DOI: https://doi.org/10.21203/rs.3.rs-2022351/v1
The evaluation of utilization and the factors influencing optimal ANC visits is critical to improving maternal and neonatal health outcomes. The goal of antenatal care is to reduce maternal and perinatal mortality and morbidity. The study's goal is to identify the factors that influence optimal ANC visits among reproductive-age women in low-income countries.
The study included a total weighted sample of 329,721 women who gave birth during the study period. For model fitness and comparison, the intra-class correlation coefficient, median odds ratio, proportional change in variance, AIC, BIC, and deviance were used. To identify the determinants of optimal ANC visits in LMICs, a multilevel multivariable logistic regression model was fitted. To declare significant determinants of optimal ANC visits, the adjusted odds ratio and its 95% confidence interval were used.
The overall prevalence of optimal ANC visits was 60.1%, and this ranged from 16.8% in Afghanistan to 97.0% in the Dominican Republic. In the Multilevel multivariable logistic regression model; age 20 to 34 (AOR = 1.267; 95%CI: 1.233–1.303)), age above 34 (AOR = 1.342; 95%CI: 1.302–1.384)), primary educated women (AOR = 1.529; 95%CI: 1.496–1.563), secondary educated women (AOR = 2.626; 95%CI: 2.561–2.694), higher educated women (AOR = 4.563; 95%CI: 4.341–4.796), middle wealth index (AOR = 1.033; 95%CI: 1.015–1.052), rich wealth index (AOR = 1.340; 95%CI: 1.284–1.399), having media exposure (AOR = 1.273; 95%CI: 1.251–1.293)), employed women (AOR = 1.252; 95%CI: 1.212–1.293), and being Central America resident (AOR = 5.967; 95%CI: 5.655–6.297) were significantly associated with optimal ANC visits.
Maternal age, maternal education level, family size, number of children, sex of household head, wealth index, marital status, husband/partner education level, husband/partner occupation, maternal occupation, media exposure, place of delivery, and region were all significant predictors of optimal ANC visits in low- and middle-income countries. This discovery assists health care providers and policymakers in implementing appropriate policies and programs to ensure optimal ANC coverage. It is critical to develop strategies to improve antenatal care access and availability.
ANC is a pregnancy-related service provided to pregnant women by health professionals; it is one of the major interventions aimed at preventing neonatal deaths and maintaining women's health during pregnancy [1–3]. The goal of antenatal care is to reduce maternal and perinatal mortality and morbidity [3]. The first intervention on the list is antenatal care, which is provided between conception and the start of labor with the goal of improving pregnancy outcomes and the health of the mother and child [4]. ANC serves as a platform for essential health care functions such as disease prevention, screening and diagnosis, and health promotion [5]. This case entails a series of assessments and appropriate treatments covering three components: monitoring the woman's and fetus' health status; provision of medical and psychosocial interventions; and support and health promotion [4].
Maternal healthcare remains a critical public health issue worldwide. Healthcare services are absolutely necessary for the well-being and survival of mothers and children during pregnancy, childbirth, and post-delivery periods [6]. For decades, the international community has been concerned with improving women's health and reducing maternal and child mortality [7, 8]. The high priority given to maternal and child health care in the Millennium Development Goals (MDGs) and, more recently, the Sustainable Development Goals (SDGs) [9] demonstrates this special interest.
Maternal mortality is one of the leading causes of death among women of reproductive age in developing countries [10]. According to the World Health Organization (WHO), the global maternal mortality rate in 2017 was 211 per 100,000 live births [11]. The United Nations (UN) set a target of less than 70 maternal deaths per 100,000 live births by 2030, with no single country exceeding 140 maternal deaths per 100,000 [12], but meeting this target will be difficult. According to the United Nations Millennium Development Goals [13], every year, at least 500,000 women and girls die unnecessarily as a result of complications during pregnancy, childbirth, or the six weeks following delivery. Almost all of these deaths (99%) occur in developing countries [14].
Globally, significant reductions in maternal and child morbidity and mortality have not been achieved, especially in low- and middle-income countries (LMICs) [15]. Despite the fact that ANC services have been linked to lower maternal morbidity and mortality [16, 17], five out of every ten women in low- and middle-income countries (LMICs) do not receive adequate ANC services [18, 19]. Sub-Saharan Africa was responsible for an estimated two-thirds of all maternal deaths in 2015, with a maternal mortality ratio (MMR) twice that of HIC, and recent data suggest that geographic disparities in maternal health are widening [20]. Common causes of maternal mortality in LMICs are classified as direct and indirect [21, 22]. Hemorrhages, hypertensive disorders, eclampsia, sepsis, abortion complications, and obstructed labor are direct causes [23, 24], while severe anemia, HIV/AIDS complications, and severe malaria are indirect causes [25].
Furthermore, the World Health Organization (WHO) has advocated for a variety of approaches to preventing pregnancy-related issues. The WHO's FANC (focused antenatal care) model, for example, emphasizes that a pregnant woman should receive comprehensive ANC visits as well as screening and treatment for anemia, malaria, HIV/AIDS, and tetanus [26]. Furthermore, the WHO recently recommended new ANC guidelines, as well as numerous other changes to the FANC model. A pregnant woman should have at least eight ANC visits, beginning with the first visit at 12 weeks' gestational age and continuing with visits to a skilled health care provider at 20, 26, 30, 34, 36, 38, and 40 weeks' gestational age [27].
The evaluation of utilization and the factors influencing optimal ANC visits is critical to improving maternal and neonatal health outcomes. Many factors associated with the use of optimal ANC visits have been reported in various studies around the world, including the mother's age, region, having a living child, women's educational level, place of residence, household wealth status, family size, sex of the household head, marital status, husband or partner, mother's occupation, current working status of the mother, place of delivery, the husband's education, and access to mass media [28–30]. Similar factors have been identified as institutional delivery determinants [31–33].
[34] were the only studies conducted in LMICs. While the authors looked at optimal ANC visits at the LMIC level, they aggregated the countries into sub-regions, failing to account for variations in prevalence and determinants of optimal ANC visits based on the regions of the countries included in their analysis. As a result, their research did not take into account regional policies that may influence the prevalence and determinants of optimal ANC visits. With ANC visits being a major determinant of maternal mortality, it is critical to use appropriate statistical techniques of analysis to distinguish the source of variation, the fundamental features of variation in the occurrence of optimal ANC visits, and the factors of optimal ANC visits. Because unsuitable results lead to confusion between deduction and involvement. However, there is no suggestion that studies done at the LMIC level can be used to determine the main factors of optimal ANC visit prevalence by taking the regional effect into account using the mixed effect model method. As a result, this study will examine the determinants of optimal ANC visits among reproductive-age women in LMICs using data from a recent demographic health survey and a multilevel binary logistic regression model. As a result, this study will provide policymakers and the general public with data that will aid in reducing the high prevalence of maternal and neonatal mortality in LMICs, which contributes to the global burden.
The study analyzed aggregate data from the most recent Demographic and Health Surveys (DHS) conducted in 55 low-income countries (LMICs) between 2010 and 2021. The DHS is a five-year national study that is being conducted in a number of LMICs across Africa, the Caribbean, Asia, and parts of South Europe and Latin America. The DHS employs consistent procedures in questionnaire design, sampling, data collection, data cleaning, coding, and analyses, enabling cross-country comparability [35]. This study included only women who had given birth in the five years preceding the survey. The information was obtained from the Measure DHS website https://www.dhsprogram.com/data/dataset_admin/login_main.cfm following approval via an online request stating the purpose of this study. This study used a weighted total sample of 329,721 women who had given birth in the five years preceding the survey.
The outcome variable is optimal ANC visits, which are dichotomized as "yes" (if a woman has at least four ANC visits) or "no" (if she does not have at least four ANC visits) (if a woman has less than four ANC visits).
According to the literature, the independent variables in this study were of two types. Individual and community level variables Region and residence are examples of community-level variables. Women's education level, husband's education level, wealth index of the household, sex of household head, marital status of women, women's age group, women's current working status, women's occupation, husband's occupation, media exposure, place of delivery, family size, and number of under-five children are the individual-level variables. Table 2 shows the independent variables listed overhead, and the parameters in Table 2 were collected through face-to-face interviews.
Region |
Country |
DHS year |
Weighted number of reproductive age women |
Antenatal care visits |
|||
---|---|---|---|---|---|---|---|
Less than 4 |
4 and above |
||||||
Frequency |
% |
Frequency |
% |
||||
Eastern Africa |
Burundi |
2016/2017 |
7573 |
3724 |
49.2 |
3849 |
50.8 |
Comoros |
2012 |
1637 |
675 |
41.2 |
962 |
58.8 |
|
Ethiopia |
2016 |
6514 |
4180 |
64.2 |
2334 |
35.8 |
|
Kenya |
2014 |
6512 |
2912 |
44.7 |
3600 |
55.3 |
|
Malawi |
2015 − 1016 |
10783 |
5241 |
48.6 |
5542 |
51.4 |
|
Rwanda |
2019–2020 |
3777 |
1164 |
15.8 |
6213 |
84.2 |
|
Tanzania |
2015–2016 |
4050 |
1354 |
33.4 |
2696 |
66.6 |
|
Uganda |
2016 |
4696 |
2081 |
44.3 |
2615 |
55.7 |
|
Middle Africa |
Angola |
2015 |
5686 |
2384 |
41.9 |
3302 |
58.1 |
Cameroon |
2018 |
4803 |
1752 |
36.5 |
3051 |
63.5 |
|
Chad |
2014–2018 |
10139 |
7265 |
71.7 |
2874 |
28.3 |
|
Congo |
2011–2012 |
5358 |
1358 |
25.3 |
4000 |
74.7 |
|
Congo DR. |
2013–2014 |
10298 |
5648 |
54.8 |
4650 |
45.2 |
|
Gabon |
2012 |
2837 |
878 |
30.9 |
1959 |
69.1 |
|
Northern Africa |
Egypt |
2014 |
15727 |
2696 |
17.1 |
13031 |
82.9 |
Southern Africa |
Lesotho |
2014 |
2238 |
544 |
24.3 |
1694 |
75.7 |
South Africa |
2016 |
5278 |
542 |
10.3 |
4736 |
89.7 |
|
Zimbabwe |
2015 |
8149 |
3192 |
39.2 |
4957 |
60.8 |
|
Namibia |
2011 |
1621 |
317 |
19.6 |
1304 |
80.4 |
|
Zambia |
2013 |
5231 |
1767 |
33.8 |
3464 |
66.2 |
|
Western Africa |
Benin |
2017 |
7738 |
3719 |
48.1 |
4019 |
51.9 |
Burkina Faso |
2010 |
10117 |
6599 |
65.2 |
3518 |
34.8 |
|
Cote d’Ivoire |
2011–2012 |
4505 |
2607 |
57.9 |
1898 |
42.1 |
|
Gambia |
2019 |
4701 |
865 |
18.4 |
3836 |
81.6 |
|
Guinea |
2018 |
4857 |
3161 |
65.1 |
1696 |
34.9 |
|
Liberia |
2019 |
2696 |
362 |
13.4 |
2334 |
86.6 |
|
Mali |
2018 |
5662 |
3230 |
57.0 |
2432 |
43.0 |
|
Mauritania |
2019–2021 |
4872 |
2823 |
57.9 |
2049 |
42.1 |
Region |
Country |
DHS year |
Weighted number of reproductive age women |
Antenatal care visits |
|||
Less than 4 |
4 and above |
||||||
Frequency |
% |
Frequency |
% |
||||
Western Africa |
Niger |
2012 |
3921 |
1772 |
29.9 |
2749 |
70.1 |
Nigeria |
2018 |
7468 |
4998 |
66.9 |
2470 |
33.1 |
|
Senegal |
2019 |
4833 |
2478 |
50.2 |
2455 |
49.8 |
|
Sierra Leone |
2019 |
3805 |
1758 |
46.2 |
2047 |
53.8 |
|
Togo |
2013-2014 |
4754 |
1112 |
23.4 |
3642 |
76.6 |
|
Ghana |
2014 |
3847 |
505 |
13.1 |
3342 |
86.9 |
|
Central Asia |
Kyrgyz Rep |
2012 |
3055 |
436 |
14.3 |
2619 |
85.7 |
Tajikistan |
2017 |
992 |
182 |
18.3 |
810 |
81.7 |
|
South – Eastern Asia |
Cambodia |
2014 |
5846 |
1501 |
25.7 |
4345 |
74.3 |
Philippines |
2017 |
5257 |
2286 |
43.5 |
2971 |
65.5 |
|
Southern Asia |
Afghanistan |
2015 |
19015 |
15823 |
83.2 |
3192 |
16.8 |
Bangladesh |
2017-2018 |
4918 |
2542 |
51.7 |
2376 |
48.3 |
|
Indonesia |
2017 |
14721 |
1621 |
11.0 |
13100 |
89.0 |
|
Maldives |
2016-2017 |
2051 |
69 |
3.4 |
1982 |
96.6 |
|
Nepal |
2016 |
1621 |
317 |
19.6 |
1304 |
80.4 |
|
Pakistan |
2017-2018 |
19787 |
8475 |
42.8 |
11312 |
57.2 |
|
Western Asia |
Armenia |
2015-2016 |
1251 |
37 |
3.0 |
1214 |
97.0 |
Jordan |
2017-2018 |
7002 |
574 |
8.2 |
6428 |
91.8 |
|
Central America |
Guatemala |
2015-2015 |
9007 |
1177 |
13.1 |
7830 |
86.9 |
Honduras |
2011-2012 |
8142 |
928 |
11.4 |
7214 |
88.6 |
|
Southern Europe |
Albania |
2017-2018 |
2280 |
573 |
25.1 |
1707 |
74.9 |
Caribbean |
Dominican republic |
2013 |
2655 |
79 |
3.0 |
2576 |
97.0 |
Haiti |
2016-2017 |
4205 |
1399 |
33.3 |
2806 |
66.7 |
|
Myanmar |
2015-2016 |
3723 |
1554 |
41.7 |
2169 |
58.3 |
|
Timor leste |
2016 |
5705 |
2909 |
51.0 |
2796 |
49.0 |
|
Oceania |
Papua NG |
2016-2018 |
8105 |
4111 |
50.7 |
3994 |
49.3 |
Overall |
329,721 |
131,656 |
39.9 |
198,065 |
60.1 |
Independent variables |
Characteristics |
Frequency (%) |
---|---|---|
Age of mother |
Less than 20 |
31618 (9.6) |
20to34 |
216402 (65.6) |
|
Greater than34 |
81701 (24.8) |
|
Types of place of residence |
Urban |
109923 (33.3) |
Rural |
219798 (66.7) |
|
mother education level |
No education |
113982 (34.6) |
Primary |
95088 (28.8) |
|
Secondary |
95103 (28.8) |
|
Higher |
25548 (7.7) |
|
Family size |
Less than 4 |
39343 (11.9) |
4 to 8 |
216296 (65.6) |
|
More than 8 |
74082 (22.5) |
|
Number of children |
Only one |
153100 (46.4) |
2 to 3 |
153099 (46.4) |
|
More than 3 |
23522 (7.1) |
|
Sex of household head |
Male |
282330 (85.6) |
Female |
47397 (14.4) |
|
Wealth index |
Poor |
101811 (30.88) |
Middle |
128126 (38.86) |
|
Rich |
99784 (30.26) |
|
Current marital status |
Married |
271605 (82.4) |
Living with partners |
49178 (14.9) |
|
Widowed |
1840 (0.6) |
|
Divorced |
1524 (0.5) |
|
Separated |
5574 (1.7) |
|
Husband/partner’s education level |
No education |
94973 (28.8) |
Primary |
89116 (27.0) |
|
Secondary |
112032 (34.0) |
|
Higher |
33600 (10.2) |
|
Husband/partner’s occupation |
Have no any work |
14288 (4.3) |
Have any work |
315433 (95.7) |
|
Currently working status |
No |
166440 (50.5) |
Yes |
163281 (49.8) |
|
Mother occupation |
Have no any work |
139690 (42.4) |
Have any work |
190031 (57.6) |
|
Media exposure |
No |
182568 (55.4) |
Yes |
147153 (44.6) |
|
Place of delivery |
Home delivery |
106,800 (32.4) |
Health facility delivery |
222,921 (67.6) |
As a result, permission to access the data was obtained via an online request from the DHS program at https://www.dhsprogram.com/data/dataset_admin/login_main.cfm, and the data used were publicly available without any personal identifier.
The data was cleaned using STATA version 17 software. Sample weighting was used for further analysis. Descriptive and inferential analyses were carried out. The descriptive analysis provided data on background characteristics, country, and antenatal care prevalence.
Because the outcome variable was binary, multilevel binary logistic regression analysis was used. A primary sampling unit, strata, and women's individual weight were used as part of a complex survey design (V005). Individual and community-level variables were checked independently in the multi-level binary logistic regression model, and variables that were statistically significant at p-value 0.20 in the multi-level mixed-effects logistic regression analysis were considered for the final individual and community-level model adjustments. In the multivariable, multilevel analysis, variables with p-values less than or equal to 0.05 were declared significant antenatal care determinants.
Four models were installed. The first was a null model with no exposure variables, which was used to assess random effects at the community level and check variation in communities. The multivariable model adjustment for individual-level variables was Model I, and Model II was adjusted for community-level factors. Potential candidate variables from both individual and community-level variables were fitted to the outcome variable in Model III.
The fixed effects (a measure of association) were used to estimate the association between antenatal care likelihood and explanatory variables at both the community and individual levels, and were expressed as odds ratios with a 95% confidence interval. Community-level variance with standard deviation, intra-cluster correlation coefficient (ICC), proportional change in community variance (PCV), and median odds ratio (MOR) were used as measures of variation (random-effects). The median odds ratio (MOR) is designed to convert area-level variance into the widely used odds ratio (OR) scale, which has a consistent and intuitive interpretation. The MOR is defined as the median value of the odds ratio between the area at the highest risk and the area at the lowest risk when randomly picking out two areas. The MOR can be conceptualized as the increased risk that an individual (in the median) would have if they moved to another area with a higher risk. It is computed by; [36]. Where Var is the area level variance and 0.6745 is the 75th percentile of the normal distribution's cumulative distribution function with a mean of 0 and a variance of 1. where the proportional change in variance is calculated as.
[36]
The model's assumptions were tested using the intra-class coefficient of correlation (ICC), which is used to measure the degree of similarity in the cluster's place of delivery prevalence, and the likelihood ratio (LR) test [36].
In 54 countries, 329,721 households were included. Overall, the prevalence of health-care delivery was 60.1%, ranging from 16.8% in Afghanistan to 97.0% in the Dominican Republic (Table 1). Pakistan had the most women (19,787 (0.06%), while Tajikistan had the fewest (992 (0.003%)). More than two-thirds, 219,798 (66.7%), of the women were from rural areas. More than one-third of the women, 113,982 (34.6%), were uneducated, and 222,921 (67.6%) were delivered in a health facility. More than half of women, 182,568 (55.4%), have access to the media, and 151,595 (46.0%) are from low-income households (Table 2).
The multilevel logistic regression model III had the lowest AIC, BIC, largest log-likelihood ratio, and lowest deviance because the models were nested. The ICC value for the empty model was 16.6% (p-value less than 0.001), indicating that cluster variability explained about 16.6% of the total variability of optimal antenatal care visits in LMICs, while individual variation explained the remaining 83.4% of the total variability. Furthermore, the MOR for an optimal ANC visit was 2.16 (95% CI: 1.52-2.81) in the null model, indicating that there is variation between communities (clustering) because the MOR is higher than the reference (MOR = 1). This revealed that there is significant heterogeneity in optimal ANC visits across communities. In the full model, the community variance (community variance = 0.124; P-value less than 0.001) remained significant but decreased (model adjusted for both individual and community-level factors). Even after accounting for some contextual risk factors, approximately 3.63% of the total variance in health facility delivery that can be attributed to contextual-level factors remained significant. The proportional change in variance (PCV) in this model was 81.07%, indicating that both community and individual level variables explained 81.07% of the community variance observed in the null model (Table 3).
Table 3. Multivariable multilevel logistic regression analysis of both individual and community-level factors associated with optimal ANC visits in LMICs.
Individual and community-level factors |
Models |
|||
Null model AOR (95%CI) |
Model I AOR (95%CI) |
Model II AOR (95%CI) |
Model III AOR (95%CI) |
|
Age of mother |
|
|
|
|
Less than 20 |
|
Ref. |
|
Ref. |
20 to 34 |
|
1.269 (1.235-1.305) |
|
1.267 (1.233 – 1.303) |
Greater than 34 |
|
1.341 (1.300-1.383) |
|
1.342 (1.302 – 1.384) |
Types of residence |
|
|
|
|
Urban |
|
|
Ref. |
Ref |
Rural |
|
|
0.417 (0.410-0.423) |
0.763 (0.749 – 1.384) |
Mother education level |
|
|
|
|
No education |
|
Ref. |
|
Ref. |
Primary |
|
1.520 (1.586-1.553) |
|
1.529 (1.496 – 1.563) |
Secondary |
|
2.611 (2.545-2.678) |
|
2.626 (2.561 – 2.694 |
Higher |
|
4.517 (4.297-4.749) |
|
4.563 (4.341 – 4.796 |
Family size |
|
|
|
|
Less than 4 |
|
Ref. |
|
Ref. |
4 to 8 |
|
0.995 (0.967-1.023) |
|
0.997 (0.969 – 1.025) |
More than 8 |
|
0.905 (0.874-0.936) |
|
0.907 (0.876 – 0.939 |
Number of children |
|
|
|
|
Only one |
|
Ref. |
|
Ref. |
2 to 3 |
|
0.789 (0.774-0.803) |
|
0.792 (0.778 – 0.807) |
4 and more |
|
0.758 (0.731-0.786) |
|
0.763 (0.736 – 0.791) |
Sex of household head |
|
|
|
|
Male |
|
Ref. |
|
Ref. |
Female |
|
1.110 (1.084-1.137) |
|
1.107 (1.081 – 1.133) |
Wealth index |
|
|
|
|
Poor |
|
Ref. |
|
Ref. |
Middle |
|
1.021 (1.011 – 1.031) |
|
1.033 (1.015 – 1.052) |
Rich |
|
1.311 (1.287 – 1.336) |
|
1.340 (1.284 – 1.399) |
Table 3. Multivariable multilevel logistic regression analysis of both individual and community-level factors associated with optimal ANC visits in LMICs (Continued).
Individual and community-level factors |
Models |
|||
Null model AOR (95%CI) |
Model I AOR (95%CI) |
Model II AOR (95%CI) |
Model III AOR (95%CI) |
|
Marital status |
|
|
|
|
Married |
|
1.471 (1.441 – 1.501) |
|
1.167 (1.090 – 1.250) |
Living with partner |
|
1.776 (1.675 – 1.883) |
|
1.239 (1.156 – 1.329) |
Widowed |
|
0.798 (0.728 – 0.875) |
|
0.898 (0.793 – 1.017) |
Divorced |
|
1.018 (0.919 – 1.128) |
|
1.043 (0.912 – 1.193) |
Separated |
|
|
|
Ref. |
Husband education level |
|
|
|
|
No education |
|
Ref. |
|
Ref. |
Primary |
|
1.238 (1.209-1.267) |
|
1.233 (1.205 – 1.263) |
Secondary |
|
1.583 (1.544-1.622) |
|
1.568 (1.530 – 1.607) |
Higher |
|
1.753 (1.685-1.823) |
|
1.742 (1.675 – 1.812) |
Husband occupation |
|
|
|
|
Have no any work |
|
Ref. |
|
Ref. |
Have any work |
|
1.112 (1.216-1.29) |
|
1.103 (1.060 – 1.147) |
Current working status |
|
|
|
|
No |
|
Ref. |
|
Ref. |
Yes |
|
1.006 (0.974-1.039) |
|
1.004 (0.973 – 1.037) |
Mother occupation |
|
|
|
|
Have no any work |
|
Ref. |
|
Ref. |
Have any work |
|
1.256 (1.216-1.297) |
|
1.252 (1.212 – 1.293) |
Media exposure |
|
|
|
|
No |
|
Ref. |
|
Ref. |
Yes |
|
1.266 (1.244-1.288) |
|
1.273 (1.251 – 1.293) |
Place of Delivery |
|
|
|
|
Home |
|
Ref. |
|
Ref. |
Health facility |
|
2.430 (2.397 – 2.474) |
|
2.437 (2.394 – 2.481) |
Table 3. Multivariable multilevel logistic regression analysis of both individual and community-level factors associated with optimal ANC visits in LMICs (Continued).
Individual and community-level factors |
Models |
|||
Null model AOR (95%CI) |
Model I AOR (95%CI) |
Model II AOR (95%CI) |
Model III AOR (95%CI) |
|
Region |
|
|
|
|
East Africa |
|
|
Ref. |
Ref. |
Middle Africa |
|
|
0.715 (0.696 – 0.735) |
1.025 (0.994 – 1.058) |
Northern Africa |
|
|
3.391 (3.240 – 3.550) |
2.930 (2.787 – 3.079) |
Southern Africa |
|
|
1.962 (1.895 – 2.030) |
2.030 (1.957 – 2.106) |
Western Africa |
|
|
0.778 (0.760 – 0.797) |
1.277 (1.244 – 1.312) |
Central Asia |
|
|
4.058 (3.715 – 4.433) |
1.581 (1.444 – 1.731) |
South-Eastern Asia |
|
|
1.499 (1.435 – 1.566) |
1.452 (1.384 – 1.523) |
Southern Asia |
|
|
0.812 (0.792 – 0.832) |
1.102 (1.072 – 1.133) |
Western Asia |
|
|
6.454 (5.927 – 7.027) |
3.672 (3.362 – 4.011) |
Central America |
|
|
5.243 (4.990 – 5.509) |
5.967 (5.655 – 6.297) |
Caribbean |
|
|
2.035 (1.845 – 2.244) |
1.525 (1.378 – 1.687) |
Oceania |
|
|
1.260 (1.214 – 1.307) |
1.394 (1.339 – 1.1452) |
Southern Europe |
|
|
0.616 (0.587 – 0.647) |
0.938 (0.888 – 0.990) |
Random effects model |
|
|
|
|
Community variance (σ2) |
0.655 |
0.128 |
0.281 |
0.124 |
ICC% |
16.60 |
3.75 |
7.87 |
3.63 |
PCV% |
Ref. |
80.46 |
57.10 |
81.07 |
MOR (95%CI) |
2.16 (1.52 – 2.81) |
1.41 (1.22 – 1.160) |
1.66 (1.35 – 1.97) |
1.40 (1.20 – 1.93) |
Model comparison |
|
|
|
|
Log-likelihood ratio |
-204228.2 |
-181824.6 |
-204169.2 |
181775.3 |
Deviance |
408456.4 |
363649.2 |
408338.4 |
363550.6 |
AIC |
408464.4 |
363703.2 |
408372.4 |
363630.6 |
BIC |
408507.2 |
363992.3 |
408554.4 |
363058.8 |
Model III, the best-fit model, had a lower AIC, BIC, deviance, and a higher log-likelihood ratio test. As a result, the fixed effects were interpreted using Model III, which was adjusted for both individual and community-level factors. As a result, in the multilevel multivariable analysis, maternal age, maternal education level, family size, number of children, sex of household head, wealth index, marital status, husband/partner education level, husband/partner occupation, maternal occupation, media exposure, place of delivery, and region were significant determinants of place of delivery in LMICs.
After controlling for other individual and community-level factors, the odds of an optimal ANC visit were 1.267 (AOR = 1.267; 95%CI: 1.233-1.303) and 1.342 (AOR = 1.342; 95%CI: 1.302-1.384) times higher among women aged 20 to 34 and greater than 34, respectively, compared to women under the age of 20. Our findings show that women's educational status is a significant positive determinant of optimal ANC visits. Women with primary, secondary, and higher education were 1.529 times (AOR = 1.529; 95%CI: 1.496-1.563), 2.626 times (AOR = 2.626; 95%CI: 2.561-2.694), and 4.563 times (AOR = 4.563; 95%CI: 4.341-4.796) more likely than women with no formal education. Not only maternal education level but also husband or partner education level were important determinants of optimal ANC visits. Husbands/partners with primary, secondary, and higher education were 1.233 times (AOR = 1.233; 95%CI: 1.205-1.263), 1.568 times (AOR = 1.568; 95%CI: 1.530-1.607), and 1.742 times (AOR = 95%CI: 1.675-1.812) more likely to have optimal ANC visits, respectively. The likelihood of optimal ANC visits was lower among reproductive-age women with more than 8 family members (AOR = 0.907; 95%CI: 0.876-0.939) when compared to less than 4 family members. The likelihood of optimal ANC visits among reproductive-age women having 2 to 3 and more than 3 children was lower by 0.792 (AOR = 0.792; 95%CI: 0.778–0.807) and 0.763 (AOR = 0.763; 95%CI: 0.736-0.791) as compared to women having only one child, respectively. The sex of the household head is an important determinant of optimal ANC visits. Households led by female heads were 1.107 times (AOR = 1.107; 95%CI: 1.081-1.133) more likely to have optimal ANC visits than households led by male heads. The odds of optimal ANC visits among women from households with middle and rich wealth status were 1.033 times (AOR = 1.033, 95%CI: 1.015–1.052) and 1.340 times (AOR = 1.340; 95%CI: 1.284–1.399), respectively, as compared to those from households with poor wealth status. The odds of optimal ANC visits among reproductive-age women who are married and living with a partner were 1.167 (AOR = 1.167; 95%CI: 1.090-1.250) and 1.239 (AOR = 1.239; 95%CI: 1.156-1.329) times higher as compared to separated women, respectively. The likelihood of optimal ANC visits among reproductive-age women whose husband or partner has an occupation was 1.103 (AOR = 1.103; 95% CI: 1.060-1.147) times higher as compared to reproductive-age women whose husband or partner has no occupation. The odds of optimal ANC visits for employed women were 1.252 (AOR = 1.252; 95% CI: 1.212-1.293) times higher as compared to unemployed women. The odds of optimal ANC visits among women who had media access were 1.273 (AOR = 1.272; 95% CI: 1.251-1.293) times higher as compared to women who had no media access. The likelihood of optimal ANC visits among women who use health facility delivery was 2.437 (AOR = 2.437; 95% CI: 2.394–2.481) times higher as compared to women who use home delivery (Table 3).
We investigated the prevalence and determinants of optimal ANC visits among reproductive-age women using data from the DHS of 54 LMICs and multilevel binary logistic regression analysis. The prevalence of optimal ANC visits was 60.1% overall. The result was significantly higher than in East Africa [37], Ethiopia [38, 39], Sub-Saharan Africa [36], and Jordan [40]. This result is also lower than those obtained in a systematic review and meta-analysis in Ethiopia [41], Ghana [42], Liberia [43], and Angola [44]. The systematic review and meta-analysis studies could explain this disparity; there is a sample size issue as well as a quality issue for articles that include the meta-analysis. Other studies are limited to a single country and are not representative of other regions. In terms of coverage of basic maternal health interventions such as antenatal care, the African Region has significant intraregional disparities [36]. According to the multilevel logistic regression analysis, maternal age, maternal education level, family size, number of children, sex of household head, wealth index, marital status, husband/partner education level, husband/partner occupation, maternal occupation, media exposure, place of delivery, and region were significant determinants of optimal ANC visits in LMICs.
Women and their partners with higher levels of education have a higher likelihood of receiving optimal ANC visits. This finding is consistent with previous research [3, 36, 45-49]. This is because education improves health-care utilization and increases knowledge about specific issues. Women's empowerment through education, household wealth, and decision-making increases their use of maternal health care services [50].
According to our findings, household wealth status is a significant predictor of optimal ANC visits. Women from the wealthiest families had significantly more ANC visits than women from the poorest families. In line with this finding, studies from Pakistan [51], Sub-Saharan Africa [52], Bangladesh [28, 30, 47], Indonesia [53], and India [54] discovered that women from the wealthiest families were more likely than women from poorer families to receive optimal ANC services. This could be because economic growth encourages health-care utilization and the ability to afford medical and non-medical costs associated with ANC services during pregnancy [55-58]. As a result of the findings, wealth status appears to be an important factor in determining optimal ANC visits. Because of their low socioeconomic status, they have less money to pay for transportation to a health facility to receive ANC services. Another possibility is that women from wealthy families have better education and access to the media than women from poor families.
We discovered that pregnant women with more children and a larger family size had a higher likelihood of receiving optimal ANC visits than pregnant women with fewer children and a smaller family size. It is comparable to previous research from Ethiopia [38, 59], Ghana [60], and Rwanda [61]. A larger family's lack of time and resources, as well as their self-confidence developed from previous pregnancy and childbirth, could be reasons for not using recommended ANC services [62-64].
In this study, married women and women who live with their partners were more likely to have optimal ANC visits than separated women. This research is related to studies [65, 66]. Separated women face a lack of or a reduced level of psychosocial support and relationship stability, according to theories that link marital status, pregnancy, and birth readiness. Unmarried pregnant women may be unplanned and/or unwanted. On the contrary, because illegitimate births are still stigmatized in many countries, social acceptance of separated status is low. As a result, separated women may differ fundamentally from married women in terms of empowerment, self-isolation, and motivation to seek health care [67-69].
According to our findings, working women and women whose husband or partner works had a higher likelihood of using optimal ANC visits than nonworking women and husbands or partners who do not work. This is supported by research from Nigeria [48], and Sub-Saharan Africa [70]. Nonworking women are unable to attend proper ANC visits and do not use health facility delivery due to financial constraints. Furthermore, women aged 20-34 and over 34 years were more likely than those aged under 20 to use optimal ANC visits. This finding corroborated previous research from Uganda [71], Ethiopia [72], and Sub-Saharan Africa [46].
In this study, access to media was also an important factor in health facility utilization. Individuals who had access to the media (reading newspapers or magazines, listening to the radio, and watching television) had a higher chance of using ANC than those who did not. This finding is in agreement with studies [49, 61, 70, 73]. Individuals with access to local media were also more likely to seek medical attention [74]. This simple fact may explain how individuals can quickly obtain various health messages, information about maternal health risk factors, and institutional delivery promotion via multiple radio or television programs [75]. According to the findings of this study, broadcasting the importance of health facility delivery on television, radio, and newspapers may help LMICs achieve maternal and child health-related goals [76]. According to the findings of this study, women who use health facility delivery have significantly higher optimal ANC visits than women who use home delivery. This finding was supported by the study from Sub-Saharan Africa [36]. One possible explanation is that women in health-care settings learned the value of ANC follow-up education. As a result, during ANC follow-up, women will exhibit behavioral changes toward health-care delivery.
Women with a female household head had a higher tendency to initiate optimal ANC visits than those with a male household head. This finding is consistent with the study's findings [70]. Seeking permission from others, especially those of the opposite gender, is a disincentive to early ANC initiation [77]. As a result, independent decision-making may be required for women to access and obtain optimal health care at ANC. Finally, these findings will help the government and stakeholders plan, design, and implement appropriate interventions, as well as address barriers to improving utilization of health facilities, thereby contributing to the reduction of maternal mortality in LMICs.
The study's findings are supported by large datasets from 54 LMICs. The information was gathered using a standard, internationally accepted methodological procedure. The findings are representative of all included countries and generalizable to women in LMICs due to the survey's representative nature. The DHS survey year variation may have an impact on this result. The data was gathered using self-reports from mothers within the five years preceding the survey, which could be a source of recall and misclassification bias.
In low-income countries, maternal age, maternal education level, family size, number of children, gender of household head, wealth index, marital status, husband/partner education level, husband/partner occupation, maternal occupation, media exposure, place of delivery, and region were found to be significant predictors of optimal ANC visits. This discovery aids health care providers and policymakers in putting in place appropriate policies and programs to ensure optimal ANC coverage. It is critical to develop strategies to improve access to and availability of antenatal care. In this study, women with a low family wealth index were less likely to attend optimal ANC visits. Financial assistance that allows mothers from low-income families to cover the costs of ANC services may be beneficial. Health campaigns targeting uneducated women and their husbands or partners are critical for raising awareness about the importance of attending at least four antenatal care services.
ANC: Antenatal Care; AOR: Adjusted Odds Ratio; CI: Confidence Interval; DHS: Demographic Health Survey; FANC: Focal Antenatal Care; ICC: Intra-class Correlation Coefficient; HIV/AIDS: Human Immunodeficiency Virus/Acquired Immunodeficiency Disease Syndrome; LLR: Log-likelihood Ratio; MGDs: Millennium Development Goals; LMICs: Low and Middle Income Countries; MMR: Maternal Mortality Ratio; MOR: Median Odds Ratio; Ref.: Reference Category; SBA: Skilled Birth Attendance; SDG: Sustainable Development Goal; UN: United Nation; WHO: World Health Organization.
Authors’ contributions
YA wrote the proposal, analyzed the data, and did the manuscript writing. AA accredited the proposal with revisions, analysis of the data, and manuscript writing. Both authors read and approved the final manuscript.
Competing interests
The authors declare that they have no competing interests.
Ethics approval and consent to participate
This study was built on the analysis of publicly available secondary data with all identifiers removed. The Institutional Review Board (IRB) of ICF Macro at Fairfax, Virginia, in the USA, reviewed and approved the MEASURE DHS Project Phase 3. The 2010–2018 DHSs are considered under that approval. The IRB of ICF Macro complied with the United States Department of Health and Human Services requirements for the "Protection of Human Subjects" (45 CFR 46). ICF Macro permitted the authors to use the data. Most importantly, the informed consent statement emphasizes that participation is voluntary; that the respondent may refuse to answer any question, decline any biomarker test, or terminate participation at any time; and that the respondent's identity and information will be kept strictly confidential. In addition, written informed consent was obtained from a parent or guardian for participants under 16 years old. The full details of the ethical approvals can be found at http://dhsprogram.com, and the data can be found at https://www.dhsprogram.com/data/dataset_admin/login_main.cfm. We confirm that all methods were carried out in accordance with the relevant guidelines and regulations.
Availability of data and materials
The data used in this article were available on http://dhsprogram.com.
Funding
The authors have no support or funding to report.
Acknowledgments
The authors are grateful to the DHS program for giving us permission to use the data for our purpose. The manuscript was edited and proofread for language by Assafaw Kelebu (MSc.), Department of English Language and Literature, Mekdela Amba University.
Consent for publication
Not applicable