Determinants of facility-based childbirth among adolescents and young women in Guinea: a secondary analysis of the 2018 Demographic and Health Survey

Introduction Maternal mortality remains very high in Sub-Saharan African countries and the risk is higher among adolescent girls. Maternal mortality occurs in these settings mainly around the time of childbirth and the first 24 hours after birth. Therefore, skilled attendance in an enabling environment is essential to reduce the occurrence of adverse outcomes for both women and their children. This study aims to analyze the determinants of facility childbirth among adolescents and young women in Guinea. Methods We used the Guinea Demographic and Health Survey (DHS) conducted in 2018. All females who were adolescents (15 -19) or young women (20-24 years) at the time of their most recent live birth in the five years before the survey were included. We examined the use of health facilities for childbirth and its determinants using multivariable logistic regression, built through the Andersen health-seeking model. Results Overall, 58% of adolescents and 57% of young women gave birth in a health facility. Young women were more likely to have used private sector facilities compared to adolescents (p<0.001). Factors significantly associated with a facility birth in multivariable regression included: secondary or higher educational level (aOR=1.81; 95%CI:1.20-2.64) compared to no formal education; receipt of 1-3 antenatal visits (aOR=8.93; 95%CI: 5.10-15.55) and 4+ visits (aOR=15.1; 95%CI: 8.50-26.84) compared to none; living in urban (aOR=2.13; 95%CI: 1.40-3.37) compared to rural areas. Women from poorest households were least likely to give birth in health facilities. There was substantial variation in the likelihood of birth in a health facility by region, with highest odds in NZerekore and lowest in Labe. Conclusion The percentage of births in health facilities among adolescents and young women in Guinea increased since 2012 but remains suboptimal. Socio-economic characteristics, region of residence and antenatal care use were the main determinants of its use. Efforts to improve maternal health among this group should target care discontinuation between antenatal care and childbirth (primarily by removing financial barriers) and increasing the demand for facility-based childbirth services in communities, while paying attention to the quality and respectful nature of healthcare services provided there.


23
Maternal mortality remains very high in Sub-Saharan African countries and the risk is higher among 24 adolescent girls. Maternal mortality occurs in these settings mainly around the time of childbirth and the 25 first 24 hours after birth. Therefore, skilled attendance in an enabling environment is essential to reduce 26 the occurrence of adverse outcomes for both women and their children. This study aims to analyze the 27 determinants of facility childbirth among adolescents and young women in Guinea.

29
We used the Guinea Demographic and Health Survey (DHS) conducted in 2018. All females who were 30 adolescents (15 -19) or young women (20-24 years) at the time of their most recent live birth in the five 31 years before the survey were included. We examined the use of health facilities for childbirth and its 32 determinants using multivariable logistic regression, built through the Andersen health-seeking model.

34
Overall, 58% of adolescents and 57% of young women gave birth in a health facility. Young women were Reducing maternal mortality to achieve Sustainable Development Goal 3 (SDG 3) is a major public health 54 challenge. Despite efforts over the past 25 years, maternal mortality remains disproportionately high in low-55 and middle-income countries where the need to address the unmet need for family planning is most 56 significant, particularly among adolescents [1]. The World Health Organization (WHO) estimated that 12 57 million adolescents aged 15-19 years give birth each year globally. The problem is most prevalent in sub-

58
Saharan Africa (SSA), with variations between countries [2]. The recent Demographic and Health Survey 59 (DHS 2018) in Guinea reported a fertility rate of 120 live births per 1,000 adolescents aged 15-19 [3]. This 60 rate is higher than the West African average of 115 per 1,000 [4]. Giving birth during adolescence carries a 61 higher risk of adverse outcomes for both the girl and the baby due in part to the mother's biological and 62 physiological immaturity [5]. An important strategy to reduce these risks is to improve skilled birth 63 attendance in an enabling environment [6]. However, despite this particular need, sub-Saharan adolescents 64 are less likely to give birth in a health facility than older women [7,8].

66
The prevalence of facility-based childbirth among sub-Saharan adolescents varies across countries. Overall,

67
an estimated 65% of adolescents give birth in a health facility [9]. This ranges from 23% in Chad [10]

71
(2018) found that the likelihood of facility-based delivery increases with maternal age [7]. However, Adde and older women in facility-based childbirth [9]. These two studies using a sample of women aged 15-49 74 years found that urban residence and higher wealth quintile were associated with a higher likelihood of 75 facility-based childbirth in SSA countries [7,9]. In Bangladesh, Shahabuddin et al. (2016) found that the 76 frequency of facility-based childbirth varied across regions [12]. Some studies among women of 77 childbearing age in SSA also found that a secondary or higher educational level was associated with a higher 78 likelihood of facility-based childbirth [9,13]. In addition, Doctor et al. found that women with at least 79 primary education were more likely to deliver in a health facility than woman with no formal education [7].

80
In Nigeria, Dahiru et al. (2015) found that women whose husbands had at least primary education were 81 more likely to deliver in a health facility [14]. Further, Adde et al. (2020) and Doctor et al. (2018) found that 82 the odds of facility-based childbirth were higher among women who had completed at least one antenatal 83 visit [7,9]. These findings had also been reported by Sanni et al. in Ethiopia in 2018 [13]. In addition, 84 frequent listening to the radio or watching television were also associated with the use of health facilities 85 for childbirth by women of childbearing age [9].

87
In a systematic review including 27 studies, Mekonnen et al. (2015) found that distance to the health facility 88 and community factors such as the proportion of educated women in the community and the rate of ANC . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted April 12, 2022 use in the community were associated with the use of maternal health services among adolescent in SSA 90 [15]. In Niger, Rai et al. (2013) found that the education of both spouses, belonging to a social group such 91 as the Deermal/Songhai, and having some decision-making autonomy increased women's chances of 92 having safe deliveries [16].

94
Guinea did not achieve the Millennium Development Goal 4 and is unlikely to achieve the Sustainable . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

124
Independent variables 125 We used Anderson's Behavioral Model of Health Services utilization [19] (Fig 1) to identify relevant and 126 available variables. We included woman's age at index birth (15-19; 20-24), woman's highest educational 127 level at survey (no formal education, any primary, and any secondary or higher). The usual age of primary 128 school completion in Guinea is 12 years [20]. We also included woman's marital status at survey 129 (married/cohabiting or not), women's religion (Muslim, Christian, other), and number of ANC visits during 130 the pregnancy preceding the index birth (none, 1-3, 4 or more). Women with 'don't know' responses on 131 number of ANC visits were included in the 1-3 category because we assumed that these women had some 132 ANC but could not recall the exact number of visits.

135
Place of residence as specified on the DHS sampling strategy (rural, urban), region of residence (Conakry,

136
Boké, Kindia, Mamou, Labé, Faranah, Kankan, and NZérékoré) and women's perception of distance to 137 the health facility (big problem or not) were used as proxies for health service availability. Exposure to 138 media (listens to radio or not) was used as a proxy access to health information. We used household wealth 139 quintile (poorest, poorer, middle, richer, and richer) to capture financial access to care. Perceived need 140 factors were mother's parity (no previous birth, 1 to 2 previous births, 3 or more births) and wantedness of 141 the index pregnancy (wanted, mistimed/unwanted).

143
The data were processed and analyzed using Stata 16.1 software (StataCorp, College Station, TX USA).

144
Two levels of statistical analysis were applied. The first level consisted of describing the characteristics of 145 the included sample. We obtained absolute numbers and percentages of the characteristics under study for 146 adolescents, young women, and the whole sample. We also compared the prevalence of facility-based 147 childbirth by type of health facility between adolescents and young women using Pearson's Chi-squared 148 test.

149
We then analyzed the determinants of facility-based childbirth among the combined sample of adolescents 150 and young women. First, we conducted a bivariate analysis to assess the relationship of each independent 151 variable with adolescents' or young women's use of health facilities for their most recent birth. We initially 152 ran one multivariable regression model for each age group, but we found no difference between the 153 characteristics associated with the outcome of interest between the two age groups. Therefore, we 154 constructed a single multivariable model with both age groups. Before fitting the final model, we assessed 155 the level of collinearity between the independent variables. The objective was to determine the relevance 156 of the variables selected for inclusion in the analyses. As a general rule, a mean-variance inflation factor 157 (VIF) score below five is tolerated. In contrast, a mean score greater than or equal to five suggests that the 158 regression coefficients could be misestimated. The VIF reported by the test remained in the tolerable limit.
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted April 12, 2022. ; https://doi.org/10.1101/2022.04.06.22273508 doi: medRxiv preprint

159
The Likelihood ratio test (LRtest) was also performed for each variable with more than two categories to 160 assess how its inclusion impacted the over model. . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted April 12, 2022 188

Proportions of facility-based delivery among adolescents and young women
189 Overall, about 58% adolescents and young women in Guinea gave birth to their most recent baby in a 190 health facility. This proportion was similar in adolescents (58%) and young women (57%). No difference 191 was noted between the two age groups in the use of government hospitals for childbirth (p=0.332).
194 Table 2. Analysis of the frequency of facility-based childbirth by level and health sector among adolescents 195 and young women for the most recent live birth in the five years prior to the 2018 DHS (N=2154).

Determinants of facility-based childbirth among adolescents and young women
201 Table 3 shows the percentage of births occurring in health facilities stratified by the independent variables.
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

202
In bivariate analysis, education level, marital status, parity at survey, exposure to media, number of ANC 203 visits, perceived distance to health facility, religion, region, residence, and household wealth quintile were 204 associated with facility-based childbirth in the combined sample of adolescents and young women.

205
In multivariable analysis ( . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted April 12, 2022

226
This study examined recent levels and determinants of facility-based childbirth among adolescents and 227 young women in the Republic of Guinea. The overall facility-based childbirth level found was 58%, 228 comparable among adolescents and young women. The study also found that about 6% of women included 229 in the analysis gave birth in the private sector. This was more frequent in young women than adolescents.

230
In addition, most adolescents and young women had initiated prenatal visits, although few reached the 231 recommended four or more ANC visits (37% of the combined sample). In multivariable analyses, education 232 level, number of ANC visits, religion, residence, administrative region, and household wealth index were 233 associated with facility-based delivery. The strongest predictors of higher facility-based childbirth were the 234 use of 4 or more ANC visits, household wealth, and residence in the NZérékoré region. In addition, 235 compared to respondents who were having their second or third child, women giving birth to their first 236 child had higher adjusted odds of giving birth in a health facility.

237
The overall levels of facility-based childbirth among adolescents and young women found in this study

238
(58%) were statistically higher than what was reported from the same DHS among all women of 239 reproductive age in Guinea (53%). In Guinea, one of the most important policies improving access to 240 maternal health services during this period is the 2011 user fee removal policy [21]. As a result, the 241 percentage of all births occurring in health facilities has increased from 40% in 2012 to 53% in 2018. This

242
increase is even more pronounced in the under 20 years age group (from 41% to 58%, p<0.001) [25,26]. A 243 similar increase has previously been noted in other sub-Saharan African countries, including Uganda [22].

244
However, despite this increase, the prevalence of facility-based childbirth among adolescents and young 245 women in Guinea is far behind the average for sub-Saharan African countries (65%) [9], including 246 neighboring countries such as Senegal (8%) and Mali (71%) [23,24]. Furthermore, we believe that this result 247 is sub-optimal in light of the challenges associated with high maternal mortality and morbidity in the country 248 and is hampering the achievement of Sustainable Development Goal 3 by 2030.

249
Our study identified large inequalities in the use of health facilities for childbirth across household wealth 250 quintiles and geographic regions. Compared to adolescents and young women living in the poorest 251 households, those from richer and richest wealth quintiles were more than five times as likely to deliver in countries [7,25,26]. In regard to geographical inequalities, we found that while adolescents and young 257 women living in Conakry had the highest bivariate odds of facility-based delivery, it was the region of

258
NZérékoré in which the highest adjusted odds were found (compared to Labé). While an in-depth 259 understanding of the reasons for this finding, we suspect that local health system actors are implementing 260 national guidelines differently across the country, depending on local challenges such as the availability of . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted April 12, 2022 of trust with the health system. It is also an opportunity for the providers to detect potential risk factors 293 and educate the woman on the benefits of continued health service utilization [32].

295
We found that primiparity was associated with a higher facility-based delivery. This is consistent with studies 296 conducted in sub-Saharan African contexts [33]. One reason for this could be the perceived higher health . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted April 12, 2022. ; https://doi.org/10.1101/2022.04.06.22273508 doi: medRxiv preprint risk attributed to the first pregnancy in such contexts [34]. We hypothesize that after the first delivery in a

298
facility where the woman may experience poor quality services and mistreatment, she would choose to 299 deliver in a private facility, if she affords, or within her community for the subsequent deliveries [35].

300
However, a study on facility-based delivery including DHS from 34 sub-Saharan African countries reported 301 that among women having their first birth, the youngest had significantly lower likelihood of delivering in 302 a facility [36]. In any case, experiencing poor quality disrespectful care at the beginning of reproductive life 303 carries negative consequences for the woman's future use of health services, her wellbeing and survival, and 304 that of her child. This is especially the case in Guinea where the total fertility rate of 4.8 [37] and high 305 maternal mortality rate (576/100,000 live births in 2017) combine into a very high lifetime risk of maternal 306 death of 1 in 35 [1].

308
This study benefited from a recent nationally representative sample of women of childbearing age.

309
However, we note some limitations. First, the study's cross-sectional nature does not allow for a temporal

317
This study showed facility-based births among adolescents and young women in Guinea are higher than in 318 the general population. In the pathway to achieve the SDG targets by 2030, there is a need to address 319 inequalities related to financial and geographic accessibility given that the poorest women are still unable to 320 afford the costs of childbirth in health facilities. Also, positive lessons for the regions achieving higher levels 321 should be documented and scaled up. Another gap identified in this study relates to the quality of care 322 across the continuum of care, which requires more investment and accountability in the health sector.

323
Extending beyond the care received at childbirth, attention to accessible, acceptable, affordable, and high-324 quality maternal care must be paid to antenatal care as the first step in the continuum of care and an 325 important determinant of the use of facilities for childbirth. This is particularly important among 326 adolescents and young women whose well-being will be affected by the care experience and health 327 outcomes experiences at the beginning of their reproductive lives.

329
Availability of data and materials 330 The datasets analyzed during the current study are accessible on https://dhsprogram.com/data/available-331 datasets.cfm . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted April 12, 2022 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted April 12, 2022. ; https://doi.org/10.1101/2022.04.06.22273508 doi: medRxiv preprint