1 Research Article

2 Assessing the geographical variation of antenatal care and maternal mortality rate

3 among pregnant women in Sierra Leone based on DHIS2 data from 2020–2022: A

4 cross-sectional study

5 Morlai Sesay1, Professor Guan Ying2

1
6 School of Public Health, Southern Medical University, 1023-1063 Shatai Nan Road,

7 Guangzhou 510515, China. Email address: [email protected]

2
8 Department of Biostatistics, School of Public Health, Southern Medical University, 1023-

9 1063 Shatai Nan Road, Guangzhou 510515, China. Email address: [email protected]

10

11 *Corresponding author

12 Professor Guan Ying: Department of Biostatistics, School of Public Health, Southern

13 Medical University, 1023-1063 Shatai Nan Road, Guangzhou 510515, China. Email address:

14 [email protected]

15

16

17

18

19

20

21

22

23

24
25 Abstract

26 Background: Sierra Leone reported a high maternal mortality ratio of 717 per 100,000 live

27 births in 2019 due to the disparities in the use, quality, and access to skilled birth attendants

28 and antenatal care (ANC) services. Our study aims to assess the geographical variation of

29 ANC and maternal mortality rate (MMR) among pregnant women in Sierra Leone based on

30 District Health Information System 2 (DHIS2) data from 2020 to 2022.

31 Methods: We employed a cross-sectional study design, which obtained data from the DHIS2

32 for all five administrative regions of Sierra Leone. The analysis of the data included all

33 pregnant women and maternal deaths captured in the DHIS2. MMR distribution was

34 evaluated using descriptive statistics, while the geographical disparity was done using a

35 factorial analysis. Multivariate regression and bivariate correlation analysis were performed

36 using SPSS 25.0 to show the relationship between ANC and MMR.

37 Results: Our findings showed an MMR of 2759.5 per 100,000 live births in 2020, which

38 significantly reduced to 848.2 in 2021 and 288.2 in 2022. However, regional disparities

39 persisted, particularly in the northern, western, and northwest. ANC utilization also varies

40 across regions, with the western region having 50.0% of births in clinics, while other regions

41 depend on maternal and child health centers. Our study revealed a significant association

42 between the regions and institutions of giving birth (p<0.001). The factorial analysis showed

43 that regional differences were less pronounced each year despite significant improvements

44 over time in early ANC visits, hemoglobin testing, syphilis screening, and intermittent

45 preventive treatment (IPT) doses. The northern and southern regions mainly utilize traditional

46 birth attendants (TBAs). Southern and northwest regions demonstrated significant variability

47 in the utilization of doctors and midwives over time. IPT 1 st dose turned out to be most
48 effective (p<0.001) in reducing MMR. Our study revealed a significant negative correlation

49 for ANC 1st visit, 8th visit, iron folate, and malaria in the 2nd or 3rd trimester treated.

50 Conclusions: This study showed significant maternal health disparities in Sierra Leone, with

51 regions having lower ANC utilization rates facing higher MMRs. Improvement in access and

52 quality of ANC services, in northern, northwest, and western regions, and engaging TBAs

53 and other trained healthcare professionals in facility delivery system in each region are

54 critical to reducing maternal mortality and achieving sustainable development goal 3.1.

55 Keywords: Geographical Variation, Maternal Mortality Rate, Antenatal/Prenatal Care,

56 Pregnant Women, DHIS2 Data, Sierra Leone

57

58

59

60

61

62

63

64

65

66

67

68

69

70

71

72
73 Background

74 Maternal mortality is still more significant in the world's poorest countries compared to the

75 richest more than a century ago, making maternal mortality a severe global health concern. In

76 2020, the global maternal mortality rate (MMR) was 223 per 100,000 live births . The WHO

77 reported that MMR stagnated or worsened in many regions between 2015 and 2020.

78 Nevertheless, MMR in Australia and New Zealand decreased by 34.6% while the sub-

79 Saharan African (SSA) region continued to bear the highest risk burden, with 201,000 (66%)

80 per 100,000 live births in 2015 .

81 The high maternal death rates in the case of SSA are governed by inequitable access to

82 healthcare, geographical disparity, and sociocultural backgrounds. Furthermore, although

83 Antenatal care (ANC) is essential for the prevention of maternal death , it is underutilized,

84 especially among geographically isolated pregnant women. ANC visits are measured by

85 indicators such as ANC 1+ or ANC4+. Frequent trans-interactions with the healthcare

86 systems only sometimes result in sufficient care delivery in deprived settings .

87 Sierra Leone's maternal mortality ratio remains among the highest in the world at an alarming

88 1360 and 717 per 100,000 live births in 2015 and 2019, respectively. Unsafe abortion,

89 hemorrhage, and sepsis are the major causes of maternal deaths and delays in getting timely

90 quality care from proximity and infrastructure resources , which are some of the contributing

91 factors. The distribution of iron folate, insecticide-treated nets, and intermittent preventive

92 treatment interventions have succeeded in reducing maternal morbidity from government

93 interventions, but disparities persist .

94 There is a low ANC attendance rate in Sierra Leone (24% had four or more visits from 2008

95 to 2013). This is due to problems in service delivery and poor access to health insurance .

96 Studies have shown the relationship between ANC attendance and institutional deliveries,
 97 with women with adequate ANC typically giving birth normally, while those with inadequate

98 ANC often undergo cesarean sections. These studies highlight maternal education, facility

99 accessibility, and decision-making as major features concerning institutional births .

100 The Global ambitions under Sustainable Development Goals (SDGs) 3.1 aim to reduce MMR

101 by 70% in 2030. Sierra Leone piloted a maternal death surveillance and response (MDSR)

102 system in 2016 to help prevent maternal deaths. However, gaps still need to be filled in

103 assessing geographical disparities in ANC and MMR using the District Health Information

104 System 2 (DHIS2), requiring data-driven strategies for equitable healthcare. To the best of

105 our knowledge, no study has been conducted to assess the geographical variation of antenatal

106 care and maternal mortality rate among pregnant women in Sierra Leone based on DHIS2

107 Data. Therefore, this study seeks to close this gap by assessing the geographical variation of

108 antenatal care and maternal mortality rate among pregnant women in Sierra Leone based on

109 the DHIS2 data from 2020–2022. Our research determined the distribution of maternal

110 mortality rates in Sierra Leone, identified and mapped geographical disparities in ANC

111 utilization, and investigated the association between ANC services, visits, and maternal

112 mortality rates. The findings might contribute to sound policy-making and the dissemination

113 of information at the global and regional levels. These findings will help shape policies and

114 programs to overcome barriers to maternal health and provide a prototype for more extensive

115 assessment in similar lower resource settings, improving understanding of maternal health

116 epidemiology.

117 Methods

118 Study design and setting

119 We did a cross-sectional study design with a quantitative approach. Administratively, Sierra

120 Leone is divided into five regions (namely: Northern, Southern, Eastern, Western Area, and

121 Northwest); each region is subdivided into a few districts, and each district is divided into
122 chiefdoms. There are 16 districts and 190 chiefdoms in the country . Sierra Leone is one of

123 the poorest countries in the world . According to 2021 mid-term census results, a total

124 population of 7,541,641 was reported, with a population of 3,716,263 males and females

125 accounting for 3,825,378 .

126 Data source and study population

127 Routine program data from the DHIS2, sourced from all five administrative regions in 16

128 districts and health facilities in Sierra Leone, was available from 2020 to 2022 via this link:

129 https://sl.dhis2.org (you must have authorization and access to the relevant organizational

130 unit to access reproductive health data). In Sierra Leone, the Management of the Health

131 Information System (HMIS) is the principal responsibility of the Directorate of Policy,

132 Planning, and Information (DPPI) in the Ministry of Health and Sanitation (MoHS) . The

133 District Health Information System 2 (DHIS2) was established in 2008 . Data were recorded

134 in eight different forms, representing the different morbidities and services at the peripheral

135 health unit (PHU) level. Copies were then sent to the district level normally every month,

136 which were entered into the DHIS2 database and synced into national for informed decisions.

137 The variables were extracted from the reporting system (DHIS2) within the period under

138 review. This data was chosen to utilize due to its reliability as a national dataset and its

139 integration into the country's disease and event surveillance system. Since Sierra Leone

140 established the MDSR system in 2016, this study targeted the 2020–2022 timeframe because,

141 by that time, Sierra Leone's MDSR system, which had been put in place in 2016, had

142 developed and produced more trustworthy maternal health data.

143 The study population included all the pregnant women and maternal deaths captured in the

144 DHIS2 database from 2020 to 2022 nationwide. All pregnant women and maternal deaths

145 reported in the DHIS2 database were included in the study. In contrast, the exclusion criteria
146 were: (i) data from districts or regions outside of Sierra Leone were excluded to maintain the

147 geographical focus of the study, (ii) data from sources other than DHIS2, (iii) data related to

148 individuals who were not pregnant during the period, (iv) records with missing or incomplete

149 information on key variables, such as ANC utilization indicators or pregnancy health

150 outcomes and (v) duplicate or redundant data entries. The MMR was calculated using the

151 formula;

152

153 Data Analysis

154 The predictors were analyzed with a 0.05 margin of error, a 95% confidential level, and a 5%

155 significance level. Inferential statistics were employed to evaluate the likelihood and

156 determine if the observed variation between groups resulted from chance or dependability. A

157 parametric and nonparametric test, specifically chi-square tests, was employed to draw

158 inferences about the outcomes based on the data distribution. This study involved different

159 times and regions in identifying and mapping geographical disparities in ANC utilization

160 across regions; therefore, factorial design was used. Multivariate regression and bi-variate

161 correlation analysis were done to determine the association between the ANC services and

162 maternal mortality rates. The data was analyzed using the Statistical Package for the Social

163 Sciences version 25.0 (SPSS 25.0).

164

165

166

167 Results
168 The distribution of maternal mortality rates in Sierra Leone

169 The MMR reported in 2020 was 2759.1 per 100,000 live births; in 2021, it was 848.2 per

170 100,000 live births, and the lowest rate (288.2 per 100,000 live births) was reported in 2022.

171 In 2020, the northern region had the highest MMR (4424.5 per 100,000 live births), followed

172 by western (3888.1), northwest (3031.7), eastern (2234.0), and southern (1159.4), which

173 reported the lowest rate. Within 2021, the northern region still has a relatively high rate

174 (2023.2 per 100,000 live births). The northwest (893.7), western (924.7), and southern

175 (487.4) regions showed considerable reductions in maternal deaths. During 2022, the

176 northern region, which had the highest in 2020, dropped sharply to 221.9 per 100,000 live

177 births. The northwest (176.9), southern (275.6), and western (279.2) regions maintained

178 relatively low MMR compared to previous years, while the eastern (437.5) recorded a higher

179 rate than in 2021, as displayed in (Figure 1).

180 MMR due to different etiologies (causes), 2020-2022

181 Figure 2 below presents the MMR varied across different aetiologies (causes) in Sierra Leone

182 from 2020 to 2022. The MMR values are significantly higher in 2020 compared to 2021 and

183 2022 for most causes. Complications (3587.1), infection deaths (3320.7), hemorrhage

184 (1244.3), and malaria (1976.7) were the major causes of maternal deaths in 2020.

185 Hypertension deaths (1290.8), hemorrhage (986.6), malaria (707.6), and infection death

186 (3208.2) top the chart of deaths in 2021, whereas hemorrhage (418.9), hypertension death,

187 and complications (361.5) bear the burden of maternal deaths in 2022. There is a steep

188 decline in mortality rate due to anemia, ectopic pregnancy, ruptured uterus, and obstructed

189 labor from 2020-2022.

190 MMR association analysis of deaths due to different etiologies, 2020-2022

191 The correlation between MMR values for all causes was significant and positive. Figure 3

192 MMR association analysis of deaths due to different aetiologies, 2020-2022

193 Geographical Disparities in ANC Utilization

194 Analysis of differences in place or institution of pregnant women giving birth in

195 different regions

196 Table 1 shows the number and percentage of each institution or place of birth in the

197 regions. The results show differences in the number of institutions where pregnant

198 women give birth in different regions; for example, 50.0% of pregnant women in the

199 western region give birth in clinics, while the other regions are mainly in maternal and

200 child health posts (MCHPs), community health centers (CHCs), and community health

201 posts (CHPs). A significant relationship exists between the regions and institutions of

202 pregnant women giving birth (P<0.001).

203 Analysis of ANC visits, ANC services and birth attendants in different regions, 2020-

204 2022

205  The original results reported the frequency, which varied between different times

206 and regions, so the principle of relative numbers was used for analysis.

207  Data analysis was normally distributed and statistically described using means and

208 standard deviations.

209  This study involved different times and regions; therefore, factorial design was

210 used for the analysis.

211

212 ANC Visits by Region: Mean Values, standard deviation (SD), F-values, and P-Values

213 for 2020-2022

214 ANC Factorial analysis for <12-week of prenatal care–There is a difference in timing, with a

215 two-fold increase in the number of pregnant women screened for prenatal care <12 weeks

216 (W12) compared to 2020 over time (28.375, P<0.001). In the same year, there is no

217 difference in the number of <W12 examined between regions (2.122, P=0.100); differences

218 between years in the same region indicate a growing awareness of the importance of W12

219 antenatal care in (western-P=0.028; northern-P=0.008; eastern-P=0.004; and northwest-

220 P=0.004).

221 1st Factorial Analysis–There is a difference in timing (4.240, P=0.021), with a significant

222 increase in the number of pregnant women attending the first antenatal visit from 2020 to

223 2022. No statistically significant difference exists in the number of pregnant women

224 attending antenatal visits in different regions in the same year (2.152, P=0.092).

225 4th Factorial Analysis–There are differences in time (12.237, P<0.001) and no differences in

226 regions (1.941, P=0.127) within any single year. Although several regions, particularly the

227 Northern (5.644, P=0.026), Eastern (8.995, P=0.016), Southern (6.186, P=0.020), and

228 northwest (6.051, P=0.036) regions, showed increases over time, there were no notable

229 regional variances within a year. The overall trend is also significant. In areas other than the

230 western part, the frequency of attendance increased by 20% in 2022 compared to 2020.

231 8th Factorial Analysis-There were differences in the timing of factorial analysis (73.823,

232 P<0.001) and no differences in regions (2.148, P=0.087) within any single year. The factorial

233 analysis of the fourth antenatal examination showed a significant decrease in eighth ANC

234 visits seen over time in the northern (24.711, P<0.001), eastern (119.111, P<0.001), southern

235 (19.038, P=0.001), and northwest regions (29.833, P=0.001), with the overall trend across all

236 regions also being highly significant (73.823, P<0.001). The results are shown in (Table 2).

237 ANC services by Region: Mean Values, SD, F-values, and P-Values for 2020-2022
238 Hb Factorial Analysis–There is a difference in time, and the amount of Hb increased

239 significantly in 2022 compared to 2020 with the change of time (7.563, P=0.002). There is no

240 statistically significant difference in Hb values between regions in the same year (1.956,

241 P=0.124). There are statistical differences in the same region and different years in the

242 western (70.574, P=0.003) and Southern (8.007, P=0.010).

243 Syphilis screening factorial analysis–There is a difference in time, with a two-fold increase in

244 the number of pregnant women undergoing prenatal syphilis screening over time (118.371,

245 P<0.001) compared to 2020. There was no difference in the frequency of syphilis screening

246 between regions in the same year (2.356, P=0.074). Differences between years in the same

247 region indicate a growing awareness of the importance of antenatal syphilis screening.

248 IPT1st dose factorial analysis–there were differences in time (5.566, P=0.008) but no

249 differences in regions (1.772, P=0.158) for intermittent prophylaxis for malaria. There were

250 no statistically significant differences in regions in the same region and different years.

251 IPT2nd Factorial Analysis–There were differences in time (10.022, P<0.0018) and no

252 differences in the region within the same year (1.613, P=0.194). The eastern (16.252,

253 P=0.004) shows statistically significant changes over time. The other regions such as western

254 (2.859, P=0.202), northern (0.343, P=0.718), southern (2.753, P=0.117), northwest (3.063,

255 P=0.121) do not show significant changes in their means across the years. Standard

256 deviations provide insight into how consistent the data is year to year, with some regions

257 showing higher variability (e.g., northwest in 2021).

258 IPT3rd Factorial Analysis–There were differences in time (11.192, P<0.001) and regional

259 differences (5.167, P=0.002). The northern (4.686, P=0.040) and eastern (14.264, P=0.005)

260 show statistically significant differences over time. This indicates that changes in these areas

261 over the three years are likely to be meaningful rather than due to random chance. The
262 western, southern, and northwest regions do not show significant changes over time, as their

263 P-values are greater than 0.05, indicating stability in their means. The total data (11.192,

264 P<0.001) show a consistent, significant overall change across all regions, which suggests that

265 while some regions remain stable, the overall trend is significant.

266 For iron folate-the factorial analysis results showed that both region (3.582, P=0.016) and

267 time (3.403, P=0.048) were influencing factors. When considering all regions together, there

268 is a statistically significant change in iron folate provision across the years. The Southern

269 region shows a significant year-over-year change (4.251, P=0.048) in iron folate provision;

270 across all regions, the overall trend is also significant, with a peak in 2021. Regional

271 differences in iron folate provision are not significant in 2020 and 2021, but they become

272 significant in 2022, suggesting an increase in regional disparities by that year. The western

273 (1.605, P=0.336), northern (3.357, P=0.081), eastern (1.279, P=0.345), and northwest (1.232,

274 P=0.356) regions do not show statistically significant differences. (Table 3) demonstrated the

275 ANC services by Region: Mean values, SD, F-values, and P-values for 2020-2022.

276 Comparison of the changes in different birth attendants in Sierra Leone from 2020 to

277 2022

278 Delivery by CHO/CHA/SECHN/MCH Aides Normal delivery factorial analysis–no

279 difference in time (0.440, P=0.648), no difference in region (1.501, P=0.225). The results

280 suggest that the measured parameter remains stable across the years for all regions except the

281 northwest (4.716, P=0.059), where some indication of variability could be explored further.

282 Overall, no substantial evidence suggests significant changes in the means from 2020 to 2022

283 across the different regions. Further investigation into the northwest region may be warranted

284 to understand the observed trends.

285 Deliveries by Doctor/Midwife/SACHO/SECHN Midwife Factorial analysis–there were

286 differences in time (37.789, P<0.001) and no differences in regions (1.677, P=0.179). The

287 southern (39.907, P<0.001) and northwest (18.058, P=0.003) regions demonstrate significant

288 variability in measurements across the years, indicating possible shifts in underlying factors

289 or conditions affecting those regions. The northern (5.203, P=0.032) and eastern (5.221,

290 P=0.049) regions also show statistically significant changes, albeit to a lesser extent than the

291 Southern region. The Western region (2.287, P=0.249) does not show significant differences

292 in its measurements over time.

293 TBAs Factorial analysis–there were differences in time (19.033, P<0.001) and no differences

294 in regions (1.309, P=0.287). The southern region shows significant changes (63.414,

295 P<0.001) in the use of traditional birth attendants over the years, indicating a possible shift in

296 practices or an increased reliance on these attendants. The northern region also exhibits a

297 statistically significant change (7.291, P=0.013), suggesting evolving circumstances or

298 practices regarding traditional birth attendance. The western (8.254, P=0.060), eastern

299 (3.232, P=0.112), and northwest (1.539, P=0.289) regions do not show significant

300 differences in means over the years, indicating that the rates of deliveries by traditional birth

301 attendants in these areas have remained relatively stable. The overall trend across all regions

302 suggests substantial changes in deliveries done by traditional birth attendants, mainly driven

303 by the southern and northern regions. This highlights the need for further investigation into

304 the factors contributing to these shifts (Table 4).

305

306 The association between the ANC and maternal mortality rates

307 Multivariate regression

308 The total MMR (3895.5) was used as the dependent variable, and the independent variables

309 included 1 variable of ANC and 4 variables of ANC services (ANC1_Hb, ANC_IPT1st,

310 ANC_IPT2nd, and ANC_deworming) for multivariate linear regression.

311 The results showed that all the variables of prenatal examination had a significant effect on

312 MMR, and the regression coefficients were all negative. In the interim period, the 1 st of IPT

313 was most effective in reducing MMR (-9.335, P<0.001). Variables that yield these results are

314 the antenatal client's first visit hemoglobin done (-3.363, P=0.001), the antenatal client 8th

315 visit (-2.753, P=0.049), and the antenatal client's IPT 1st dose (-9.335, P<0.001), the antenatal

316 client's IPT 2nd dose (-5.337, P=0.027), and the antenatal client given deworming dose (-

317 2.513, P=0.001). The results are shown in (Table 5).

318 Bivariate correlation

319 The independent variables included 4 variables of ANC and 12 ANC services, 4 variables of

320 the mode of delivery, and 3 variables of personnel who do the delivery for bivariate

321 correlation. Statistically significant negative correlations were observed for:

322 ANC1_Hb (First ANC–Haemoglobin testing; -0.619, P=0.041); ANC1_<12W (First ANC

323 before 12 Weeks; -0.720, P=0.018), ANC8th visits (-0.601, P=0.048); ANC_LLITN given

324 (Long-lasting Insecticide-treated Nets; -0.204, P=0.026); ANC_iron folate (Iron

325 Supplementation; -0.515, P= 0.031); malaria in 2nd or 3rd trimester treated (Second or third

326 Maternal treated; -0.345, P=0.016). Highly significant positive correlation: Delivery by

327 Traditional Birth Attendant (TBA; 0.245**, P<0.001). The results are presented in (Table 6).

328 Discussion

329 The distribution of maternal mortality rates in Sierra Leone

330 The findings illustrate a significant decrease in MMRs in Sierra Leone from 2020 to 2022.

331 The MMRs reported in 2020 were high at 2759.5 maternal deaths per 100,000 live births;

332 however, this figure improved to 848.2 and later decreased further to 288.2 by the end of the

333 years 2021 and 2022, respectively. Such trends are similar to other study findings, such as the

334 results from sub-Saharan Africa (SSA) countries that indicated reductions in maternal

335 mortality through effective healthcare service delivery interventions . Similarly, significant

336 reductions in MMR have been reported in Rwanda concerning policies that increased ANC

337 access, skilled birth attendance, and facility-based delivery, with a reported drop rate from

338 1070 per 100,000 live births in 2000 to 290 in 2015 . According to TZ Zelelie, DS Gebreyes, AT

339 Tilahun, HA Craddock and NZ Gishen [22], a study in Ethiopia revealed that the MMR fell from

340 676 per 100,000 live births in 2011 to 412 in 2016. Here, access to ANC and region-specific

341 interventions, especially in rural areas, closely resembled Sierra Leone's efforts to reduce

342 maternal mortality rates in high-risk areas.

343 While these positive trends align with global efforts toward reducing maternal mortality, the

344 study also points to uneven regional disparities. The regions that had consistently higher

345 MMRs at first were the northern, western, and northwest, at 4424.5, 3888.1, and 3031.7 in

346 the year 2020, but they fell considerably down to 221.9, 279.2, and as low as 176.9 in the

347 year 2022, respectively! The results are consistent with other previous studies. For example,

348 in a study by M-L Bats, B Rucheton, T Fleur, A Orieux, C Chemin, S Rubin, B Colombies, A Desclaux,

349 C Rivoisy and E Mériglier [23] , substantial regional variation was observed in India. MMR was

350 higher in central and northern states at 700 per 100,000 live births, while southern states had

351 less than 100 per 100,000 live births. The factors contributing to this were literacy rate,

352 limited healthcare accessibility, and ANC frequency. A study by EN Chege [24] in Sierra

353 Leone cited regional differences where the western region had the highest number of

354 maternal deaths (n=227, 35%), followed by the northern region with 162 (25%) pregnant
355 women who died during pregnancy or childbirths. In contrast, the eastern region had the

356 fewest maternal deaths (n=98, 15.1%).

357 On the other hand, based on the causes or aetiologies of MMR, the majority of the MMR

358 occurred due to malaria, hemorrhage, hypertension, infections, and other complications

359 developed during pregnancy. This is consistent with the global and regional (SSA) causes of

360 maternal deaths. The causes contributing to maternal deaths region-wise, as identified by the

361 UN's Maternal Mortality Estimation Inter-Agency Group, inform that postpartum

362 hemorrhage is the main reason, as it causes approximately 34% of the overall maternal deaths

363 in Sub-Saharan Africa (SSA), whereas, in Southern Asia, the figure is closer to 30%. Pre-

364 eclampsia and eclampsia are the two health conditions that are complications of high blood

365 pressure (hypertension), resulting in around 10-12% of the deaths in both these regions.

366 Sepsis (infection death) is responsible for approximately 10-15% of maternal deaths all over

367 the world, but it has a higher incidence in areas with deficient healthcare infrastructure . In

368 women, malaria serves as background illnesses that are exacerbated by pregnancy, resulting

369 in approximately 20% more maternal deaths . The correlation between MMR values for all

370 causes was significant and positive.

371 The geographical disparities in ANC utilization and maternal mortality rates

372 The results revealed differences in the number of institutions for pregnant women giving

373 birth in different regions, as 50.0% of pregnant women in the western region have childbirth

374 at the clinics, while the other regions are mainly in MCHPs, CHPs, and CHCs. Our study

375 finds a significant relationship (P<0.001) between the regions and institutions of pregnant

376 women giving birth. Compared with other countries in SSA, the results of the current study

377 agree with those of other related studies. According to a Nigerian study by YM Adamu, IA

378 Abdulkarim, FiI Sheshe and OI Babatimehin [28] , there are notable regional differences in the
379 kinds of medical facilities that expectant mothers use to give birth. Most deliveries in rural

380 and underdeveloped regions were at CHCs, while many women delivered in clinics or

381 hospitals from urban and affluent areas. As observed in the western and other regions in

382 Sierra Leone, a statistically significant association (P<0.001) between region and facility type

383 was also reported, indicating that socio-economic and geographic factors affect facility

384 choice. Research conducted in India revealed a substantial correlation (P<0.001) between the

385 region and the type of facility pregnant women utilized to give birth. About 60% of women

386 gave birth in private clinics in wealthier states and urban regions . A Tanzania study also

387 revealed that the majority of births in rural areas happened in smaller health centers and

388 maternity and child health posts (MCHPs), while 55% of births in urban areas were in clinics

389 or hospitals. Their study shows a P-value <0.001, which supports the current findings about

390 regional differences in the utilization of institutions or places of birth and a strong correlation

391 between geographic location and the kind of medical facility utilized for delivery .

392 The factorial analysis of this study indicated that regional differences were less pronounced

393 in each given year despite significant improvement over time on several ANC indicators,

394 such as early ANC visits, Hb testing, syphilis screening, and IPT doses. By 2022, however,

395 regional variation in iron folate and malaria treatment had become more pronounced,

396 showing where focused regional health strategies and policies were necessary. The results of

397 the antenatal care increase in syphilis screening are similar to a study done in Zambia and

398 Ghana that also identified similar improvements in prenatal syphilis screening due to more

399 public awareness and policies . Current evidence suggests that re-implementing periodic

400 screening during ANC improves maternal health outcomes and reduces congenital syphilis.

401 Studies conducted in Malawi and Uganda reported that early ANC visits are crucial for

402 lowering maternal and newborn risks . Similar to this finding of increasing early ANC

403 attendance with time but with no discernible geographical differences, a Ugandan study, for
404 instance, revealed that early ANC commencement was essential for recognizing health risks

405 and implementing preventative measures. The significant increase over time in Hb testing

406 and iron folate distribution by geographical region of Sierra Leone is consistent with

407 improved maternal outcomes often associated with a reduction in maternal anemia and

408 improved birth outcomes, according to studies from Ethiopia and India . Studies conducted in

409 Tanzania and Nigeria demonstrate the beneficial effects of IPT on mother and newborn

410 health outcomes , similar to the noted increases in IPT dosages with time in Sierra Leone.

411 The association between increased IPT coverage and fewer malaria-related complications

412 supports these findings of better IPT3 administration in some regions over time.

413 With regards to the birth attendants according to this study, the findings revealed consistent

414 delivery rates over time by Maternal and Child Health (MCH) Aides, Community Health

415 Officers (CHO), and Community Health Assistants (CHA) across all regions. This pattern

416 reflects the critical but sometimes limited function of CHOs and aides in low-resource

417 settings, consistent with work in Ghana and Kenya reporting similar stable rates across rural

418 and urban domains . Because most CHOs in Ghana augment primary healthcare delivery in

419 physician-shortage areas, the overall rates of deliveries by CHOs remained relatively high

420 over time. The results revealed significant variation in births performed by doctors and

421 midwives over time. This indicates that access to trained birth attendants is uneven across

422 Sierra Leone. The Southern and Northwest have become more reliant on doctors and

423 midwives for childbirth over time. This result is consistent with studies from Nigeria and

424 Uganda , where midwife and doctor deliveries showed comparable variation trends with time

425 and location, significantly impacted by urban-rural disparities. Similar to the disparities

426 observed in Sierra Leone, some parts of Nigeria had more reliable access to doctors and

427 midwives, especially in metropolitan areas. Among other results, our study from Sierra Leone

428 suggests that TBA use is predominately regional (northern and southern), especially in rural
429 areas of these two regions, which have become increasingly dependent on TBAs sectional

430 over the years. In Tanzania and India, this reliance on TBAs is reported from rural

431 populations that still primarily turn to TBAs due to practical barriers, cultural preferences for

432 TBAs, and trust in local practitioners. Similar to the northern and Southern regions of Sierra

433 Leone, where TBAs continue to play a role, TBA engagement in deliveries is still substantial

434 in some parts of Tanzania despite efforts to promote skilled attendance . Similarly, research

435 indicates that rural areas in India are more dependent on TBAs and that this dependence can

436 differ significantly between states , much like the regional variations in Sierra Leone.

437 The association between the ANC and maternal mortality rates

438 The multivariate regression analysis revealed that all the variables of prenatal examination

439 measured had a significant effect on MMR, and the regression coefficients were all negative,

440 also showing the interim period of the antenatal client given Intermittent Preventive

441 Treatment (IPT) 1st dose for malaria was most effective in reducing MMR. Research

442 conducted in Uganda found that enhancing maternal outcomes required comprehensive ANC

443 care, which included preventive malaria medications . Consistently, the IPT for malaria (first

444 and second dosages) reported significant reductions in maternal mortality. Moreover, routine

445 testing for hemoglobin (Hb) and deworming at prenatal visits were found to be crucial in

446 reducing the risk related to anemia, thus promoting a reduction in maternal mortality .

447 Maternal mortality and the frequency of prenatal visits were also found to be negatively

448 correlated in an Ethiopian study . Similar to this study, there was a substantial correlation

449 between significantly reduced MMRs and multiple ANC visits, especially four or more. This

450 study emphasized that regular and quality antenatal care (ANC) services, including

451 hemoglobin (Hb) tests, deworming, and malaria control, were crucial in reducing maternal

452 risks during pregnancy. A Nigerian study demonstrated a reduction in MMR, attributed to

453 multiple ANC visits and targeted prevention against malaria and event management of
454 anemia . In the study, pregnant women who received malaria prophylaxis medicine were less

455 likely to experience maternal morbidity and death. The role of each antenatal care service

456 concerning MMR is independent of the others; some interventions are more successful in this

457 respect.

458 In this study, bivariate correlations revealed six statistically significant negative associations

459 for most of the variables observed, such as prenatal client's first visit, hemoglobin screening,

460 prenatal client for below 12 weeks, 8 th prenatal client visit, antenatal care recipients receiving

461 LLITN (long-term insecticide-treated netting), antenatal client receives iron folate (iron

462 supplementation), and treating malaria in the 2nd or 3rd trimesters. Traditional Birth Attendants

463 (TBAs) observed a significant positive correlation in delivery. Studies conducted in Nepal

464 have also revealed that maternal death rates were considerably lower in regions where TBAs

465 were trained and included in a system for referring patients to medical facilities. According to

466 S Maru, I Nirola, A Thapa, P Thapa, L Kunwar, W-J Wu, S Halliday, D Citrin, R Schwarz and I Basnett

467 [44], educated TBAs decreased maternal mortality by identifying signs of complications early

468 and sending pregnant women to local medical facilities for expert care. This supports the

469 findings of this study and suggests that TBAs could support improvements in maternal health

470 indicators if adequately trained and incorporated into the health care system.

471 Several other studies have also shown the significance of ANC service utilization. Research

472 conducted in Uganda in 2021 reported that fewer malaria-related problems were connected to

473 lower maternal death rates when LLITNs and antimalarial therapy were distributed during the

474 second or third trimester . This supports our conclusion that a negative correlation exists

475 between maternal mortality in Sierra Leone and the use of insecticide-treated nets and

476 malaria prevention. Study conducted in Ethiopia reported that higher ANC visits, especially

477 four or more have been linked to lower maternal death rates. For example, an Ethiopian study

478 showed that higher interactions with ANC negatively affect maternal mortality. It was noted
479 that ANC registration was as early as possible, and interventions such as iron

480 supplementation and anemia screening helped boost the outcomes .

481 Limitations of the Study

482 However, this study has limitations. First, the fact that the study is cross-sectional implies

483 that the data so far collected was collected within a specific time frame, that is, 2020-2022.

484 This restricts the capability to determine causal relationships and does not consider MMR and

485 ANC use trends over the years. Future studies could employ a longitudinal design to monitor

486 variables such as maternal mortality and ANC over a specified period. In addition, it may be

487 helpful to combine qualitative approaches, such as interviews or focus groups, with

488 quantitative data to understand better the dynamics associated with ANC and maternal

489 outcomes over a period. Secondly, the data is primarily secondary and obtained from DHIS2.

490 This makes the study obtained only part of the story behind the ANC uptake levels and MMR

491 determinants. Additional exploratory data, particularly survey data and interviews conducted

492 with practicing health providers and pregnant women, would underscore the barriers and

493 enablers of ANC services utilization. Local-level engagement will also offer more focused

494 perspectives on maternal health issues, and such engagement will help analyze the collected

495 data.

496

497

498 Conclusions

499 Regional disparities in ANC utilization and MMR are still significant problems in Sierra

500 Leone. Strengthening ANC services, particularly in the northern and western regions, is

501 crucial for reducing maternal deaths. Policymakers should prioritize equal resource allocation
502 and bridge the infrastructural gaps to ensure universal access to quality maternal healthcare in

503 order to achieve SDG 3.1.

504 List of abbreviations

505 ANC Antenatal care

506 CHA Community Health Assistant

507 CHO Community Health Officer

508 CS Cesarean Section

509 DHIS2 District Health Information System 2

510 DPPI Directorate of Policy, Planning, and Information

511 IPT Intermittent Preventive Treatment

512 LLITN Long-Lasting Insecticide-Treated Nets

513 MCH AidesMaternal Child Health Aides

514 MDSR Maternal Death Surveillance and Response

515 MMR Maternal Mortality Rate

516 MoHS Ministry of Health and Sanitation

517 SACHO Senior Assistant Community Health Officer

518 SECHN State Enrolled Community Health Nurse

519 SSA sub-Saharan African

520 SDGs Sustainable Development Goals

521 TBAs Traditional Birth Attendants

522 WHO World Health Organization

523 Ethical approval and consent to participate

524 The Sierra Leone Ethics Review Board (MoHS, Freetown) approved the study proposal.

525 Informed consent is not required because the study employed a secondary dataset from the

526 DHIS2, although confidentiality and ethical behavior were carefully considered at every

527 stage.

528 Consent for publication

529 Not applicable

530 Data availability

531 The datasets generated and/or analyzed during the current study are not publicly available

532 due to privacy, policy, and governance but are available from the corresponding author upon

533 reasonable request.

534 Competing Interests

535 The authors declare that they have no competing interests

536 Funding

537 No funding was obtained from external sources

538

539 Authors’ contributions

540 M.S. was responsible for the background, study design, and conceptualization, drafted the

541 original manuscript, developed the methodology, and contributed to interpretation and

542 discussion. G.Y. supervised the study, reviewed the original draft, contributed to the

543 methodology and data analysis, approved the final version of the manuscript, and is the

544 corresponding author.

545 Acknowledges

546 The authors would like to acknowledge the Ministry of Health and Sanitation, Sierra Leone,

547 for data access from DHIS2 and Southern Medical University (China) for their institutional

548 support.

549 Authors’ information

550 School of Public Health, Southern Medical University, 1023-1063 Shatai Nan Road,

551 Guangzhou 510515, China. Email address: [email protected]

552 Department of Biostatistics, School of Public Health, Southern Medical University, 1023-

553 1063 Shatai Nan Road, Guangzhou 510515, China. Email address: [email protected]

554 Morlai Sesay, Professor Guan Ying

563 References