Our study highlights that low socio-economic status, characterised by poor education
and being employed in unskilled/less-skilled occupations is an independent risk factor
for unsafe abortion. However, it is not the most influencing factor for unsafe abortions
leading to complications, but many other aspects in relation to their family formation
characteristics that are not confounded by their low socio-economic status, but act
as independent risk factors for unsafe abortion. These factors include being unmarried,
pregnant for the first time, not having decided on the desired family size, absence
of a girl child and longer average birth intervals.
In comparison, in high-income countries where abortion is legalised, unsafe abortion
is a risk among single, primigravid women of younger age 24], 25], while the low- and middle-income countries provide less-consistent evidence on these
risks 3], 14]. Based on studies that have used comparative groups, the determinants of unsafe abortion
identified in the low- and middle-income region are: younger 16] or older 17], 26] age at the time of abortion, younger age at first sexual act 17], unmarried 16], fewer 16] or many 17] living children, previous abortions 16], 1 partner 17], poverty 12], poor 12] or better 26] education, and use of contraception 26]. In the context of wide variation seen in these countries in relation to the data
sources, the legal status of abortion and accessibility to safe abortion services,
these findings need cautious application. In addition, the risks are not reported
with adjustments made for their confounders.
In Sri Lanka, the determinants of induced abortion identified in one of the recent
studies were older age, being married, young age of last child, having grown-up children,
completed family and socio-economic constraints 27]. Our findings were in contrast to these. This was a community-based study on women
who claimed to have undergone unsafe abortions (defined by any pregnancy terminated
by an unskilled person or by a qualified doctor at a private place during last 18 months).
Inferences drawn from this study are therefore less applicable for unsafe abortions,
since the majority of women in this study would have undergone safely induced abortions
or received relatively safe abortifacients from unskilled abortionists 28], 29]. Therefore, our study provides new knowledge on risk factors of unsafe abortion specifically
among women with post-abortion complications.
Previous research has shown that low socio-economic status is the main risk factor
for resorting to unsafe abortions 11]–14]. This may be owing to the difficulty in affording more costly yet safer abortion
services in countries where abortion is illegal. In concurrence, our study highlights
the risk of abortion associated with poor economic status of women, characterised
by their poor education (adjusted-OR?=?1.5; 95 % CI?=?1.1–2.4) and low-income occupations
(adjusted-OR?=?2.3; 95 % CI?=?1.4–3.6). Difficulty in raising a child as well as fear
of losing employment that had minimum job security could have been the predisposing
factors associated with low socio-economic status for abortion. Previous studies have
shown that less-urban districts having marginalised populations report higher proportions
of unsafe abortion, reflecting the unavailability of safe abortion services 27]. Such studies conducted in hospitals or clinics providing free health care are known
to represent patients predominantly of low socio-economic class, which could over-estimate
the prevalence of low socio-economic status among abortion 14], 20], 30], 31]. We minimised this type of selection bias in our study, by comparing the cases with
a comparable group of controls recruited from the same hospitals using similar selection
criteria, when assessing the risk factors of unsafe abortion.
Unmarried status may prompt a woman to abort owing to difficulties in raising a child
as a single mother. In contrast, we have shown in our study that this vulnerability
of unmarried women for abortion was independent of their social status (adjusted-OR?=?9.3;
95 % CI?=?4.0–21.6). This reflects the social-stigma associated with ‘unwed mothers’
in Sri Lanka. Co-habitation outside marriage is not an accepted norm in the country
5]. In the current national family health programme, only the women in customary/legal
marriage are entitled to domiciliary or clinic care by the public health midwife,
unless they are pregnant or already having a child 5]. This excludes women not living together with their partners who have difficulty
in accessing contraception services for preventing unintended pregnancies. Findings
reiterate the importance of extending such services also to the unmarried women without
causing any stigma.
Primigravid women in our study imparted a two-fold risk for unsafe abortions (adjusted-OR?=?2.2;
95 % CI?=?1.2–4.2), which indicated women’s desire for delaying their first childbearing.
However, this finding is not in concurrence with the usual preference of Sri Lankan
women, which is to bear the first child early, as the preference is shown to fall
drastically from 77 % after marriage to 26.1 % after the first child 7]. It should be noted that this fact is relevant only for married women and that the
behaviour of unmarried women following an unplanned pregnancy may not be the same.
According to our study, unmarried status was also an independent risk factor for unsafe
abortions, thus having some bearing on their desire for delaying their first pregnancy.
On the other hand, the risk of abortion associated with being primigravid was independent
of their poor socio-economic status, implying factors other than poor economic status
for delay. Past decades have seen women changing attitudes on ideal family size 32] and becoming independent as much as their partners towards contributing actively
to the family earnings and social commitments. This strife may be a likely reason
for the risk of unsafe abortions in primis.
Previous research suggests that poverty predisposes a woman to restrict her family
size, by resorting to abortion when faced with an unintended pregnancy 11]–14]. In contrast, our study provides different views on their desired family size. Most
disturbingly, women at risk of unsafe abortion were seen to be indecisive of their
desired family size (adjusted OR?=?2.2; 95 % CI?=?1.4–3.5). This highlights a deficiency
in the family planning programme in Sri Lanka. Primary objective of the family planning
policy is to facilitate families to make informed decisions about their desired family
size and to control their fertility through contraceptives 5]. This highlights the need to access newly-married couples for pre-pregnancy counselling
and contraception services. Although eligible couple registration is almost 100 %
in Sri Lanka 5], motivating couples to access pre-pregnancy services has been challenging for the
public-health-midwife. The services need to be re-designed to make it more attractive
and tailor-made, especially for working couples.
Young age of the last child that reflects a short last birth interval is a well-known
risk factor for abortion 15], 27]. In contrast, our findings did not demonstrate this relationship. Instead, women
having generally longer birth intervals, including the last one were at increased
risk for abortion. This may imply the risk of abortion associated with the older age
of the last child, which is likely to be a social stigma within the Sri Lankan society.
Also, though family completion (26.5 %) has been a common risk for abortion worldwide
14], it was not so according to our study. In contrast, inability to make an informed
decision on their desired family size has been a strong risk factor for unsafe abortions
(adjusted OR?=?2.2; 95 % CI?=?1.4–3.5). These findings may be an indication of the
unmet need for contraception among women who have no definitive plans on family formation.
Despite a remarkable new acceptor rate of 90 % for temporary contraceptive methods
and a contraceptive prevalence of 64.4 % 5] in Sri Lanka, unmet needs do exist particularly among women on temporary methods
for a long period, that compel them to discontinue abruptly with no alternative protection.
Over the past 15–20 years in Sri Lanka, age at marriage has been increasing along
with fertility reductions observed in all age groups across all socio-economic strata,
demonstrating women’s inclination towards limiting the number of children relatively
early in marriage 7]. This further highlights an unmet need for protection from unplanned pregnancies
until such women reach their menopause. Unless they are protected with long-term or
permanent contraceptive methods, unsafe abortions would be a recurring problem in
Sri Lanka.
Strengths and limitations of the study
Obtaining a diagnosis of unsafe abortion is crucial in settings where reporting is
deterred by legal, ethical and moral concerns. Particularly, population-based studies
suffer from ‘misclassification’, as the diagnosis is made solely dependent on woman’s
recall of past events 12], 33]. Hospital-based studies are also notorious for under-reporting 17], 34] but has its advantage of making a valid diagnosis based on clinical evidence following
examination. In our study, we addressed under-reporting of unsafe abortions by restricting
the cases to women in whom the diagnosis was confirmed by definitive clinical evidence
and by obtaining data using medical graduates who were not involved in participants’
care but well-trained in conducting in-depth interviews. Furthermore, morbidity and
mortality statistics owing to complications of unsafe abortion are reported predominantly
from the state hospitals in Sri Lanka 5], 7]. Therefore, using state hospitals as the study setting provided a representative
sample of women who were most at risk of morbidity consequences following an unsafe
abortion. Though abortion services are not provided, state-owned hospitals provide
post-abortion care, which are liberally accessed by women in the event of post-abortion
complications, owing to free health services provided and the high health seeking
behaviour among females in Sri Lanka. Only a small proportion would access non-state
private health facilities for treatment of such complications following unsafe abortion.