Ethical review for this research was completed by the Allendale Investigational Review Board in the United States and the National Health Research Ethics Committee in Nigeria. Study staff obtained verbal informed consent from all participants 18 years and older, and verbal informed assent from all participants 17 years and younger. Due to concerns about participant privacy and anonymity, the IRB and ethics committee waived the need for parental consent for minors. Funders of the study had no role in the study design, data collection, analysis, interpretation, writing of the report, or the decision to submit the paper for publication.

This analysis pooled data collected in two separate studies: (1) the Studying Accompaniment Feasibility and Effectiveness (SAFE) study [16, 19] and (2) the Abortion Service Quality (ASQ) study [17]; both are prospective, observational cohort studies that enrolled people who contacted safe abortion hotlines/accompaniment groups for information and support related to self-managed abortion. Researchers followed participants for approximately four weeks to assess pregnancy and quality of care outcomes; the primary outcomes for the SAFE study were self-managed abortion safety and effectiveness, and the primary outcomes for the ASQ study focused on quality of care; main outcomes from these studies have been published elsewhere [16, 17]. While the SAFE and ASQ studies collected data in several countries, this analysis presents results from Nigeria only.

Between October 2019 and September 2020 we recruited callers to a safe abortion hotline in Lagos, Nigeria. When a person contacted the safe abortion hotline seeking support with ending a pregnancy, counselors at the hotline helped the person verify the duration of the pregnancy, rule out contraindications for using the medications, and then provided World Health Organization-endorsed information on how to take pills to safely end a pregnancy, provided emotional and informational support on pain and symptom management, as well as if, when, and where to seek medical care if needed. Anyone who contacted the group during the study period requesting information about abortion for their own pregnancy was counted in a call log, and then screened for eligibility by study staff. Anyone ≥ 13 years of age (SAFE study) or ≥15 years of age (ASQ study), starting a new medication abortion process, with no contraindications to medication abortion [20], with a pregnancy up to 12 weeks was eligible to participate [19]. We excluded those experiencing ongoing symptoms (bleeding, cramping) of abortion or miscarriage, experiencing symptoms suggestive of ectopic pregnancy (low back pain, pain on side of the abdomen/pelvis, shoulder pain, dizziness/fainting); or anyone who did not want to be contacted by study staff.

At enrollment, hotline counselors administered a baseline questionnaire over the phone at the end of the first call with the caller. Subsequently, trained study coordinators contacted all participants at two additional time points (for SAFE:~7 days and ~21 days after taking the first medication abortion pills, for ASQ: ~7 days and ~30 days) to complete additional telephone surveys. Participants were provided between $10–25 USD in phone credit depending on how many follow-up surveys they completed.

The primary outcomes for this analysis included direct questions as to whether the COVID-19 pandemic and related lockdowns influenced participants ability to or comfort in various parts of the abortion process, including confirming their pregnancy, getting the pills, taking the pills, seeking medical care, and more. Specifically, questions followed this structure: “Did COVID-19 and the lockdowns affect your ability to confirm your pregnancy? (For example, to obtain a pregnancy test, to take a pregnancy test in privacy, to get an ultrasound, other.)” Additional COVID-19 related questions asked as to whether COVID-19 and related lockdowns influenced where participants had their abortion, the number of people who stayed in their home, and whether anyone found out about their abortion. Participants in both SAFE and ASQ studies answered these questions.

We also included three sets of secondary outcomes: the number of calls received by the hotline in each month, burdens experienced during the abortion process, and abortion completion. Burdens faced during the self-managed abortion process were only measured in the SAFE study, not in ASQ. To measure burdens, SAFE study coordinators asked participants “In preparing for, during, or in the weeks since you ended the pregnancy, did you have to do any of the following as part of the abortion process?” Answer choices included taking time off of work, forfeiting lost wages for time off from work, arranging childcare, finding lodging outside the home, borrowing money, selling something, traveling more than 30 minutes, or a participant-reported burden. Other variables measured in both studies included participant age and education, duration of pregnancy at enrolment, and medication regimen.

To measure abortion completion, both SAFE and ASQ studies utilized self-report. SAFE study participants self-reported abortion completion at the one and three week follow-ups, and ASQ study participants self-reported abortion completion at the one month follow-up. In the SAFE study, study coordinators asked participants two questions to ascertain abortion completion: “Do you feel that your abortion process is complete?”, as well as whether they reported receiving a manual vacuum aspiration or dilation and curettage procedure. In the ASQ study, study coordinators asked participants “Do you believe that you may still be pregnant today?” Participants who answered “yes” were classified as “not complete / not sure”, while those who responded with “no” were classified as “complete, inclusive of surgical intervention”.

For analysis, we compared frequencies and percentages for each outcome overall and across phases of the COVID-19 lockdown, and tested for differences in outcomes using t-tests, Chi-square tests, and tests of proportion. We also evaluated differences in the adjusted odds of each abortion-related burden by time period using the Mantel-Haenszel formula [21] to calculate stratum-specific odds ratios (ORs) for each education level (less than high school, high school graduate, and university graduate), adjusted for age (centered and continuous), and tested for homogeneity across stratum-specific estimates to evaluate for effect measure modification by education. Where no heterogeneity was found, we estimated a weighted average of the stratum-specific ORs using mhodds in Stata; where heterogeneity was found, we report adjusted ORs separately by strata of education to highlight effect measure modification [21]. We explored differences by age and education as these characteristics differed across pre versus during COVID-19 samples.

We defined COVID-19 pandemic related lockdowns in two ways: a binary formulation of “pre” versus “during” COVID-19, defined as any data gathered from anyone enrolled up to and including March 30, 2020 as “pre” (a five month period from October through March), and all data among those enrolled on March 31st and after as “during” (a six month period from March 31 through September). In an additional time category formulation, we defined COVID-19 related lockdowns in phases: Phase 1 from March 31-June 1, Phase 2 from June 2 –September 3, and Phase 3 from September 4 –December 31, 2020. These dates correspond with the dates that various movement and other pandemic related restrictions were enacted or lifted in Nigeria [22].

We analyzed data in STATA (version 15.1). While the ASQ and SAFE study samples sizes were powered to evaluate abortion completion and quality outcomes, not for the outcomes evaluated in this analysis, we calculated the minimum detectable absolute difference in proportions for a core outcome of this analysis, burdens faced prior to versus during the pandemic within the recruited sample sizes, and found the minimum detectable absolute difference to be 5.1% with 80% power and an alpha of 0.05.

Overall, most participants (63.1%) were between 20–29 years old, and over half completed post-secondary education (Table 1). At baseline, most participants (n = 468, 58%) were <7 weeks into their pregnancies; and most (n = 793, 98.3%) took misoprostol alone to end their pregnancies. Participants recruited after the onset of the COVID-19 pandemic were slightly younger and less well educated than were participants recruited prior to the onset of the pandemic.

Of all the burdens measured, participants most frequently reported having to borrow money as part of the abortion process and did so at a higher level during the COVID-19 pandemic as compared to before (Table 2). Similarly, participants were more likely to report having to find lodging outside of the home to self-manage their abortion during the pandemic than they were prior. After the onset of the pandemic, participants were less likely to report having to take time off from work or school than they were prior to the pandemic, and among participants with a university-level education or more, participants had lower odds of needing to travel more than 30 minutes as part of self-managing their abortion during the pandemic than they did prior to it.

Across the three specified phases of COVID-19 related lockdowns in Nigeria, participants reported the greatest difficulty accessing and taking pills in the second phase (Table 3). No substantial differences across phases were observed in where participants had their abortions or where they would have preferred to have their abortion–in all cases, the majority of participants had and preferred to have the abortion in a private room in their home. Across all other measured impacts of the pandemic, participants reported the greatest difficulties in the first and strictest phase of the lockdown. During phase I, participants most frequently reported a change in the number of people staying in the home, someone finding out about their abortion, changes in their ability to confirm their pregnancy, as well as their comfort or ability in seeking medical care during or after their abortion. Across all phases, participants reported lost wages as a direct result of the COVID-19 pandemic, most frequently in the first phase.

At last follow-up, the majority (n = 766/768, 99.7%) of participants who reported an outcome reported a complete abortion, most without the need for surgical intervention (Table 4). Only two participants (0.2%) reported an incomplete abortion, or uncertainty about abortion completion. Missingness for self-managed abortion outcomes was highest in the third phase of the pandemic and was due to loss to follow-up among participants in the ASQ study (the ASQ study protocol implemented fewer follow-up contacts than did the SAFE study, and did not prioritize abortion completion as an outcome). If we assume that all missing values were incomplete abortions, then the conservative estimate of abortion completion is 94.9%.