Substance Use Among Women: Associations with Pregnancy, Parenting, and Race/Ethnicity

Article

DOI: 10.1007/s10995-008-0375-8

Cite this article as:
Muhuri, P.K. & Gfroerer, J.C. Matern Child Health J (2009) 13: 376. doi:10.1007/s10995-008-0375-8

Abstract

Objective To examine the association of pregnancy and parenting (0- to 2-year-old child) statuses with past 30-day use of alcohol, cigarettes and marijuana, psychotherapeutics and cocaine among American women aged 18–44, overall and by race/ethnicity. Methods Five years of data (2002–2006) from the National Survey on Drug Use and Health (NSDUH) were pooled. Binary logistic regression analyses stratified by race/ethnicity were performed to examine the relationships adjusting for age, marital status, education and family income. Results Overall, past 30-day alcohol, cigarette, marijuana, psychotherapeutic or cocaine use was substantially lower among pregnant women, particularly in their second or third trimesters, than among their parenting or non-pregnant counterparts. Logistic regression analysis suggested a strong negative association between pregnancy status and substance use, with no considerable variations in the magnitude of the relationship by race/ethnicity for most measures. The relationship between parenting status and substance use was in the same direction, but relatively weak, and not statistically significant for non-Hispanic (NH) blacks for any measures except for alcohol use. Regardless of pregnancy and parenting statuses, NH white women reported substance use at the highest rates, followed by Hispanics and NH blacks. Conclusions There is indirect evidence of postpregnancy resumption in substance use. Pregnant and parenting women, regardless of race/ethnicity, could benefit from prevention efforts focusing on cessation rather than temporary abstinence from substance use.

Keywords

Substance use Alcohol Cigarettes Marijuana, Psychotherapeutics Cocaine Pregnancy Parenting Race/ethnicity 

Introduction

In the United States (US), substance use among women who are pregnant or parenting young children has remained a public health concern. Prenatal alcohol, cigarette and illicit drug use have been shown to be associated with poor pregnancy and neonatal outcomes [1, 2, 3, 4, 5, 6, 7, 8, 9] as well as behavioral and developmental problems at early childhood [10, 11, 12]. Parental smoking through environmental tobacco exposure has been linked to respiratory problems in children [13, 14]. Children of mothers abusing alcohol and illicit drugs are likely to be at increased risk of long-term physical, cognitive, behavioral and academic problems including their own substance use [15, 16]. Substance abuse by parents of young children has been linked to impaired parenting capacity [17] and child abuse and neglect [18, 19, 20, 21].

Prevalence estimates of substance use during pregnancy vary depending on the population sampled, the exposure interval of pregnancy covered and the reference period used in prevalence measures [22, 23]. Despite these differences, previous studies have consistently shown that women reduce their substance use after becoming pregnant. The 1992–1993 National Pregnancy & Health Survey by the National Institute on Drug Abuse reported reductions in the prevalence rates of alcohol, cigarette, marijuana, and cocaine use by the second trimester, compared with use in the 3 months before pregnancy [24]. Analyses from the 1994 to 1995 National Household Surveys on Drug Abuse (NHSDA) showed lower prevalence rates of past 30-day use of alcohol, cigarettes and any illicit drugs including marijuana among pregnant women (i.e., those who reported being pregnant at the time of the interview) than among non-pregnant counterparts. Women living with at least one child under age 2 were twice as likely as pregnant women to report alcohol and illicit drug use [25]. Similar results are echoed in later analyses of NHSDA/NSDUH [26, 27, 28, 29].

Results from these analyses pointed to postpregnancy resumption in use. However, when comparing the results between pregnant women vs. recent mothers, these prior studies did not account for differences in demographic and socioeconomic characteristics. Some of these studies reported lower prevalence of any past month substance use among pregnant white, black and Hispanic women than among non-pregnant women in these racial/ethnic groups [28, 29], but did not examine whether the association of recent motherhood or parenting status with use of specific substances varied by race/ethnicity. An important gap in the literature concerns the examination of the prevalence of substance use by pregnancy and parenting statuses under a common analytic framework that would address racial/ethnic differences while adjusting for socioeconomic differences. Understanding the nature of these associations in different race/ethnicity groups may help target preventive efforts toward the most vulnerable and thereby reduce the adverse maternal and child health consequences of substance use. The aim of this study was to (1) provide estimates of substance use for US women aged 18–44 by pregnancy and parenting statuses and race/ethnicity based on new nationally representative data from large population-based surveys and (2) examine whether the associations of pregnancy and parenting statuses with substance uses are still evident and if they vary by race/ethnicity after controlling for socioeconomic and demographic characteristics.

Data and Methods

The analysis was based on the cross-sectional data with no missing information on the variables of interest from 94,483 women aged 18–44 years, pooled from the NSDUH conducted in 2002, 2003, 2004, 2005 and 2006. The survey methods were identical in these years. Averaging for the five survey years, the annual household screening response rate was 91% and the interview response rate for women aged 18–44 years was 82%. The NSDUH is an annual survey administered by SAMHSA, and collects detailed information on the use of illicit drugs, alcohol and tobacco and its correlates from the civilian, non-institutionalized US population aged 12 years or older. Respondents in the survey are selected though multistage area probability sampling, and data are collected using computer-assisted interviewing methods [30].

For this analysis, several measures of use during the past 30 days were defined. These included any use of alcohol (yes/no), cigarettes (yes/no), marijuana/hashish (yes/no), psychotherapeutics (i.e., pain relievers, tranquilizers, stimulants or sedatives used non-medically) (yes/no), and cocaine (yes/no). Other past 30-day measures, which involved intense use, were binge alcohol use (i.e., five or more drinks at the same time or within a couple hours) (yes/no), daily cigarette smoking (yes/no), use of marijuana/hashish on six or more days (yes/no), and use of three or more of the eight substances that included alcohol, cigarettes, marijuana, psychotherapeutics, cocaine, heroin, hallucinogens, and inhalants (yes/no).

Pregnancy and parenting statuses were the key predictors in relation to which past 30-day substance use measures were examined. The pregnancy status was determined at the time of the interview by asking female survey respondents whether they were pregnant and how many months pregnant they were. The parenting status was derived from responses to a series of questions on all household members and their relationships. Pregnant women not reporting trimester data and women for which the parenting status could not be ascertained because of missing information were excluded from the analysis. Parenting women were defined as the mother of at least a 0- to 2-year-old child. “Not parenting” women included those not having a 0- to 2-year-old child even if they had an older child in the household. This construct was used as an indirect indicator identifying postpregnancy status of women to address the issue of resumption in use by postpartum women. For bivariate analysis, the pregnancy/parenting status was used as a composite variable with the following three mutually exclusive categories: (a) not parenting, non-pregnant; (b) parenting, non-pregnant; and (c) pregnant, regardless of parenting status. The pregnancy category was further broken down into three subcategories—pregnancy trimesters 1, 2 and 3. Race/ethnicity was another key variable used in the analysis to compare the prevalence of substance use among NH whites, NH blacks and Hispanics and examine the associations of pregnancy and parenting statuses with substance use in logistic regression framework for each group separately.

In logistic regression analysis, a total 36 unadjusted models were first estimated—nine models (each containing a separate binary dependent variable with two common predictor variables, pregnancy and parenting), overall and for NH whites, NH blacks and Hispanics. Then adjusted models were estimated after adding women’s age, marital status, education and annual family income to each of the 36 unadjusted models. These variables were used as statistical controls because of their associations with substance use found in prior research [23, 25, 29, 31, 32] and because of their possible association with pregnancy and parenting statuses. In the model, these controls were expressed in terms of indicator variables. Because these surveys employed complex designs, their design characteristics (i.e., stratum, replicate and analysis weight) were specified in Stata survey commands when estimating substance use prevalence rates and logistic regression models [33]. Throughout the analysis, results not significant at the 0.05 level were considered nonsignificant.

Results

Averaging for 5 years (2002–2006), the study estimated the prevalence of pregnancies, regardless of parenting a 0- to 2-year-old child, among women aged 18–44 on the day of interview to be 4.3% (95% CI: 4.1–4.5). The prevalence of parenting a 0- to 2-year-old child (not including the mother of a 0- to 2-year-old child who reported being pregnant) was 17.4% (95% CI: 17.0–17.8). Of them, 99.1% were biological mothers and the remaining 0.9% were mothers of a step, adoptive or foster child. Table 1 shows descriptive statistics of race/ethnicity, age, marital status, education and family income for the three pregnancy/parenting groups and combined. Overall, the mean age of the sample women covered by this analysis was 31.4 years. One-half of them were married. More than half at least had some college. Two-fifths of the women had annual family income $50,000 or higher. A large majority of the women were NH whites followed by Hispanics, NH blacks and NH others (NH Native American/Alaska Native, Native Hawaiian/other Pacific Islanders, Asian, NH more than one race).
Table 1

Percentages of women aged 18–44 in categories of selected characteristics by pregnancy/parenting status, NSDUH 2002–2006, United States

Selected characteristics

Not parentinga, non-pregnant

n = 71,732

Parentinga, non-pregnant

n = 17,734

Pregnant (parentinga or not parentinga)

n = 5,017

Total

n = 94,483

Race/ethnicity

    Non-Hispanic White

64.6

60.7

60.2

63.8

    Non-Hispanic Black

13.7

13.0

12.2

13.5

    Hispanic

14.4

19.7

20.1

15.5

    Non-Hispanic Othersb

7.3

6.6

7.5

7.2

    Total

100.0

100.0

100.0

100.0

    Design-based F (5.54, 9980.77) = 22.1; P < 0.001

Age (in years)

    18–21

15.4

10.6

18.1

14.7

    22–25

12.4

18.9

20.7

13.9

    26–34

72.2

70.5

61.2

71.4

    Total

100.0

100.0

100.0

100.0

    Design-based F (3.53, 6355.83) = 170.0; P < 0.001

Marital status

    Married

45.0

69.1

65.8

50.1

    Widow/divorced/separated

14.4

7.2

6.4

12.8

    Never married

40.6

23.7

27.8

37.1

    Total

100.0

100.0

100.0

100.0

    F (3.90, 7020.67) = 405.7; P < 0.001

Education

    Less than high school

13.7

17.3

18.9

14.5

    High school graduate

29.5

28.6

27.7

29.3

    Some college or higher

56.8

54.1

53.4

56.2

    Total

100.0

100.0

100.0

100.0

    Design-based F (3.90, 7018.59) = 24.8; P < 0.001

Annual family income (in dollars)

    <20,000

22.4

22.8

23.1

22.5

    20,000–49,999

35.9

35.7

36.6

35.9

    50,000 or higher

41.7

41.5

40.3

41.4

    Total

100.0

100.0

100.0

100.0

    F (5.79, 10415.98) = 1.4; P = 0.2234

aParenting women were defined as the mother of at least a 0- to 2-year-old child. “Not parenting” women included those not having a 0- to 2-year-old child even if they had an older child in the household

bNon-Hispanic others included non-Hispanic Native American/Alaska Native, Native Hawaiian/other Pacific Islanders, Asian and non-Hispanic more than one race

When compared across pregnancy/parenting groups, there emerged some striking compositional differences in these characteristics. Both pregnant and parenting women were significantly younger and composed of relatively more Hispanics, and a larger proportion of them were married. In contrast, the majority of not parenting, non-pregnant women were never married or widow/divorced/separated. The disparity across groups was minimal in regard to education and family income. The design-based F-test suggests that the associations between the pregnancy/parenting status and selected characteristics were statistically significant (P < 0.001) except for annual family income.

Table 2 presents prevalence estimates of past 30-day substance use for the pregnancy/parenting categories. In the study population, the most reported substance was alcohol (1 in 2), followed by cigarettes (1 in 3), marijuana (1 in 15), psychotherapeutics (1 in 28), and cocaine (1 in 91). Results for measures of more intense use were as follows: binge alcohol (1 in 4), cigarettes daily (1 in 5), marijuana on six or more days (1 in 29) and three or more of the eight substances (1 in 17).
Table 2

Prevalence estimates of past 30-day alcohol, cigarette and selected illicit drug use and 95% confidence interval (CI) by pregnancy/parenting status among women aged 18–44, NSDUH 2002–2006, United States

Substance used in past 30 days preceding interview

Not parentinga, non-pregnant

n = 71,732

Percent [95% CI]

Parentinga, non-pregnant

n = 17,734

Percent [95% CI]

Pregnant (parentinga or not parentinga)

Total

n = 94,483

Percent [95% CI]

All

n = 5,017

Trimester 1

n = 1,442

Trimester 2

n = 1,899

Trimester 3

n = 1,676

Percent [95% CI]

Percent [95% CI]

Percent [95% CI]

Percent [95% CI]

Alcohol

57.4

49.0***

10.5***

18.5

7.3†***

6.9†***

53.9

[56.7,58.0]

[47.8,50.2]

[9.2,12.0]

[15.6,21.9]

[5.7,9.4]

[4.9,9.4]

[53.4,54.5]

Binge alcohol

25.8

17.1***

3.2***

7.6

1.6†***

1.0†***

23.3

[25.3,26.3]

[16.3,17.9]

[2.6,3.9]

[6.0,9.6]

[1.0,2.5]

[0.5,2.2]

[22.9,23.7]

Cigarettes

32.6

25.3***

16.8***

22.5

14.1†***

14.7†***

30.6

[32.0,33.1]

[24.4,26.3]

[15.4,18.4]

[19.5,25.9]

[12.2,16.2]

[12.4,17.3]

[30.1,31.1]

Daily cigarettes

20.1

16.3***

10.6***

12.4

10.3

9.3†*

19.0

[19.7,20.6]

[15.5,17.0]

[9.4,11.8]

[10.4,14.7]

[8.6,12.3]

[7.4,11.6]

[18.6,19.5]

Marijuana or hashish

7.3

4.6***

2.8***

4.5

2.6†**

1.5†***

6.6

[7.0,7.6]

[4.2,5.1]

[2.3,3.3]

[3.5,5.7]

[1.8,3.6]

[0.9,2.4]

[6.4,6.9]

Marijuana on six or more days

3.7

2.7***

1.8***

3.1

1.3†***

1.1†***

3.5

[3.6,3.9]

[2.3,3.0]

[1.4,2.2]

[2.3,4.2]

[0.9,1.9]

[0.6,2.0]

[3.3,3.6]

Psychotherapeutics

4.0

2.6***

1.1***

2.1

0.8†**

0.6†**

3.6

[3.8,4.2]

[2.3,3.0]

[0.8,1.5]

[1.4,3.3]

[0.5,1.2]

[0.3,1.4]

[3.5,3.8]

Cocaine (including crack)

1.2

0.6***

0.3***

NR

NR

NR

1.1

[1.1,1.3]

[0.4,0.8]

[0.2,0.4]

NR

NR

NR

[1.0,1.2]

Three or more of eight itemsb

6.5

3.9***

1.4***

2.8

1.1†***

0.4†***

5.8

[6.2,6.7]

[3.5,4.3]

[1.1,1.7]

[2.1,3.9]

[0.7,1.6]

[0.1,1.1]

[5.6,6.0]

*** The difference between this estimate and the estimate for not parenting, non-pregnant women was significant at the 0.001 level

†*** The difference between this estimate and the estimate for trimester 1 was significant at the 0.001 level

†** The difference between this estimate and the estimate for trimester 1 was significant at the 0.01 level

†* The difference between this estimate and the estimate for trimester 1 was significant at the 0.05 level

aParenting women were defined as the mother of at least a 0- to 2-year-old child. “Not parenting” women included those not having a 0- to 2-year-old child even if they had an older child in the household

bThe eight items included (1) alcohol, (2) cigarettes, (3) marijuana/hashish, (4) psychotherapeutics (i.e., pain relievers, tranquilizers, stimulants and sedatives used nonmedically), (5) cocaine (including crack), (6) heroin, (7) hallucinogens and (8) inhalants

NR = Estimate not reported because of low precision

Of the three pregnancy/parenting groups, women who were neither parenting a young child nor pregnant reported alcohol, cigarette, marijuana, psychotherapeutic, cocaine use or three or more substances at the highest rates. Compared to this group, pregnant women reported use of all these substances at significantly lower rates (P < 0.001). A sizeable proportion of pregnant women reported using substances in the first trimester. Women in the second trimester were less likely than those in the first trimester to report past 30-day use of alcohol (P < 0.001), binge alcohol (P < 0.001), marijuana (P < 0.01), marijuana on six or more days (P < 0.001), psychotherapeutics (P < 0.01) and three or more substances (P < 0.001). Results were similar for the third trimester. One notable exception to this general finding was that the difference in the prevalence of daily cigarette smoking between the first and second trimesters was not significant although the second and third semester difference was significant (P < 0.05). Among pregnant women, cigarette smoking was the most prevalent substance use behavior.

Parenting, non-pregnant women also reported substance use at lower rates (P < 0.001), but their rates were much higher than pregnant women’s. Additional tabulations were done to examine rates among mothers of older children in the household. When compared with women parenting a 0- to 2-year-old child, those not parenting a 0- to 2-year-old child but living with at least a 3–5, 6–11 or 12- to 17-year-old child in the household reported substance use at higher rates: alcohol (52.5%, 52.8% and 51.8%, respectively) and cigarette use (29.1%, 29.7% and 34.3%, respectively), but slightly lower rates of marijuana use (3.9%, 3.5% and 3.4%, respectively). The rate of marijuana use was 10.7% among women living with a ≥18-year-old child or no child. For each pregnancy/parenting group, past 30-day alcohol, cigarette and marijuana use rates were examined across survey year and no evidence of trend in the prevalence rate was found (results not shown).

Table 3 shows the prevalence estimates of past 30-day cigarette, alcohol and selected illicit drug use among women by race/ethnicity. In general, NH whites reported use at the highest rate, followed by NH blacks and Hispanics. In regard to seven of the nine measures, NH blacks reported use significantly at lower rates than NH whites did (P < 0.001); the black-white difference for the remaining two measures (marijuana use on six or more days and cocaine use) were not statistically significant. On each measure, Hispanics were less likely than NH whites to report substance use (P < 0.001). They were also less likely than NH blacks to report use of these substances (P < 0.001) although the difference was not significant for use of binge alcohol, psychotherapeutics and cocaine.
Table 3

Prevalence estimates of past 30-day alcohol, cigarette and selected illicit drug use and 95% confidence intervals [CI] by race/ethnicity among women aged 18–44, NSDUH 2002–2006, United States

Substance used in past 30 days preceding interview

Percent [95% CI]

P-value for the difference in prevalence estimates between

Non-Hispanic White

n = 61,747

Non-Hispanic Black

n = 12,282

Hispanic

n = 13,640

Totala

n = 94,483

Non-Hispanic White vs. Non-Hispanic Black

Non-Hispanic White vs. Hispanic

Non-Hispanic Black vs. Hispanic

Alcohol

61.3

43.8

39.0

53.9

0.000

0.000

0.000

[60.7,62.0]

[42.4,45.2]

[37.6,40.5]

[53.4,54.5]

Binge alcohol

26.5

18.8

18.0

23.3

0.000

0.000

0.264

[25.9,27.0]

[17.8,20.0]

[17.0,19.0]

[22.9,23.7]

Cigarettes

35.7

25.5

19.3

30.6

0.000

0.000

0.000

[35.1,36.3]

[24.2,26.8]

[18.2,20.4]

[30.1,31.1]

Daily cigarettes

24.0

13.8

7.3

19.0

0.000

0.000

0.000

[23.5,24.5]

[12.8,14.9]

[6.6,8.0]

[18.6,19.5]

Marijuana or hashish

7.7

6.5

3.8

6.6

0.000

0.000

0.000

[7.4,8.0]

[6.0,7.1]

[3.4,4.3]

[6.4,6.9]

Marijuana on six or more days

3.9

3.9

1.7

3.5

0.811

0.000

0.000

[3.7,4.1]

[3.5,4.3]

[1.5,2.0]

[3.3,3.6]

Psychotherapeutics

4.2

2.6

2.8

3.6

0.000

0.000

0.486

[4.0,4.4]

[2.2,3.1]

[2.4,3.3]

[3.5,3.8]

Cocaine (including crack)

1.2

1.0

0.7

1.1

0.231

0.000

0.215

[1.1,1.3]

[0.7,1.3]

[0.6,1.0]

[1.0,1.2]

Three or more of eight itemsb

6.9

4.8

3.2

5.8

0.000

0.000

0.000

[6.6,7.2]

[4.3,5.3]

[2.8,3.6]

[5.6,6.0]

aThis included non-Hispanic Whites, non-Hispanic Blacks, Hispanics and non-Hispanic others (i.e., non-Hispanic Native American/Alaska Native, Native Hawaiian/other Pacific Islanders, Asian and non-Hispanic more than one race). No results were separately reported for non-Hispanic others

bThe eight items included (1) alcohol, (2) cigarettes, (3) marijuana/hashish, (4) psychotherapeutics (i.e., pain relievers, tranquilizers, stimulants and sedatives used nonmedically), (5) cocaine (including crack), (6) heroin, (7) hallucinogens and (8) inhalants

As part of logistic regression analysis, the odds ratios (OR’s) associated with the pregnancy and parenting predictors were compared between unadjusted (not shown) and adjusted models for each race/ethnicity group. The OR’s remained almost unchanged even after adjusting for socioeconomic and demographic characteristics. Table 4 presents results from the 36 adjusted models. The adjusted OR’s were examined as measures of the association of pregnancy and parenting statuses with past 30-day substance use. Overall, consistent with the bivariate analysis, the pregnancy effect sizes in regard to all nine outcome measures were low, suggesting lower alcohol, cigarette, marijuana, psychotherapeutic and cocaine use associated with pregnancy. However, the odds of cigarette and marijuana use associated with pregnancy were not as low as the odds of alcohol use, suggesting relatively low level of abstinence from use of these substances. There were no large variations in the OR associated with pregnancy across race/ethnicity with two exceptions. The OR’s for psychotherapeutics for NH blacks (Model 27) and cocaine use for Hispanics (Model 32) were not statistically significant, suggesting no significant pregnancy-related abstinence from use of the particular substance by NH blacks or Hispanics.
Table 4

Adjusted odds ratios (OR) and 95% confidence intervals (CI) of past 30-day alcohol, cigarette or selected illicit drug use (36 modelsa) associated with pregnancy and parenting at least a 0- to 2-year-old child by race/ethnicity among women aged 18–44, NSDUH 2002–2006, United States

Dependent variables and covariates

All racesb

OR [95% CI]

Non-Hispanic White

OR [95% CI]

Non-Hispanic Black

OR [95% CI]

Hispanic

OR [95% CI]

Alcohol use

Model 1

Model 2

Model 3

Model 4

    Pregnant

.08***

0.07***

0.11***

0.08***

[0.07,0.10]

[0.06,0.09]

[0.07,0.15]

[0.05,0.13]

    Parenting at least a 0- to 2-year-old

0.74***

0.73***

0.82**

0.78**

[0.70,0.79]

[0.68,0.78]

[0.71,0.95]

[0.66,0.91]

Binge alcohol use

Model 5

Model 6

Model 7

Model 8

    Pregnant

0.10***

0.07***

0.21***

0.11***

[0.08,0.12]

[0.06,0.10]

[0.13,0.34]

[0.06,0.19]

    Parenting at least a 0- to 2-year-old

0.64***

0.64***

0.72***

0.71***

[0.60,0.68]

[0.59,0.69]

[0.60,0.86]

[0.58,0.85]

Cigarette use

Model 9

Model 10

Model 11

Model 12

    Pregnant

0.41***

0.43***

0.43***

0.27***

[0.37,0.46]

[0.38,0.49]

[0.33,0.57]

[0.19,0.40]

    Parenting at least a 0- to 2-year-old

0.73***

0.79***

0.92

0.62***

[0.69,0.78]

[0.74,0.85]

[0.78,1.09]

[0.53,0.74]

Daily cigarette use

Model 13

Model 14

Model 15

Model 16

    Pregnant

0.48***

0.49***

0.45***

0.32***

[0.42,0.54]

[0.42,0.57]

[0.31,0.66]

[0.20,0.52]

    Parenting at least a 0- to 2-year-old

0.80***

0.88**

0.91

0.64***

[0.75,0.86]

[0.81,0.95]

[0.73,1.13]

[0.50,0.81]

Marijuana use

Model 17

Model 18

Model 19

Model 20

    Pregnant

0.37***

0.36***

0.41***

0.45*

[0.31,0.45]

[0.29,0.46]

[0.28,0.59]

[0.22,0.94]

    Parenting at least a 0- to 2-year-old

0.69***

0.68***

0.87

0.67*

[0.62,0.77]

[0.59,0.78]

[0.67,1.12]

[0.49,0.91]

Use of marijuana on six or more days

Model 21

Model 22

Model 23

Model 24

    Pregnant

0.46***

0.49***

0.41***

0.50*

[0.36,0.58]

[0.36,0.66]

[0.26,0.64]

[0.26,0.95]

    Parenting at least a 0- to 2-year-old

0.75***

0.75**

0.83

0.85

[0.65,0.87]

[0.63,0.90]

[0.59,1.17]

[0.54,1.33]

Use of psychotherapeutics

Model 25

Model 26

Model 27

Model 28

     Pregnant

0.28***

0.25***

0.78

0.22***

[0.20,0.39]

[0.18,0.34]

[0.33,1.88]

[0.09,0.54]

    Parenting at least a 0- to 2-year-old

0.69***

0.73***

0.80

0.69

[0.59,0.81]

[0.60,0.88]

[0.48,1.33]

[0.45,1.07]

Cocaine use

Model 29

Model 30

Model 31

Model 32

    Pregnant

0.24***

0.28***

0.19*

0.30

[0.14,0.40]

[0.14,0.52]

[0.04,0.89]

[0.09,1.04]

    Parenting at least a 0- to 2-year-old

0.60***

0.67*

1.22

0.28**

[0.44,0.80]

[0.49,0.92]

[0.44,3.38]

[0.13,0.62]

Use of three or more of eight itemsc

Model 33

Model 34

Model 35

Model 36

    Pregnant

0.21***

0.22***

0.18***

0.21***

[0.16,0.27]

[0.17,0.30]

[0.06,0.49]

[0.09,0.45]

    Parenting at least a 0- to 2-year-old

0.66***

0.69***

0.88

0.53***

[0.59,0.73]

[0.60,0.79]

[0.64,1.21]

[0.38,0.74]

Reference Categories for pregnant and parenting covariates were non-pregnant and not parenting a 0- to 2-year-old, respectively

*** P < .001; ** P < .01; and * P < .05

aEach Model adjusted for age, marital status, education and annual family income

bThis included non-Hispanic Whites, non-Hispanic Blacks, Hispanics and non-Hispanic others (i.e., non-Hispanic Native American/Alaska Native, Native Hawaiian/other Pacific Islanders, Asian and non-Hispanic more than one race). No results were separately reported for non-Hispanic others

cThe eight items included (1) alcohol, (2) cigarettes, (3) marijuana/hashish, (4) psychotherapeutics (i.e., pain relievers, tranquilizers, stimulants and sedatives used nonmedically), (5) cocaine (including crack), (6) heroin, (7) hallucinogens and (8) inhalants

Overall, the parenting effect sizes were significantly low (P < 0.001) but not as low as the pregnancy effect sizes, reflecting its relatively weak relationship with reported past 30-day use. This was true for NH whites. This was also true for Hispanics except that the parenting effect sizes with respect to marijuana use on six or more days (Model 24) and use of psychotherapeutics (Model 28) were not significant. For NH blacks, statistically significant results associated with parenting a young child were limited to alcohol use and the results on the remaining eight measures were not significant.

Discussion

The study provides corroborating evidence from a large, nationally representative sample that American women who were pregnant were substantially less likely than those who were neither pregnant nor parenting a young child (0–2 years) to report past 30-day alcohol, cigarette, marijuana, psychotherapeutic and cocaine use. However, it is troubling to note that cigarette use was reported by nearly a quarter of pregnant women in their first trimester and that one in seven reported cigarette use in the second or third trimester. A recent analysis attributed nicotine dependence as an important barrier to smoking cessation among pregnant women [32].

Logistic regression analysis suggested a strong negative association between pregnancy status and substance use. The association between parenting status and substance use was in the same direction although of much lower magnitudes. These results were consistent with the bivariate results reported in previous studies [25, 26, 27, 28, 29]. In this study, the postpregnancy substance use behavior of the pregnant women could not be directly measured, but was assumed to be similar to those of parenting women, particularly in light of the statistical controls applied in the logistic regression models. Thus, based on the comparison of substance use behavior of pregnant and parenting women, the study provides indirect evidence that pregnancy-related abstinence from substance use wanes out, with resumption in use in less than 3 years after child birth. Distinguishing between any versus intense use (i.e., any alcohol versus binge alcohol use; any cigarette versus daily cigarette smoking; any marijuana versus marijuana on six or more days) did not produce any considerable difference in the magnitude of the association with pregnancy or parenting.

Regardless of the pregnancy or parenting status, NH whites reported alcohol, cigarette and other substance use at highest rates followed by Hispanics and NH blacks with two exceptions. There was no significant difference between NH white and NH black women in regard to marijuana use on six or more days and cocaine use and no difference between black and Hispanic women in regard to binge alcohol, psychotherapeutic and cocaine use. In contrast to the general findings as referred to above, for most measures, there were no large variations in the magnitude of the relationship between pregnancy and substance use by race/ethnicity. Lower odds of substance use associated with parenting were significant for both whites and Hispanics. However, for NH blacks, significant results associated with parenting were limited to alcohol use. This suggested that NH blacks were more likely than NH whites or Hispanics to resume use of most substances.

The strengths of the study include the large, nationally representative sample, permitting the estimation of prevalence of relatively rare substance use behaviors in pregnancy/parenting categories and by race/ethnicity. Several important limitations should be noted. First, the data did not permit us to examine the span of the pregnancy interval that was not exposed to substance use and did not allow us to use direct prepregnancy and postpregnancy substance use measures. Second, although previous research has found underreporting of pregnant women’s illicit drug use and smoking on surveys [34], respondents’ self-reports used in this analysis could not be directly verified. However, differentials in underreporting of substance use between pregnant and non-pregnant women were unlikely because pregnancy-related questions were asked and answered by the respondent after substance use-related questions during interview. Third, because this was an observational study by design, no causal inferences were made. One may question whether self-selection bias could be introduced to our results as the analysis relies on cross-sectional data. More specifically, lower rates of substance use among pregnant women could be the result of a greater likelihood of becoming pregnant among non-drug users. To address this issue, the percentages of women who reported being pregnant at the time of interview among lifetime users and non-users were compared: alcohol (4.1% vs. 5.3%; P = 0.001); cigarettes (4.1% vs. 4.6%; P = 0.031); and marijuana (4.0% vs. 4.6%; P = 0.004). The between-group differences in the percentage pregnant at a given day were small although significant, suggesting no strong evidence of selection bias. Fourth, higher rates of substance use for women who were parenting young children (but not pregnant) than those who are pregnant could be due to increases in initiation (i.e., first-time use) occurring after childbirth rather than resumption in use among former users. To address this concern, past-year initiation rates among parenting women were estimated as follows: 0.9% for alcohol, 0.4% for cigarette, and 0.3% for marijuana use. These initiation rates are too small to explain the differences in the past-30-day use rates between pregnant and parenting women, supporting the conclusion of a postpregnancy resumption in use.

In conclusion, there is indirect evidence of postpregnancy resumption in substance use overall and in each of the three race/ethnicity groups examined. The resumption in use appears to be most prominent among NH black women, with respect to most substances. Implementing effective prevention programs to prevent the resumption in use could improve the overall health and well-being of mothers and their children. Prevention messages targeting expectant mothers should not be limited by a narrow focus on protecting just the fetus from exposure to harmful substances. The consequences of parental substance abuse on child growth and development should be emphasized. Treatment programs that provided integrated prenatal care and substance abuse treatment services, compared with program providing just prenatal care, produced improved prenatal outcomes [35]. Prenatal care-seeking behavior of drug using pregnant women is often impaired because of their poor psychosocial conditions and unstable living arrangement [36], and some previous research suggested targeting this group when planning a drug education screening strategy during pregnancy [37]. Given the findings of this study that sizeable proportions of women who reported being in the first trimester of pregnancy used cigarettes, alcohol and illicit drugs, prenatal care and substance abuse treatment services should include screening for nicotine dependence, and alcohol abuse and drug abuse treatment. The goal of the substance use prevention programs for American women in childbearing ages should be cessation rather than temporary abstinence from substance use.

Acknowledgements

The authors gratefully acknowledge Jessica Campbell Chambers, James Colliver and three anonymous reviewers for their thoughtful comments on earlier versions of this manuscript.

Copyright information

© Springer Science+Business Media, LLC 2008

Authors and Affiliations

  1. 1.Division of Population Surveys, Office of Applied StudiesSubstance Abuse & Mental Health Services AdministrationRockvilleUSA
  2. 2.Division of Population Surveys, Office of Applied StudiesSubstance Abuse & Mental Health Services AdministrationRockvilleUSA

Personalised recommendations