Leaving no one behind: the impact of disability and socioeconomic status on maternal continuum of care

Abstract

Background

To ensure that women with disabilities (WwD) have access to essential maternal health services, understanding their service utilization within the continuum of care (CoC) framework is vital. However, the influence of women’s disability status on maternal CoC has not been fully explored. Hence, this paper examines the completion level and inequality of basic maternal CoC, as well as its association with women’s disability status.

Methods

We conducted analyses on demographic and health survey data of nine low- and middle-income countries collected between 2016 and 2022. Disability among reproductive-age women was assessed using the Washington Group Short Set questionnaires. The maternal CoC was defined to include receiving four or more antenatal visits, skilled birth attendance and obtaining timely postnatal care. Concentration indices were used to measure wealth-related inequalities in completing CoC. Multivariable logistic regression was used to identify factors associated with inequalities in the CoC completion.

Results

A total of 14.0% of women had a disability of at least some difficulty in one domain of function. Among women who made their first antenatal care contact, only 35.8% completed CoC; this percentage was lower among women with disability (32.7%). The odds of completing CoC was lower among WwD (AOR = 0.89, 95% CI: 0.83–0.95). Higher maternal education (AOR = 1.63–2.27), female-headed household (AOR = 1.14, 95% CI: 1.07–1.22), currently working (AOR = 1.29, 95% CI:1.22–1.37) and wealth quintile (increasing from poor to the richest (AOR = 1.24–2.18) were positively associated with higher odds of completing the CoC. We found overall pro-rich inequality in CoC completion (CI 0.27: 95%CI: 0.26–0.29). Higher inequalities were observed in countries with lower coverage of maternal healthcare services.

Conclusion

Maternal CoC completion was lower among WwD, especially those with lower socioeconomic status. Effective strategies that ensure disability-friendly maternal health care services will play a pivotal role. Maternal health service programs should prioritize women’s disability status alongside other key socioeconomic factors and address health care barriers to ensure more equitable and comprehensive maternal health care.

Background

As of 2023, around 16% of the world’s population have some form of disability [1]. Disability affects people of all socioeconomic backgrounds; however, prevalence is higher among females than males (i.e., 19% compared to 12% respectively) and older people [2, 3]. Three-quarters of women with disabilities (WwD) live in low- and middle-income countries (LMICs) [4].

Contrary to the era of eugenics where forced sterilization, violence and rape were common practices against WwD, WwD today, like those without disability, nowadays have the right to access reproductive health services which is both secure and fulfilling [5]. According to the United Nations Convention on the Rights of Persons with Disability (UNCRPD), tailored actions should be taken to ensure access for WwD to health services that are disability-friendly, including sexual and reproductive health (SRH) [6].

The World Health Organization (WHO) and their partners have launched a multipronged and inclusive initiative to improve efforts to curb the maternal mortality ratio in LMICs [7–9]. The Sustainable Development Goals (SDGs), through the agenda of “to leave no one behind”, address the rights and needs of WwD as a matter of priority [10]. However, WwD are one of the most excluded, poorest and most disadvantaged clusters of people in the world [11] and existing services are found to be not friendly to accommodate persons with disabilities [12]. In most cases, the cost of maternal health services for WwD may be higher due to the need for additional accommodation and specialised care. For instance, a woman with a physical impairment may require a wheelchair-accessible facility and a woman who is unable to see may require additional support and assistance during prenatal care and childbirth [13].

WwD face a higher likelihood of negative birth outcomes compared to women without disability (WwoD). Evidence indicates that newborns and infants born to WwD are more prone to low birth weight, low Apgar score, preterm birth, and perinatal death [14, 15]. Pregnancy complications such as gestational diabetes mellitus, gestational hypertension, postpartum depression and postpartum hospitalisation are more prevalent among WwD compared to their counterparts [15].

Given the link between adverse feto-maternal outcomes and disability, WwD should receive tailored maternal health services. However, the typical experience of WwD is poor maternal health service, with few receiving the key components of quality maternal services [16].

Ensuring positive pregnancy and newborn outcomes for WwD requires access to a continuum of maternal care underscored by notions of availability, accessibility, affordability, equity and quality among other actions [17]. The notion of continuum of care (CoC) first came to effect in the 1970 s in the integration of research and practice for provision of a continuum of care for elderly people [18]. Since then, the concept has been extensively used in health care systems, particularly nursing, palliative care, and mental health, with less focus on public health [19]. The maternal CoC outlines the steps that mothers go through from pregnancy to childbirth and beyond [17]. It is an ideal model tailored to the unique needs of special populations like WwD, aiming to optimize their health and their newborn, and to make sure they are not overlooked due to gaps in a fragile system. Maternal CoC is defined in two dimensions: time and place of care. The time dimension includes the care provided from pre-pregnancy through the postpartum period [17]. The place dimension includes care provided at home, first level facility, and in hospital.

Evidence suggests that the implementation of CoC concepts within maternal health services has led to a decrease in maternal and neonatal mortality [20]. Despite substantial progress in reducing maternal mortality and morbidity in the past couple of decades in LMICs, socioeconomic inequalities in access to maternal CoC are significant and persistent [21]. While there has been some research on maternal CoC among WwoD [22, 23], quantitative research on maternal CoC among WwD remains scarce [24, 25].

To ensure that WwD can effectively access essential maternal health services, understanding WwD’s key maternal service use within the CoC framework is crucial. Therefore, the present study aims to explore socioeconomic inequality in CoC completion among WwD in LMICs through the lens of the intersectionality of disability and wealth status. The predictors of inequality in completing CoC among women with and without disability were also evaluated. The research findings will provide evidence-based insight for maternal, newborn and child health programs that will support policy decisions toward strengthening the CoC for WwD in LMICs.

Methods

Data sources, study design, and study population

This study utilized the most recent DHS data from nine countries: six in sub-Saharan Africa, namelyMauritania (2019-21), Rwanda (2019-20), South Africa (2016), Mali (2018), Senegal (2019), and Uganda (2016); and three countries in South and Southeast Asia, namely Cambodia (2021-22), Pakistan (2017-18), and Timor-Leste (2016). Data from these countries were cleaned differently and appended together for the final analysis. Only countries that collected data related to disability status were included. The DHS is a population based cross-sectional study which aims to collect country-specific data on population characteristics and healthcare utilisation. Data includes main demographic indicators such as household socioeconomic status, sexual and reproductive health, family planning utilization, fertility preferences, disability status, domestic violence, childhood vaccination, nutritional status of children and women, HIV/AIDS, and other sexually transmitted infections [26]. The demographic health survey uses the multistage sampling techniques. The sampling frame used depends on each country’s population and housing census. The census frame is a complete list of enumeration areas (EAs) created for the respective country’s population and housing census. Women of reproductive age (aged 15 to 49 years) who gave birth in the last five years preceding the survey were included in this study.

Variables of the study

Outcome variable

We examined maternal CoC completion among women with and without disability which was constructed from the following three maternity services data items: (I) Antenatal care: receiving a minimum of four ANC during her last childbirth [8]; (II) Skilled birth attendance: giving birth with support of qualified professionals considered as skilled delivery attendants by respective country [27]; (III) postnatal care [9]. In this study, postnatal care refers to healthcare provided within 48 hours, regardless of the place of delivery, as evidence suggests that incidence of maternal mortality is highest within the first forty-eight hours following delivery [28]. Women who received these three comprehensive maternity care services were categorized as completed CoC and those who failed to receive any one of the components (i.e. received fewer than three components) are categorized as not completed CoC or dropout from continuum of care (Fig. 1) [29–31].

Fig. 1.

Continuation and dropout along the maternal continuum of care pathway

*Only women withcomplete information on disability status are included; **Weighted samples for both WwD and WwoD

Independent variables

The main explanatory variables in this study are disability status and socioeconomic status.

Disability status was assessed in demographic health survey using the Washington Group Short Set (WG-SS) questionnaire on functioning [32]. There were six sets of WG-SS on six domains of function to define disability. These questions include whether the women have any difficulty in each domain of function, i.e., vision, hearing, communication, cognitive, physical, and selfcare. For instance, “Do you have difficulty in seeing?”. The answer to these questions is: “no difficulty (a)”, “some difficulty (b)”, “a lot of difficulties (c)” or “cannot do at all (d)”. In our analysis, an individual was categorized as having a disability (‘Yes’) if they experienced at least some difficulty (b to d) in at least one domain of function, or as having no disability (‘No’) if they reported no difficulty (a) in all domains of function [33]. Additionally, a three-level disability severity variable was created: ‘severe disability’ for participants who reported ‘a lot of difficulty’ or ‘cannot do at all’ (c to d) in any domain of function; ‘some difficulty’ for those who reported some difficulty (b) in at least one domain of function; and ‘no disability’ for those who reported no difficulty (a) in any domain of function [33].

Given that socioeconomic status refers to the social rank or an indicator of living standards in a particular community or society, we employed household wealth quintile and educational attainment as indicators of socioeconomic status. In the DHS, the wealth index was derived from household asset information using principal component analysis. To create this index, each household member is assigned a score based on their assets. The individuals are then ranked according to these scores, and the overall population is divided into five equal groups, from the poorest to the wealthiest, with each group representing 20% of the population.

The educational level of the mother, another socioeconomic indicator, was categorized into three groups: no education, primary education, and secondary or higher education.

The analysis also considered other covariates, including sociodemographic factors such as the respondent’s age, marital status, employment status. Other factors included the WHO region (i.e., Sub-Saharan Africa or South and Southeast Asia), reproductive characteristics (parity, age at first pregnancy and first birth, pregnancy intendedness, history of pregnancy termination, and mode of delivery), and media exposure. The choice of these variables was informed by the Andersen-Newman Behavioral Model of Health care Utilization [34], the WHO Commission on Social Determinants of Health framework [35], and relevant literature [36].

Community level variables

Community-level variables include media exposure, wealth status, residence (urban/rural), and WHO region (sub-Saharan Africa or South and Southeast Asia). Community level media exposure and wealth status were derived by aggregating individual characteristics within each community (cluster). Median were used as cut-off points for the categorization based on the distribution as follows: Community level poverty was categorized as low or high based on the median value.

Community-level media exposure was measured by the proportion of women who had been exposed to at least one media (television, radio or newspaper/magazine). Community level media exposure was classified as high if the proportion was equal to or greater than median value, and as low if it was below the median value.

Data management and statistical analysis

Data analysis was performed by using Stata version 18 software [37]. Both descriptive statistics and multilevel logistic regression were employed. Weighting was applied to restore representativeness of statistical estimates for all analyses. The missing values were managed by excluding them from both the numerator and the denominator or leaving them as they were if the data in some variables was not affected.

The DHS exhibits a hierarchical structure, meaning that participants within a cluster may be more likely to share similar characteristics compared to those in the other clusters. This condition violates the assumption of independence of observation and variability between the clusters. Hence, we have employed multilevel mixed-effects logistic regression to address cluster level variability and to determine factors influencing completing maternal CoC [38]. Initially, the relationship between individual and community-level variables and the completion of CoC was analysed independently using bi-variable multilevel mixed-effects logistic regression models. Variables with a p-value of ≤ 0.20 in the bivariate analysis were included in the multivariable model [39, 40]. Factors with a p-value of < 0.05 were identified as significant predictors of maternal CoC completion in a multivariable multilevel mixed-effects analysis. Intraclass correlation coefficient (ICC) was computed to quantify the proportion of overall variation in completing CoC attributable to clustering.

To identify the factors influencing CoC completion, four distinct models were applied. The null model serves as a baseline intercept-only model, with no explanatory variables included. It is used to assess random variability in the intercept and to estimate the ICC. Model I incorporates individual-level variables to determine their impact on CoC completion. Model II utilizes community-level variables to evaluate their effects. The final model, Model III, combines both individual and community variables to assess their collective influence on CoC completion.

The potential multicollinearity was assessed by variance inflation factor. To assess the variability of maternal CoC completion between clusters, both ICC and proportional change in variance (PCV) were calculated. Regarding model fit, the model with the lowest AIC, model III, was the best-fitted model [41].

We utilized the concentration curve alongside with concentration index (CCI) to assess wealth-related inequalities in CoC completion among women, both with and without disabilities, within each country. The concentration curve graphs the cumulative percentage of CoC completion on y-axis against the cumulative percentage of the population, ranked by householdwealth on x-axis [42]. When the curve lies above the equality line (45^°), proportion of completing CoC is disproportionately concentrated on the poor. When the curve lies below the equality line, the proportion of completing CoC is disproportionately concentrated among the rich. The CCI quantifies the degree of socioeconomic-related inequality in completing maternal CoC. The sign and magnitude of the CCI indicate the direction and degree of inequality in the health variable. The sign of the CCIconcentration index indicates the direction of relationship between the CoC completion and cumulative percentage of the population ranked by living standards, and its magnitude reflects both the strength of the relationship and the degree of variability in CoC completion [42, 43].

Results

Sociodemographic characteristics

A total of 53,615, including 7,507 (14.0%) WwD and 46,108 (86.0%) WwoD, were included in the final analysis. The median (IQR) ages of WwD and WwoD were 32 (± 12) and 29 (± 10) years respectively. 30.2% of WwD and 32.2% of WwoD were urban residents. Regarding the wealth index, 17.0% and 21.7% of the WwD were from the richest and poorest households respectively. Regarding the educational status, 28.0% of WwD and 31.9% of WwoD had no formal education (Table 1). Regarding the disability status, 14.0% of women were living with at least some difficulty in at least one domain of function. The vision domain of function reported the highest proportion of “some difficulty” at 52.8% (Fig. 2).

Table 1.

Sociodemographic characteristics of the study participants

Variables*	Disability status		Total n = 53,615(100%)	f++ P-value
	WwoD n = 46,108 (85.5%)	WwD n = 7,507(14.5%)
Age in 5-year groups
15–19	2883(6.3)	192(2.6)	3075(5.7)	p < 0.001
20–24	9732(21.1)	1081(14.4)	10,813(20.2)
25–29	12,338(26.8)	1526(20.3)	13,864(25.9)
30–34	10,292(22.3)	1754(23.4)	12,047(22.5)
35–39	6866(14.9)	1568(20.9)	8434(15.7)
40–44	3054(6.6)	1020(13.6)	4074(7.6)
45–49	944(2.0)	365(4.9)	1309(2.4)
Type of place of residence
Urban	14,826(32.2)	2265(30.2)	17,091(31.9)	p < 0.001
Rural	31,282(67.8)	5242(69.8)	36,525(68.1)
Highest educational level
No education	14,710(31.9)	2106(28.0)	16,816(31.4)	p < 0.001
Primary	15,593(33.8)	3328(44.3)	18,921(35.3)
Secondary	12,833(27.8)	1707(22.7)	14,539(27.1)
Higher	2972(6.4)	367(4.9)	3339(6.2)
Sex of household head
Male	35,328(76.6)	5224(69.6)	40,552(75.6)	p < 0.001
Female	10,780(23.4)	2283(30.4)	13,063(24.4)
Wealth index combined
Poorest	9707(21.1)	1632(21.7)	11,339(21.1)	P < 0.001
Poorer	9353(20.3)	1635(21.8)	10,987(20.5)
Middle	9106(19.7)	1509(20.1)	10,614(19.8)
Richer	9124(19.8)	1459(19.4)	10,583(19.7)
Richest	8819(19.1)	1274(17.0)	10,092(18.8)
Frequency of reading newspaper or magazine
Not at all	37,700(81.8)	6271(83.5)	43,971(82)	P < 0.001
Less than once a week	4968(10.8)	758(10.1)	5726(10.7)
At least once a week	3438(7.5)	478(6.4)	3915(7.3)
Frequency of listening to radio
Not at all	21,952(47.6)	3525(47.0)	25,477(47.5)	p < 0.001
Less than once a week	8782(19.0)	1260(16.8)	10,042(18.7)
At least once a week	15,374(33.3)	2722(36.3)	18,096(33.8)
Frequency of watching television
Not at all	22,503(48.8)	4216(56.2)	26,719(49.8)	p < 0.001
Less than once a week	7930(17.2)	1128(15.0)	9058(16.9)
At least once a week	15,675(34)	2163(28.8)	17,838(33.3)
WHO region
South and south-east Asia	14,173(30.7)	1855(24.7)	16,028(29.9)	p < 0.001
Sub-Saharan Africa	31,936(69.3)	5652(75.3)	37,588(70.1)
Marital Status
Not Married/Not in union	5418(11.8)	993(13.2)	6411(12.0)	p < 0.001
Married/living with partner	40,690(88.2)	6514(86.8)	47,204(88.0)
Respondent currently working				P = 0.001
No	24,267(52.6)	3235(43.1)	27,502(51.3)
Yes	21,841(47.4)	4272(56.9)	26,113(48.7)
Covered by health insurance				P < 0.001
No	30,034(82.8)	5631(82.1)	35,666(82.7)
Yes	6249(17.2)	1226(17.9)	7475(17.3)

^** Design-based f-statistics; ^*Percentages are generated from the column total

Fig. 2.

Percentage of women with disabilities who recently gave childbirth by domain of function

Forty percent of the study participants (54.2% among WwD and 37.6% among WwoD) had a parity of four or more, while only 11.5% had their most recent birth by caesarean section. (Table 2). One in six had a history of termination of pregnancy, with a higher proportion among WwD (24.6%). Additionally, 56.1% of women (52.7% among WwD and 56.6% among WwoD) had their first antenatal visit within 12 weeks of their pregnancy. Most respondents (73.6%) reported their most recent pregnancy as intended, with 64.1% among WwD and 75.1% among WwoD.

Table 2.

Reproductive characteristics and maternal continuum of care of women with and without disability

Variables	Disability status
	WwoD n = 46108 (85.5%)	WwD n = 7,507 (14.5%)	Total n = 53,615 (100.0%)	Design-based-f P -value
Age at first birth
<20	25690(55.7)	4207(56)	29896(55.8)	P=0.080
20-30	19253(41.8)	3067(40.9)	22320(41.6)
≥31	1166(2.5)	233(3.1)	1399(2.6)
Pregnancy wanted				P<0.001
Intended	34627(75.1)	4811(64.1)	39438(73.6)
Unintended	11457(24.9)	2696(35.9)	14153(26.4)
Number of ANC visit				P=.001
No ANC visit	4553(9.9)	635(8.5)	5188(9.7)
1 to 3	15453(33.5)	2697(35.9)	18150(33.8)
≥4	26103(56.6)	4175(55.6)	30278(56.5)
Timing of first ANC visit				P<0.001
≤12 weeks	23951(56.6)	3698(52.7)	27649(56.1)
>12 weeks	18352(43.4)	3324(47.3)	21676(43.9)
Birth order				P<0.001
1st	10828(23.5)	1002(13.3)	11830(22.1)
2nd	10235(22.2)	1248(16.6)	11483(21.4)
3rd	7688(16.7)	1192(15.9)	8880(16.6)
≥4th	17357(37.6)	4065(54.2)	21422(40.0)
Skilled birth attendance				P=0.985
No	12102(26.2)	1969(26.2)	14071(26.2)
Yes	34007(73.8)	5538(73.8)	39544(73.8)
Delivery by caesarean section				P=0.329
No	40812(88.6)	6601(88.1)	47413(88.5)
Yes	5247(11.4)	891(11.9)	6138(11.5)
Sex of child				P=0365
Male	23519(51)	3780(50.4)	27299(50.9)
Female	22589(49)	3727(49.6)	26316(49.1)
Ever had a terminated pregnancy				P<0.001
No	38393(83.3)	5614(74.8)	44007(82.1)
Yes	7715(16.7)	1893(25.2)	9608(17.9)
Received PNC				P<0.001
No	17862(38.7)	3410(45.4)	21272(39.7)
Yes	28247(61.3)	4097(54.6)	32344(60.3)
Completed CoC				P<0.001
No	29344(63.6)	5053(67.3)	34397(64.2)
Yes	16764(36.4)	2454(32.7)	19218(35.8)

Percentages calculated from the column total

Continuum of care coverage among women with and without disability

Nine in ten of women surveyed had a minimum of one ANC contact; however, only 19,218 (35.8%) had completed maternal CoC (32.7% and 35.8% among WwD and WwoD respectively) (Table 3). Among the women included in the study, 56.5% (55.6% and 56.6% among WwD and WwoD, respectively) had four or more ANC visits. There is no significant difference in skilled birth attendance during childbirth between WwD (73.8%) and WwoD (73.8%). Three in five women (54.6% of WwD and 61.3% of WwoD) received postnatal care within two days of delivery (Table 2).

Table 3.

Completion of continuum of care among the domains of disabilities

Variables	Categories	Continuum of care completed++		A design-based f-test P -Value
		No	Yes
Overall disability status	No disability	29344 (63.6)	16764 (36.4)	f= 21.73 p < 0.001
	Have disability	5053 (67.3)	2454 (32.7)
Domain of function**
Vision	Have no disability	31854 (64)	17931 (36)	f=7.80 p = 0.0053
	Have disability	2650 (66.8)	1315 (33.2)
Hearing	Have no disability	33550 (64)	18901 (36)	f=18.0 p = 0.0000
	Have disability	931 (70.8)	384 (29.2)
Communication	Have no disability	34322 (64.1)	19213 (35.9)	f=4.90 p = 0.027
	Have disability	263 (70.3)	111 (29.7)
Cognitive	Have no disability	32896 (63.9)	18550 (36.1)	f=11.73 p = 0.0006
	Have disability	1685 (68.6)	773 (31.4)
Physical	Have no disability	33281 (64)	18700 (36)	f=6.38 p = 0.012
	Have disability	1299 (67.5)	625 (32.5)
Self-care	Have no disability	34380 (64.1)	19224 (35.9)	f=0.58 p = 0.4455
	Have disability	202 (66.9)	100 (33.1)
Level of severity	no difficulty	29344 (63.6)	16764 (36.4)	f=12.38 p = 0.0000
	Some difficulties	4412 (67.1)	2167 (32.9)
	A lot of difficulties or can't do at all	641 (69.1)	287 (31)

Figure 3 provide maternal CoC among WwD across countries. Maternal CoC completion among WwD ranged from 19.4% in Mauritania to 71.3% in Cambodia. Receiving four or more ANC visit among WwD ranged from 40.5% in Mauritania to 86.0% in Cambodia. Skilled birth attendance among WwD ranged from 50.6% in Mali to 96.5% in Cambodia.

Fig. 3.

Coverage of key maternal health services among women with disabilities

Completion of continuum of care among the domains of disabilities

In the context of domains of disability, except self-care domain, women with at least some difficulty in each domain of disability are less likely to complete maternal CoC. CoC completion ranges from 29.2% among women with “at least some difficulty of hearing” to 33.2% among women with “at least some difficulty of seeing” (Table 3).

Inequalities in completion of continuum of care among WwD

Among WwD, the absolute difference in completing CoC between the poorest and richest was 22.4% (24.5% among the poorest quintile compared to to 46.9% among the richest quintiles). Overall, the absolute inequality in completing the maternal CoC varies significantly among all women across different countries, with a big difference observed within WwD in Pakistan (56.0%). (Fig. 4) The findings from concentration indices revealed high wealth-related inequality in completing CoC across countries, with highest inequality in countries with lower proportion of CoC completion. The highest wealth-related inequality among WwD were reported in Pakistan (CCI = 0.41, 95% CI 0.32 to 0.50) and Mali (CCI = 0.41, 95% CI 0.32 to 0.50) whereas the lowest wealth related inequality was observed in South Africa (CCI=−0.05, 95% CI −0.21 to 0.12) (Fig. 5). Wealth-related disparities in CoC completion among WwD in LMICs were compared between rural and urban areas, as illustrated in Fig. 5. The positive CCIs show pro-rich coverage in completing CoC. While not completing CoC showed disproportionate concentration among women with lower socioeconomic status (Fig. 6).

Fig. 4.

Absolute inequality in completing continuum of care among women with disability

Fig. 5.

Socioeconomic inequalities (concentration index and 95% confidence interval) in completion of maternal continuum of care among women with disabilities in nine LMICs

Fig. 6.

Concentration curves for completion of maternal continuum of care for WwD across nine LMICs

Predictors of continuum of care completion

In multilevel mixed-effects logistic regression, disability status, household wealth status, level of educational, sex of head of household, marital status, working status, pregnancy intention, delivery by caesarean, ever terminated pregnancy, residence, exposure to media, and WHO region of countries were associated with completing maternal CoC. Having a minimum of “some difficulty” in any domain of function is associated with lower odds of CoC completion (AOR 0.89, 95%CI: 0.83, 0.95) compared to WwoD. The odds of CoC completion increased with level of maternal education (AOR 1.63, (95%CI:1.51,1.76) for primary education to (AOR 2.27, 95% CI:2.09, 2.46) for secondary or higher education), working status (AOR 1.29, 95% CI: 1.22, 1.37), wealth (increasing from poor to the richest quintile, AOR: 1.24 to 2.18), caesarean delivery (AOR 2.69, 95% CI:2.46, 2.93), urban residence (AOR 1.36, 95% CI:1.25, 1.49) and facing big problem in access to health care (AOR 0.81, 95% CI:0.76, 0.85). The odds of completing continuum of care were higher among women from south and southeast Asia (AOR 1.68, 95% CI:1.56, 1.82) compared to those women from sub-Saharan Africa (Table 4).

Table 4.

Determinants of continuum of care completionamong women with and without disability in LMICs

Variables	Model I AOR (95% CI)	Model II AOR (95% CI)	Model III AOR (95% CI)	P-Value
Disability status
No disability	1.0		1.0(reference)
Have disability	0.88 (0.82, 0.94)		0.89 (0.83, 0.95)
Age of respondents
15–24	1.0(reference)
25–29	1.09 (1.02, 1.16)		1.05 (0.98, 1.12)	0.163
30–34	1.13 (1.05, 1.22)		1.08 (1.00, 1.16)	0.044
35–49	1.00 (0.93, 1.07)		0.96 (0.89, 1.03)	0.219
Educational status
No education	1.0(reference)		1.0(reference)
Primary	1.61 (1.49, 1.74)		1.63 (1.51, 1.76)	0.000
Secondary or higher	2.57 (2.37, 2.79)		2.27 (2.09, 2.46)	0.000
Sex of head of household
Male	1.0(reference)		1.0(reference)
Female	1.11 (1.04, 1.18)		1.14 (1.07, 1.22)	0.000
Wealth Index
Poorest	1.0(reference)		1.0(reference)
Poorer	1.25 (1.15, 1.35)		1.24 (1.14, 1.35)	0.000
Middle	1.51 (1.38, 1.65)		1.46 (1.33, 1.60)	0.000
Richer	1.88 (1.70, 2.07)		1.72 (1.55, 1.90)	0.000
Richest	2.48 (2.24, 2.75)		2.18 (1.95, 2.43)	0.000
Pregnancy wanted
Intended	1.0(reference)		1.0(reference)
Unintended	0.77 (0.72, 0.82)		0.83 (0.78, 0.89)	0.000
Respondents currently working
No	1.0(reference)		1.0(reference)
Yes	1.15 (1.09, 1.22)		1.29 (1.22, 1.37)	0.000
Marital Status
Not married/Not in union	1.0(reference)		1.0(reference)
Married/Live with partner	1.10 (1.01, 1.19)		1.02 (0.94, 1.11)	0.648
Delivery by caesarean section
Yes	2.84 (2.60, 3.10)		2.69 (2.46, 2.93)	0.000
No	1.0(reference)		1.0(reference)
Ever had terminated pregnancy
Yes	1.19 (1.11, 1.27)		1.15 (1.08, 1.23)	0.000
No	1.0(reference)		1.0(reference)
Health care barrier
Not a big problem	1.0(reference)		1.0(reference)
Big problem	0.84 (0.79, 0.88)		0.81 (0.76, 0.85)	0.000
Community level predictors
Region
Sub-Saharan Africa		1.0(reference)	1.0(reference)
South and south-east Asia		1.72 (1.58, 1.87)	1.68 (1.56, 1.82)	0.000
Residence
Urban		2.46 (2.26, 2.68)	1.36 (1.25, 1.49)	0.000
Rural		1.0(reference)	1.0(reference)
Community level media exposure
High		1.11 (1.02, 1.22)	1.03 (0.95, 1.12)	0.466
Low		1.0(reference)
Community level poverty
High		1.01 (0.92, 1.11)	1.06 (0.97, 1.15)	0.183
Low		1.0(reference)
Model parameters	Null model	Model I	Model II	Model III
ICC	7.8%	6.4%	8.0%	6.0%
PCV	Reference	0.1926	−0.0342	0.2419
Log likelihood	−34533.2	−30176.0	−33298.8	−29880.7
AIC	69070.4	60390.1	66609.62	59807.4
BIC	69088.1	60558.4	66,663	60011.2

Table 4. also shows that the ICC in the null model showed that 7.8% of the variability in CoC completion was attributed due to differences between clusters. The highest PCV in model III revealed that 24.2% of the variability in CoC completion was explained by the combined multilevel individual and community level factors. Similarly, the analysis of intersectionality of disability, household wealth, education status, and healthcare barrier showed the disparities in CoC completion were influenced by intersection of disadvantaged groups. When a WwD faces compounded disadvantage i.e. having a disability and being poor, being rural resident, having big problem of healthcare barrier, and having no education, their odds of CoC completion is low compared to their counterparts. However, when a WwD have a double advantage i.e. non poor or educated her CoC completion level increases. The difference between these groups remain significant after controlling other covariates (Fig. 7).

Fig. 7.

Intersectionality of Disability and socioeconomic status and CoC Completion. * Each model was adjusted for all variables included in Table 4

Discussion

Access to the maternal and newborn CoC has a positive impact on both mother and newborn health. However, the use of these services is fragmented across the care pathways in many LMICs. Our findings present the pattern of maternal health service utilization along the maternal CoC in nine LMICs, comparing WwD with WwoD. About one in seven (14.0%) women in our sample have a disability which involves at least some difficulty in one of the domains of function. This is comparable with the global report of people living with at least some difficulty (15%) [3].

In this study, 90.3% of women had at least one ANC visit during pregnancy, yet only 35.8% fully completed the CoC, with a slightly lower percentage among WwD (32.7%) compared to WwoD (36.4%). The odds of completing CoC was 11% lower among WwD compared to WwoD after controlling for other covariates. There might be several reasons for the shortfall of WwD from CoC. Physically, WwD may encounter difficulties in accessing healthcare facilities due to inadequate infrastructure, lack of transportation options, or inaccessible medical equipment [44]. Additionally, they may face discrimination or stigma from healthcare providers, leading to substandard care or denial of services. Systemically, healthcare systems may not be equipped to accommodate the specific needs of WwD, such as providing sign language interpreters, offering accessible information, or ensuring that medical personnel are trained in disability-inclusive care [45]. These factors contribute to a higher likelihood of WwD experiencing complications during pregnancy and childbirth, as well as higher rates of maternal and infant mortality [14, 15].

Only 55.6% of women had the minimum of 4 ANC contacts, with a slightly lower percentage among WwD (54.8%) compared to WwoD (55.8%). The gap in the continuity of care from first ANC contact to ANC4 + is probably due to delays in their first visit and inadequate quality of care provided during early visits [46]. Nearly half of WwD (47.3%) and 43.4% of WwoD experienced delayed initiation of ANC visit. This is lower than previous studies in LMICs, that report the proportion of women with delayed ANC initiation from 63.0% [47] to 68.9% [22]. This difference might be due to the countries included in the studies. For instance, three Asian countries with improved maternal health services are part of this study, in which 71.4% of them had timely ANC initiation compared to 49.6% in SSA. Given, antenatal contact is a window of opportunity to sign up women into the maternal CoC, this finding highlights the urgent need for a roadmap of strategic actions to end preventable maternal and neonatal death by ensuring access to the maternal CoC, which should prioritize WwD.

The highest dropout along the CoC is detected between childbirth and postnatal follow-up with the largest dropout observed among WwD. Given the sizable number of maternal deaths that occur during childbirth and shortly after delivery, the postnatal period would be a potential time to curb the death rate of the maternal-baby dyad. It is also a potential period to provide counselling for nutrition and family planning methods. Our finding of a large gap between entering the maternal care pathway and completing the CoC is inline with similar study from sub-Saharan Africa [22].

Our inequality analysis highlighted that WwD from a lower socioeconomic background experienced a lower completion of maternal CoC. For instance, the rate of CoC completion is higher among the richest quintile by more than one-quarter in countries such as Mali (47.8%), Senegal (31.6%), and Pakistan (56.0%) as compared to WwD from lower wealth quintile. The inequalities between the countries could be attributed to difference in countries’ performance to ensure compliance of their laws and policies with the UNCRPD. For instance, UNCRPD identified serious concerns about unequal access to education, employment, and healthcare services, including services related to HIV/AIDS, for women and girls with disabilities in Senegal [48]. Similarly, in Rwanda, the UNCRPD has identified insufficient access to public health education, healthcare facilities and services, including emergency care and sexual and reproductive health services, particularly in remote rural regions. Additionally, UNCRPD identified that there was insufficient training for health professionals on the human rights of persons with disabilities, including issues related to free and informed consent [49].

Likewise, in Pakistan, where our findings indicated one of the highest levels of inequality, UNCRPD committee identified a lack of proper mechanisms at the national and local level to implement policies developed for the wellbeing of disadvantaged people [50]. In contrast, in countries like South Africa, where there has been an improvement in socioeconomic status [51], and Cambodia, where maternal service coverage is high [52], less wealth related disparity was observed in the completion of CoC among WwD.

Overall, WwD in sub-Saharan African countries such as Mali, Senegal, Mauritius, Rwanda, Uganda, and South-Africa recorded a lower percentage completing CoC as compared to south Asian countries such as Cambodia, Pakistan, and Timor-Leste (29.2% versus 43.4%). Even though both WHO regions are the highest shareholders of maternal mortality, the lower coverage of key maternal health services has been witnessed in sub-Saharan Africa [53]. The gap could be attributed to the lower universal health care service coverage index and the smaller proportion of the population spending over 10% of their household budget on healthcare in sub-Saharan Africa. Additionally, in sub-Saharan Africa more than two-thirds of the rural population is affected by lack of access, which needs a holistic approach to achieve the SDG agenda of leaving no one behind by 2030 [54].

In our multilevel analysis, beside disability status, other covariates such as educational status, female headed household, wealth status, working status, caesarean delivery, ever had terminated pregnancy, WHO region, and exposure to media were found to be associated with maternal CoC completion.

We found that residence (urban) and region (South and Eastern Asia) were positively associated with maternal CoC completion. The fact that rural women are less likely to complete CoC could be attributed to the interplay of factors such as low acceptability of service, low socioeconomic status, barriers in terms of geography, and finance. This can be addressed by increasing healthcare workers coverage in rural and remote areas aligning with the SDGs of leaving no one behind, universal health coverage, the reaffirmation of a global dedication to primary health care, and an improved evidence base to support cost-effective interventions to improve WwD’s access to maternal health care services [55].

Women who attained secondary or higher education were more likely completing CoC. This observation aligns with earlier research conducted in LMICs [56, 57] that demonstrates a positive association between education attainment and CoC completion. Maternal education and media exposure may enhance access to maternal healthcare through the synergistic effect of education on health literacy fostering positive attitudes and improving health-seeking behavior [58].

Our finding indicates that women from lower socioeconomic status and those who are not working are less likely to complete CoC. Women from lower socioeconomic status are affected through multifaceted ways: limited financial resources, limited health insurance coverage, transportation and geographic barriers, and psychosocial stressors [59]. Hence, strategies such as community-based health workers, conditional cash transfers and removing persistent barriers to WwD should be considered by all countries to address inequality in access to care.

We also found that women who are unmarried, i.e., divorced, widowed or separated were more likely to receive no or few of maternal health services. In contrast female headed household, delivery by caesarean section, and having terminated pregnancy are positively related with maternal CoC completion. The positive association to caesarean delivery is likely to reflect the longer hospital length of stay, ensuring access to postnatal care within the birth hospitalisation as found in other LMICs studies [60, 61].

Pregnancy intention was negatively associated with completing maternal CoC. Women with unintended pregnancy were less likely to complete maternal continuum of care. This finding is in line with studies conducted in LMICs [31, 62]. Several factors contribute to this negative association. Unintended pregnancy is common among women with lower income and education status. In this study, unintended pregnancy was higher among WwD which creates the compounded disadvantage for WwD to navigate through maternal healthcare services.

As a strength, our study is the first to evaluate socioeconomic inequality along maternal continuum of care for women with disability in nine LMICs using DHS. DHS surveys cover a wide range of health and demographic indicators using standardized data collection methods and questionnaires across different countries, ensuring comparability of data over time and between countries. Our study is based on a large sample size which enhances the reliability and precision of the estimates. Additionally, this study used concentration curves and indexes to offer more comprehensive evidence to evaluate inequalities in completing the continuum of care relative to socioeconomic indicators such as wealth and educational status. It also examines other significant determinants of these inequalities through a multilevel analysis.

Our study has some limitations. Given that our study relies on self-reported information, there may be recall bias, especially for events that occurred in the distant past, ranging from weeks to five years. Despite large sample sizes, sampling errors can still affect the precision of estimates, particularly for small sub-populations.

Policy implications

This study comprehensively investigated the socioeconomic related inequalities in completion of continuum of care by including women living with disabilities. Leaving no one behind is a core principle of the SDGs to identify and lift up those who are left furthest behind. Hence, to reduce preventable maternal and child mortality, which is unequivocally higher among the socioeconomically disadvantaged women and women with disability, equitable and comprehensive access to the maternity continuum of care plays a paramount role.

Achieving equity in access to maternal CoC must involve more than just targeting those with the lowest income levels but also necessitates combating discrimination and rising inequalities within and amongst WwD, and their root causes. A comprehensive approach that considers both the physical and financial barriers is crucial to ensure WwD receive equal access to maternal CoC. Removing physical barriers should involve a continuity of care strategy between various caregiving locations, including households, communities, outpatient and outreach services, clinical-care settings, and residential places [17].

Additionally, reducing financial barriers through creating effective community support networks and implementing policies that provide financial assistance or subsidies specifically targeted at poor and WwD is paramount. These policies might include conditional cash transfers, which serve as a targeted approach to improve the well-being and empower WwD through a combination of financial support and incentivized actions [63].

Conclusion

The completion of the maternal CoC was low in LMICs, particularly among women with disabilities and those from lower socioeconomic backgrounds, who are often left behind in care pathways. The main contributing factors for these inequalities were socioeconomic status, in which the wealthier, educated, urban residents, and those who have media exposure were more likely to complete maternal CoC. Comprehensive and inclusive policies are vital to address the unique healthcare needs of women with disabilities. This involves ensuring accessible maternal health services, providing tailored information and education, and incorporating disability-inclusive practices into existing healthcare frameworks. Additionally, promoting equal access to maternal health services and addressing both physical and financial barriers are crucial. Overall, these recommendations emphasize the importance of fostering inclusivity, awareness, and targeted support within reproductive health policies for women with disabilities.

Abbreviations

AIC

Akaike Information Criterion

ANC

Antenatal Care

AOR

Adjusted Odds Ratio

BIC

Bayesian Information Criterion

CCI

Concentration Curve Index

CI

Confidence Interval

COC

Continuum Of Care

DHS

Demographic Health Surveys

ICC

Intraclass Correlation Coefficients

LMICs

Low-And Middle-Income Countries

SDG

Sustainable Development Goals

UNCRPD

United Nations Convention on Rights of People with Disability

WG-SS

Washington Group Short Set Questionnaires

WHO

World Health Organization

WwoD

Women Without Disabilities

WwD

Women With Disabilities

Authors’ contributions

ET, JAO, LG, and HL conceived the study. ET carried out the data analysis and interpretation. ETJ drafted the article. ET, JAO, LG, HL, EL, and HM involved in critical revision of the article. All authors read and approved the final manuscript.

Funding

ET is supported by the Deakin University Postgraduate Research Scholarship (DUPRS). However, the author(s) received no specific funding for this work.

Data availability

This study employed an analysis of an existing dataset from the DHS repository, which is freely accessible online with all identifying information removed (https://www.dhsprogram.com/Data/).

Declarations

Ethical approval and consent to participate

The study was conducted in accordance with the Declaration of Helsinki. Ethical approval to use the secondary data from DHS was secured from the Deakin University Research Ethics (2023 − 102). Permission to access the DHS data was obtained from MEASURE DHS International program following the submission of a brief outlining the study’s objectives. Consent to participate is not applicable as the data used were publicly available.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.