This article will discuss the nature of the association between abortion and mental health problems. Studies arguing about both sides of the debate as to whether abortion per se is responsible will be presented. The prevalence of various psychiatric disorders will be outlined and where there is dispute between studies, these will be highlighted. The impact of abortion on other areas such as education, partner relationships and sexual function will also be considered. The absence of specific interventions will be highlighted. Suggestions for early identification of illness will be made.
Induced abortion is the most common surgical procedure in the Western world with an estimated 42 million pregnancy terminations taking place worldwide each year, according to the Alan Guttmacher Institute. The relationship between abortion and subsequent life outcomes, particularly those relating to mental health outcomes, is one that has been discussed, often acrimoniously, in peer-reviewed journals for several decades, although this was hampered by significant methodological flaws in the earlier studies. It is an important question to resolve since an increase in the incidence of mental health problems post-abortion would impact significantly on resources as well as on personal well-being.
A number of questions present themselves in relation to the subsequent life outcome for women who terminate their pregnancies. These are
- 1.
What is the usual emotional reaction to abortion?
- 2.
Is there a specific psychiatric disorder-associated abortion?
- 3.
Is abortion associated with an increase in mental health problems and, if so, is the relationship a causal one?
- 4.
What psychiatric disorders are associated with abortion?
- 5.
Does abortion impact on social outcomes such as education, employment and relationships?
- 6.
Does terminating an unwanted pregnancy help women’s mental health?
- 7.
What therapeutic interventions are available when a woman has an adverse reaction and do these have an impact on subsequent life outcomes?
Each of these questions will be considered in this article. However, it is important to outline problems in presenting and interpreting the data that answering these questions entails. In addition, there are significant gaps in our knowledge, particularly in relation to treatments and this will be highlighted.
Presenting and interpreting the data
Presenting the data from studies on abortion and mental health problems is difficult. Due to the nature of the statistical analysis and the controls for confounding that this demands, prevalence and incidence rates may not be available and instead odds ratios (ORs), risk ratio (RR) and population attributable risks (PARs) (a measure of the proportion of total morbidity contributed by the variable of interest) are presented.
A further problem in interpreting the data rests with the variable quality and methodology of the studies. For example, some use clinical diagnosis obtained from general practitioners’ records , while others are based on diagnosis derived from structured interviews conducted in the general population. Some are cohort studies, which identify women seeking abortion and carry out evaluations after a set period thereafter , while others use a nested case-control design. Sourcing the data varies also with some using public databases such as the National Longitudinal Survey of Youth or record linkage methods, for example, abortion registered variously linked to suicide registers and to Medicaid claims. These considerations are important since they address different aspects of the question relating to abortion and subsequent mental health.
A further issue is the problem of confounding with some factors associated with both the decision to abort and poor subsequent mental health such as prior psychiatric history while there may be underlying differences between women who abort and those who choose to continue a pregnancy, for example, socioeconomic status. Controlling for these possible confounders and sources of bias is thus essential. Most of the large-scale studies for over 2 decades have controlled for socio-demographic variables and some personal variables such as prior mental health history but a recent crop of studies has also begun to control for variables connected to the decision to abort such as partner violence and childhood abuse and even more recently for the wantedness of the pregnancy , resulting in controls for over 30 variables in the latter, the largest number ever.
In this article, only studies that have controlled for confounding will be included, with the exception of one qualitative study.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
