking trajectory from pre- to post-pregnancy. However, secondary follow-up analyses that explored moderation by age (presented in text here) revealed that caregiver alcohol use interacted with pregnancy age (B=.04, SE=.02, p=.024) to distinguish the postpartum initiation and low-risk groups. Regions of significance analyses revealed that a higher caregiver AUDIT score was associated with increased probability of membership in the postpartum initiation group, but only for women who were at least 22 years of age at pregnancy.

Table 4.

Fractional logistic regressions predicting the probability of membership in each perinatal risky drinking trajectory class

Model 1 (N=432) Class 2 (postpartum initiation)
 Class 3 (moderate risk)
 Class 4 (high risk)
B SE p 95% CI B SE p 95% CI B SE p 95% CI
 Caregiver AUDIT score .01 .03 .760 [−0.05, 0.08] .03 .02 .814 [−0.06, 0.08] .02 .07 .806 [−0.11, 0.15]
 Positive alcohol expectancies .01 .03 .690 [−0.05, 0.07] −.15 .03 .203 [−0.03, 0.13] .10 .07 .157 [−0.04, 0.24]
 Negative alcohol expectancies −.03 .04 .427 [−0.10, 0.04] .01 .04 .726 [−0.10, 0.07] .01 .08 .857 [−0.14, 0.17]
 Peer alcohol before pregnancy −.11 .13 .374 [−0.36, 0.14] .05 .04 <.001 [0.28, 0.83] 1.13 .35 .001 [0.45, 1.82]
 Peer alcohol during pregnancy .15 .10 .137 [−0.05, 0.35] −.02 .04 <.001 [0.21, 0.70] .91 .27 .001 [0.39, 1.44]
 Peer alcohol after pregnancy .34 .10 .001 [0.14, 0.53] .55 .14 .021 [0.05, 0.57] .45 .21 .029 [0.05, 0.86]
Model 2 (N=145)
 Coping motivations −.07 .11 .500 [−0.28, 0.14] .07 .09 .454 [−0.11, 0.24] .16 .12 .197 [−0.08, 0.40]
 Enhancement motivations −.20 .10 .057 [−0.40, 0.01] .06 .10 .567 [−0.14, 0.25] .23 .15 .130 [−0.07, 0.53]
Open in a new tab

Note. Reference group = Class 1 (low-risk). Model 1 includes the full sample, whereas Model 2 focuses on the subset of mothers who were already drinking (at least 5 times a year) before pregnancy. Models covaried for minority race, age of alcohol initiation, linear and quadratic effects of age at T0, and the month of pregnancy at T0.

Perceived peer alcohol use significantly differentiated trajectory classes. Reporting more friends who drank alcohol post-pregnancy increased probability of membership in the postpartum initiation trajectory relative to the low-risk trajectory. Perceived peer drinking before, during, and after pregnancy each independently increased probability of membership in the moderate- and high-risk trajectory classes relative to the low-risk class. Compared to the moderate-risk class, the high-risk class had marginally more friends who drank pre-pregnancy (B=.58, SE=.32, p=.074) and significantly more friends who drank during pregnancy (B=.46, SE=.23, p=.001), although post-pregnancy peer drinking did not differ (p>.10). These effects persisted even after adjusting for linear and quadratic effects of age, age of alcohol initiation, and other demographic variables, and they were not moderated by age.

Drinking motives.

Model 2 in Table 4 displays the fractional logistic regression results for the subset of mothers (N=145) who endorsed recurring alcohol use before pregnancy. Adjusting for the linear and quadratic effects of age, age of alcohol initiation, and race, neither coping nor enhancement motives had a main effect on trajectory membership. Instead, age at pregnancy interacted with coping motives to predict membership in the moderate-risk (B=.21, SE=.07, p=.004) and high-risk trajectories (B=.23, SE=.08, p=.006). Regions of significance analyses revealed that a pre-pregnancy tendency to drink to cope with negative emotions predicted greater probability of membership in the moderate- and high-risk trajectories (vs. low-risk group), but only if women were 19 years or older at pregnancy.

Discussion

In a longitudinal sample of adolescent and young adult mothers prospectively assessed since childhood, we used person-centered latent growth analyses to identify distinct trajectories of risky alcohol use from pre- to post-pregnancy. We capitalized on data that were prospectively collected prior to conception to improve on existing studies that relied on retrospective reporting. Four latent trajectories of risky peripartum drinking emerged, including three relatively stable trajectories ranging in severity (‘low-risk’, ‘moderate-risk’, and ‘high-risk’) and one additional group characterized by increases in risky drinking during the postpartum period (‘postpartum initiators’). Our study had several unique strengths: first, we focused on mothers in adolescence and early adulthood, an understudied group with elevated risk for alcohol-related problems during and after pregnancy. Second, we identified pre-pregnancy cognitive (alcohol expectancy, motives) and social environmental predictors (caregiver/peer alcohol use) of perinatal alcohol use trajectories to inform preventative interventions. Our predictive models revealed that perceiving a greater number of friends who drink alcohol during and after pregnancy were key risk factors that differentiated higher- from lower-risk alcohol use trajectories in young pregnant women. Furthermore, for young adult mothers specifically, pre-pregnancy exposure to caregiver alcohol use and a pre-existing tendency to drink to cope with negative emotions increased probability for higher-risk trajectories, whereas these associations were not observed for the younger adolescent mothers.

The four identified trajectory patterns were partially consistent with hypotheses; almost half of individuals maintained low to no risky drinking from pre- to post-pregnancy (‘low-risk’), about a quarter engaged in a moderate frequency of risky drinking throughout this period (‘moderate-risk’), and a small group exhibited more frequent and recurring risky drinking that persisted from pre- to post-pregnancy (‘high-risk’). Interestingly, we additionally detected a sizeable group (21%) that rarely/never engaged in risky drinking pre-pregnancy but began regular risky drinking during the postpartum period (‘postpartum initiators’). Given that adolescence and the transition to early adulthood is a developmental period characterized by normative increases in alcohol initiation and risky use (Chartier et al., 2010), this additional group may represent a risky trajectory unique to mothers in adolescence and early adulthood.

When examining predictors of trajectory membership, the major differentiating factor between trajectories was perceived peer alcohol use. Reporting more friends who drank alcohol before, during, and after pregnancy significantly differentiated the high/moderate-risk groups from the low-risk group, even after accounting for linear and quadratic effects of age on risky drinking and age of alcohol initiation. Perceiving more peer alcohol use specifically during pregnancy also uniquely differentiated high-risk drinkers from moderate-risk drinkers. Similarly, whereas both the low-risk and postpartum-initiators engaged in little to no risky drinking pre-pregnancy, the postpartum-initiator group significantly increased frequency of risky drinking during the first year after childbirth and was differentiated from the low-risk group by postpartum reporting of peer drinking. These results are consistent with prior studies of (non-pregnant) adolescents that have found peer drinking behaviors to prospectively predict initiation of drinking during adolescence (Trucco et al., 2011). Thus, having friends who drink alcohol during and after pregnancy may be associated with adolescent and young adult mothers’ drinking patterns through risk pathways, such as peer reinforcement and modeling of risky alcohol use (McCann et al., 2019).

Adjusting for peer drinking, neither caregiver alcohol use nor early alcohol expectancies directly differentiated between trajectory groups when examining the full sample, highlighting the potentially unique role of peer relationships in peripartum drinking patterns for these young mothers. However, secondary analyses testing for interactions and regions of significance by age revealed that prior exposure to caregiver alcohol use significantly increased risk for the postpartum initiation (vs. low-risk) trajectory for young adult mothers (≥22 years), whereas these effects were not observed for mothers 21 years or younger during pregnancy. The specificity of these associations to young adult (vs. adolescent) mothers may be related to the legal drinking age (21 years) in the US. Prior exposure to parental alcohol use may have modeled acceptable caregiving behaviors (Brown et al., 1999) and contributed to increased vulnerability to risky drinking after childbirth, but these effects may depend on age-related access to alcohol during the peripartum period.

Similarly, when examining pre-pregnancy drinking motives among the subsample of women who reported recurring alcohol use before pregnancy, neither enhancement motives (drinking to increase positive emotions) nor coping motives (drinking to reduce negative emotions) had an overall main effect on trajectory membership. Instead, effects were moderated by age, such that women with pre-existing tendencies to drink to cope with negative emotions were more likely to follow a moderate- or high-risk trajectory of risky drinking across the perinatal period, but only if they were 19 years or older at pregnancy. Compared to same-aged peers, women who become pregnant during the transition to early adulthood may experience unique stressors including sudden changes in financial instability and increased responsibilities as a parent (Rutherford and Mayes, 2019). Our findings suggest that young adult women with a history of using alcohol to cope with negative emotions may be particularly vulnerable to engaging in risky perinatal drinking. These effects may not be as evident for high-school aged mothers (≤ 18 years), who are more likely to be living with caregivers (i.e., grandparents of the new infant) and may experience more childcare and financial support. Follow-up studies are needed to directly test these hypotheses to identify potential mediators underlying the pattern of age-specific vulnerabilities that emerged in the present study.

Overall, our results suggest that many of the same risk factors traditionally implicated in risky drinking in the general adolescent/young adult population, such as peer and parental alcohol use (Chassin et al., 2002), can be generalized and integrated into preventative care for adolescent and young adult mothers. However, interventions may need to be tailored to the adolescent/young adult pregnancy context to address unique factors experienced by young women transitioning to motherhood (Bottorff et al., 2014), including unique contextual changes related to age of pregnancy. Importantly, extant prenatal alcohol use screenings typically focus exclusively on symptoms of alcohol use disorders (Chang et al., 1999; Floyd et al., 2009). Few screeners focus on identifying developmentally-sensitive precursors for the development of risky drinking patterns during the prenatal/postpartum periods, an important step to initiating preventative interventions. Existing interventions that aim to reduce problem drinking during pregnancy often emphasize psychoeducation about the risks of alcohol use (Floyd et al., 2009), reflecting the tendency of public health interventions to focus on increasing knowledge. Our results suggest that for pregnant adolescent and young adult women, a critical step to reducing risky trajectories of perinatal drinking is to directly screen and shift the peer drinking context during adolescence and early adulthood (NIAAA, 2012). In addition, our findings suggest that risk factors for adolescent mothers may differ from risk factors for young adult mothers, potentially due to differences in access to alcohol or pregnancy-related changes in social-economic context. Future studies are needed to examine how these factors influence risky peripartum drinking for young mothers transitioning from adolescence to early adulthood.

These findings should be interpreted in the context of study limitations. First, although this study employed a longitudinal framework, the directionality of effects for measures obtained within the same year remain unclear. For example, research on peer drinking and adolescent alcohol use suggests that in addition to peer socialization effects, peer selection effects account for significant variability in risky drinking outcomes (Mercken et al., 2009). Similarly, perception of peer drinking may differ from actual rates of peer drinking, which may have unique associations with risky drinking outcomes (Iannotti and Bush, 1992; Simons-Morton et al., 2018). Thus, studies that measure each of these factors across multiple time points during the perinatal period are needed to tease apart the complex dynamics underlying the association between peer drinking and risky perinatal drinking. Second, consistent with other community-based studies, the average rates of risky drinking in our sample were low, which may have limited the ability to test some hypotheses. Although several moderate-higher risk groups emerged in our LCGA, including a small group (6%) with high risk for risky drinking, future studies that recruit from clinical samples are needed to further differentiate between higher-risk groups of perinatal drinkers. Third, because this study drew pregnancy data from a larger longitudinal study structured by annual assessments, the timing of the T0 assessment during pregnancy varied across participants. Our analyses covaried for the month of pregnancy at T0; however, future studies are needed that measure risky alcohol use across multiple time points throughout pregnancy to increase precision regarding timing effects. Finally, findings are based on a predominantly minority race sample of adolescent and young adult mothers from low-income, urban neighborhoods and may not generalize to other groups.

In conclusion, the present findings highlight significant heterogeneity in trajectories of alcohol use change from pre- to post-pregnancy for adolescent and young adult mothers. Despite elevated risk associated with the normative increases of risky drinking during this developmental period, most of the young mothers in our sample did not drink alcohol during pregnancy, whereas a small subgroup of women reported regularly engaging in moderate-heavy drinking during pregnancy and/or the first two years postpartum. Adolescent and young adult women who perceive more friends drinking may have the greatest risk for engaging in risky drinking during and after pregnancy. Prior exposure to caregiver alcohol use and a pre-existing tendency to drink to cope with negative emotions may also increase risk for postpartum drinking, although effects were specific to women in early adulthood (vs. adolescents). Our results support a need to develop preventative interventions that directly address the unique developmental needs and precursors of perinatal drinking among pregnant adolescent and young adult women as they transition to early motherhood.

Acknowledgements:

This project was supported by the National Institute of Mental Health (R01MH056630, K01MH071790), the Eunice Kennedy Shriver National Institute of Child Health and Human Development (R01HD067185), the National Institute of Drug Abuse (R01DA012237), and the National Institute of General Medical Sciences (R01GM113243). IT was supported by a postdoctoral training grant from the National Institute on Alcohol Abuse and Alcoholism (T32AA007453). Thanks to Gale Richardson, PhD for reading and providing feedback on a previous draft of this manuscript. Special thanks to the families of the Pittsburgh Girls Study for their participation in this research, and to our dedicated research team for their continued efforts.

Footnotes

Conflicts of Interest: None of the authors have any conflicts of interest to declare.

References

  1. Alati R, Baker P, Betts KS, Connor JP, Little K, Sanson A, Olsson CA (2014) The role of parental alcohol use, parental discipline and antisocial behaviour on adolescent drinking trajectories. Drug Alcohol Depend 134:178–184. [DOI] [PubMed] [Google Scholar]
  2. Bandura A (1969) Principles of behavior modification, Principles of behavior modification. Oxford, England, Holt, Rinehart, & Winston. [Google Scholar]
  3. Bottorff JL, Poole N, Kelly MT, Greaves L, Marcellus L, Jung M (2014) Tobacco and alcohol use in the context of adolescent pregnancy and postpartum: A scoping review of the literature. Health Soc Care Community 22:561–574. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Brown SA, Tate SR, Vik PW, Haas AL, Aarons GA (1999) Modeling of alcohol use mediates the effect of family history of alcoholism on adolescent alcohol expectancies. Exp Clin Psychopharmacol 7:20–27. [DOI] [PubMed] [Google Scholar]
  5. Buis M (2017) FMLOGIT: Stata module fitting a fractional multinomial logit model by quasi maximum likelihood, Statistical Software Components. [Google Scholar]
  6. Cavazos-Rehg PA, Krauss MJ, Spitznagel EL, Schootman M, Cottler LB, Bierut LJ (2011) Substance use and the risk for sexual intercourse with and without a history of teenage pregnancy among adolescent females. J Stud Alcohol Drugs 72:194–198. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. CDC (2018) Alcohol Use in Pregnancy. Cent Dis Control Prev. Available at: https://www.cdc.gov/ncbddd/fasd/alcohol-use.html Accessed April 27, 2020. [Google Scholar]
  8. Chang G, Orav EJ, Jones JA, Buynitsky T, Gonzalez S, Wilkins-Haug L (2011) Self-reported alcohol and drug use in pregnant young women: A pilot study of prevalence and associated factors. J Addict Med 5:221–226. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Chang G, Wilkins-Haug L, Berman S, Goetz MA (1999) The TWEAK: Application in a prenatal setting. J Stud Alcohol 60:306–309. [DOI] [PubMed] [Google Scholar]
  10. Chapman SLC, Wu L-T (2013) Substance use among adolescent mothers: A review. Child Youth Serv Rev 35:806–815. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Chartier KG, Hesselbrock MN, Hesselbrock VM (2010) Development and vulnerability factors in adolescent alcohol use. Child Adolesc Psychiatr Clin 19:493–504. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Chassin L, Pitts SC, Prost J (2002) Binge drinking trajectories from adolescence to emerging adulthood in a high-risk sample: Predictors and substance abuse outcomes. J Consult Clin Psychol 70:67–78. [PubMed] [Google Scholar]
  13. Cooper ML (1994) Motivations for alcohol use among adolescents: Development and validation of a four-factor model. Psychol Assess 6:117–128. [Google Scholar]
  14. Cooper ML, Kuntsche E, Levitt A, Barber LL, Wolf S (2016) Motivational models of substance use: A review of theory and research on motives for using alcohol, marijuana, and tobacco In: The Oxford Handbook of Substance Use and Substance Use Disorders (Sher KJ ed), pp 375–421. Oxford University Press. [Google Scholar]
  15. Cooper ML, Russell M, George WH (1988) Coping, expectancies, and alcohol abuse: A test of social learning formulations. J Abnorm Psychol 97:218–230. [DOI] [PubMed] [Google Scholar]
  16. Cranford JA, Zucker RA, Jester JM, Puttler LI, Fitzgerald HE (2010) Parental alcohol involvement and adolescent alcohol expectancies predict alcohol involvement in male adolescents. Psychol Addict Behav 24:386–396. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Curran PJ, Stice E, Chassin L (1997) The relation between adolescent alcohol use and peer alcohol use: A longitudinal random coefficients model. J Consult Clin Psychol 65:130–140. [DOI] [PubMed] [Google Scholar]
  18. De Genna NM, Goldschmidt L, Marshal M, Day NL, Cornelius MD (2017) Maternal age and trajectories of risky alcohol use: A prospective study. Alcohol Clin Exp Res 41:1725–1730. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. De Genna NM, Larkby C, Cornelius MD (2007) Early and adverse experiences with sex and alcohol are associated with adolescent drinking before and during pregnancy. Addict Behav 32:2799–2810. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Denny CH (2019) Consumption of alcohol beverages and binge drinking among pregnant women aged 18–44 years — United States, 2015–2017. MMWR Morb Mortal Wkly Rep 68. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Donovan JE (2009) Estimated blood alcohol concentrations for child and adolescent drinking and their implications for screening instruments. Pediatrics 123:e975–e981. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Floyd RL, Weber MK, Denny C, O’Connor MJ (2009) Prevention of fetal alcohol spectrum disorders. Dev Disabil Res Rev 15:193–199. [DOI] [PubMed] [Google Scholar]
  23. Gadow KD, Sprafkin J (1994) Child Symptom Inventories Manual. Stony Brook, NY, Checkmate Plus. [Google Scholar]
  24. Gupta KK, Gupta VK, Shirasaka T (2016) An update on fetal alcohol syndrome: Pathogenesis, risks, and treatment. Alcohol Clin Exp Res 40:1594–1602. [DOI] [PubMed] [Google Scholar]
  25. Hipwell AE, Loeber R, Stouthamer-Loeber M, Keenan K, White HR, Kroneman L (2002) Characteristics of girls with early onset disruptive and antisocial behaviour. Crim Behav Ment Health 12:99–118. [DOI] [PubMed] [Google Scholar]
  26. Hornik R, Maklan D, Cadell D, Prado A, Barmada C, Jacobsohn R, Orwin R, Sridharan S, Zador P, Southwell B, Sanutto E, Baskin R, Chu A, Morin C, Taylor K, Steel D (2002) Evaluation of the national youth anti-drug media campaign: Fourth semi-annual report of findings. Rockville, MD, Westat. [Google Scholar]
  27. Iannotti RJ, Bush PJ (1992) Perceived vs. actual friends’ use of alcohol, cigarettes, marijuana, and cocaine: Which has the most influence? J Youth Adolesc 21:375–389. [DOI] [PubMed] [Google Scholar]
  28. Keenan K, Hipwell A, Chung T, Stepp S, Stouthamer-Loeber M, Loeber R, McTigue K (2010) The Pittsburgh Girls Study: Overview and initial findings. J Clin Child Adolesc Psychol Off J Soc Clin Child Adolesc Psychol Am Psychol Assoc Div 53 39:506–521. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Kuntsche E, Cooper ML (2010) Drinking to have fun and to get drunk: Motives as predictors of weekend drinking over and above usual drinking habits. Drug Alcohol Depend 110:259–262. [DOI] [PubMed] [Google Scholar]
  30. Kuntsche E, Knibbe R, Gmel G, Engels R (2005) Why do young people drink? A review of drinking motives. Clin Psychol Rev 25:841–861. [DOI] [PubMed] [Google Scholar]
  31. Leung RK, Toumbourou JW, Hemphill SA (2014) The effect of peer influence and selection processes on adolescent alcohol use: A systematic review of longitudinal studies. Health Psychol Rev 8:426–457. [DOI] [PubMed] [Google Scholar]
  32. Loeber R, Farrington DP, Stouthamer-Loeber M, Kammen WBV (1998) Antisocial behavior and mental health problems: Explanatory factors in childhood and adolescence. Psychology Press. [Google Scholar]
  33. Lohr MJ, Gillmore MR, Gilchrist LD, Butler SS (1992) Factors related to substance use by pregnant, school-age adolescents. J Adolesc Health 13:475–482. [DOI] [PubMed] [Google Scholar]
  34. Maisto SA, Carey KB, Bradizza CM (1999) Social learning theory In: Psychological Theories of Drinking and Alcoholism, pp 106–163. New York, Guildford Press. [Google Scholar]
  35. McCann M, Jordan J-A, Higgins K, Moore L (2019) Longitudinal social network analysis of peer, family, and school contextual influences on adolescent drinking frequency. J Adolesc Health 65:350–358. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Mercken L, Candel M, Willems P, de Vries H (2009) Social influence and selection effects in the context of smoking behavior: Changes during early and mid adolescence. Health Psychol 28:73–82. [DOI] [PubMed] [Google Scholar]
  37. Meyer-Leu Y, Lemola S, Daeppen J-B, Deriaz O, Gerber S (2011) Association of moderate alcohol use and binge drinking during pregnancy with neonatal health. Alcohol Clin Exp Res 35:1669–1677. [DOI] [PubMed] [Google Scholar]
  38. Muthén BO (2004) Latent variable analysis: Growth mixture modeling and related techniques for longitudinal data In: The Sage Handbook of Quantitative Methodology for the Social Sciences , pp 345–68. Thousand Oaks, CA, Sage Publications. [Google Scholar]
  39. NIAAA (2012) Alcohol screening and brief intervention for youth: A practitioner’s guide. Natl Inst Alcohol Abuse Alcohol NIAAA. Available at: https://www.niaaa.nih.gov/alcohol-screening-and-brief-intervention-youth-practitioners-guide Accessed December 3, 2019. [Google Scholar]
  40. Oxford ML, Gilchrist LD, Morrison DM, Gillmore MR, Lohr MJ, Lewis SM (2003) Alcohol use among adolescent mothers: Heterogeneity in growth curves, predictors, and outcomes of alcohol use over time. Prev Sci 4:15–26. [DOI] [PubMed] [Google Scholar]
  41. Pandina RJ, Labouvie EW, White HR (1984) Potential contributions of the life span developmental approach to the study of adolescent alcohol and drug use: The Rutgers Health and Human Development Project, a working model. J Drug Issues 14:253–268. [Google Scholar]
  42. Patrick ME, Schulenberg JE (2014) Prevalence and predictors of adolescent alcohol use and binge drinking the United States. Alcohol Res Curr Rev 35:193–200. [PMC free article] [PubMed] [Google Scholar]
  43. Patton KA, Gullo MJ, Connor JP, Chan GCK, Kelly AB, Catalano RF, Toumbourou JW (2018) Social cognitive mediators of the relationship between impulsivity traits and adolescent alcohol use: Identifying unique targets for prevention. Addict Behav 84:79–85. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Preacher KJ, Curran PJ, Bauer DJ (2006) Computational tools for probing interactions in multiple linear regression, multilevel modeling, and latent curve analysis. J Educ Behav Stat 31:437–448. [Google Scholar]
  45. Rutherford HJ, Mayes LC (2019) Parenting stress: A novel mechanism of addiction vulnerability. Neurobiol Stress 11:100172. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Salas-Wright CP, Vaughn MG, Ugalde J, Todic J (2015) Substance use and teen pregnancy in the United States: Evidence from the NSDUH 2002–2012. Addict Behav 45:218–225. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Saunders JB, Aasland OG, Babor TF, Fuente JRDL, Grant M (1993) Development of the Alcohol Use Disorders Identification Test (AUDIT): WHO collaborative project on early detection of persons with harmful alcohol consumption-II. Addiction 88:791–804. [DOI] [PubMed] [Google Scholar]
  48. Simons-Morton B, Haynie D, Bible J, Liu D (2018) Prospective associations of actual and perceived descriptive norms with drinking among emerging adults. Subst Use Misuse 53:1771–1781. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Skagerstróm J, Chang G, Nilsen P (2011) Predictors of drinking during pregnancy: A systematic review. J Womens Health 20:901–913. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Spears GV, Stein JA, Koniak–Griffin D (2010) Latent growth trajectories of substance use among pregnant and parenting adolescents. Psychol Addict Behav 24:322–332. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Torres R, Goyal D, Burke-Aaronson AC, Gay CL, Lee KA (2017) Patterns of symptoms of perinatal depression and stress in late adolescent and young adult mothers. J Obstet Gynecol Neonatal Nurs JOGNN 46:814–823. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Tran NT, Najman JM, Hayatbakhsh R (2015) Predictors of maternal drinking trajectories before and after pregnancy: evidence from a longitudinal study. Aust N Z J Obstet Gynaecol 55:123–130. [DOI] [PubMed] [Google Scholar]
  53. Trucco EM, Colder CR, Wieczorek WF (2011) Vulnerability to peer influence: A moderated mediation study of early adolescent alcohol use initiation. Addict Behav 36:729–736. [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Tung I, Beardslee J, Pardini D, Chung T, Keenan K, Hipwell AE (2019) Adolescent childbirth, miscarriage, and abortion: Associations with changes in alcohol, marijuana, and cigarette use. J Child Psychol Psychiatry. [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. U.S. Department of Health and Human Services and U.S. Department of Agriculture (2015) 2015–2020 Dietary Guidelines ∣ health.gov. [Google Scholar]
  56. White IR, Royston P, Wood AM (2011) Multiple imputation using chained equations: Issues and guidance for practice. Stat Med 30:377–399. [DOI] [PubMed] [Google Scholar]
  57. Windle M (2016) Drinking over the lifespan: Focus on early adolescents and youth. Alcohol Res Curr Rev 38:95–101. [PMC free article] [PubMed] [Google Scholar]

RESOURCES