Abortion, Reproductive History and Substance Abuse

Editor's Note: The following five studies examine the association between substance use/abuse and reproductive history, including elective abortion. Their findings sometimes have been misinterpreted as suggesting that substance abuse is an effect of abortion. With the exception of Reardon & Ney, none of these researchers reached this conclusion. The evaluation of each study indicates why the assumption that abortion causes substance abuse is unwarranted. In sum, data were primarily retrospective, potential confounding variables were not considered or, when considered, the relationship between abortion and substance use disappeared. It should be noted that the Amaro, Zuckerman & Cabral (1989) and Frank et al. (1988) articles analyze different samples from the same data set, and the Klassen & Wilsnack (1986) and the Wilsnack, Wilsnack & Klassen (1984) articles are also from the same general data set.

Citation: Reardon, R. C. [Elliot Institute, P.O. Box 7348, Springfield, IL 62791-7348], & Ney, P.G. (2000). Abortion and subsequent substance abuse. American Journal of Drug and Alcohol Abuse, 26, 61-75.

Introduction: Past studies suggest a possible association between abortion and substance abuse but the reasons for this relationship are not clear. This study attempted to clarify the relative risk of post-abortion substance abuse.

Method: An exploratory survey was sent to a national sample of 4929 women aged 24-44 was randomly selected from a mailing house list. The questionnaire centered on reproductive history and asked about age at each pregnancy and pregnancy outcome. Also included were items measuring whether drugs or alcohol had ever been abused (yes-no), ages of onset and termination of substance abuse, demographic characteristics and emotional reaction to the survey. Fourteen percent of the surveys (700) were returned; non-white women were substantially underrepresented in the sample. Groups (substance abusers and non-substance abusers, women who had abortions and women with other pregnancy outcomes) were compared using chi-square.

Results: Fifteen percent of all pregnancies in the sample were reported to terminate in abortion, which suggests that some respondents concealed their abortions and/or women who had abortions were under-represented among survey completers. The analysis, which focused on outcome of first pregnancy, was restricted to 491 white women who either never had an abortion or terminated their first pregnancy by abortion. Results indicated that women who abused alcohol or drugs were younger at first intercourse and first pregnancy and more likely to have had an abortion. Thirty-five percent of participants with a history of substance abuse reported onset of substance abuse "subsequent" to first pregnancy, defined as occurring in the same year or later (27 women). When only those with "subsequent" substance abuse were included, women who had an elective abortion were found to be 4-5 times more likely to report post-pregnancy onset of substance abuse than women with births or pregnancy loss (16.5% versus 3.7%). Women who had an abortion or reported substance abuse also indicated they experienced more emotional difficulty in answering the survey.

Evaluation: This study's limited measurement of onset, type and degree of abuse, along with its retrospective design, undermine its ability to adequately examine the temporal ordering of abortion and substance abuse. Other significant limitations include its failure to consider demographic and reproductive history differences between the comparison groups, small sample of substance abusers, very low response rate and analysis limited to white women. This research certainly does not support the contention that having an abortion causes alcohol or drug abuse. Moreover, it provides no data to support the authors' assumption that abortion is a psychological stressor that can be relieved through substance use. Better-designed research is necessary to understand the circumstances under which an association exists between history of abortion and substance abuse and the reasons for such an association. Practical applications recommended by the authors include evaluation of women seeking prescription drugs and women entering substance abuse treatment programs for issues related to unresolved post-abortion grief, and informing women seeking abortion about the increased risk of post-abortion substance abuse. As there is no reputable research evidence showing that abortion causes substance abuse, these recommendations are unwarranted. Rather it is desirable to screen all women seeking reproductive services (including pregnancy termination and prenatal services) for alcohol and drug abuse and promptly provide treatment for those in need.

Citation: Wilsnack, R.W. [Dept. of Sociology, University of North Dakota, Grand Forks, ND, 58202], Wilsnack, S.C., & Klassen, A.D. (1984). Women's drinking and drinking problems: Patterns from a 1981 national survey. American Journal of Public Health, 74, 1231-1238.

Introduction: At the time of this data collection (1981), increased attention had been focused on women's drinking without clear indication that drinking among women had actually increased. This study addresses some limitations of previous research by using a large national representative sample of women to examine drinking levels, drinking behavior, and related health behaviors and symptoms of alcohol dependence.

Methods: Questionnaires were administered to 917 women and 396 men (as a control group), and included items about alcohol consumption, drinking contexts, problems resulting from drinking, and symptoms of alcohol dependence. Also included were items on role performances, life stressors, symptoms of anxiety and depression, and physical health. No demographic characteristics were reported for the sample, although some of these variables were considered in the analysis.

Results: No clear increase in drinking was found, although more women over age 50 were abstinent and women between 35 and 64 drank more than previously. Drinkers or heavy drinkers were more likely to be of higher education or income, Caucasian, and of no indicated religious preference. Marital status appeared relatively salient, with divorced, separated, never married, or cohabitating women being more likely to drink heavily. It was also noted that women often shared the drinking habits of their husband or partner and close friends. Patterns also indicated that many women may experience problems or symptoms related to heavy drinking and abstain temporarily in response to these problems. Current drinking was not related to serious injury or illness, although drinking in the year preceding pregnancy was related to more premature deliveries, miscarriages, stillbirths, birth defects, and failure to become pregnant. No mention was made of prevalence of elective abortion. Chronic heavy drinking was also related to depressive symptoms.

Evaluation: As these data are relatively dated, especially with regard to the research question, no suggestions about current drinking trends among women can be made. Whatever the current trends and relationships, however, the authors make a solid point that conclusions about direction of effect with regard to any correlates of heavy drinking are impossible with this type of data. This is especially true of what the authors term "obstetrical and gynecological problems", which in this study includes premature delivery, miscarriage, and stillbirth, but not abortion. Recency of the data aside, this study is a good methodological template for future research, and sheds light on characteristics of women and their partners/companions that may be salient when assessing for alcohol abuse.

Citation: Amaro, H. [School of Public Health, Boston University School of Medicine, 85 E. Newton St., M-840, Boston, MA 02118], Zuckerman, B., & Cabral, H. (1989). Drug use among adolescent mothers: Profile of risk. Pediatrics, 84, 144-151.

Introduction: Although drug use and pregnancy are known to be related in the adolescent population, few studies have systematically and prospectively examined this relationship. This study was conducted to identify drug use patterns in currently pregnant adolescents and to compare characteristics between drug users and non-drug users.

Methods: Data was collected on 253 pregnant females between the ages of 13 and 19. The sample was mostly African-American and Latina, unmarried, and low-income. The majority were not living with the father of the baby. A combination of self-report and urine assay was used to determine lifetime drug use, use within the past year, and use during pregnancy. Data was also collected regarding demographics, current depression, life stress, social support, history of violence, and drug use by the father of the baby. Data for users and non-users was comparison analyzed using chi-square, t-tests, and regression analysis.

Results: The majority of this sample has used alcohol and marijuana both before and during pregnancy, with cocaine use occurring to a lesser extent. Drug users here were more likely to be older, married, or living with the father of the baby. They were also more likely to have had an elective abortion in the past, to have had a sexually-transmitted disease, and to report negative life events that led to negative effects for them. Further analysis using regression to control for age, ethnicity, and other psychosocial variables showed no significant relationship between drug use and history of elective abortion. Amount of social support between the two groups did not differ, but drug users were more likely to receive their social support from their male partners (who are more likely to be drug users). Drug users were also more likely to be threatened, abused, or involved in fights during pregnancy. No differences in current depression were found.

Evaluation: This study represents a strong contribution to the literature given its combination of self-reported drug use and urine assay to address previous underidentification of drug use. Elective abortion was examined as a variable here due to previous findings of drug use among teens who become pregnant compared to teens who are not sexually-active. One weakness of the paper is that elective abortion is mentioned as something that provides added risk for poor outcomes for future pregnancies, but not supported in the text by any literature references. Practical applications provided by the authors are strong, and consist of including an adolescent's male partner (likely also a drug user) in her drug treatment, recognizing drug use as a potential risk factor in any adolescent pregnancy, and training staff to adequately assess for adolescent drug use.

Citation: Frank, D.A. [Division of Developmental and Behavioral Pediatrics, FGH Bldg 3, Boston City Hospital, 818 Harrison Ave., Boston, MA 02118], Zuckerman, B.S., Amaro, H., Aboagye, K., Bauchner, H., Cabral, H., Fried, L., Hingson, R., Kayne, H., Levenson, S.M., Parker, S., Reece, H., and Vinci, R. (1988). Cocaine use during pregnancy: Prevalence and correlates. Pediatrics, 82, 888-895.

Introduction: Much research has targeted the potential negative effects of cocaine during pregnancy, although methodological problems have prevented strong recommendations from being made. This study uses a non-clinical (i.e., not seeking drug treatment) sample of women seeking routine prenatal care and also targets potentially confounding variables that have not been addressed previously. The study was designed to assess prevalence and patterns of cocaine use, identify other correlates of cocaine use, and to compare women's self-report of cocaine use to urine assay identification.

Methods: Interviews were conducted with 679 women prenatally and in the immediate postpartum period. Participants were primarily low-income, African-American or Latina, and unmarried. A minority of women in this sample (comprised of drug users and non-drug users) also had histories of what the authors termed "risk factors", including sexually-transmitted diseases (STDs--19%), low birth weight infants (9%), stillborn infants (9%), and previous spontaneous (22%) or induced abortion (36%). Interview questions centered around use of cocaine, marijuana, alcohol, and tobacco during pregnancy, as well as the substance use of the father of the baby. Participants were also asked for a urine sample at each interview, which was tested for marijuana and cocaine metabolites. Cocaine users and non-users were compared using Chi-square and t-tests.

Results: Data from self-report and urine assay combined showed the following prevalence rates: cocaine 17%, marijuana 28%, tobacco 44%, alcohol 59%, opiates 4%, and other illicit drugs 3%. Of the cocaine users, 24% denied use in self-report, but were detected by urine assay. Compared cocaine nonusers, cocaine users were more likely to be unmarried, Caucasian, and less well nourished. Users also reported significantly more STDs, prior low birth weight infants, spontaneous abortions (one or more), elective abortions (two or more), and greater use in frequency and amount of alcohol, tobacco, marijuana, opiates, and other illicit drugs during pregnancy.

Evaluation: This study's particular strength is that it surveyed women who were merely seeking routine obstetrical care and not drug treatment. It also solidifies the importance of using urine testing to identify cocaine users, at least with regard to conducting further research in this area. The authors' conclusion that cocaine use is common enough to include in routine screening for prenatal patients is certainly justified. Given the co-occurrence of other substances with cocaine in prenatal patients, this study provides important information on patients who might be at risk for cocaine use as well as important confounding variables to consider for future research in this area. With regard to rates of elective abortion, it is important to note that this study found that users and non-users were not statistically different in their likelihood of having had one elective abortion. The increased likelihood was only found for cocaine users with regard to two or more abortions. This might indicate that cocaine users are simply more likely to experience repeated unintended pregnancies. As this study did not report data on number of previous pregnancies, pregnancy rates for users versus nonusers are unknown.

Citation: Klassen, A.D. [Department of Sociology, University of North Dakota, Grand Forks, N.D., 58202], and Wilsnack, S.C. (1986). Sexual experience and drinking among women in a U.S. national survey. Archives of Sexual Behavior, 15, 363-392.

Introduction: Research examining the relationship between alcohol and sexual behavior has consisted of the almost exclusive study of male subjects, and has also been limited by small non-representative samples. This study used a large representative sample of women to address sexual feelings, behaviors, and dysfunctions according to differing levels of alcohol consumption.

Methods: Questionnaires were administered to 917 women and 396 men (as a control group), and included items on current and lifetime alcohol consumption, demographics, sexual experience, and other psychosocial variables. Several measures for drinking levels were used to categorize women as moderate-to-heavy, light-drinking, abstaining, or temporarily-abstaining women. Sexual experience was measured by using various techniques based on general principles of Kinsey's sex research work. No demographic information was provided for the sample, and was apparently not considered in the analysis.

Results: A majority of women reported that drinking reduces sexual inhibitions and makes sexual activity more pleasurable. A minority of women reported that drinking increases sexual assertiveness and less particular choice of sexual partners, while more than half of all women here reported that someone who had been drinking became more sexually aggressive with them. Moderate-to-heavy drinkers were more likely to report premarital intercourse, masturbation to orgasm, possible interest in homosexuality, and abortion. Relationships between drinking and sexual dysfunction were relatively weak, although this was a non-clinical (i.e., non-alcoholic) sample.

Evaluation: The representative nature of the study is certainly a strength. While this was a retrospective study, and thus cannot imply direction of effect, the results here are important to consider in any women's health care setting. Acknowledging that alcohol use can be an antecedent, concomitant, or consequence of sex-related or reproductive issues can be useful. In terms of alcohol treatment programs, it is also important to recognize the role of alcohol in sexual experience and behavior. What is found here, though, is an overall theme of more "sexually liberated" behavior among the moderate-to-heavy drinkers that encompasses a wide range of behaviors. The authors propose either a disinhibiting effect of alcohol, a different moral value system among these women, or both as an explanation. The study certainly does not provide any reason to assume that having an abortion causes women to consume more alcohol. Obviously a major limitation of the study is its recency, as the data was collected in 1981.