Sesay et al. BMC Women’s Health (2023) 23:44 BMC Women’s Health https://doi.org/10.1186/s12905-023-02175-9 RESEARCH Open Access Determinants of induced abortion among women of reproductive age: evidence from the 2013 and 2019 Sierra Leone Demographic and Health Survey Foday Robert Sesay*, Emmanuel Anongeba Anaba, Adom Manu, Ernest Maya, Kwasi Torpey and Richard M. K. Adanu Abstract Background Worldwide, pregnancy termination due to unintended pregnancy is crucial in maternal health, par- ticularly in settings where abortion laws are restrictive. Presently, there is a paucity of literature on determinants of induced abortion among women of reproductive age in Sierra Leone. The study findings could be used to improve the country’s maternal mortality indices and inform health programs and reproductive health policies geared toward tackling induced abortion. Methods We analyzed secondary data from the 2013 and 2019 Sierra Leone Demographic and Health Surveys. The surveys were nationally representative, with weighted samples comprising 16,658 (2013) and 15,574 (2019) women of reproductive age. Descriptive statistics, including frequencies and percentages, were computed, while Chi-square and Binomial Logistics Regression were employed to identify correlates of induced abortion. Results The results showed that a minority (9%) of the participants had induced abortion in both surveys. Abor- tion was significantly associated with age, marital status, employment status, education, parity, and frequency of listening to the radio and watching television (p < 0.05). For instance, women aged 45–49 years (AOR = 7.91; 95% CI: 5.76–10.87), married women (AOR = 2.52; 95% CI: 1.95–3.26), and working women (AOR = 1.65; 95% CI: 1.45–1.87) had a higher likelihood of induced abortion compared to their counterparts. Moreover, women with primary educa- tion (AOR = 1.27; 95% CI:1.11–1.46) and those who watch television once a week (AOR = 1.29; 95% CI: 1.11–1.49) were more likely to terminate a pregnancy. Women with six or more children (AOR = 0.40; 95% CI: 0.31–0.52) were less likely to terminate a pregnancy compared to those with no child. Conclusion The study revealed that a minority of the women had induced abortions. The prevalence of induced abortion did not change over time. Induced abortion was influenced by age, marital status, employment status, edu- cation, parity, and exposure to mass media. Therefore, policies and programs to reduce unwanted pregnancies should focus on increasing access to modern contraceptives among women of lower socio-economic status. Keywords Induced abortion, Determinants, Women of reproductive age, Sierra Leone *Correspondence: Department of Population, Family and Reproductive Health, School Foday Robert Sesay of Public Health, College of Health Sciences, University of Ghana, Accra, fodrobert24@gmail.com Ghana © The Author(s) 2023. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licens es/ by/4. 0/. The Creative Commons Public Domain Dedication waiver (http:// creat iveco mmons. org/ publi cdoma in/ zero/1. 0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Sesay et al. BMC Women’s Health (2023) 23:44 Page 2 of 10 Background this law (Section 58 of the 1861 offenses against the Per- Worldwide, pregnancy termination due to unintended son Act) in the Safe Abortion Bill, allowing abortion on pregnancy is a crucial factor in maternal health, par- request [15]. However, because of concerns raised by reli- ticularly in settings where abortion laws are restrictive. gious leaders in the whole of Sierra Leone, this bill was Unsafe abortion affects both the individual and society not signed into law. in terms of health and economic implication [1]. Most The quality of health care services in Sierra Leone has women resort to induced abortion because they lack been a significant problem coupled with limited access to a partner’s support, are financially unstable, a victim of sexual and reproductive health services, partly due to the rape or incest, and have untimely pregnancies [2]. The eleven years of civil war followed by the Ebola outbreak World Health Organization (WHO) defines unsafe abor- in 2014. In addition, there is a severe shortage of trained tion as a process of terminating a pregnancy by someone medical personnel to provide the needed health services. lacking the required skills or in a setting that does not For example, in contrast to the WHO recommendation meet the minimum medical standards or both [3]. Abor- of 23 skilled health providers per 10,000 population, the tion is classified by WHO as safe when it is done with a country has about two qualified skilled health providers method recommended by the WHO, that is appropriate (physicians, midwives, and nurses) per 10,000 popula- for the gestational age of the pregnancy and the person tion [16]. Furthermore, the situation of women having providing or supporting the abortion is trained [4]. Abor- induced abortions is made worse by the low modern tion is less safe when either the method or provider crite- contraceptive uptake (24%) among women of reproduc- rion is met, but not both, and least safe when they meet tive age [17]. Sierra Leone’s maternal mortality ratio of neither the provider nor method criterion [4, 5]. 717 maternal deaths per 100, 000 live births is one of The global estimate for abortion annually is 73.3 mil- the worst worldwide [18]. Of the direct causes of mater- lion, corresponding to a worldwide abortion rate of 39 nal mortality in Sierra Leone, unsafe abortion is ranked abortions per 1000 women aged 15–49  years [6]. The fifth, trending behind obstetric hemorrhage, hyperten- global yearly estimate for unsafe abortion is around 25 sion, obstructed labor, and sepsis. Moreover, unsafe abor- million, and the majority of them (97%) occur in the tion contributes to about 10% of Sierra Leone’s maternal developing world [7]. Globally, the proportion of unin- mortality ratio [19]. In addition, a study [2] estimated the tended pregnancies ending in abortion increased from cost of treatment and impact of unsafe abortion in Sierra 51% in 1994 to 61% in 2019 [8]. Data from 2010 to 2014 Leone as $35 for simple post-abortion care (PAC) with indicated that approximately 55% of abortions world- $166 and $272 for moderate and severe complications, wide were considered safe, 31% less safe, and 14% least respectively [2]. safe [9]. About a million women of reproductive age are Induced abortion in Sierra Leone has not been exten- hospitalized yearly due to unsafe abortion globally [10] sively investigated. Previous studies among women of and unsafe abortion accounts for approximately thirteen reproductive age have sought to examine why women percent of global maternal deaths [11]. In Sub-Saharan resort to abortion, especially from unskilled providers, as Africa, unsafe abortion is estimated to have killed one well as their knowledge and use of contraceptives [2, 20]. woman every eight minutes in 2015 [12]. In addition, a study on the influence of international and It has been established that countries with restrictive regional human rights treaties on domestic abortion poli- abortion laws have higher maternal mortality [4]. Sierra cies in Sierra Leone revealed that women are dying from Leone is among those countries with restrictive abortion unsafe abortion [21]. Notwithstanding, there is a paucity laws. According to the Center for Reproductive Rights of literature on the determinants of induced abortion (2009), the country is classified in category three in the among women of reproductive age in Sierra Leone. In world’s abortion law, meaning abortion is performed order to address the obstacles to obtaining safe abortion when it is geared toward saving the mother’s life. The treatment in Sierra Leone, this study examines factors country’s law on abortion was inherited from the British influencing induced abortion among women. colonial government. It states that women who attempt to abort are guilty of a crime and " shall be liable and sen- Methods tenced to life" [13]. The above unfortunate situation, cou- Study location, design, and data source pled with the fact that women want to avoid the stigma Sierra Leone is located on the west coast of Africa and created by religious and cultural influences have caused covers an area of 72,000 square kilometers [18]. It shares many women to resort to unsafe abortion practices. It a border with Guinea on the north and northeast, Libe- has also caused a dilemma among caregivers to perform ria on the east and southeast, and the west by the Atlan- their duty as health care providers or obey the law [14]. tic Ocean [18]. According to the 2015 Population and In December 2015, Sierra Leone attempted to revoke Housing Census, the country has a total population of S esay et al. BMC Women’s Health (2023) 23:44 Page 3 of 10 7,092,113 with just over half being female (50.8%) [24]. children = 3; 6 or more children = 4). Other independent This study analyzed the women’s data from the two most variables were current contraceptive use (no method = 1; recent 2013 and 2019 Sierra Leone Demographic and modern method = 2; traditional method = 3), knowledge Health Surveys (SLDHS) [18, 22]. The DHS is a house- about ovulation, correct (halfway between two menstrual hold-based, nationally representative survey. It uses a periods) = 1; incorrect = 2; don’t know = 3), frequency of two-stage sample design. For instance, in the 2013 DHS, reading newspaper, listening to radio and watching tel- the first stage involved selecting 435 enumeration areas evision (not at all = 1; less than once a week = 2; at least from 27 strata with probability proportional to size, using once a week = 3). the 2004 Population and Housing Census report [23], while the second comprised the selection of 30 house- Statistical analyses holds from each cluster. A total number of 13,006 house- All analyses were carried out using STATA/SE version 16 holds within the enumeration areas were selected. We (Stata Corp, College Station., Texas, USA). Descriptive obtained 16,658 women as the weighted sample size of statistics of the background characteristic of respondents women aged 15–49 years. were computed and summarized (Table 1). At the bivari- Similarly, in the 2019 DHS, the first stage comprised ate level, the Chi-squared test was used to determine the selection of 578 enumeration areas from 31 strata, the association between variables under study and the proportional to size employing the 2015 Population outcome of interest. Similarly, at the multivariable level, and Housing Census report [24], while the second stage binary logistics regression was used to determine the involved the selection of 24 households from each cluster, predictors of induced abortion among women of repro- resulting in a total sample size of approximately 13,872. A ductive age. In all, three models were computed. Model total of 15,574 women aged 15–49 years were obtained as 1 looked at predictors of induced abortion in 2013, while a weighted sample. The target population was women of model 2 focused on predictors of induced abortion in reproductive age who had ever terminated a pregnancy 2019. The third model (model 3) focused on predictors and passed the night before the survey in the selected of induced abortion in 2013 and 2019 (combined) while households. adjusting for the survey year. The significance for the The anonymized data was cleaned, missing values analysis was set at p < 0.05, while the strength of associa- were dropped and adjusted for the complex nature of tion was examined using odds ratios and their 95% confi- the survey. Permission to use the DHS data was sought dence interval. from Measure DHS. The anonymized datasets were only downloaded on approval of the request to undertake this Results analysis. The data analysed in this study were saved on Descriptive statistics of participant characteristics a password-protected personal computer. The data was The study analyzed data from 16,658 women and 15,1574 declared survey data using sampling weight, weight, and women in the 2013 and 2019 SLDHS respectively. In strata or employing the ’svy’ STATA command. Detailed both surveys, the prevalence of induced abortion was information about the 2013 and 2019 DHS is included 9%. In the 2013 survey, 36% of the participants resided elsewhere [18, 22]. in urban areas compared to 46% in 2019 survey. In addi- tion, a higher proportion of the participants in the 2019 Measurements survey (37%) had secondary education compared to the The dependent variable in this study was ever termi- 2013 survey (27%). The use of modern contraceptives had nated a pregnancy (induced abortion), coded as yes = 1 increased from 20% in 2013 to 24% in 2019. Similarly, and no = 0. The independent variables mentioned in the accurate knowledge about ovulation had increased from literature include those characteristics of the women 29% in 2013 to 51% in 2019 (Table 1). who attest to having terminated a pregnancy. These include women’s age (15–19 = 1; 20–24 = 2; 25–29 = 3; Association between participant characteristics 30–34 = 4; 35–39 = 5; 40–44 = 6; 45–49 = 7), educational and termination of pregnancy status (no education = 1; primary = 2; secondary = 3; In both 2013 and 2019, induced abortion was signifi- higher = 4), employment status (not working = 1; work- cantly associated with age, marital status, employment ing = 2), wealth index (poorest = 1; second = 2; mid- status, education and parity, (p < 0.05). In 2019, knowl- dle = 3; fourth = 4; richest = 5), religion (Christianity = 1; edge about ovulation, frequency of listening to the radio, Muslim = 2; others religion = 3), place of residence reading newspapers were significantly associated with (urban = 1; rural = 2), marital status (never in union = 1; abortion (p < 0.05). In the combined analysis, induced married/in union = 2; single (formerly married/in abortion was associated with age, marital status, employ- union) = 3), and parity (none = 1; 1–2 children = 2; 3–5 ment status, educational status, parity, frequency of Sesay et al. BMC Women’s Health (2023) 23:44 Page 4 of 10 Table 1 Participant characteristics Characteristic 2013 (N = 16,658) 2019 (N = 15,574) 2013–2019 n (%) n (%) (N = 32,232) n (%) Age group 15–19 3878(23) 3427 (22) 7305(23) 20–24 2683(16) 2629 (17) 5312(16) 25–29 2843(17) 2728 (18) 5571(17) 30–34 2287(14) 1942 (12) 4229(13) 35–39 2260(14) 2224 (14) 4484(14) 40–44 1362(8) 1337 (9) 2699(8) 45–49 1344(8) 1288 (8) 2632(8) Marital status Never in union 4730(28) 5058(32) 9788(30) Married/in a union 10,903(65) 9715(62) 20,618(64) Single 1025(6) 801(5) 1826(6) Employment status Not working 5319(32) 4831(31) 10,150(31) Working 11,339(68) 10,743(69) 22,082(69) Residence Urban 5933(36) 7163 (46) 13,096(41) Rural 10,725(64) 8411 (54) 19,136(59) Educational status No education 9293(56) 7081(45) 16,375(51) Primary 2331(14) 2103(14) 4433(14) Secondary 4533(27) 5724(37) 10,257(32) Higher 501(3) 666(4) 1167(4) Wealth index Poorest 3089(19) 2738(18) 5828(18) Second 3046(18) 2831(18) 5877(18) Middle 3140(19) 2954(19) 6093(19) Fourth 3388(20) 3385(22) 6773(21) Richest 3994(24) 3666(24) 7660(24) Religion Christianity 3527(21) 3616(23) 7143(22) Muslim 13,032(78) 11,953(77) 24,985(78) Others 99(1) 6(0) 105(0) Parity None 4168(25) 4361(28) 8529(26) 1–2 4459(27) 5224(34) 9683(30) 3–5 5119(31) 4893(31) 10,011(31) 6 or more 2912(17) 1097(7) 4008(12) Current contraceptive use No method 12,982(78) 11,794(76) 24,776(77) Modern method 3361(20) 3696(24) 7058(22) Traditional method 315(2) 83(1) 398(1) Knowledge of ovulation Correct 4882(29) 7867(51) 12,748(40) Incorrect 6694(40) 5865(38) 12,558(39) Don’t know 5083(31) 1842(12) 6925(21) Frequency of reading newspaper Not at all 14,844(89) 14,320(92) 29,164(90) S esay et al. BMC Women’s Health (2023) 23:44 Page 5 of 10 Table 1 (continued) Characteristic 2013 (N = 16,658) 2019 (N = 15,574) 2013–2019 n (%) n (%) (N = 32,232) n (%) Less than once a week 666(4) 851(5) 1517(5) At least once a week 1149(7) 403(3) 1552(5) Frequency of listening to the radio Not at all 6325(38) 8653(56) 14,978(46) Less than once a week 3674(22) 3182(20) 6856(21) At least once a week 6659(40) 3739(24) 10,399(32) Frequency of watching television Not at all 13,217(79) 11,143(72) 24,359(76) Less than once a week 1090(7) 2109(14) 3199(10) At least once a week 2351(14) 2322(15) 4674(15) Ever had a pregnancy terminated No 15,099(91) 14,246(91) 29,345(91) Yes 1559(9) 1328(9) 2887(9) reading newspaper and frequency of listening to radio Discussion (p > 0.05) (Table 2). The prevalence of women who ever had a pregnancy terminated was 9% in both the 2013 and 2019 SLDHS, which is consistent with studies reported in Mozambique Predictors of termination of pregnancy among women [25] and Ethiopia [26] but lower (25%) than a study done of reproductive age in Sierra Leone in Ghana [25]. The reason for the difference between In the adjusted analysis for model 1, we found that the Sierra Leone and Ghana might be the differences in the respondent’s age, marital status, employment status, study period, target population, and the increased access parity, and exposure to radio were significant predictors to maternal health care services over the years. How- of induced abortion in the 2013 SLDHS (p < 0.05). For = ever, the prevalence of induced abortion in our study was example, women aged 45–49 years (AOR 4.60; 95%CI: found to be higher than in a study done among female 3.05–6.94) were about four times more likely to termi- university students in Wolaiytasodo, Ethiopia [27]. A nate a pregnancy compared to those aged 15–19. Also, = possible explanation might be the difference in the study women who were employed (AOR 1.71; 95% CI: 1.45– population. We utilized national-level data based on 2.02) were about twice more likely to terminate a preg- SLDHS, while the study in Wolaiytasodo Ethiopia was nancy compared to those who were unemployed. In the conducted among a particular population (female univer- adjusted analysis for model 2, respondent age, marital sta- sity students). tus, employment status, education, parity, and frequency Our study found a statistically significant relation- of listening to the radio and reading newspapers were ship between pregnancy termination and age, with the significant predictors of induced abortion. For example, odds higher among women 45–49  years. This finding women who listen to the radio (AOR = 1.57; 95% CI: is congruent with prior studies conducted in Ethiopia 1.23–2.01) had high odds of terminating a pregnancy [27], Ghana [25], and Mozambique [25], where older compared with those who do not listen to the radio. Also, = women experienced more abortion occurrences com-women who had primary education (AOR 1.37; 95% pared to their younger counterparts. This could be partly CI: 1.12–1.69) were more like to terminate a pregnancy explained by the fact that older women are predisposed compared to those with no education. In the adjusted to medical and pregnancy-related complications like car- analysis for model 3, the significant predictors of induced diovascular disease, diabetes mellitus and chromosomal abortion were age, marital status, employment status, abnormality, which could complicate the pregnancy and education, parity, and exposure to the radio. For instance, = result in a poor prognostic outcome [28]. Similarly, they women aged 45–49  years (AOR 7.91; 95% CI: 5.76– = may have attained their desired family size. On the con-10.87), married women (AOR 2.52; 95% CI: 1.95–3.26), = trary, a study in Ethiopia [29] reported that ever having working women (AOR 1.65; 95% CI: 1.45–1.87) had a a pregnancy terminated was higher in youth and young higher likelihood of terminating a pregnancy compared adults than in older women. with their counterparts (Table 3). Sesay et al. BMC Women’s Health (2023) 23:44 Page 6 of 10 Table 2 Cross-tabulation of participant characteristics and abortion among women of reproductive age in Sierra Leone Characteristic Ever had a pregnancy terminated 2013 SLDHS 2019 SLDHS 2013–2019 SLDHS No Yes Χ2 No Yes Χ2 No Yes Χ2 Age group 15–19 3786(98) 92(2) 48.06* 3386(99) 40(1) 56.86 * 7172(98) 133(2) 102.43 * 20–24 2501(93) 182(7) 2486(95) 143(5) 4986(94) 325(6) 25–29 2523(89) 320(11) 2476(91) 252(9) 4999(90) 572(10) 30–34 2022(88) 265(12) 1715(88) 227(12) 3737(88) 492(12) 35–39 1944(86) 316(14) 1936(87) 288(13) 3880(87) 604(13) 40–44 1170(86) 193(14) 1154(86) 183(14) 2323(86) 376(14) 45–49 1155(86) 190(14) 1093(85) 195(15) 2247(85) 385(15) Marital status Never in union 4574(97) 157(3) 81.02* 4932(98) 126(2) 86.46* 9506(97) 282(3) 167.74 * Married/in union 9644(88) 1258(12) 8619(89) 1096(11) 18,263(89) 2355(11) Single 881(86) 144(14) 695(87) 106(13) 1577(86) 250(14) Employment status Not working 5054(95) 265(5) 118.15 * 4633(96) 198(4) 120.59* 9687(95) 463(5) 236.58 * Working 10,045(89) 1294(11) 9613(89) 1130(11) 19,658(89) 2424(11) Residence Urban 5388(91) 545(9) 0.08 6589(92) 575(8) 1.44 11,977(91) 1120(9) 1.19 Rural 9711(91) 1014(9) 7657(91) 753(9) 17,369(91) 1767(9) Educational status No education 8289(89) 1004(11) 17.95* 6342(90) 740(10) 27.59* 14,631(89) 1744(11) 44.92 * Primary 2114(91) 217(9) 1897(90) 206(10) 4011(90) 423(10) Secondary 4260(94) 273(6) 5428(95) 296(5) 9688(94) 569(6) Higher 436(87) 65(13) 580(87) 86(13) 1016(87) 151(13) Wealth index Poorest 2832(92) 258(8) 0.83 2462(90) 276(10) 1.55 5294(91) 534(9) 0.30 Poorer 2739(90) 307(10) 2585(91) 246(9) 5324(91) 553(9) Middle 2832(90) 308(10) 2729(92) 225(8) 5561(91) 533(9) Richer 3089(91) 299(9) 3095(91) 290(9) 6184(91) 589(9) Richest 3608(90) 387(10) 3375(92) 291(8) 6983(91) 678(9) Religion Christianity 3239(92) 288(8) 2.56 3302(91) 314(9) 0.08 6541(92) 601(8) 1.10 Muslim 11,769(90) 1264(10) 10,939(92) 1014(8) 22,708(91) 2277(9) Others 91(93) 7(7) 5(88) 1(12) 96(92) 8(8) Parity None 3992(96) 176(4) 47.94 * 4165(96) 196(4) 30.75* 8157(96) 372(4) 75.58 * 1–2 4076(91) 383(9) 4732(91) 491(9) 8808(91) 875(9) 3–5 4499(88) 619(12) 4362(89) 531(11) 8861(89) 1150(11) 6 or more 2532(87) 380(13) 987(90) 110(10) 3519(88) 489(12) Current contraceptive use No method 11,738(90) 1243(10) 1.17 10,779(91) 1015(9) 0.92 22,518(91) 2258(9) Modern method 3070(91) 291(9) 3394(92) 303(8) 6464(92) 594(8) Traditional method 291(92) 24(8) 73(88) 10(12) 364(91) 35(9) Knowledge of ovulation Correct 4403(90) 479(10) 0.47 5337(91) 681(9) 3.07* 11,589(91) 1160(9) 0.46 Incorrect 6086(91) 608(9) 7186(91) 527(9) 11,423(91) 1135(9) Don’t know 4611(91) 472(9) 1722(93) 120(7) 6333(91) 592(9) S esay et al. BMC Women’s Health (2023) 23:44 Page 7 of 10 Table 2 (continued) Characteristic Ever had a pregnancy terminated 2013 SLDHS 2019 SLDHS 2013–2019 SLDHS No Yes Χ2 No Yes Χ2 No Yes Χ2 Frequency of reading newspaper Not at all 13,422(90) 1422(10) 2.91 13,096(91) 1224(9) 9.72* 26,518(91) 2646(9) 3.03 * Less than once a week 610(92) 56(8) 805(95) 46(5) 1415(93) 102(7) At least once a week 1068(93) 81(7) 344(85) 59(15) 1412(91) 139(9) Frequency of listening to a radio Not at all 5791(92) 534(8) 2.34 8020(93) 633(7) 8.05* 13,810(92) 1168(8) 10.65 * Less than once a week 3308(90) 365(10) 2865(90) 317(10) 6173(90) 682(10) At least once a week 6000(90) 659(10) 3361(90) 378(10) 9361(90) 1037(10) Frequency of watching television Not at all 11,959(90) 1258(10) 0.61 10,208(92) 935(8) 1.32 22,167(91) 2192(9) 1.03 Less than once a week 999(92) 91(8) 1944(92) 165(8) 2943(92) 256(8) At least once a week 2141(91) 210(9) 2094(90) 229(10) 4235(91) 439(9) * p-value < 0.05; SLDHS Sierra Leone demographic and health survey, X2 Chi-square The current study showed that maternal education was women with no children were more likely to terminate a a significant predictor of induced abortion. Women with pregnancy than those with parity four and above. It was primary education were more likely to have a terminated argued that women with no children are most likely to be pregnancy than uneducated women. This relationship is adolescents. They face challenges of unmet family plan- consistent with the report from a study done in Ethiopia ning and unintended pregnancies. [27]. Educated women are more likely to afford abortion It was observed that the prevalence of induced abortion services or more knowledgeable about abortion service was low among unmarried women compared to women providers and laws [30]. with other marital statuses. This corroborates the find- It was also observed that the odds of terminating preg- ings reported in previous studies done in Ethiopia [33] nancy were higher among working women than women and Nigeria [34]. Contrary to our findings, a study in who were not employed. This is consistent with previ- Nepal [35] explained that the high prevalence of abortion ous studies done in Mozambique [25] and Ghana [25]. among unmarried women is expected due to the unde- The high prevalence of pregnancy termination among sirable attitude of medical personnel, society, and fam- employed women can be partly explained by the fact that ily members towards never-married women. Similarly, they are financially empowered and can afford the cost the current study also found that pregnancy termination involved in terminating a pregnancy compared to their was high among single women. A possible explanation unemployed counterparts. In addition, it might be due might be that these women are without husbands, hence to the increase in knowledge and self-responsibilities as they are more likely to be single parent. Besides, there is a working woman. stigma associated with having children out of wedlock in In this study, media exposure was a significant predic- most Africa countries. tor associated with increased odds of induced abortion. These findings concur with studies from Ethiopia [26], Implications of the findings in this study Ghana [25], and Mozambique [25]. It could be due to the The findings from this study have implications for abor- reason that the media serve as an important channel of tion policy, programming and research. Induced abortion providing information about abortion care. Furthermore, constitutes a health problem among women of reproduc- women who have access to mass media may be knowl- tive age. Therefore, the relevant authorities must pro- edgeable about abortion laws and abortion pills [31, 32]. vide comprehensive and culturally appropriate sexual Regarding parity, the current study found that women and reproductive health services for women. Programs with parity of six and above were less likely to terminate addressing women’s education and livelihood should be a pregnancy than women with no children. This find- set up to help them make informed choices like contra- ing confirms what was found in a study done in Ghana ceptive use and prevention of unwanted pregnancies. [25] and Mozambique [25]. These studies reported that Presently, there is a paucity of literature in Sierra Leone Sesay et al. BMC Women’s Health (2023) 23:44 Page 8 of 10 Table 3 Logistic regression analysis of predictors of abortion among women of reproductive age in Sierra Leone Model 1 (2013 SLDHS) Model 2 (2019 SLDHS) Model 3 (2013–2019 SLDHS) Characteristic AOR 95% CI AOR 95% CI AOR 95% CI Age group 15–19 1(ref ) 1(ref ) 1(ref ) 20–24 2.39(1.65–3.47)* 3.97(2.51–6.26)* 2.91(2.19–3.87)* 25–29 3.74(2.56–5.46)* 6.58(4.01–10.78)* 4.74(3.58–6.36)* 30–34 3.72(2.52–5.49)* 9.64(5.87–15.83)* 5.68(4.22–7.66)* 35–39 4.67(3.11–6.99)* 11.37(6.96–18.56)* 6.98(5.15–9.45)* 40–44 4.71(3.09–7.17)* 13.05(7.98–21.35)* 7.45(5.44–10.21)* 45–49 4.60(3.05–6.94)* 15.06(9.04–25.08)* 7.91(5.76–10.87)* Marital status Never in union 1(ref ) 1(ref ) 1(ref ) Married/in union 2.09(1.48–2.95)* 3.00(2.03–4.42)* 2.52(1.95–3.26)* Single 2.25(1.51–3.35)* 2.84(1.69–4.76)* 2.59(1.88–3.55)* Employment status Not working 1(ref ) 1(ref ) 1(ref ) Working 1.71(1.45–2.02)* 1.54(1.26–1.88) * 1.65(1.45–1.87)* Residence Urban 1(ref ) 1(ref ) 1(ref ) Rural 0.99(0.73–1.35) 0.97(0.76–1.24) 0.97(0.79–1.19) Educational status No education 1(ref ) 1(ref ) 1(ref ) Primary 1.19(0.98–1.44) 1.37(1.12–1.69)* 1.27(1.11–1.46)* Secondary 1.16(0.91–1.47) 0.98(0.74–1.29) 1.04(0.86–1.24) Higher 1.62(0.93–2.83) 1.16(0.81–1.66) 1.34(0.98–1.84) Wealth index Poorest 1(ref ) 1(ref ) 1(ref ) Second 1.19(0.93–1.53) 0.87(0.69–1.09) 1.02(0.86–1.20) Middle 1.18(0.92–1.52) 0.81(0.63–1.03) 0.98(0.82–1.17) Fourth 1.15(0.84–1.57) 0.99(0.73–1.34) 1.06(0.85–1.33) Richest 1.46(0.94–2.26) 0.76(0.51–1.14) 1.06(0.78–1.44) Religion Christianity 1(ref ) 1(ref ) 1(ref ) Muslim 1.13(0.92–1.38) 1.02(0.85–1.22) 1.06(0.93–1.21) Others 0.92(0.39–2.15) 1.73(0.18–16.20) 0.92(0.41–2.03) Parity None 1(ref ) 1(ref ) 1(ref ) 1–2 0.73(0.53–1.01) 0.47(0.36–0.62)* 0.59(0.47–0.72)* 3–5 0.70(0.51–0.97)* 0.30(0.23–0.40)* 0.45(0.37–0.56)* 6 or more 0.69(0.47–1.01)* 0.23(0.17–0.33)* 0.40(0.31–0.52)* Current contraceptive use No method 1(ref ) 1(ref ) 1(ref ) Modern method 0.97(0.79–1.19) 1.08(0.91–1.29) 1.01(0.89–1.16) Traditional method 0.76(0.46–1.26) 1.43(0.69–2.94) 0.91(0.60–1.38) Knowledge of ovulation Correct 1(ref ) 1(ref ) 1(ref ) Incorrect 0.93(0.76–1.14) 0.99(0.84–1.16) 0.96(0.85–1.10) Don’t know 1.04(0.86–1.26) 0.04(0.78–1.39) 1.06(0.91–1.24) S esay et al. BMC Women’s Health (2023) 23:44 Page 9 of 10 Table 3 (continued) Model 1 (2013 SLDHS) Model 2 (2019 SLDHS) Model 3 (2013–2019 SLDHS) Characteristic AOR 95% CI AOR 95% CI AOR 95% CI Age group Frequency of reading newspaper Not at all 1(ref ) 1(ref ) 1(ref ) Less than once a week 1.05(0.72–1.52) 0.67(0.45–0.99) * 0.85(0.64–1.12) At least once a week 0.74(0.49–1.13) 1.72(1.11–2.65) * 1.00(0.74–1.35) Frequency of listening to a radio Not at all 1(ref ) 1(ref ) 1(ref ) Less than once a week 1.27(1.01–1.60) * 1.57(1.23–2.01) * 1.39(1.18–1.64) * At least once a week 1.27(1.04–1.54) * 1.24(0.97–1.59) 1.29(1.11–1.49) * Frequency of watching television Not at all 1(ref ) 1(ref ) 1(ref ) Less than once a week 0.85(0.63–1.15) 0.89(0.69–1.15) 0.87(0.72–1.06) At least once a week 0.97(0.74–1.28) 1.30(0.96–1.77) 1.10(0.90–1.35) Year of survey 2013 1(ref ) 2019 0.89(0.76–1.04) * p-value < 0.05, SLDHS Sierra Leone demographic and health survey, AOR Adjusted odd ratio, CI Confidence interval on the sociodemographic correlates of induced abortion abortion should focus on reducing unwanted pregnan- among women of reproductive age. This study set the cies through increasing access to modern contracep- platform for future research on the subject matter to aid tives among women of low socio-economic status. policymakers and programmers in decision-making and program planning. Abbreviations CI Confidence interval Strengths and limitations of the study DHS D emographic health survey A major strength of this study is that the analysis used MMR Maternal mortality ratioTOP Termination of pregnancy nationally representative data, following international SA S pontaneous abortion standards in every country. This study is the first study SLDHS S ierra Leone Demographic and Health Survey in Sierra Leone to assess the sociodemographic deter- WHO World Health Organization minants of induced abortion. These findings should be Acknowledgements interpreted with caution because cross-sectional studies FRS is receiving funding as a Ph.D. candidate from the HRP Alliance, part of cannot confirm causal relationships. Also, since abortion the UNDP/UNFPA-UNICEF-WHO-World Bank Special Programme of Research, Development and Research Training in Human Reproduction (HRP), a co- is a culturally sensitive issue and is based on self-report- sponsored program executed by the World Health Organization (WHO), to ing, there may be the possibility of social desirability bias complete the study. The authors would like to thank the DHS Program for that led to under-reporting. providing us with the data. Author contributions Conclusion FRS conceptualized the topic, designed and implemented the study, EAA assisted in analyzing the data, AM and EM helped put together the manu- This study revealed that a minority of Sierra Leo- script for publication, KT and RMKA supervised the study. All authors read and nean women of reproductive age had ever termi- approved the final manuscript. nated a pregnancy. Older age, higher education, being Funding employed, exposure to mass media, being single, and The authors received no funding for this research. low parity were significant determinants of induced abortion. Our study findings provide relevant informa- Availability of data and materialsAll data generated or analysed during this study are included in this published tion for maternal health policy and planning. We rec- article [and its supplementary information files]. ommend that interventions aimed at reducing induced Sesay et al. BMC Women’s Health (2023) 23:44 Page 10 of 10 Declarations 20. November L, Sandall J. ‘Just because she’s young, it doesn’t mean she has to die’: exploring the contributing factors to high maternal mortality Ethics approval and consent to participate in adolescents in Eastern Freetown; a qualitative study. Reprod Health. The study took approval online from the Sierra Leone Demographic Health 2018;15(1):1–18. Survey (DHS) ethics committee, and participants provided an informed 21. Redmond-Sovie MM. The Impact of regional and international law on consent form before the interviews began. All methods were carried out in domestic abortion law: a case study of Sierra Leone. Northeastern Univer- accordance with relevant guidelines and regulations. sity ProQuest Dissertations Publishing; 2020. 28025178. 22. Statistics Sierra Leone (SSL): Sierra Leone Demographic and Health Sur- Consent for publication vey (SLDHS 2013), Survey Findings Report. In Freetwon: Statistics Sierra Not applicable. Leone. 2013. 23. Statistics Sierra Leone (SSL). 2004 population and housing census report. Competing interests Freetown: Statistics Sierra Leone; 2004. The authors declare that they have no competing interests. 24. Statistics Sierra Leone (SSL). 2015 poulation and housing census report. Freetown: Statistics Sierra Leone; 2015. 25. Dickson KS, Adde KS, Ahinkorah BO. Socio–economic determinants of Received: 1 July 2022 Accepted: 13 January 2023 abortion among women in Mozambique and Ghana: evidence from demographic and health survey. Arch Public Health. 2018;76(1):1–10. 26. Tesema GA, Mekonnen TH, Teshale AB. Spatial distribution and determi- nants of abortion among reproductive age women in Ethiopia, evidence from Ethiopian Demographic and Health Survey 2016 data: Spatial and References mixed-effect analysis. PLoS ONE. 2020;15(6):e0235382. 1. Sahoo H, Stillman M, Frost J, Acharya R, Hussain R. Availability, practices 27. Gelaye AA, Taye KN, Mekonen T. Magnitude and risk factors of abortion and acceptance of postabortion contraceptive services in health facilities: among regular female students in Wolaita Sodo University, Ethiopia. BMC a study in six states of India. Contraception. 2019;101(2):106–11. Womens Health. 2014;14(1):1–9. 2. Paul M, Gebreselassie H, Samai M, Benson J, Kargbo S, Lazzarino M. 28. Ryan E, Savu A, Yeung R, Moore L, Bowker S, Kaul P. Elevated fasting vs Unsafe abortion in Sierra Leone: an examination of costs and bur- post-load glucose levels and pregnancy outcomes in gestational diabe- den of treatment on healthcare resources. J Women’s Health Care. tes: a population-based study. Diabet Med. 2020;37(1):114–22. 2015;4(1000228):2167–420. 29. Gebreselassie H, Fetters T, Singh S, Abdella A, Gebrehiwot Y, Tesfaye S, 3. Word Health Organization. Abortion. In. Geneva: WHO; 2020. Geressu T, Kumbi S. Caring for women with abortion complications in 4. World Health Organization. Factsheet: preventing unsafe abortion. Ethiopia: national estimates and future implications. Int Perspect Sex Geneva: WHO; 2018. Reprod Health. 2010;36:6–15. 5. Huber D, Curtis C, Irani L, Pappa S, Arrington L. Postabortion care: 20 years 30. Yaya S, Uthman OA, Ekholuenetale M, Bishwajit G. Women empower- of strong evidence on emergency treatment, family planning, and other ment as an enabling factor of contraceptive use in sub-Saharan Africa: programming components. Glob Health Sci Pract. 2016;4(3):481–94. a multilevel analysis of cross-sectional surveys of 32 countries. Reprod 6. Bearak J, Popinchalk A, Ganatra B, Moller A-B, Tunçalp Ö, Beavin C, Kwok Health. 2018;15(1):1–12. L, Alkema L. Unintended pregnancy and abortion by income, region, and 31. LaRoche KJ, Foster AM. Exploring Canadian women’s multiple abortion the legal status of abortion: estimates from a comprehensive model for experiences: implications for reducing stigma and improving patient- 1990–2019. Lancet Glob Health. 2020;8(9):e1152–61. centered care. Womens Health Issues. 2018;28(4):327–32. 7. Sultana A. Induced abortion: rural Women’s escape against unintended 32. Vázquez-Quesada L, Shukla A, Vieitez I, Acharya R, RamaRao S. Abortion pregnancies. Pak Vis. 2020;21(1):466. self-care: a forward-looking solution to inequitable access. Int Perspect 8. Guttmacher Institute. Unintended pregnancy and abortion worldwide. Sex Reprod Health. 2020;46(Supplement 1):91–5. New York: Guttmacher Institute; 2020. 33. Gilano G, Hailegebreal S. Determinants of abortion among youth 9. Singh S, Remez L, Sedgh G, Kwok L, Onda T. Abortion worldwide 2017: 15–24 in Ethiopia: a multilevel analysis based on EDHS 2016. PLoS ONE. uneven progress and unequal AccessAbortion worldwide 2017: uneven 2021;16(3):e0248228. progress and unequal access. 2018. https:// www.g uttm acher.o rg/r eport/ 34. Ibrahim I, Onwudiegwu U. Sociodemographic determinants of compli- abort ion-w orld wide- 2017. cated unsafe abortions in a semi-urban Nigerian town: a four-year review. 10. World Health Organization: Safe abortion: technical and policy guidance West Indian Med J. 2012;61(2):163. for health systems. Geneva: World Health Organization; 2012. 35. Puri M, Lamichhane P, Harken T, Blum M, Harper CC, Darney PD, Hender- 11. World Health Organization: Preventing unsafe abortion: fact sheet. In son JT. “Sometimes they used to whisper in our ears”: health care workers’ Geneva: WHO; 2014. perceptions of the effects of abortion legalization in Nepal. BMC Public 12. World Health Organization. World health statistics, monitoring health for Health. 2012;12(1):1–9. the SDGs. Geneva: WHO; 2015. 13. UK Legislation. Offences against the person act of 1861. In: UK public Publisher’s Note general act 100 c (Regnal 24 and 25 Vict) sec58. 2017. Springer Nature remains neutral with regard to jurisdictional claims in pub- 14. Mutua MM, Manderson L, Musenge E, Achia TNO. Policy, law and post- lished maps and institutional affiliations. abortion care services in Kenya. PLoS ONE. 2018;13(9):e0204240. 15. Guilbert K. Strict laws drive West African women to dangerous, illegal abortions. 2016. https:// news. trust. org/i tem/ 20160 41809 1818- qx2zt. 16. World Health Organization. Country health statistics, Freetown, Sierra Leone. Freetown, Sierra Leone: WHO; 2013. 17. Statistics Sierra Leone (SSL), ICF International. Sierra Leone demographic and health survey. Freetown: Ministry of Health and Sanaitation & Statis- tics Sierra Leone; 2019. 18. Statistics Sierra Leone (Stats SL), ICF International. Sierra Leone demo- graphic and health survey 2019. Sierra Leone: Staistics Staistics Sierra Leone (SSL); 2019. 19. Ministry of Health and Sanitation Sierra Leone. Maternal death surveil- lance and reponse. Freetown: Directorate of Reprodutive & Child Health; 2016.