Inequities in safe abortion: women’s care trajectories in Abuja and Lagos, Nigeria

Abstract

Background

Unsafe abortion remains a significant cause of maternal morbidity and mortality in many African countries, including Nigeria. This study aims to fill gaps in understanding of how abortion safety has evolved since previous estimates. The study also offers new insight on pregnant people’s experiences along the abortion care trajectory and how sociodemographic factors correlate with these experiences and the overall safety of abortion.

Methods

Data presented are drawn from a larger study that was conducted in Lagos and Abuja, Nigeria in 2023 using a total market assessment approach. This analysis drew exclusively on data from a quantitative survey of women of reproductive age with a 5-year history of induced abortion(s). Abortion safety was operationalised based on two dimensions in line with a previous analysis of population-based survey data from Nigeria: [1] whether the method(s) used included any non-recommended methods and [2] whether the source(s) used were clinical or non-clinical. We combined source and method information to categorize a woman’s abortion into one of four safety categories. Analyses include frequencies of abortion safety and experiences along the abortion care trajectory, as well as bivariate and multivariate assessments of sociodemographic and reproductive history correlates.

Results

Two hundred women completed the survey of which 197 reported their abortion method and were included in this analysis. Almost half (44.7%) of respondents’ reported abortions were categorised as most unsafe involving non-recommended methods and non-clinical source(s). Those living in rural areas and those living in severe poverty were significantly more likely to have had the most unsafe abortions. Those living in poverty were also at increased likelihood of being unable to access an abortion from their preferred source and of experiencing any complications.

Conclusion

These findings confirm that unsafe abortion in Nigeria is a public health concern and an issue of social inequity. Efforts to improve equitable access to safe, high-quality and client-centred services are needed. National health policies must address both the legal and practical barriers to safe abortion access. Ensuring expanded access to in-facility surgical procedures and post-abortion care is critical for management of complications from continued unsafe abortion. Simultaneously, harm reduction efforts - including increasing awareness of quality medication abortion drugs for safer self-induction and training lower-cadre providers on the medical management of abortion - can help mitigate the toll of abortion-related morbidity and mortality.

Background

Unsafe abortion remains a leading cause of maternal mortality in low- and middle-income countries, contributing to 8–15% of maternal deaths [1]. Despite progress in the availability of safe abortion care and a revolution in access to quality medical abortion products over the last two decades, the prevalence of unsafe abortions remains high [2]. While there are limited variations in abortion rates by geographic setting alone, abortion safety is greatly impacted by the legal environment [2]. Most unsafe abortions worldwide take place in legally restrictive settings where 40% (approximately 753 million) of women of reproductive age live [3]. Although abortions performed in accordance with medical standards have a high safety profile [4], unsafe abortion remains a significant threat to maternal health because access to the required standard of care is limited.

Nigeria has one of the most restrictive laws in the region. Induced abortions are only legally permitted to save the woman’s life and when performed by a qualified practitioner [5]. The legal framework varies regionally, with the penal code applying in the North where Abuja is located, while the criminal code applies in the Southern part of Nigeria where Lagos is situated. In Lagos, induced abortion is slightly less restricted, being permitted to preserve the woman’s life and physical health [5]. Despite this restrictive environment in Nigeria, induced abortion is a common reproductive health experience, with an estimated 46 abortions per 1000 women aged 15 to 49 years, according to a 2017 confidante method-estimated incidence [6]. This amounts to 2.4 million abortions annually when applied to 2023 estimates of the population of women of reproductive age [7]. This marks an increase from a 2012 estimate of 1.25 million abortions annually, drawn from an estimated 33 abortions per 1,000 women aged 15–49 from an abortion incidence complication measurement approach [8]. The only national study of abortion incidence in Nigeria preceding Bankole et al. (2015) estimated a rate of 25 abortions per 1,000 women aged 15–44 in 1996 [9], suggesting that abortion incidence may have increased over the past two decades.

Previous evidence indicates that two-thirds of abortions in Nigeria are most unsafe, involving non-recommended methods (i.e., methods other than surgery (manual vacuum aspiration (MVA) or dilation and evacuation (D&E) or medication abortion drugs) from non-clinical providers [10]. Estimates from 2012 indicate that these predominantly unsafe abortions resulted in nearly 500,000 women experiencing serious health consequences, with less than half (212,000) receiving treatment for these complications [8]. These numbers are likely much higher today with rapid population growth in Nigeria. Unsafe abortions continue to contribute to Nigeria’s high maternal mortality ratio, estimated at 1047 maternal deaths per 100,000 live births in 2020 [11]. A recent analysis of a population-based and nationally representative survey of women aged 15–49 in Nigeria found that unsafe abortion is an issue of health inequity. Women aged 15–19, those who have never attended school, and women within the poorest wealth quintile are most likely to experience an unsafe abortion [6].

This study draws on data from a larger study that employed a total market assessment approach to assess the context of abortion and post-abortion care in Lagos and Abuja, Nigeria. The primary objective of this analysis is to determine the safety of reported abortions in our sample and the social determinants of experiencing a less safe abortion, including both sociodemographic and reproductive history characteristics. An improved understanding of factors that may contribute to women accessing a less safe abortion care option in this setting can inform harm reduction efforts that seek to reduce morbidity and mortality due to unsafe abortion.

The secondary aim is to explore how the same social determinants can influence abortion care-seeking trajectories, from taking a pregnancy test to experiencing any complications or having social support throughout one’s abortion care pathway. While it is important to understand the social determinants of access to safe abortion, a more granular exploration of how any inequities in access may play out at different stages of one’s abortion care pathway can help inform targeted harm reduction efforts and help improve access to safer, client-centred services and ultimately reduce morbidity and mortality due to unsafe abortion.

Methods

Study design

This paper draws on an analysis of data from one component of a larger, cross-sectional study in Lagos and Abuja, Nigeria employing a total market assessment approach and combining qualitative and quantitative methods. The primary quantitative data collection included a (1) a survey conducted at a sample of health-facilities and individual providers (both informal and formal providers) providing either induced abortions or post-abortion care in the last 3-months (2), a mystery client survey at all health facility sites/providers surveyed, and (3) a quantitative survey of women who had an induced abortion in the five years preceding the survey. The qualitative component included key informant interviews with strategic stakeholders, including policymakers, representatives of client groups, and programme delivery organisations. All data collection was completed in October 2023.

The analysis for this paper draws exclusively on the data from the quantitative survey of women with a history of an induced abortion in the 5-years preceding the survey. The survey employed a convenience sampling approach, targeting a purposive sample of 200 women aged 15–49 in Lagos and Abuja with an equal sample of 100 respondents per study location. The sample size of 100 participants per location (N = 200) was determined based on feasibility constraints while ensuring sufficient statistical power to detect meaningful differences in unsafe abortion prevalence. This approach balances logistical considerations with the need for adequate representation, aligning with previous research employing similar convenience sampling methods in health studies when random sampling is not possible given the study population and/or resource constraints [12, 13].

Eligible women had a history of at least one induced abortion in the five years preceding the study. While the study locations were Abuja and Lagos, clients could reside anywhere in Nigeria, provided they received their abortion care in these two locations. Survey participants were recruited through (1) referrals by abortion and PAC providers (2), initial seeds identified at abortion and PAC service delivery sites who became peer recruiters, and (3) community health volunteer referrals. Abortion and PAC providers in facilities selected for the health facility survey(HFS), who introduced the research to women they cared for and provided any interested women the study information sheet and a ‘coupon’ with the study team’s number with guidelines to call, text, or flash-message the team to call them back if they wished to participate in the survey. The peer-recruitment involved survey respondent’s who agreed to support with recruitment of other women in their networks who had received abortion care. Lastly, community health volunteers recruited women and girls to the study team. Like the provider and peer-to-peer recruitment strategies, community health volunteers approached and discussed the study with potential participants and gave interested, potential respondents coupons to contact the study team.

All respondents provided written, informed consent to participate before the beginning the survey. Participants aged 15 to 18 years, considered “emancipated minors” under National Human Research Ethics Committee (NHREC) guidelines due to life experiences like pregnancy or marriage, were included in the study after providing informed consent [14]. Parental consent was avoided as we anticipated that such minors may have procured abortion without the consent of their parents, and attempting to obtain parental consent may jeopardise the objective of the study and add undue risks to the participants. This approach was used to gather crucial data on abortion among younger age groups.

Interviews were conducted face-to-face by trained interviewers in Hausa, Yoruba or English. The questionnaire was professionally translated into Hausa and Yoruba from English and piloted with native speakers before data collection. Data was collected using a survey questionnaire programmed on OpenDataKit (ODK) software for administration using smartphones. The questionnaire was developed previously for a recent study in Kenya with minor adaptations to ensure relevance in the Nigerian context [15]. Through this survey approach, we explored several questions related to women’s experiences with abortion and post-abortion care including the pathway to care, cost, incidence of complications, source of care, methods used, quality of care, stigma, and social support.

Measures

In the questionnaire, women could report all abortion methods and provider types (referred to as source) used if they reported doing multiple things to end the pregnancy. Respondents described the place, methods used, safety and experience of all reported abortions. The abortion method and source measures in this analysis draw on the woman’s least safe abortion method or source in the case that they reported multiple abortion methods or sources.

In this analysis, we categorised abortion methods into surgery, medication abortion (MA) drugs, other pills, or pills without sufficient information to categorise as MA drugs, and traditional or other methods (i.e. herbal drinks, injections, alcohol, or other traditional remedies). Any respondents with missing values for abortion method were excluded from the analysis (n = 3). We further classified abortion methods as recommended and non-recommended (i.e., other than WHO-approved surgical procedures or MA drugs) that put the woman at potentially high risk of abortion related morbidity or mortality [4]. For recommended surgical procedures, we excluded all surgical procedures reported other than WHO-approved methods of electric vacuum aspiration (EVA), manual vacuum aspiration (MVA) and dilation and curettage (D&C) [4]. For recommended MA methods, we included Misoprostol-alone as well as the combination regiment of Misoprostol and Mifepristone [4]. If a respondent reported doing multiple things and used a method other than a WHO-approved surgical method or MA drugs at any point in the termination, we categorised the abortion as non-recommended.

We categorised abortion method sources into three categories: (1) public facilities and private facilities (including non-governmental organisations and private doctors) (2), pharmacies or chemist shops and (3) traditional or other non-medical sources (including shops, markets, friends or relatives or home). We further classified abortion source as clinical or non-clinical. Clinical sources included all public and private facility types while non-clinical sources included pharmacies, drug shops and traditional or other non-medical sources. Missing values for the abortion method source variable (n = 87) were classified as ‘other’. All respondents with missing values had reported using only a traditional or other non-medical method for their pregnancy termination. In line with the abortion method classification, we categorised an abortion as non-clinical if at any point the respondent reported using a source other than a public or private facility.

Using this data, we operationalised abortion safety into four dimensions in line with a previous analysis of population-based abortion data from representative samples of reproductive age women in Nigeria, India and Cote D’Ivoire [10].: (1) recommended method(s) involving only clinical source(s); (2) recommended method involving non-clinical source(s); (3) non-recommended method(s) involving clinical source(s) and (4) non-recommended method(s) involving non-clinical source(s). For our binary outcome variable, abortions in category four were deemed the most unsafe with the first three categories grouped together as safer services.

For our secondary outcomes of experiences along the abortion care trajectory, we included the use of a pregnancy test or other medical examination (binary: yes/no), abortion decision making role (binary: sole decision/other person involved in decision), the abortion method used (binary: recommended/non-recommended), source of method used (binary: clinical/non-clinical source), was able to access preferred source (binary: yes/no), experienced any complications (binary: yes/no), had social support in terms of someone who provided advice about abortion care (binary: yes/no).

In this analysis, we also examined several respondent sociodemographic and reproductive history characteristics at the time of the abortion. Sociodemographic characteristics included age (categorical), school attendance, marital status, religious affiliation, employment status (binary), urban, peri-urban or rural residence, and poverty status using the global Multidimensional Poverty Index (MPI) and categorised women both as living in multidimensional poverty and severe multidimensional poverty [16]. Reproductive history characteristics of the participants included the number of children, the number of previous abortions, and any contraceptive use.

Data analysis

We first used descriptive statistics to examine the sociodemographic and reproductive history characteristics, safety of abortion, and experiences along the abortion care pathway. To address our primary aim, we explored differences in sociodemographic and reproductive history characteristics between women who accessed the most unsafe abortions, assessed via bivariate and multivariate logistic regressions. All sociodemographic and reproductive history covariates that were significant in the bivariate logistic regression were included in the multivariate logistic regression, along with study location which was maintained to control for any effect of geographic location where the respondent accessed their abortion. Individual multivariate regressions were run for multidimensional poverty and severe multidimensional poverty, given the collinearity between these variables.

To address the secondary study aim, bivariate logistic regressions were run with each of the abortion care pathway experience measures which were treated as the dependent variable, with the suite of sociodemographic and reproductive history covariates as the independent variables. As with the primary outcome analysis, all covariates that were significant in the bivariate logistic regression along with study location were included in the multivariate logistic regression.

Statistical significance was determined to be at p < 0.05. We conducted a post-hoc power analysis to evaluate whether the sample size (N = 197) was sufficient to detect a significant difference in unsafe abortion prevalence. Comparing our observed proportion (44.7%) to (63.4%) [6] yielded a Cohen’s h effect size of 0.37, indicating a moderate effect. Power was assessed across proportions ranging from 40% to 70% in 5% increments. The analysis demonstrated that at N = 197, power remained above 85% for the observed difference, exceeding the 80% threshold for adequate power. These results confirm that the study had sufficient power to detect meaningful differences, minimizing the risk of a Type II error and supporting the robustness of the findings.

All analyses were conducted using Stata 16.1 (College Station, TX).

Ethical approval and consent to participate

The study was conducted in accordance with the ethical principles outlined in the Declaration of Helsinki. The study protocol was approved by MSI’s independent Ethical Review Committee (application reference number: 003–23). Further, IRB approvals were also sought and received from the Federal Capital Territory Health Research Ethics Committee in Abuja (approval number: FHREC/2023/01/136/25-07-23) and Lagos State University Teaching and Hospital Health Research Ethics Committee in Lagos (approval number: LREC/06/10/2210).

Results

Sample characteristics

A total of 200 women of reproductive age completed the survey, with 197 (98.5%) providing information on the method of their abortion and were thus included in this analysis. The average age of respondents was 32 years old; most had attended some higher education (60.4%) and the majority were currently married or cohabiting (52.8%) (Table 1). Respondents primarily identified as Christian (73.6%) and lived in urban (43.7%) or peri-urban areas (44.7%). Many respondents were nulliparous (46.7%), yet nearly one in five had 3 or more children (18.8%). Two-thirds of respondents were not living in multi-dimensional poverty, with 32.5% living in poverty and 13.7% living in severe poverty. Over half (61.4%), had been using some form of family planning at the time of their most recent pregnancy that they terminated. However, only 2.5% of those using any form of family planning (4.8% of those using a modern contraceptive method) reported using a long-acting method (IUD, n = 1; Implant, n = 3). Of note, almost a third (30.6%) reported using emergency contraception. Another third (38%) reported using condoms (22.3%) or traditional methods (15.7%).

Table 1.

Demographic and reproductive history sample characteristics

a: Demographic Characteristics				b: Reproductive History Characteristics
	No.	% (Std.Err)			No.	% (Std.Err)
Mean age	31.7	(0.53)		Parity
Age				No children	92	46.7%
15–19	10	5.1%		1–2 children	68	34.5%
20–24	30	15.2%		3 + children	37	18.8%
25–29	40	20.3%		Abortions in past 5 years
30–34	46	23.4%		1	147	74.6%
35–49	71	36.0%		2	35	17.8%
Education				3+	15	7.6%
Primary	7	3.6%		Contraceptive use at abortion
Secondary	71	36.0%		Non-use	76	38.6%
Higher	119	60.4%		Use	121	61.4%
Marital status				Type of FP used
Married/cohabiting	104	52.8%		Sterilisation	0	0.0%
Divorced/separated/widowed	18	9.1%		IUD	1	0.8%
Never married	75	38.1%		Implant	2	1.7%
Religion				Injectable	8	6.6%
Christian	145	73.6%		Oral contraceptive pills	27	22.3%
Islam	51	25.9%		Emergency contraception	37	30.6%
No religion	1	0.5%		Condoms	27	22.3%
Employment				Traditional methods	19	15.7%
Unemployed	24	12.2%
Employed	173	87.8%
Residence
Urban	86	43.7%
Peri-urban	88	44.7%
Rural	23	11.7%
Study location
Abuja	98	49.7%
Lagos	99	50.3%
MPI poor
Not poor	133	67.5%
Poor	64	32.5%
MPI severe poor
Not poor	170	86.3%
Severely poor	27	13.7%

Primary outcome: abortion safety

About half (51.8%) terminated their pregnancy using recommended methods, with a larger proportion using recommended surgical methods (28.4%) than medical (22.8%). Only 28.4% only received care from clinical sources with the remaining 71.6% accessing non-clinical care. Combining both criteria, almost half the respondents’ (44.7%) abortions fell into the most unsafe category, involving non-recommended methods dispensed by non-clinical or non-medical providers (Table 2). Only about 1 in 5 (22.3%) abortions were categorised as most safe, involving recommended method(s) dispensed in clinical settings, while 26.4% involved recommended methods involving non-clinical source(s) and 6.6% involved non-recommended method(s) involving clinical source(s).

Table 2.

Safety characteristics of the abortion care accessed

	No.	%
Method used
Non-recommended method	96	48.7%
Recommended method - surgical	56	28.4%
Recommended method - medical	45	22.8%
Method source
Non-clinical	141	71.6%
Clinical	56	28.4%
Safety category
Recommended method(s) involving only clinical source(s)	44	22.3%
Recommended method involving non-clinical source(s)	52	26.4%
Non-recommended method(s) involving clinical source(s)	13	6.6%
Non-recommended method(s) involving non-clinical source(s)	88	44.7%
Safer vs. most unsafe
Safer services (safety categories 1–3)	109	55.3%
Most unsafe (safety category 4)	88	44.7%

Examining abortion safety by sociodemographic and reproductive history characteristics, respondents residing in rural areas were significantly more likely to have had the most unsafe abortions (OR 6.91, 95% CI: 2.44–19.57) as were those from peri-urban areas compared to urban areas (OR 2.67, 95% CI: 1.43–5.00). Those living in poverty (OR 2.20, 95% CI: 1.20–4.04) and severe poverty (OR 2.86, 95% CI: 1.21–6.73) were also more likely to have experienced the most unsafe abortions. Previous use of family planning methods was a protective factor. Respondents who reported using family planning at the time of their most recent pregnancy that was terminated had a decreased likelihood of a most unsafe abortion (OR 0.50, 95% CI: 0.28–0.89).

Results from the multivariable logistic regression indicated that place of residence was independently associated with a most unsafe abortion among respondents. Those living in rural areas had the greatest likelihood (aOR 4.87, 95% CI: 1.64–14.44) of most unsafe abortion (Table 3). Living in poverty was also associated with an increased odds of having experienced a most unsafe abortion in the multivariable model (aOR 2.08, 95% CI: 1.08–4.00). In the multivariate model, living in severe poverty was only marginally significant (p < 0.01) associated with having increased odds of having experienced a most unsafe abortion. Family planning use remained significantly associated with lower odds (aOR 0.51, 95% CI: 0.26–0.99) of having experienced a most unsafe abortion (Table 3).

Table 3.

Bivariate and multivariate regressions of characteristics associated with experiencing an unsafe abortion

a: Demographic Characteristics
	OR	95% CI		aOR	95% CI
Age
15–19	1.00
20–24	1.00	0.23	4.31
25–29	0.36	0.09	1.49
30–35	0.67	0.17	2.68
35–49	0.41	0.11	1.58
Education
Primary	1.00
Secondary	0.97	0.20	4.65
Higher	0.44	0.09	2.06
Marital status
Married/cohabiting	1.00
Divorced/separated/widowed	1.48	0.54	4.03
Never married	1.44	0.79	2.62
Religion
Christian	1.00
Islam	1.39	0.73	2.64
No religion	1.00
Employment
Unemployed	1.00
Employed	0.53	0.22	1.27
Residence
Urban	1.00			1.00
Peri-urban	2.67	1.43	5.00	2.30	1.20	4.40
Rural	6.91	2.44	19.57	4.87	1.64	14.44
Study location
Abuja	1.00			1.00
Lagos	0.83	0.47	1.46	0.93	0.48	1.78
MPI poor
Not poor	1.00			1.00
Poor	2.20	1.20	4.04	2.08	1.08	4.00
MPI severe poor
Not poor	1.00
Severely poor	2.86	1.21	6.73	2.42	0.94	6.20

b: Reproductive History Characteristics
	OR	95% CI		aOR	95% CI
Parity
No children	1.00
1–2 children	0.83	0.44	1.57
3 + children	1.13	0.53	2.42
Previous abortions
1	1.00
2	0.73	0.34	1.55
3+	1.84	0.62	5.44
Any FP use
No	1.00			1.00
Yes	0.50	0.28	0.89	0.51	0.26	0.99

Bold values indicate statistical significance at the p < 0.05 level

Bold italicised values indicate statistical significance at the p < 0.01 level

OR Odds ratio, aOR Adjusted odds ratio

Study location was not significantly associated with a most unsafe abortion in either the bivariate of multivariate model.

Secondary outcomes: experiences along the abortion care pathway

Most respondents (74.6%) had used a pregnancy test or accessed a medical examination to determine their pregnancy. In terms of the respondent’s role in deciding to terminate their pregnancy, almost half (40.1%) reported that it was mainly their decision, half (50.3%) reported that it was a joint decision, 7.6% reported that it was mainly their husband or partner’s decision, and a remaining 2.0% reported that it was someone else’s decision. Almost one out of three (30.9%) were not able to access their preferred source, mainly those who had accessed a non-clinical source but would have preferred a clinical one. Further, one in ten (10.7%) respondents reported experiencing any complications, and two-thirds (67.0%) reported having someone who could provide them advice about abortion care (Table 4).

Table 4.

Frequencies of experiences along the abortion care pathway

	No.	%
Used any pregnancy test/medical examination
No	50	25.4%
Yes	147	74.6%
Role in decision making to terminate pregnancy
Joint decision	99	50.3%
Mainly husband/partner’s decision	15	7.6%
Mainly respondent’s decision	79	40.1%
Someone else’s decision	4	2.0%
Least safe method used
Surgical procedure	56	28.4%
Medical procedure	44	22.3%
Other pills	37	18.8%
Traditional/other methods	59	29.9%
Least safe source of method accessed
Public facility	0	0.0%
Private facility	56	28.4%
Pharmacy/chemist shops	39	19.8%
Traditional/other non-medical	102	51.8%
Able to access preferred source
No	56	28.4%
Yes	125	63.5%
Experienced any complications
No	176	89.3%
Yes	21	10.7%
Had social support available for advice on abortion
No	65	33.0%
Yes	132	67.0%

Examining experiences along the abortion care pathway by sociodemographic and reproductive history characteristics, respondents who were divorced, separated, or never married were significantly more likely to be the sole decision-maker. Those divorced and separated were also significantly more likely to have experienced any complications (Table 5).

Table 5.

Bivariate and multivariate regressions of characteristics associated with experiences along the abortion care journey

a: Pre-service experiences
	Used a pregnancy test or other medical examination
	OR	95% CI		aOR	95%CI		OR	95% CI		aOR
Marital status
Currently married/cohabiting	1.00						1.00			1
Divorced or separated/widowed	0.96	0.31	2.93				21.71	4.69	100.53	36.55	6.96	191.79
Never Married	1.26	0.63	2.51				2.38	1.27	4.45	2.86	1.45	5.65
Religion
Christian	1.00						1.00
Islam	0.87	0.42	1.80				1.34	0.70	2.56
No religion	1.00						1.00
Residence*
Urban	1.00			1.00			1.00
Peri-urban	2.88	1.42	5.85	4.34	1.91	9.85	0.63	0.34	1.15
Rural	3.95	1.09	14.35	8.84	2.05	38.09	0.61	0.24	1.60
Study location
Abuja	1.00			1.00			1.00			1.00
Lagos	2.76	1.40	5.43	4.68	2.01	10.87	1.71	0.96	3.04	2.97	1.46	6.03
MPI poor
Not poor	1.00			1.00			1.00
Poor	0.57	0.29	1.11	0.54	0.24	1.17	0.94	0.51	1.73
MPI severely poor
Not poor	1.00						1.00
Severely poor	0.64	0.27	1.52				0.86	0.37	1.99
Previous abortions
1	1.00						1.00
2	0.90	0.40	2.05				1.18	0.56	2.50
3+	5.06	0.64	39.73				1.38	0.48	4.02
Previous FP use
No	1.00			1.00	0.36	1.82	1.00			1.00
Yes	0.47	0.23	0.95	0.81	0.53	4.30	2.19	1.19	4.03	4.05	1.89	8.66

b: During service experiences
	Used an approved medical or surgical method						Used a clinical source						Accessed preferred source of abortion method used
	OR	95% CI	aOR		95%CI		OR	95% CI		aOR	95%CI		OR	95% CI		aOR	95%CI
Marital status
Currently married/cohabiting	1.00						1.00						1.00
Divorced or separated/widowed	0.61	0.22	1.67				0.49	0.13	1.83				0.59	0.19	1.81
Never Married	0.63	0.35	1.15				1.09	0.57	2.09				0.80	0.41	1.56
Religion
Christian	1.00						1.00						1.00
Islam	0.81	0.43	1.53				0.52	0.24	1.14				0.92	0.45	1.88
No religion	1.00						1.00						1.00
Residence*
Urban	1.00			1.00			1.00						1.00			1.00
Peri-urban	0.31	0.17	0.58	0.32	0.17	0.63	0.97	0.51	1.85				2.03	1.04	3.99	2.19	1.09	4.42
Rural	0.04	0.01	0.18	0.04	0.01	0.23	0.49	0.15	1.57				3.10	0.96	10.05	4.15	1.21	14.24
Study location
Abuja	1.00			1.00			1.00			1.00			1.00			1.00
Lagos	0.87	0.50	1.52	0.61	0.31	1.21	1.47	0.79	2.75	1.28	0.68	2.43	0.87	0.46	1.63	0.73	0.37	1.45
MPI poor
Not poor	1.00						1.00			1.00			1.00			1.00
Poor	0.64	0.35	1.17				0.41	0.19	0.85	0.43	0.20	0.91	0.40	0.20	0.77	0.31	0.15	0.63
MPI severely poor
Not poor	1.00			1.00			1.00						1.00
Severely poor	0.42	0.18	1.00	0.63	0.22	1.77	0.53	0.19	1.48				0.88	0.35	2.20
Previous abortions
1	1.00						1.00						1.00
2	1.14	0.54	2.39				1.23	0.55	2.73				1.67	0.67	4.16
3+	0.84	0.29	2.44				1.34	0.43	4.15				0.62	0.20	1.91
Previous FP use
No	1.00			1.00			1.00						1.00
Yes	1.67	0.94	2.98	1.22	0.62	2.41	1.19	0.62	2.25				0.75	0.39	1.45

c: Post service experiences
	Experienced any complications						Had social support available for advice on abortion
	OR	95% CI		aOR	95%CI		OR	95% CI		aOR	95%CI
Marital status
Currently married/cohabiting	1.00			1.00			1.00
Divorced or separated/widowed	5.28	1.60	17.43	6.55	1.684098	25.45449	0.80	0.28	2.24
Never Married	0.92	0.31	2.70	0.43	0.06	3.08	1.15	0.61	2.17
Religion
Christian	1.00			1.00			1.00
Islam	2.97	1.18	7.49	1.92	0.66	5.63	1.12	0.56	2.21
No religion	1.00						1.00
Residence*
Urban	1.00						1.00			1.00
Peri-urban	2.34	0.69	7.89				2.53	1.32	4.83	2.11	1.03	4.34
Rural	10.93	2.92	40.94				2.24	0.81	6.24	2.49	0.77	8.06
Study location
Abuja	1.00			1.00			1.00			1.00
Lagos	1.10	0.44	2.72	1.66	0.53	5.17	1.06	0.59	1.93	0.78	0.38	1.62
MPI poor
Not poor	1.00			1.00			1.00
Poor	2.55	1.02	6.37	1.01	0.23	4.39	1.01	0.54	1.91
MPI severely poor
Not poor	1.00			1.00			1.00
Severely poor	5.09	1.87	13.84	6.96	1.29	37.62	0.98	0.42	2.33
Previous abortions
1	1.00						1.00			1.00
2	0.83	0.23	3.02				0.20	0.09	0.43	0.20	0.08	0.47
3+	2.20	0.56	8.69				0.50	0.17	1.51	0.39	0.12	1.28
Previous FP use
No	1.00						1.00			1.00
Yes	0.66	0.27	1.64				0.31	0.16	0.61	0.28	0.13	0.61

Bold values indicate statistical significance at the p < 0.05 level

Bold italicised values indicate statistical significance at the p < 0.01 level

OR Odds ratio, aOR Adjusted odds ratio

The only significant difference by religion was whether a respondent had experienced any complications, with those of the Islamic faith being significantly more likely to have experienced complications (Table 5).

Residence status was significantly associated with using a pregnancy test or other medical examination to confirm pregnancy, using an approved surgical or medical method, accessing their preferred source of abortion, and having social support available for advice on abortion. Those living in peri-urban or rural areas were significantly more likely to have used a pregnancy test, more likely to have accessed their preferred source of their abortion method used, and those in peri-urban areas were more likely to have had social support, as compared to urban areas. Those in peri-urban and rural areas were also less likely to have used an approved surgical or medical method of abortion.

When looking at poverty status, those living in poverty were less likely to use a clinical source or have accessed their preferred source of abortion. They were also more likely to have experienced any complications. Those living in severe poverty were less likely to have used an approved method (this did not remain independently significant in the multivariate model) and far more likely to have experienced any complications (aOR 6.96, 895%CI: 1.29–37.62).

In terms of previous abortions, those with more than 1 previous abortion were less likely to have any social support for advice.

Lastly, previous use of family planning was significantly associated with a lower likelihood of using a pregnancy test, an increased likelihood of being the sole decision maker, and a decreased likelihood of having had any social support for advice.

There were no significant associations with any of the abortion care pathway experiences by age, education, and employment status.

Discussion

Abortion safety

This study describes the distribution of abortion safety in Lagos and Abuja, Nigeria and provides new insights into the sociodemographic correlates with women’s experiences along their abortion care trajectory. Almost half of the abortions (44.7%) in our sample were most unsafe, with those living in rural areas and those living in poverty at the greatest risk of having the most unsafe abortions. Previous use of family planning was a protective factor, with those who reported any use of family planning at the time of their most recent pregnancy that was terminated having a decreased likelihood of having a most unsafe abortion. These findings are consistent with previous literature suggesting that the most disadvantaged women are those most likely to resort to unsafe means of pregnancy termination [6, 17–20]. While, like previous evidence, we found that already vulnerable and disadvantaged women were most likely to experience a most unsafe abortion, other sociodemographic factors that have been found to be associated with unsafe abortion, including being younger, unmarried, and uneducated, were not in our sample. This could be due to various factors, including a limited sample size to have the necessary power to detect the association and a sample that was skewed towards older and more affluent women, likely due to the peer-to-peer recruitment approach.

Our estimate of the proportion of most unsafe abortions was also lower than the previous estimate of 63.4% in 2017 (and 73.7% from the same study for the past five years) [6]. These 2017 estimates are also lower than the unsafe abortion estimates for Western Africa in 2010–2014 at 84.7% [17]. This could be due to various reasons, including that our sample may not be representative of the national population of women who access abortions and is likely biased towards a more affluent population. The potential bias in our sample is due to the sampling process whereby providers linked interviewers to women who had received abortion care from them. Such women are likely to have received safer care involving clinic-based providers which may not be representative of the care for the general population. Furthermore, clinic-based providers may have been more likely to refer women who had a safer and successful abortion without any complications. Lastly, the most unsafe category did not include abortions involving non-recommended methods in clinical settings (6.6% in our sample), which would meet the WHO’s classification of ‘unsafe’ abortion.

It is also probable that safe abortion options are becoming more available in Nigeria, largely due to the increased availability of medication abortion drugs (Mifepristone and Misoprostol) in recent years. Mifepristone was registered for use in 2017, and Misoprostol has been on the country’s Essential Medicines List for incomplete and spontaneous abortion since 2010 [21]. A 2018 review of referral hospital medical record data showed an increase in Misoprostol use for induced abortions over a nine-year study period and this was associated with a reduction in severe morbidity [22]. Other studies have indicated that women increasingly have the option to self-manage their abortion by purchasing the drugs through pharmacies, chemist shops, abortion hotlines and accompaniment models [23, 24]. This trend toward out-of-facility medical management of abortion appears to have been captured in our study as well, with 26.4% of respondents reporting using recommended methods (medication abortion drugs) from non-clinical sources, compared to just 5.4% in 2017 [6]. We also observed a decrease in the use of recommended method(s) involving only clinical source(s) compared to the 2017 data, at 22.3% versus 29.1% [6]. This indicates that while the overall safety of abortion may be improving, there may have been a shift away from clinical sources of abortion which may mean future gaps in needed access to clinical care for surgical options and management of complications. The apparent shift away from clinical sources captured in this study may also reflect more recent effort to train pharmacist to provide safer medication abortion [25, 26].

Experiences along the abortion care trajectory

Women’s decisions regarding where to seek abortion care, the method(s) and source(s) that they eventually use, the quality of care they receive, and their access to or use of any follow-up care are influenced by several factors. Coast et al. developed a comprehensive framework that encompasses the full range of factors that shape women’s abortion care-seeking trajectories, which includes the abortion-specific context, women’s individual context, and the international or sub-national context [27]. Previous evidence from the Nigerian context found that non-clinical providers (55.0%) were more often used than clinical providers (45.0%); however, clinical providers were preferred by most women (55.6%) [28]. While, we found a lower use of clinical sources (28.4%), we saw a similar trend in preference for clinical providers, with 90.7% of those who accessed a clinical source reporting that they had accessed their preferred source compared to just 59.8% of those who had accessed non-clinical sources. We also found that those who were living in poverty had a significantly decreased likelihood of accessing their preferred source and of accessing a clinical source. This finding suggests important wealth inequities in ability to exercise choice in source of abortion and importantly in access to facility-based care. We also found that women living in poverty and severe poverty were significantly more likely to have experienced any complications (MPI poor OR: 2.55, 95% CI:1.02–6.37; MPI severe poor OR: 5.09, 95% ci: 1.87–13.84). When adjusting for other demographic covariates, severe poverty remained significantly associated with experiencing any complications, demonstrating the public health imperative of ensuring improved equitable access to safe sources of care, with options for both facility and out-of-facility.

Furthermore, our study indicates that the use of contraception at the time of the most recently terminated pregnancy serves as a protective factor against experiencing the most unsafe abortions. Contraceptive use is associated with increased reproductive autonomy, enabling individuals to make informed decisions about their reproductive health [29]. This autonomy may empower them to seek safer abortion services when necessary. Moreover, individuals utilizing contraception may have better knowledge of and access to reproductive healthcare services, which can facilitate safer abortion care when needed [30]. Access to comprehensive reproductive health services, including contraception and safe abortion care, is crucial in preventing unintended pregnancies and reducing the reliance on unsafe abortion practices. Integrating sexual and reproductive health services, such as contraception counselling and abortion or post-abortion care, within broader healthcare provision is essential. Such integration ensures that individuals have continuous access to the necessary resources to manage their reproductive health effectively, thereby reducing the incidence of unsafe abortions, reducing related morbidity and mortality, and promoting overall reproductive well-being.

Limitations and strengths of the study

Our study has some limitations. First, we relied on women’s self-reporting of their abortion experiences in a social context where such information is highly sensitive and in a legal context where one can be prosecuted for securing abortion. Self-reporting under such circumstances could introduce social desirability bias, leading to underreporting or misreporting experiences. Previous studies have highlighted the challenge of underreporting and misreporting of abortion (31). Apart from the information being sensitive, laypeople may be unable to accurately report medical information. Further to social desirability bias, recall bias, introduced because of the study’s retrospective nature, could also limit the reliability of the results; it is realistically challenging for people recall with exactitude the experiences that may have happened in the past five years.

Another source of potential bias is misclassification. Regarding surgical abortion procedures, we determined that respondents could not provide details on the specific procedure conducted. As such, we categorized any surgical abortion as recommended, which would include dilation and curettage [4]. Similarly, women could not report on the training of a given provider, thus we relied on information about the source or location of the abortion method and whether providers at the source of abortion would be clinicians or not. Additionally, a substantial number of respondents were unable to provide sufficient details to categorize the type of pills they used, and among those who reported use of MA drugs, we did not try to determine whether the correct dosage based on clinical guidelines and for gestational age was used. However, while these limitations in our categorisation of abortion safety would lead to misclassifications, they are not likely to be influenced by any specific characteristic of the study population in a systematic manner.

The further limitation relates to the sample and sampling approach. Our sample of 197 included in this analysis may not be representative of the broader population of women who have had abortions in Abuja and Lagos, let alone Nigeria. We also used a peer-to-peer sampling approach which may create an underrepresentation of certain demographic groups or an overrepresentation of well-connected individuals, affecting the generalizability of the findings. To further compound this limitation, we identified the seed samples from facility-based providers, who are likely to have introduced the research team to women likely to receive care from formal healthcare settings. As noted above, this may also have biased the sample towards people who had received safer abortion services.

Notwithstanding the limitations discussed, this study has significant strengths. Our utilization of a peer-to-peer sampling approach ensured the resulting sample included women who received abortions from a wide spectrum of sources and providers, thereby capturing a broad range of abortion services and experiences. This approach allowed us to capture women who received services from formal and informal providers, highlighting service characteristics outside the formal health sector and comparing the varied sources and experiences. Further, the study examined abortion safety based on the WHO guidelines and examined the method(s) used and the outcomes of abortion care [4]. This study provides valuable contributions by categorizing abortions into safety categories and assessing the determinants of safe abortion practices to highlight necessary programmatic intervention for improvement in quality of care and improved health outcomes. Finally, the sociodemographic characteristics and reproductive histories of respondents explored in this study provide an understanding of disparities in access to safe abortion care and the resulting health outcomes. This allows us to situate access to abortion as socially determined, especially based on social status and income.

Conclusions

Our study suggests that, while the overall safety of abortion could have marginally improved over the last five years, nearly half (44.7%) of abortions in Lagos and Abuja, Nigeria in our study were the most unsafe. This finding underscores the urgent need for national health policies that prioritize access to safe abortion services and harm reduction strategies.

Our results further indicate that economically disadvantaged women in rural areas, with limited ability to navigate and access safe abortion in this legally restrictive setting, are most at risk of experiencing a most unsafe abortion. This highlights abortion in Nigeria as a public health concern and an issue of social inequity. Efforts to expand the legal conditions for abortion in Nigeria are critical. Previous evidence has shown that restrictive abortion laws negatively impact abortion safety without reducing overall abortion incidence rates [17]. Without legal expansion, women will continue to seek services from unregulated providers who may lack appropriate training or resort to sub-quality products to self-manage their abortion.

In this context, national health policies must address both the legal and practical barriers to safe abortion access. In the absence of legal reform, harm reduction efforts - including increasing awareness of quality MA drugs for safer self-induction and training lower-cadre providers on the medical management of abortion - can help mitigate the toll of abortion-related morbidity and mortality.

Additionally, expanding contraceptive services, including counselling to counteract fears of contraceptive-related infertility, is essential to reduce reliance on unsafe abortion. Our findings suggest that previous contraceptive use was associated with a decreased likelihood of experiencing a most unsafe abortion, potentially due to factors such as improved access to health information and trusted healthcare providers.

Furthermore, policies and on-the-ground implementation must prioritise the expansion of quality, facility-based abortion care, including surgical abortion and PAC, to reduce abortion-related morbidity and mortality. Given the existing inequities in access to these services, policy and implementation efforts must ensure that already disadvantaged persons and marginalised populations are not left behind. Addressing these disparities through targeted public health interventions and policy reforms is crucial to improving abortion safety and reproductive health outcomes in Nigeria.

Abbreviations

D&C

Dilation and Curettage

EVA

Electric Vacuum Aspiration

IRB

Institutional Review Board

MA

Medication Abortion

MVA

Manual Vacuum Aspiration

PAC

Post-abortion Care

SAC

Safe abortion Care

SRH

Sexual and Reproductive Health

SRHR

Sexual and Reproductive Health and Rights

WHO

World Health Organisation

Authors’ contributions

MR led the study design, supported the development of the protocol and data collection tools, led the analysis and interpretation, and drafted the manuscript.BU and AA contributed to the study design, developed the protocol and data collection tools, supported the tool scripting, testing, and surveyor training, supervised data collection, contributed to the analysis and interpretation, and substantially contributed to the manuscript.AT, OO and OI contributed to the conception and design of the study, provided oversight to data collection training, contributed to the interpretation of data, and contributed to the manuscript.All authors read and approved the final manuscript.

Funding

The development of the study design, protocol and ethical review was supported by a large anonymous donor.

Data availability

The datasets used and analysed during the current study are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

This study protocol was approved by MSI’s Ethical Review Board as part of MSI’s larger study entitled “A total market assessment for safe abortion and post-abortion care in Abuja & Lagos, Nigeria” with application reference number 003–23. Further IRB approvals were sought and received from the Federal Capital Territory Health Research Ethics Committee in Abuja (approval number: FHREC/2023/01/136/25-07-23) and the Lagos State University Teaching Hospital Health Research Ethics Committee (approval number: LREC/06/10/2210.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.