Depression associated with abortion and childbirth: A long-term analysis of the NLSY cohort

Abstract
 

Available research indicates that although abortion is generally a stressful event, most women do not suffer from severe negative reactions at the time of the initial post-abortion assessment. Among the few studies examining reactions over a year after the abortion, however, Major et al (2000) and Miller et al (1998), it has been discovered that delayed reactions are not uncommon. In addition, the trend line over time is toward increasing negative emotions and regret. Longitudinal studies examining a period of time in excess of one year have been recommended.

Our study employed the National Longitudinal Survey of Youth (NLSY), a general purpose study, which has interviewed 6283 women since 1979. It contains psychological assessments administered in 1979 (the Rotter Internal-External Locus of Control scale) and 1992 (the Center for Epidemiological Studies Depression scale) and variables related to pregnancy outcome.

Compared to post-childbirth women, aborting women (n=735) were found to have significantly higher depression scores as measured an average of10years after their pregnancy outcome. Higher depression scores could not be accounted for by confounding variables.

Controlling for age, total family income, and locus of control scores prior to the first pregnancy event, post-abortive women were found to be 41% more likely than non-aborting women to score in the "high-risk" range for clinical depression. Furthermore, in response to a self-assessment question administered in 1998, aborting women were 73% more likely to complain of "depression, excessive worry, or nervous trouble of any kind" an average of 17 years post-abortion.

A weakness of this study is that only 40% of the expected abortions are reported the NLSY population. Since concealment would tend to dilute the observed effects, the actual rate of depression in the "high-risk" range may be considerably higher than suggested by this data. Concealment of abortions in the NLSY was lowest among whites and married women. As expected, analysis by race and marital status confirmed that the differences in depression scores were highest among the groups with the least concealment.

Pre-abortion and pre-childbirth locus of control scores were analyzed. As predicted, the locus of control scores correlated well with depression scores for both nulliparous and childbearing women but this correlation disappeared for aborting women. This suggests that aborting women may either (1) experience a shift toward a more external locus of control, or (2) experience a different type of depression than other women.

Additional longitudinal research is warranted. Implications for future research are discussed.
 

Background

• Post-abortion adjustment has been studied extensively, but most investigations have been limited in scope and have significant methodological limitations. (Koop, 1989)

• Recent research suggests the need for more long-term studies of post-abortion adjustment:

1. In a longitudinal investigation with women involved in a clinical trial of RU-486, Miller, Pasta & Dean (1998) found that regret increased from 2 weeks to 6-8 after the abortion. In regard to delayed reactions, they concluded that "there is very likely a time course associated with the coping process that follows abortion and this time course almost certainly does not show a straight line of improvement. In other words, the low point following the abortion may not occur for days, weeks, or even months" (p. 262).

2. Most recently, Major and colleagues (2000) analyzed the psychological state of women one hour pre-abortion, and one hour, one month, and two years post-abortion. They found an increase in negative emotions and a decrease in relief and positive emotions between the assessments at one hour post-abortion and two years post-abortion. In addition, depression increased and satisfaction with the abortion decision decreased over time.


 
• Two of the largest and most extensive studies of post-abortion adjustment were carried out using the National Longitudinal Survey of Youth (Russo & Zierk, 1992; Russo & Dabul, 1997). The authors found comparable Rosenberg self-esteem scores between aborting and non-aborting women. From this single finding, the authors drew the broad conclusion that abortion has no independent effect on women's well-being. That sweeping conclusion, however, is not actually supportable using the data presented by the authors for a number of reasons:

1. The Rosenberg Self Esteem scale is not a sensitive measure of post-abortion adjustment. The results of Miller and colleagues' (1998) investigation suggested that self-esteem scores remained virtually unchanged at three different time points while other emotional states varied significantly. Similarly, Major and colleagues (2000) found that, among post-abortive women, self-esteem scores increased over time even while negative reactions were increasing.

2. Analysis of the abortion rate reported by the NLSY cohort compared to national statistics compiled by the Alan Guttmacher Institute has shown that women in the NLSY report approximately 40 percent of abortions that would be expected (Jones & Forrest, 1992). Significant differences were found in reporting among married and unmarried women (74% vs 30%). White women (45%) were also much more likely to report an abortion than Blacks (27%) and Hispanics (19%).

3. Post-abortion research has found that women who conceal their abortion experience from others are more likely to suppress thoughts of the abortion, experience more intrusive abortion-related thoughts, and feel greater psychological distress (Major & Gramzow, 1999). This suggests that the women who admitted having an abortion in the NLSY study may have been less likely to experience psychological distress than those who concealed their abortions.

4. The authors neglected to examine other psychiatric measures available in the NLSY.

Method

• The NLSY was used to examine post-abortion and post-childbirth adjustment among 1,971 women.
• The twenty item Center for Epidemiological Studies Depression scale (CES-D; Radloff, 1977) was presented to women by NLSY interviewers in 1992. This measure was used in our analysis to gauge post-abortion and post-childbirth depression. An abbreviated version of the Rotter Internal-External Locus of Control scale (Rotter, 1966) was given to respondents in 1979 and was used as a measure of pre-pregnancy mental state. Among all women in the NLSY, 1979 locus of control scores correlated with 1992 depression scores (r(4423)=.135, p<.0001)
• The sample used for our analysis included women who had their first pregnancy events between 1980 and 1992. Three hundred and four had at least one abortion and 1,667 had at least one childbirth and no abortions. The average year of both aborting and delivering women's first pregnancy event was 1984. Among all women with a history of abortion, 23.0% were Hispanic, 24.7% were Black, and 52.3% were Non-Black and Non-Hispanic. Among all women with a history of childbirth, 20.8% were Hispanic, 23.9% were Black, and 55.4% were Non-Black and Non-Hispanic. In order to examine the effects of abortion concealment, univariate and logistic regression analyses were carried out with special attention to race and marital status. Prior locus of control scores, age in 1992, and income in 1992 were used as covariates throughout.
• In 1998, two items relating to mental health were presented to a subset of the NLSY cohort. These items were also analyzed in regard to post-abortion and post-childbirth adjustment.

• Univariate analyses revealed that 1992 depression scores were significantly higher for aborting women (M=11.94, SE=0.53) than delivering women (M=9.85, SE=0.22; F(1,1966)=13.46, p<.0003).
• Among aborting women, 27.6% were found to be "at risk" of depression (CES-D score 16 or greater) compared to 21.9% of delivering women. Logistic regression analyses revealed that the odds ratio of severe depression scores among aborting women compared to delivering women was 1.40 (95% CI: 1.05 to 1.86; Wald=5.20, p<.023).
• Hispanic and black aborting women (M=13.27, SE=0.79) scored significantly higher in depression than women of the same ethnicity who carried to term (M=11.25, SE=0.35). Among whites, women with a history of abortion (M=10.89, SE=0.70) had significantly higher depression scores than delivering women (M=8.69, SE=0.29).
• Significant differences in rates of depression emerged between aborting and delivering women when examining different demographic groups. The results of these analyses are given in Table 1.
• The 1979 Rotter scale correlated significantly with scores on the 1992 CES-D scale for women who delivered with no reported history of abortion (r(1667)=.177, p<.0001). Among aborting women, however, the 1979 Rotter scale did not correlate with scores on the 1992 depression scale (r(304)=.050, p>.05). A significant correlation reflected the fact that greater external locus of control in 1979 was associated with higher scores of depression in 1992.
• A subset of the cohort women were asked two items pertaining to mental health in 1998: "Has a doctor ever told you that you had emotional, nervous, or psychiatric problems?" and "Are you currently experiencing depression or excessive worry or nervous trouble of any kind?"
• The sample was too small to include only women with Rotter scores prior to their first pregnancy event. Therefore, all women with an abortion (n=108) or delivery (n=394) prior to 1998 were included in these analyses. Both items correlated with 1992 depression scores (Diagnosis: r(480)=.215, p<.0001; Depression: r(480)=.246, p<.0001). The average year of the first abortion among post-abortive women was 1983. The average year of the first childbirth among non-aborting mothers was 1984. 
• In regard to the first question (doctor's diagnosis), 37 (9.4%) childbirth women answered in the affirmative compared to 8 (7.4%) aborting women. Controlling for age and total family income in 1998, logistic regression analyses revealed no significant differences (OR=0.77, 95% CI: 0.35 to 1.71, Wald=.41, p>.05).
• In regard to the second question (depression, excessive worry or nervous trouble), 27 (25.0%) aborting women answered in the affirmative compared to 64 (16.2%) mothers without a history of abortion. A logistic regression analysis revealed significant differences when controlling for age and income (OR=1.73, 95% CI: 1.03 to 2.89; Wald=4.36, p<.037).

Unintended Pregnancies Analyses

• In 1992, the same year as the CES-D assessment, women were asked if they intended to become pregnant before their first delivery. Women who responded "yes" or "didn't matter" were excluded from our sample. The following additional analysis was conducted to compare depression scores of women who carry an unintended pregnancy to term to women who abort.
• The final sample of women used in these analyses includes only women who had their first abortion or first unintended delivery between 1980 and 1992 and completed both the 1979 Rotter scale and the 1992 CES-D scale (n=434). On average, aborting women had their first pregnancy in 1984 at age 22, and delivering women gave birth in 1986 at age 24.
• The results of univariate and logistic regression analyses are shown in Tables 2 and 3.
• Overall, women who aborted a first unintended pregnancy had significantly higher depression scores compared to similar women who carried an unintended first pregnancy to term and did not report a history of abortion. As before, the groups in which there was the least concealment of past abortions revealed the highest relative risk for depression scores in the "high risk" range.
• Among women having unintended births, the Rotter scores correlated with depression scores (r(130)=.176, p<.045), as expected. Among aborting women, however, Rotter and CES-D scores did not significantly correlate (r(304)=.050, p>.05).

 

• The present study revealed that women with a prior history of abortion had significantly higher scores of depression than childbearing women without a history of abortion. In addition, when controlling for age, total family income, and locus of control scores prior to the first pregnancy event, post-abortive women were found to be 40% more likely to score in the "high-risk" range for clinical depression compared to mothers without a history of abortion.
• Compared to women who carried a first unintended pregnancy to term, women who aborted a first unintended pregnancy had significantly higher depression scores as measured an average of eight years after their abortions.
• Correlation between prior locus of control scores and post-pregnancy event depression was only evidenced for non-aborting mothers. These two scales did not correlate women who had at least one abortion in the intervening years (1980-1992). This is most likely due to a shift in locus of control among aborting women. It may also be attributed to a different type of depression in this group.
• The odds ratios for high-risk depression scores were highest within those groups who were most likely to report an abortion (married and White). Conversely, among the groups with the highest concealment rate (all unmarried and all Blacks and Hispanics) significant differences were lost. A number of possible explanations exist for these findings:
1. Among groups which have the highest concealment rate, women who admit an abortion are less likely to experience depression than their counterparts who conceal their abortions. This explanation is supported by the finding of Major and Gramzow (1999) that women who conceal their abortion are more likely to suppress thoughts of the abortion and feel greater psychological distress.

2. Abortion concealment may simply introduce the misclassification of aborting women as non-aborting women with high depression scores.

3. Unmarried mothers may have higher depression scores due to the inherent stress of raising a child without support.

• The results of our investigation run counter to conclusions that abortion has no independent effect on women's "well-being" drawn from prior NLSY research (Russo & Zierk, 1992; Russo & Dabul, 1997). These contradictory findings can be attributed to the insufficiency of the Rosenberg self-esteem scale in measuring emotional reactions. Future post-abortion research would benefit from the inclusion of more state-specific measures.
• Our results do not support the hypothesis that depression following unintended pregnancies is independent of outcome (birth or abortion) and primarily or solely related to prior mental history (Major et al, 2000). Women who carried an unintended first pregnancy to term are at less risk of subsequent depression compared to similar women who aborted.
• While Rotter scores correlate well with depression (Costello, 1982), our investigation was limited in that pre-pregnancy depression scores were not available. Another shortcoming is that the high rate of abortion concealment in the NLSY data set (60%)would tend to dilute the observed effects. Future longitudinal research which specifically examines reproductive events, life circumstances, and emotional health year by year is warranted.
• The present study's findings suggest that future research efforts should be directed toward the examination of long-term post-abortion reactions. Such research efforts would be beneficial in attempts to understand why abortion, or the circumstances surrounding abortion, may be problematic for some women. A greater knowledge of various abortion-related stressors would be helpful in pre- and post-abortion counseling.

Funding: Elliot Institute 

Conflict of interests: None 

Table 1
Women scoring in "high-risk" depression range who had their first abortion or first childbirth between 1980-1992 - by race and marital status.

	Delivering Women		Aborting Women		Adjusted Odds-Ratio (95%CI)*	P
	Total/N	%	Total/N	%
Minority-Unmarried	102/265	38.5%	23/79	29.1%	0.69 (0.40 to 1.21)	NS
Minority-Married	94/468	20.1%	20/60	33.3%	2.23 (1.21 to 4.11)	<.011
Minority-All	201/744	27.0%	46/145	31.7%	1.28 (0.86 to 1.90)	NS
White-Unmarried	42/146	28.8%	14/50	28.0%	1.30 (0.60 to 2.81)	NS
White-Married	115/753	15.3%	23/104	22.1%	1.79 (1.07 to 3.02)	<.028
White-All	164/923	17.8%	38/159	23.9%	1.56 (1.03 to 2.37)	<.036
All Unmarried	144/411	35.0%	37/129	28.7%	0.84 (0.54 to 1.30)	NS
All Married	209/1221	17.1%	43/164	26.2%	1.93 (1.30 to 2.85)	<.002
All races	365/1667	21.9%	84/304	27.6%	1.40 (1.05 to 1.86)	<.023

*Adjusted for age, family income in 1992, and Rotter score in 1979.

Note: Concealment of abortions was lowest among whites and married women. Among groups with the highest concealment rates, no significant findings were observed.
 
 

Table 2
Mean and adjusted depression score means of women who had their first abortion or first unintended childbirth between 1980-1992.
 

	Unintended birth women (n=130)				Aborting women  (n=304)
	Mean (SD)	Adj. Mean* (SE)	95% CI*		Mean (SD)	Adj. Mean* (SE)	95% CI*	p
Married  (n=239)	9.15 (8.4)	8.2 (1.12)	6.07 to 10.47		10.65 (10.0)	11.05 (0.73)	9.61 to 12.18	0.047
Unmarried  (n=182)	11.96 (8.1)	11.86 (1.62)	8.67 to 15.05		13.14 (12.3)	13.18 (1.00)	11.20 to 15.16	NS
All Women	10.29 (8.4)	9.57 (0.91)	7.78 to 11.37		11.89 (11.9)	12.20 (0.58)	11.05 to 13.35	0.019

* adjusting for family income, education, race, age at first pregnancy, and 1979 Rotter score
 
 

Table 3
Women scoring in "high-risk" range for clinical depression (CES-D score >15) who had their first abortion or first unintended childbirth between 1980-1992.

	Unintended birth women  high risk / total	Aborting women  high risk / total	Adjusted Odds-Ratio*	95% CI*	p
Married	13 / 75	43 / 164	2.38	1.09 to 5.22	0.031
Unmarried	16 / 53	37 / 129	0.94	0.43 to 2.03	NS
All Women	30 / 130	84 / 304	1.55	0.92 to 2.59	0.099

* adjusting for family income, education, race, age at first pregnancy, and 1979 Rotter score
 

References

Costello, E. J. (1982). Locus of control and depression in students and psychiatric outpatients. Journal of Clinical Psychology, 38, 340-343.

Jones, E. F., & Forrest, J. D. (1992). Under reporting of abortion in surveys of U.S. women: 1976 to 1988. Demography, 29, 113-126.

Koop, C. (January 9, 1989). Letter to President Ronald Reagan concerning the health effects of abortion. Medical and Psychological Impact of Abortion (pp. 68-71). Washington, DC: U.S. Government Printing Office.

Major, B., & Cozzarelli, C. (1992). Psychological predictors of adjustment to abortion. Journal of Social Issues, 48, 121-142.

Major B., Cozzarelli C., Cooper M. L., Zubek J., Richards C., Wilhite M., & Gramzow R. H. (2000). Psychological responses of women after first-trimester abortion. Archives of General Psychiatry, 57(8), 777-84.

Miller, W. B., Pasta, D. J., & Dean, C. L. (1998). Testing a model of the psychological consequences of abortion. In L. J. Beckman and S. M. Harvey (eds.), The new civil war: The psychology, culture, and politics of abortion. Washington, DC: American Psychological Association.

Radloff, L. S. (1977). The CES-D scale: A self-report depression scale for research in the general population. Applied Psychological Measurement, 1, 385-401.

Rotter, J. B. (1966). Generalized expectancies for internal versus external control of reinforcement. Psychological Monographs: General and Applied, 80 (1, Whole No. 609).

Russo, N. F., & Dabul, A. J. (1997). The Relationship of Abortion to Well-Being: Do Race and Religion Make a Difference? Professional Psychology: Research and Practice, 28.

Russo, N. F., & Zierk, K. (1992). Abortion, childbearing, and women's well-being. Professional Psychology: Research and Practice, 23, 269-280.