The decision to seek an abortion is never easy. Women have different reasons for choosing an abortion and their social, economic and religious background may influence how they cope. Furthermore, once pregnant, the alternatives of childbirth and adoption or keeping the baby may not be psychologically neutral. Research studies in this area have been hampered by methodological problems, but most of the better-quality studies have shown no increased risk of mental health problems in women having an abortion. A consistent finding has been that of pre-existing mental illness and subsequent mental health problems after either abortion or childbirth. Furthermore, studies have shown that only a minority of women experience any lasting sadness or regret. Risk factors for this include ambivalence about the decision, level of social support and whether or not the pregnancy was originally intended. More robust, definitive research studies are required on mental health after abortion and alternative outcomes such as childbirth.
It has been claimed that the decision to terminate an unwanted pregnancy can lead to mental health problems for women. However, once a woman is in the situation of having an unwanted pregnancy, there is no magical state of ‘un-pregnancy’ and the alternative courses of action of childbirth and raising a child or adoption may also pose a psychological threat. Childbirth can be a physically and emotionally demanding time for mothers and many studies have demonstrated an increase in depression and anxiety post-partum. In Scotland in 2002, data collected from a subset of general practitioner (GP) practices for each face-to-face patient with a GP consultation revealed that 27% of mothers were diagnosed with depression or anxiety within 12 months of childbirth compared with 19% in women who had not given birth in the same year. Furthermore, there is evidence that if women are denied an abortion, their children are at an increased risk of mental health problems. A longitudinal study that followed up children born in 1961–63 in the former Czechoslovakia (now Czech republic), of women who were denied abortion and forced to continue with the pregnancy, found that compared to matched controls (children from wanted pregnancies), offspring from unwanted pregnancies were at increased risk of negative psychosocial development and wellbeing. The negative effects on mental health continued into adulthood and, compared to controls, the children of unwanted pregnancies were more likely as adults to have a psychiatric illness, less job satisfaction and females were more likely to be single or divorced at age 26–28 years.
The question whether abortion has a negative effect on mental health of the woman is a recurring one. In 1989, the American Psychological Association (APA) reported the results of a systematic review of the published literature at that time. They reported that most methodologically sound studies indicated that severe negative reactions after legal non-restrictive first trimester abortion were rare and could best be understood in the framework of coping with a normal life stress.
In more recent times, however, there has been renewed interest in the mental health outcomes of induced abortion, exemplified by a headline for an article in the New York Times in 2007 titled ‘Is there a Post abortion Syndrome? The term ‘post-abortion syndrome’ is not recognised by any medical or psychological society, but was coined by Vincent Rue. It was used to imply post-traumatic stress disorder following the stress of abortion, where post-traumatic stress disorder is a severe and ongoing emotional reaction to an extreme psychological trauma.
Given this renewed interest in mental health after abortion and the fact that there had been new relevant publications since 1989, the APA convened a task force on mental health and abortion in 2007 to review the published literature since 1989. A further systematic review of published articles over a similar time period (1989 and 2008) was conducted by Charles et al. in 2008. The conclusions of both systematic reviews by APA and Charles et al. in 2008 were in agreement with the conclusions of the earlier report of the APA, and are outlined below. Both systematic reviews highlighted the numerous methodological flaws with research in this area.
Methodological problems with research on mental health and abortion
The systematic reviews of APA and Charles et al. in 2008 of published studies, which examined mental health and abortion, observed recurring methodological problems with the studies in this area.
(i) Comparative groups
Few studies have included appropriate comparative groups such as those women denied an abortion, who give the baby for adoption and women delivering and raising an unwanted child. The general population of women who deliver a baby are not an appropriate comparison group since women who plan a pregnancy and deliver a wanted baby may differ in important characteristics from women, such as level of social support, for whom the pregnancy was clearly unintended.
(ii) Co-occurrence of risk factors
There is good evidence that factors such as poverty, personality or behaviour (e.g., smoking, alcohol and drugs) can predispose a woman to unplanned pregnancy and abortion as well predispose to mental health problems. Few studies assessed or adequately controlled for confounding factors such as the co-occurrence of unwanted pregnancy with adverse circumstances and adverse circumstances with mental health problems.
(iii) Sampling
Some studies used volunteer samples that can introduce bias, since women who agree to participate (e.g., in response to a mailed questionnaire) may report different psychological experiences after abortion to those who do not agree to participate. Some studies have also been small in terms of sample size or have performed secondary analyses of data sets that were not designed to examine relative risks of mental health after abortion. Furthermore, in some studies, there was differential exclusion of women who had a child but had a subsequent abortion. This introduces selection bias and limits the generalisability of the findings to the general population.
(iv) Reproductive history and under-reporting
A major confounder in studies were history of abortion is self-reported is that of under-reporting due to the stigma associated with termination of pregnancy.
If women who experience most psychological problems following an abortion do not report that they had an abortion, this leads to an underestimation of any negative effect of abortion on mental health and, similarly, this leads to overestimation such women are more likely to report that they had an abortion. Many studies have not specified or clearly reported the gestations that women had abortions, nor the reason for the abortion or whether the pregnancy was originally intended. This is important because abortion at later gestations may be associated with more pain and increased risk of complications with a greater likelihood of being a more distressing experience. A late abortion may also reflect an underlying ambivalence about terminating the pregnancy or the diagnosis of a foetal anomaly in what was originally a planned and wanted pregnancy.
(v) Outcome measures and statistical analysis
Some studies used poor or unvalidated measures of mental health. In some studies the timing of the measurement of mental health relative to the time of the abortion was unspecified or varied. It is generally accepted that the closer one is to an event, the more accurate ones’ reporting of health or emotion is likely to reflect health/emotion at the time of the event. Many studies focussed only on negative mental health outcomes and neglected to consider possible positive outcomes. Some studies reported outcomes that were based upon results of multiple statistical testing and thus may have arisen by chance alone. For many studies the loss to follow-up was considerable. This can affect the validity of results, since if most psychologically disturbed women are lost to follow-up, this will lead to an underestimation of any effect of abortion on mental health and vice versa.
Studies indicating a neutral effect of abortion on mental health
One of the studies deemed to be of best quality by both the APA and the systematic review of Charles et al. in 2008 was a prospective, longitudinal, cohort study conducted in England by the Royal College of General Practitioners and Royal College of Obstetricians and Gynaecologists. The study cohort comprised 13 261 women recruited between 1976 and 1987, who presented to their GP with an unplanned pregnancy, of which 6410 proceeded with an abortion, 6151 did not request an abortion, 379 were denied an abortion and 321 who initially requested abortion but then decided to continue with the pregnancy. At recruitment, socioeconomic data were recorded including history of previous pregnancies and of pre-existing psychiatric illness. Every 6 months until the end of the study, GPs provided data on new episodes of illness and any further pregnancies. Morbidity on psychoses, depression and anxiety was coded using the World Health Organization codes. There was no significant difference in the rates of total psychiatric disorder or depression or anxiety between women who underwent pregnancy termination or childbirth. An important finding was that women with a previous history of psychiatric illness were most at risk of psychiatric disorder at the end of the pregnancy regardless of its outcome. Women without a history of mental illness had an apparently lower relative risk (RR) of psychosis after abortion than childbirth (RR 0.4, 95% confidence interval (CI): 0.3–0.7). The authors did note that rates of deliberate self-harm were higher in women refused an abortion (RR 2.9, 95% CI: 1.3–6.3) or having an abortion (RR 1.7, 95% CI: 1.1–2.6) compared to the childbirth group, but added that this may have been due to possible confounding factors that they were unable to account for such as co-existing social difficulties associated with the request for termination and also with self-harm behaviour. This study had the advantages of being large, prospective, with appropriate comparison groups that account for pregnancy intention and with validated outcome measures (physician diagnosis). It also had the advantage of comparing across groups at multiple time points and controlling for pre-existing mental health.
Another study that suggested no overall effect of abortion on mental health was that of Russo and Zierk, based upon data from the National Longitudinal Survey of Youth (NLSY) from USA. This survey involved annual interviews with a stratified sample of the population aged 14–21 years in 1979, with oversampling of black, Hispanic and poor white populations. In this study, out of a cohort of 5295 women, interviewed in 1987, those women who reported ever having an abortion ( n = 773) had no different measures of self-esteem than other women who did not report a history of abortion. This study also reported that having repeated unwanted pregnancies (birth or abortion) was significantly correlated with poverty and low education.
Studies from Norway, which followed up women from 10 days to 5 years after a first trimester abortion ( n = 80) or miscarriage ( n = 40), reported no significant differences between the groups in measures of anxiety, depression or wellbeing. While this study inferred similar levels of psychological stress with both miscarriage and abortion; nevertheless, miscarriage is not an appropriate comparison group for induced abortion as it is not an alternative that can be chosen by women with an unplanned pregnancy.
Studies indicating a neutral effect of abortion on mental health
One of the studies deemed to be of best quality by both the APA and the systematic review of Charles et al. in 2008 was a prospective, longitudinal, cohort study conducted in England by the Royal College of General Practitioners and Royal College of Obstetricians and Gynaecologists. The study cohort comprised 13 261 women recruited between 1976 and 1987, who presented to their GP with an unplanned pregnancy, of which 6410 proceeded with an abortion, 6151 did not request an abortion, 379 were denied an abortion and 321 who initially requested abortion but then decided to continue with the pregnancy. At recruitment, socioeconomic data were recorded including history of previous pregnancies and of pre-existing psychiatric illness. Every 6 months until the end of the study, GPs provided data on new episodes of illness and any further pregnancies. Morbidity on psychoses, depression and anxiety was coded using the World Health Organization codes. There was no significant difference in the rates of total psychiatric disorder or depression or anxiety between women who underwent pregnancy termination or childbirth. An important finding was that women with a previous history of psychiatric illness were most at risk of psychiatric disorder at the end of the pregnancy regardless of its outcome. Women without a history of mental illness had an apparently lower relative risk (RR) of psychosis after abortion than childbirth (RR 0.4, 95% confidence interval (CI): 0.3–0.7). The authors did note that rates of deliberate self-harm were higher in women refused an abortion (RR 2.9, 95% CI: 1.3–6.3) or having an abortion (RR 1.7, 95% CI: 1.1–2.6) compared to the childbirth group, but added that this may have been due to possible confounding factors that they were unable to account for such as co-existing social difficulties associated with the request for termination and also with self-harm behaviour. This study had the advantages of being large, prospective, with appropriate comparison groups that account for pregnancy intention and with validated outcome measures (physician diagnosis). It also had the advantage of comparing across groups at multiple time points and controlling for pre-existing mental health.
Another study that suggested no overall effect of abortion on mental health was that of Russo and Zierk, based upon data from the National Longitudinal Survey of Youth (NLSY) from USA. This survey involved annual interviews with a stratified sample of the population aged 14–21 years in 1979, with oversampling of black, Hispanic and poor white populations. In this study, out of a cohort of 5295 women, interviewed in 1987, those women who reported ever having an abortion ( n = 773) had no different measures of self-esteem than other women who did not report a history of abortion. This study also reported that having repeated unwanted pregnancies (birth or abortion) was significantly correlated with poverty and low education.
Studies from Norway, which followed up women from 10 days to 5 years after a first trimester abortion ( n = 80) or miscarriage ( n = 40), reported no significant differences between the groups in measures of anxiety, depression or wellbeing. While this study inferred similar levels of psychological stress with both miscarriage and abortion; nevertheless, miscarriage is not an appropriate comparison group for induced abortion as it is not an alternative that can be chosen by women with an unplanned pregnancy.
Studies indicating a negative effect of abortion on mental health
In 2006 a study from New Zealand reported a negative effect of abortion upon the mental health of young women who had an induced abortion. This was a longitudinal study that followed up a cohort born in Christchurch in 1977. The cohort included 630 women who reported their reproductive history between 15 and 25 years of age. The researchers reported higher rates of depression, suicidal ideation and illicit drug dependence in those undergoing abortion. While the study did measure validated outcomes and was able to account for confounding factors linked to women’s family and childhood, a study flaw was that it relied upon self-reported abortion (thus likely under-reporting) and did not separate single from multiple abortions. A more recent publication from this same study cohort, which followed 500 women up to age 30 years, reported similar findings of a small increase in mental health disorders (e.g., depression, anxiety, suicidal ideation and illicit drug dependence) in those undergoing abortion. In the systematic review by Charles et al., the earlier study was deemed to be of ‘fair’ quality only since it did not account for pregnancy intentions and compared women having an abortion with women who were not pregnant and women who were delivering a child. In addition, the study only controlled for pre-existing mental health up to 15 years of age, although the majority of pregnancies occurred more than 3 years later. The more recent study was of recent similar design and could be criticised for many of the same reasons.
Using the data set from the NLSY, Cougle et al. examined first pregnancy outcome (abortion or childbirth) among a sub-sample of 1884 women, and reported that those whose first pregnancy ended in abortion were significantly more likely to exceed the depression score for clinical depression (on average 8 years later) than those who gave birth. This study has been criticised for inappropriate comparison groups as there was no accounting for intended-ness of pregnancy and women were excluded from the delivery group if they went on to have subsequent abortions. In the systematic review by Charles et al., this study was rated as being of poor quality. In medical record linkage studies conducted in Finland, higher rates of death of any cause (including violent death, homicide and suicide) within 1 year of the pregnancy were reported for women who had an abortion compared to those delivering a baby. However, subsequent analyses showed that the abortion group actually had lower rates of death from causes aggravated or related to pregnancy. In addition, when abortions conducted for medical (therapeutic) reasons were excluded, there was no difference in deaths of any cause in the abortion group compared to the childbirth group. The most recent study from this group was rated as ‘very poor’ quality by the systematic review of Charles et al. as there were inappropriate comparison groups, no control for pre-existing mental health or other confounders and the study design using record linkage being able to provide contextual information only.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
