Factors associated with early initiation of antenatal care in Bangladesh: a survival analysis using Bangladesh Demographic and Health Survey 2022

Abstract

Objectives

This study aims to identify factors associated with early antenatal care (ANC) initiation using a survival analysis approach applied to nationally representative data.

Design, setting and participants

This study used a cross-sectional design based on data from the nationally representative 2022 Bangladesh Demographic and Health Survey. The survey was conducted at the community level across all administrative divisions of Bangladesh. A total of 5128 ever-married women aged 15–49 years who had a live birth within 5 years prior to the survey were included in the analysis. Women with missing or incomplete information regarding the timing of their first ANC visit were excluded from the study.

Outcome

The primary outcome was early initiation of ANC, defined as the first ANC contact within the first trimester.

Materials and methods

The study applied survival analysis methods, including Kaplan-Meier survival curves, log-rank tests and an Accelerated Failure Time model, to assess the determinants of early ANC initiation.

Findings

Only 37.9% (95% CI 36.0% to 39.9%) of women in Bangladesh initiated ANC within the first trimester. Early ANC initiation was associated with higher maternal age, education, skilled employment, wealthier households, media exposure, higher decision-making autonomy, higher husband’s education and urban residence. Women who reported that distance to a health facility was not a big problem had initiated ANC earlier than those who considered distance a major barrier. Regional disparities were also evident, with women from Barishal, Chattogram, Rajshahi, Khulna and Rangpur accessing ANC later than those in Dhaka.

Conclusions

Persistent inequalities in early ANC initiation highlight the need for targeted policies to reduce financial barriers, improve healthcare accessibility and strengthen awareness campaigns to ensure equitable maternal healthcare in Bangladesh.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• This study used nationally representative data from the 2022 Bangladesh Demographic and Health Survey, ensuring the generalisability of the findings to the broader population.

• The use of survival analysis, particularly the Accelerated Failure Time model, allowed for a more precise examination of the time to antenatal care (ANC) initiation than traditional regression methods.

• Data collection was conducted by trained personnel under the authority of National Institute of Population Research and Training, ensuring high data quality and consistency.

• The cross-sectional nature of the survey limits causal interpretations between the explanatory variables and early ANC initiation.

• Reliance on self-reported data introduces the potential for recall and social desirability biases, particularly regarding the timing of ANC visits.

Introduction

In 2020, an estimated 287 000 women died from pregnancy-related complications, with 95% of these deaths occurring in low-income and middle-income countries (LMICs).¹ Proper maternal healthcare utilisation plays a crucial role in reducing these deaths and promoting positive pregnancy outcomes.² Maternal health complications, including hypertension, gestational diabetes, preeclampsia, miscarriage, depression and infections, pose significant risks during pregnancy.³ Access to early antenatal care (ANC) is essential for mitigating these risks, yet disparities persist in healthcare utilisation.⁴ Moreover, the global burden of stillbirth remains high, with nearly 1.9 million cases in 2021, many of which could have been prevented through proper maternal healthcare utilisation.⁵ Therefore, strengthening ANC services is key to reducing maternal and neonatal mortality, improving birth outcomes and achieving global health targets.⁶

Globally, significant inequalities exist in ANC access and utilisation. Only 14% of women in LMICs receive comprehensive maternal healthcare services, including ANC, skilled birth attendance and postnatal care.⁷ Maternal mortality remains disproportionately high in LMICs, reaching 430 deaths per 100 000 live births compared with just 13 per 100 000 in high-income countries.¹ The WHO recommends at least eight ANC visits and first contact in the first 12 weeks of gestation for optimal care, yet many LMICs struggle to meet this standard.⁸ In Ethiopia, only 20% of pregnant women seek ANC in the first trimester,⁹ and in Uganda, 73% of women lack awareness of the appropriate gestational age to begin ANC.¹⁰ Socioeconomic factors significantly influence ANC initiation, with wealthier and more educated women more likely to seek care early.¹¹ A study in India found that early ANC initiation was higher among educated women, yet only 60.15% attended at least one ANC visit, with a median delay of 4 months.¹² Similarly, in Cameroon, higher parity was associated with delayed ANC attendance.¹³

Bangladesh faces similar maternal healthcare challenges. The maternal mortality ratio stands at 156 per 100 000 live births, while the neonatal mortality rate remains at 20 per 1000 live births.¹⁴ However, ANC utilisation has increased, with 82% of women accessing services in 2017/2018; only 34% completed four or more visits as recommended by national guidelines.¹⁵ Moreover, only 37% of women in Bangladesh initiate ANC in the first trimester, highlighting significant disparities in timely care access.¹⁶ Despite nearly all healthcare facilities offering ANC services, only 4% are adequately equipped to meet WHO quality standards, restricting access to comprehensive maternal care.¹⁷ Barriers such as poor service quality, financial constraints and infrastructure challenges continue to hinder ANC initiation.¹⁸ Additionally, socioeconomic factors play a crucial role in determining ANC initiation. Women’s education and household wealth consistently emerge as strong predictors of ANC utilisation, while urban residence, media exposure and decision-making autonomy also positively influence access to care.19,21 Community-level factors further contribute to disparities, with rural areas reporting significantly lower ANC utilisation rates.²² Moreover, geographical location, religious beliefs and awareness of pregnancy-related complications influence ANC-seeking behaviour.²³

While previous studies have explored ANC utilisation in Bangladesh, most have focused on overall service coverage, socioeconomic determinants or the impact of ANC on maternal and neonatal outcomes. A few recent studies have explored early ANC initiation using logistic regression or Cox proportional hazards models; however, these approaches may not fully capture the distributional characteristics of time-to-first ANC.24,26 Moreover, none have applied accelerated failure time (AFT) models to identify factors influencing the timing of ANC initiation. Also, most prior studies relied on older DHS rounds or regional samples, limiting their current national representativeness. The present study advances this body of work by (1) employing AFT models to identify factors associated with early ANC initiation, which directly quantify the acceleration or delay in timing and (2) using the most recent nationally representative data from the Bangladesh Demographic and Health Survey (BDHS) 2022. This approach provides a more comprehensive and temporally sensitive understanding of when women initiate ANC and the determinants influencing this timing. By generating updated, policy-relevant evidence, the study contributes to improving maternal healthcare access and supports Bangladesh’s progress toward Sustainable Development Goal 3, ensuring good health and well-being for all by 2030.

Methods

Data source and sampling technique

The study used data from the 2022 BDHS, which was conducted between August and December 2022. A two-stage stratified sample strategy was used in BDHS 2022, with the Bangladesh Bureau of Statistics establishing the sampling frame derived from the 2011 Population Census. Initially, 675 enumeration areas (EAs) comprising 237 urban and 438 rural were chosen using probability proportionate to size. The household listings inside these EAs became the foundation for the second stage, when 45 households per EA were carefully selected to guarantee dependable demographic and health data for both urban and rural populations. Among the chosen households, 30 were administered a comprehensive individual questionnaire, while 15 were provided with a shortened form, facilitating the collection of reproductive health data across Bangladesh. More details of the sampling methodology can be found in the final report of the BDHS 2022.²⁷ The survey included 30 330 households, of which 30 149 were approached, and 30 018 were successfully interviewed, achieving a household response rate of 99.6%. Of the ever-married women aged 15–49 identified in these households, 30 078 were interviewed, yielding a response rate of 99.1%. Women with no live births (3145), whose previous birth took place more than 5 years before the survey (21 574). Finally, the most recent birth of 5128 (weighted) ever-married women aged 15–49 during the 5 years before the survey has been included for further analysis in this study. The sampling procedures are visually represented in figure 1.

Figure 1. Schematic representation of the sampling procedures in the study of time to first ANC visit. ANC, antenatal care; BDHS, Bangladesh Demographic and Health Survey.

Study variables and measurement

The time to the first ANC visit, measured in months, was the primary outcome of interest. A woman who attended her first ANC visit within the first trimester (within 3 months of pregnancy) was classified as having early ANC initiation, whereas those who initiated ANC during the second or third trimester were classified as having late ANC initiation.

Operational definitions

• Event: If a woman attended at least one ANC visit during the pregnancy (coded as ‘1’). If she did not attend, the event was coded as ‘0’.

• Censored: If a woman attended an ANC visit after 3 months, with gestational age recorded during delivery.

• Survival time: The duration (in months) from the start of pregnancy to the first ANC visit (gestational age at the first ANC visit).²⁶

Furthermore, this study considered various explanatory variables to determine predictors of early ANC initiation, as shown in table 1. These variables were selected based on theoretical relevance and evidence from previous studies highlighting their potential influence on maternal healthcare utilisation and the timing of ANC initiation.^{17 26}

Table 1. The list of explanatory variables of the study.

Variable	Descriptions
Respondent’s age	The respondents’ ages were categorised as 15–24, 25–34 and 35–49.53
Husband’s age	The ages of husbands were classified as 15–24, 25–34, 35–44 and 45+.53
Respondent’s education	Women’s level of education was classified as no education, primary, secondary or higher.
Husband’s education	Husband’s education was categorised as no education, primary, secondary or higher.
Respondent’s occupation	Respondent’s occupation was categorised as housewife, agricultural, non-agricultural and skilled worker and others.
Husband’s occupation	The husband’s occupation was categorised as jobless, agricultural, non-agricultural and skilled worker and others.
Religion	Religion was classified as Islam or others.
Wealth index	The wealth index was categorised as poorest, poorer, middle, richer and richest.
Media exposure	Media exposure was categorised as no or yes.
Sex of household head	The household head’s sex was categorised as male or female.
Birth order	Birth order was classified as first, second, or third and above.
Household size	Household size was classified as less than or equal to four and more than 4.54
Last pregnancy intention	Pregnancy intention was recorded as no or yes for the last pregnancy.
Distance to the health facility	Distance to the health facility was categorised as a big problem or not a big problem.
Women’s autonomy	Women’s autonomy in major household decision-making was classified as low, medium or high.55
Place of residence	The place of residence was classified as urban or rural.
Division	The division was categorised as Barishal, Chattogram, Dhaka, Khulna, Mymensingh, Rajshahi, Rangpur or Sylhet.

Statistical analysis

Data cleaning, recoding and analysis were conducted using Stata V.17.0 (StataCorp) and R V.4.3.0. The sample weights were applied to maintain the survey’s representativeness, according to the DHS-8 analysis guidelines.²⁸ The analysis adhered to the Strengthening the Reporting of Observational Studies in Epidemiology cross-sectional reporting guidelines.²⁹

Exploratory data analysis was performed with a 95% CI to examine the dataset’s characteristics, which involved reporting frequencies and percentages for the study variables. The Kaplan-Meier survival curve was constructed to estimate the time to the first ANC booking. The log-rank test was utilised to assess significant variations in the timing of ANC intention across diverse socioeconomic and demographic attributes. A notable log-rank test outcome signifies statistically substantial disparities in survival among the groups. Subsequent to the log-rank test, the Cox proportional hazards (PH) model was initially used to identify the determinants influencing early ANC initiation. The PH assumption was checked by the analysis of Schoenfeld residuals to confirm the model’s validity. Alongside the Cox model, an AFT model was fitted using Weibull, log-logistic, exponential, lognormal and Gompertz distributions. The AFT model assumes that the effect of covariates accelerates or decelerates the life course of an event by a constant factor. This characteristic is particularly relevant for understanding how various socioeconomic and demographic factors may impact a woman’s intention to seek early ANC. Mathematically, it can be expressed as:

$$\begin{array}{}\text{(1)}& {\displaystyle {\displaystyle \log (T)={\beta }_0+{\beta }_1{X}_1+{\beta }_2{X}_2+\cdots +\epsilon }}\end{array}$$

where T represents the survival time (time to the first ANC visit), β is the coefficients, X are the covariates, and ϵ is the error term.

The Akaike information criterion (AIC) and the Bayesian information criterion (BIC) were calculated to determine the best-fitting model. The model with the lowest AIC and BIC values was selected for the final analysis. Multicollinearity among the covariates was examined using the variance inflation factor (VIF), with the highest VIF recorded for the respondent’s age being 2.28 and a mean VIF of 1.37. These values indicated no significant multicollinearity issues. Significant variables from the log-rank test were included in the final model. In the final multivariable AFT model, exponentiated coefficients ($e^{\mathrm{\beta }}$) representing time ratios as adjusted acceleration factor (AAF) with 95% CIs were used to present the associations with early intention. Variables with a p≤0.05 were considered statistically significant.

Ethical approval

This research used a secondary dataset obtained from the National Institute of Population Research and Training, Bangladesh, and the Monitoring and Evaluation to Assess and Use Results Demographic and Health Surveys. We obtained authorisation from ‘The DHS programme’ to access and use the dataset for this investigation. All BDHS participants provided written informed consent before participation, and all information was collected confidentially. Participants gave informed consent to participate in the survey before taking part.

Patient and public involvement

None.

Results

Background characteristics

Among 5128 reproductive-aged women (15–49 years) in Bangladesh who received ANC services and had a live birth within the 5 years prior to the survey, 47% were aged 15–24, and over a half (54.3%) had secondary education. The majority (73.4%) were housewives, and 52.3% of their husbands were employed in non-agricultural jobs. Furthermore, 92.5% of the women followed Islam, while 38.2% belonged to the richer or richest wealth categories. About 43.2% had media access, and 62.6% reported high autonomy in household decision-making. Approximately 38 in 100 were first-time mothers, 80.8% wanted their last pregnancy, and 67% lived in households with more than four members. Distance to health facilities was a major issue for 45.2%. Most respondents (73.2%) resided in rural areas, with 24.5% from the Dhaka division (table 2).

Table 2. Distribution of study variables among reproductive-aged women in Bangladesh who had given birth during the 5 years before the survey (n=5128).

Variables	Total		Early ANC
	Weighted frequency	Weighted percentage	% (95% CI)
Overall	5128	100.0	37.9 (36.0 to 39.9)
Age of women
15–24	2409	47	35.3 (32.9 to 37.8)
25–34	1411	27.5	40.2 (37.0 to 43.5)
35–49	1308	25.5	40.3 (36.9 to 43.9)
Age of husband
15–24	501	9.8	26.4 (22.3 to 31.0)
25–34	2346	45.7	38.4 (35.7 to 41.1)
35–44	1797	35	41 (38.0 to 44.0)
45+	432	8.4	38.1 (33.0 to 43.4)
Missing	52	1
Women’s education
No education	275	5.4	20.2 (14.6 to 27.1)
Primary	1160	22.6	25.7 (22.8 to 28.9)
Secondary	2786	54.3	37.4 (35.3 to 39.6)
Higher	907	17.7	60.6 (56.6 to 64.4)
Husband’s education
No Education	764	14.9	22.1 (18.8 to 25.7)
Primary	1510	29.4	28.1 (25.5 to 30.9)
Secondary	1796	35	40.5 (37.8 to 43.4)
Higher	996	19.4	60.8 (56.9 to 64.5)
Missing	63	1.2
Women’s occupation
Housewife	3764	73.4	38.4 (36.3 to 40.6)
Agricultural	853	16.6	29.7 (26.2 to 33.3)
Non-agricultural	359	7	39.3 (33.2 to 45.8)
Skilled worker and others	152	3	69.2 (60.1 to 77.0)
Husband’s occupation
Jobless	157	3.1	41.8 (33.5 to 50.6)
Agricultural	970	18.9	23.5 (20.5 to 26.9)
Non-agricultural	2684	52.3	36.2 (34.1 to 38.4)
Skilled worker and others	1311	25.6	51.6 (48.0 to 55.2)
Missing	6	0.1
Religion
Islam	4743	92.5	37.7 (35.6 to 39.8)
Others	385	7.5	41.2 (35.8 to 46.7)
Wealth index
Poorest	1048	20.4	22 (19.3 to 24.9)
Poorer	1069	20.8	26.3 (23.2 to 29.6)
Middle	1057	20.6	33.7 (30.4 to 37.2)
Richer	1018	19.9	45 (41.2 to 48.8)
Richest	936	18.3	66.2 (61.8 to 70.3)
Mass media exposure
No	2913	56.8	31.9 [29.7,34.2]
Yes	2215	43.2	45.9 [43.2,48.6]
Sex of household head
Male	4516	88.1	37.5 (35.5 to 39.5)
Female	612	11.9	41.5 (37.1 to 46.0)
Birth order
1st	1939	37.8	41.5 (38.5 to 44.5)
2nd	1744	34	38.8 (35.9 to 41.8)
3 or more	1445	28.2	32.1 (29.3 to 35.1)
Number of household members
Less than or equal to 4	1691	33	39.9 (36.9, 43.1)
4+	3437	67	37 (34.7, 39.2)
Wanted last pregnancy
No	983	19.2	33.8 (30.5, 37.3)
Yes	4145	80.8	38.9 (36.7, 41.1)
Distance to the health facility
Big problem	2317	45.2	33.5 (31.1, 36.0)
Not a big problem	2811	54.8	41.6 (39.1, 44.1)
Women’s autonomy in household decision making
Low	781	15.2	27.6 (24.1 to 31.4)
Medium	1138	22.2	36.7 (33.3 to 40.3)
High	3209	62.6	40.9 (38.7 to 43.2)
Place of residence
Urban	1373	26.8	50.9 (46.5 to 55.3)
Rural	3755	73.2	33.2 (31.2 to 35.2)
Division
Dhaka	1257	24.5	48.6 (43.4 to 53.9)
Barishal	314	6.1	32.3 (28.4 to 36.5)
Chattogram	1132	22.1	33.9 (29.8 to 38.2)
Khulna	530	10.3	37.6 (33.3 to 42.0)
Mymensingh	446	8.7	34.7 (30.0 to 39.7)
Rajshahi	541	10.6	36.8 (31.8 to 42.1)
Rangpur	576	11.2	28.5 (25.0 to 32.4)
Sylhet	330	6.4	39.7 (34.0 to 45.7)

ANC, antenatal care.

Distribution of early ANC initiation across study variables

Table 2 also presents the bivariate analysis of study variables and early ANC initiation. The log-rank test was used to assess the statistical significance of differences in early ANC initiation across various groups. The results indicate that 37.9% of women received their first ANC visit within the first 3 months of pregnancy.

The analysis shows notable differences in early ANC initiation across sociodemographic groups. Women with higher education (60.6%) and those whose spouses had higher education (60.8%) had higher early ANC initiation rate compared with those with lower educational attainment. Early ANC initiation also varied across wealth quintiles, with women from the richest households showing substantially higher uptake than those from poorer or poorest households. Media exposure was associated with increased early ANC initiation, as women with access to media tended to seek care earlier. Urban women and those reporting no difficulty with distance to health facilities also demonstrated higher rates of early ANC initiation. Greater autonomy in household decision-making corresponded with higher early ANC initiation. Regional differences were evident, with the highest early ANC initiation observed in Dhaka and the lowest in Rangpur.

Kaplan-Meier survival scenario

Figure 2 shows the Kaplan-Meier survival plot for overall ANC uptake and demonstrates the timing of initial ANC visits among women throughout pregnancy. The curve shows a slow decline initially, indicating that few women receive ANC in the first month. The most rapid decline occurs between 2 and 4 months, with around half of the women initiating ANC by around 4 months. As the curve continues, most women (90%) have started ANC by 7 months. However, the survival probability does not approach zero until after 8 months, indicating that some women delay ANC visits much later into pregnancy.

Figure 2. The Kaplan-Meier survival plot for overall ANC uptake. ANC, antenatal care.

Table 3 presents Kaplan-Meier survivor probabilities at 3 months across sociodemographic subgroups. Since the survivor probability reflects the proportion who have not yet initiated ANC by 3 months, lower values indicate earlier initiation, while higher values indicate delays. Early ANC initiation was more common among women with higher education (S(3)=0.38), from the richest households (S(3)=0.33) and those residing in urban areas (S(3)=0.46). Conversely, delayed ANC initiation was observed among women with no schooling (S(3)=0.73), women in the poorest wealth quintile (S(3)=0.73), and rural residents (S(3)=0.64). Significant differences were also seen across husbands’ education, household wealth, occupation, decision-making autonomy and perceptions of distance to health facilities.

Table 3. Kaplan-Meier survivor probabilities at 3 months by background characteristics among reproductive-aged women in Bangladesh who gave birth in the 5 years before the survey (n=5128).

Variables	Survivor probability at 3 months, S(3) (95% CI)	P value
Overall	0.59 (0.57 to 0.60)
Age of women		<0.001
15–24	0.62 (0.60 to 0.64)
25–34	0.56 (0.53 to 0.59)
35–49	0.55 (0.52 to 0.58)
Age of husband		<0.001
15–24	0.71 (0.67 to 0.75)
25–34	0.59 (0.57 to 0.61)
35–44	0.55 (0.53 to 0.58)
45+	0.55 (0.50 to 0.60)
Women’s education		<0.001
No education	0.73 (0.66 to 0.78)
Primary	0.70 (0.67 to 0.73)
Secondary	0.60 (0.58 to 0.62)
Higher	0.38 (0.35 to 0.41)
Husband’s education		<0.001
No education	0.73 (0.70 to 0.77)
Primary	0.69 (0.66 to 0.71)
Secondary	0.57 (0.55 to 0.59)
Higher	0.38 (0.35 to 0.41)
Women’s occupation		<0.001
Housewife	0.58 (0.57 to 0.60)
Agricultural	0.67 (0.64 to 0.70)
Non-agricultural	0.58 (0.52 to 0.63)
Skilled worker and others	0.27 (0.20 to 0.34)
Husband’s occupation		<0.001
Jobless	0.54 (0.45 to 0.62)
Agricultural	0.72 (0.68 to 0.75)
Non-agricultural	0.61 (0.59 to 0.63)
Skilled worker and others	0.46 (0.43 to 0.49)
Religion		0.418
Islam	0.59 (0.57 to 0.60)
Others	0.56 (0.51 to 0.61)
Wealth index		<0.001
Poorest	0.73 (0.70 to 0.76)
Poorer	0.71 (0.68 to 0.73)
Middle	0.64 (0.61 to 0.67)
Richer	0.54 (0.50 to 0.56)
Richest	0.33 (0.30 to 0.36)
Mass media exposure		<0.001
No	0.64 (0.63 to 0.66)
Yes	0.52 (0.50 to 0.54)
Sex of household head		0.11
Male	0.59 (0.58 to 0.61)
Female	0.56 (0.52 to 0.60)
Birth order		0.001
1st	0.57 (0.54 to 0.59)
2nd	0.58 (0.56 to 0.61)
3 or more	0.62 (0.60 to 0.65)
Number of household members		0.172
1–4	0.57 (0.55 to 0.60)
4+	0.60 (0.58 to 0.61)
Wanted last pregnancy		0.034
No	0.62 (0.58 to 0.65)
Yes	0.58 (0.57 to 0.60)
Distance to the health facility		<0.001
Big problem	0.63 (0.61 to 0.65)
Not a big problem	0.55 (0.53 to 0.57)
Women’s autonomy in decision-making		<0.001
Low	0.69 (0.65 to 0.72)
Medium	0.60 (0.57 to 0.63)
High	0.56 (0.54 to 0.58)
Place of residence		<0.001
Urban	0.46 (0.44 to 0.49)
Rural	0.64 (0.62 to 0.65)
Division		<0.001
Barishal	0.65 (0.59 to 0.70)
Chattogram	0.63 (0.60 to 0.66)
Dhaka	0.48 (0.46 to 0.51)
Khulna	0.61 (0.56 to 0.65)
Mymensingh	0.60 (0.55 to 0.65)
Rajshahi	0.61 (0.56 to 0.65)
Rangpur	0.69 (0.65 to 0.72)
Sylhet	0.54 (0.48 to 0.59)

Figure 3 further illustrates these disparities through group-wise Kaplan-Meier curves. Women with higher education and those whose husbands have higher education show faster declines in survival probability, indicating earlier ANC initiation. Economic gradients are clear, with wealthier women seeking care earlier than the poorest. Women exposed to mass media initiate ANC earlier than those with no exposure, highlighting the role of information and awareness. Distance to health facilities also shapes timing; women who report distance as ‘not a big problem’ seek ANC earlier than those who perceive it as a major barrier. A notable urban–rural divide is visible, with urban women initiating ANC earlier throughout pregnancy.

Figure 3. The Kaplan-Meier survival plot for early ANC initiation by different socioeconomic and demographic variables. ANC, antenatal care.

Model selection

The study initially applied the Cox proportional hazards model to identify factors influencing early ANC initiation among women in Bangladesh. The proportional hazards assumption was violated (p<0.01), indicating that the Cox model was not suitable for these data; global tests based on Schoenfeld residuals are presented in online supplemental materials figure S1, and the corresponding Cox model estimates are reported in online supplemental materials table S1. For this reason, parametric AFT models were then considered as an alternative framework. Five AFT specifications with different lifetime distributions (exponential, Weibull, log-logistic, lognormal and Gompertz) were fitted, and their fit statistics (AIC and BIC) are presented in table 4. All five models are shown for transparency; however, based on the lowest AIC and BIC values, the lognormal AFT model was identified as the best-fitting model for assessing factors associated with early ANC initiation in Bangladesh.

Table 4. Factors associated with early initiation of ANC services among reproductive-aged women in Bangladesh based on the AFT Models (n=5128).

	Lognormal		Exponential		Weibull		Log-logistic		Gompertz
Covariates	AAF (95% CI)	P value	AAF (95% CI)	P value	AAF (95% CI)	P value	AAF (95% CI)	P value	AAF (95% CI)	P value
Age of women
15–24 (ref)	—	—	—	—	—	—	—	—	—	—
25–34	0.93 (0.85 to 0.99)	0.044	0.93 (0.81 to 1.06)	0.27	0.94 (0.86 to 1.04)	0.248	0.94 (0.85 to 1.03)	0.195	0.94 (0.87 to 1.02)	0.12
35–49	0.90 (0.80 to 1.02)	0.101	0.91 (0.77 to 1.07)	0.246	0.92 (0.81 to 1.04)	0.174	0.91 (0.80 to 1.03)	0.125	0.88 (0.79 to 0.98)	0.021
Age of husband
15–24 (ref)	—	—	—	—	—	—	—	—	—	—
25–34	0.81 (0.71 to 0.92)	0.001	0.74 (0.61 to 0.90)	0.002	0.79 (0.69 to 0.91)	0.001	0.79 (0.69 to 0.91)	0.001	0.86 (0.78 to 0.96)	0.005
35–44	0.71 (0.61 to 0.82)	<0.001	0.64 (0.51 to 0.79)	<0.001	0.71 (0.60 to 0.83)	<0.001	0.70 (0.60 to 0.82)	<0.001	0.74 (0.66 to 0.84)	<0.001
45+	0.67 (0.56 to 0.81)	<0.001	0.59 (0.45 to 0.77)	<0.001	0.66 (0.54 to 0.81)	<0.001	0.65 (0.53 to 0.80)	<0.001	0.78 (0.67 to 0.92)	0.004
Women’s education
No education (ref)	—	—	—	—	—	—	—	—	—	—
Primary	0.96 (0.79 to 1.16)	0.675	0.95 (0.71 to 1.28)	0.743	0.95 (0.77 to 1.19)	0.67	0.96 (0.78 to 1.19)	0.729	0.94 (0.80 to 1.10)	0.452
Secondary	0.83 (0.68 to 0.99)	0.048	0.75 (0.56 to 1.00)	0.047	0.78 (0.63 to 0.96)	0.02	0.81 (0.66 to 0.99)	0.042	0.90 (0.77 to 1.05)	0.191
Higher	0.71 (0.58 to 0.88)	0.002	0.63 (0.46 to 0.86)	0.003	0.67 (0.53 to 0.84)	0.001	0.70 (0.56 to 0.87)	0.002	0.80 (0.66 to 0.95)	0.014
Husband’s education
No education (ref)	—	—	—	—	—	—	—	—	—	—
Primary	0.86 (0.76 to 0.97)	0.015	0.83 (0.69 to 1.00)	0.053	0.87 (0.76 to 1.00)	0.048	0.85 (0.75 to 0.97)	0.015	0.90 (0.81 to 0.99)	0.033
Secondary	0.77 (0.68 to 0.87)	<0.001	0.68 (0.57 to 0.82)	<0.001	0.74 (0.65 to 0.85)	<0.001	0.74 (0.65 to 0.85)	<0.001	0.86 (0.78 to 0.96)	0.004
Higher	0.65 (0.56 to 0.75)	<0.001	0.54 (0.44 to 0.66)	<0.001	0.61 (0.52 to 0.71)	<0.001	0.62 (0.54 to 0.72)	<0.001	0.80 (0.71 to 0.91)	<0.001
Women’s occupation
Housewife (ref)	—	—	—	—	—	—	—	—	—	—
Agricultural	1.01 (0.92 to 1.12)	0.776	1.01 (0.87 to 1.16)	0.906	1.00 (0.90 to 1.11)	0.973	1.01 (0.91 to 1.12)	0.843	1.05 (0.96 to 1.14)	0.272
Non-agricultural	1.03 (0.90 to 1.17)	0.696	1.01 (0.85 to 1.21)	0.88	1.01 (0.89 to 1.16)	0.849	1.01 (0.88 to 1.15)	0.907	1.11 (0.99 to 1.24)	0.076
Skilled worker and others	0.81 (0.68 to 0.97)	0.024	0.72 (0.58 to 0.89)	0.002	0.75 (0.64 to 0.88)	<0.001	0.82 (0.69 to 0.97)	0.019	1.05 (0.88 to 1.24)	0.59
Wealth index
Poorest (ref)	—	—	—	—	—	—	—	—	—	—
Poorer	1.02 (0.91 to 1.15)	0.729	1.01 (0.84 to 1.20)	0.945	1.01 (0.88 to 1.15)	0.904	1.01 (0.89 to 1.15)	0.863	1.01 (0.92 to 1.12)	0.759
Middle	0.95 (0.84 to 1.07)	0.388	0.92 (0.77 to 1.10)	0.384	0.95 (0.83 to 1.08)	0.441	0.92 (0.81 to 1.05)	0.215	0.99 (0.90 to 1.09)	0.826
Richer	0.80 (0.71 to 0.90)	<0.001	0.73 (0.61 to 0.88)	0.001	0.79 (0.69 to 0.90)	<0.001	0.77 (0.68 to 0.88)	<0.001	0.85 (0.77 to 0.94)	0.002
Richest	0.63 (0.55 to 0.72)	<0.001	0.51 (0.42 to 0.62)	<0.001	0.58 (0.50 to 0.66)	<0.001	0.61 (0.53 to 0.70)	<0.001	0.79 (0.70 to 0.89)	<0.001
Media exposure
No (ref)	—	—	—	—	—	—	—	—	—	—
Yes	0.91 (0.85 to 0.98)	0.010	0.89 (0.81 to 0.98)	0.019	0.91 (0.85 to 0.98)	0.013	0.90 (0.84 to 0.97)	0.005	0.96 (0.91 to 1.03)	0.242
Sex of household head
Male (ref)	—	—	—	—	—	—	—	—	—	—
Female	0.94 (0.85 to 1.04)	0.249	0.95 (0.83 to 1.09)	0.457	0.97 (0.87 to 1.07)	0.487	0.94 (0.85 to 1.05)	0.269	0.91 (0.84 to 1.00)	0.039
Birth order
1st (ref)	—	—	—	—	—	—	—	—	—	—
2nd	1.20 (1.10 to 1.32)	<0.001	1.25 (1.11 to 1.42)	<0.001	1.20 (1.10 to 1.32)	<0.001	1.20 (1.09 to 1.31)	<0.001	1.20 (1.11 to 1.30)	<0.001
3 or more	1.28 (1.13 to 1.45)	<0.001	1.28 (1.08 to 1.52)	0.004	1.22 (1.07 to 1.38)	0.002	1.27 (1.12 to 1.44)	<0.001	1.32 (1.19 to 1.46)	<0.001
Intended pregnancy
No (ref)	—	—	—	—	—	—	—	—	—	—
Yes	1.03 (0.95 to 1.13)	0.455	1.05 (0.93 to 1.19)	0.401	1.05 (0.95 to 1.15)	0.342	1.03 (0.94 to 1.13)	0.466	1.03 (0.96 to 1.11)	0.408
Distance to health facility
Big problem (ref)	—	—	—	—	—	—	—	—	—	—
Not a big problem	0.93 (0.87 to 0.99)	0.034	0.92 (0.84 to 1.01)	0.066	0.94 (0.88 to 1.01)	0.072	0.93 (0.87 to 1.00)	0.044	0.95 (0.90 to 1.01)	0.101
Women’s decision-making autonomy
Low (ref)	—	—	—	—	—	—	—	—	—	—
Medium	0.82 (0.73 to 0.92)	0.001	0.75 (0.63 to 0.89)	0.001	0.80 (0.71 to 0.91)	<0.001	0.80 (0.71 to 0.91)	<0.001	0.89 (0.80 to 0.98)	0.014
High	0.74 (0.67 to 0.82)	<0.001	0.65 (0.56 to 0.76)	<0.001	0.71 (0.64 to 0.79)	<0.001	0.72 (0.65 to 0.80)	<0.001	0.82 (0.76 to 0.89)	<0.001
Place of residence
Urban (ref)	—	—	—	—	—	—	—	—	—	—
Rural	1.09 (1.01 to 1.19)	0.030	1.16 (1.04 to 1.29)	0.008	1.14 (1.05 to 1.23)	0.001	1.10 (1.02 to 1.20)	0.018	1.00 (0.93 to 1.07)	0.991
Division
Dhaka (ref)	—	—	—	—	—	—	—	—	—	—
Barishal	1.20 (1.03 to 1.40)	0.021	1.31 (1.05 to 1.63)	0.017	1.25 (1.06 to 1.47)	0.007	1.21 (1.03 to 1.42)	0.023	1.10 (0.96 to 1.25)	0.158
Chattogram	1.20 (1.09 to 1.33)	<0.001	1.29 (1.13 to 1.48)	<0.001	1.23 (1.12 to 1.36)	<0.001	1.21 (1.09 to 1.34)	<0.001	1.16 (1.06 to 1.26)	0.001
Khulna	1.23 (1.09 to 1.39)	0.001	1.34 (1.13 to 1.58)	0.001	1.27 (1.12 to 1.43)	<0.001	1.24 (1.09 to 1.40)	0.001	1.14 (1.03 to 1.27)	0.016
Mymensingh	1.06 (0.92 to 1.21)	0.406	1.11 (0.92 to 1.33)	0.287	1.10 (0.96 to 1.26)	0.167	1.05 (0.91 to 1.21)	0.478	1.11 (0.99 to 1.25)	0.073
Rajshahi	1.20 (1.06 to 1.35)	0.005	1.28 (1.09 to 1.52)	0.004	1.23 (1.09 to 1.39)	0.001	1.21 (1.07 to 1.38)	0.003	1.08 (0.97 to 1.21)	0.145
Rangpur	1.36 (1.20 to 1.55)	<0.001	1.52 (1.27 to 1.83)	<0.001	1.41 (1.23 to 1.61)	<0.001	1.38 (1.20 to 1.58)	<0.001	1.30 (1.16 to 1.45)	<0.001
Sylhet	0.93 (0.80 to 1.08)	0.320	0.89 (0.73 to 1.08)	0.244	0.91 (0.79 to 1.06)	0.229	0.91 (0.79 to 1.06)	0.249	0.99 (0.87 to 1.13)	0.899
AIC	9389.4		9973.1		9897.4		9614.3		9920.2
BIC	9402.3		9979.6		9910.3		9627.2		9933.1

AAF, adjusted acceleration factor; AFT, accelerated failure time; AIC, Akaike information criterion; ANC, antenatal care; BIC, Bayesian information criterion; CI, Confidence interval.

Determinants of early initiation of ANC

Table 4 presents the estimated AAF and 95% CIs for each covariate under all fitted AFT models, while the narrative interpretation focuses exclusively on the lognormal AFT model, which provided the best fit. The study found that women aged 25 to 34 have a 7% reduction in time to receive early ANC (AAF=0.93; 95% CI 0.85 to 0.99) compared with women aged 15–24. Women with secondary education experienced a 17% shorter time to initiate early ANC compared with illiterate women (AAF=0.83; 95% CI 0.68 to 0.99). Similarly, those with higher education showed a 29% reduction in time to receive early ANC (AAF=0.71; 95% CI 0.58 to 0.88). Women in skilled work tended to take 19% (AAF=0.81; 95% CI 0.68 to 0.97) less time to receive early ANC than the women who were housewives. In the case of the husband’s profile, the husband’s age and educational qualification were found to be significant factors. Women whose husbands were aged 25–34 were likely to receive ANC service 19% (AAF=0.81; 95% CI 0.71 to 0.92) faster compared with those whose husbands were aged 15–24. The time to receive early ANC was even reduced for women whose husbands were aged 35–44, with these women accessing services 29% (AAF=0.71; 95% CI 0.61 to 0.82) faster. Women with husbands aged more than 45 years had 33% (AAF=0.67; 95% CI 0.56 to 0.81) reduction in time to receive early ANC than women with husbands aged 15–24. Additionally, women whose husbands had completed secondary and higher education took 23% (AAF=0.77; 95% CI 0.68 to 0.87) and 35% (AAF=0.65; 95% CI 0.56 to 0.75) less time, respectively, to receive early ANC compared with women whose husbands had no formal education.

Regarding the wealth index, women from richer and richest households tend to take 20% (AAF=0.80; 95% CI 0.71 to 0.90) and 37% (AAF=0.63; 95% CI 0.55 to 0.72) less time, respectively, to receive early ANC than women from the poorest household. The time to early ANC visit was 9% (AAF=0.91; 95% CI 0.85 to 0.98) less for the women who had media access compared with women who had no media exposure. Additionally, women with their second child were 1.20 (AAF=1.20; 95% CI 1.10 to 1.32) times more likely to delay their first ANC visit compared with first-time mothers. In fact, women with three or more children were 1.28 (AAF=1.28; 95% CI 1.13 to 1.45) times more likely to delay receiving early ANC.

The time to receive early ANC was 7% (AAF=0.93; 95% CI 0.87 to 0.99) less for those women whose distance from home to health facility was not a big problem than for women who claimed that distance to health facility was a big problem for them. Additionally, women with medium and high autonomy in household decision-making access early ANC 18% (AAF=0.82; 95% CI 0.73 to 0.92) and 26% (AAF=0.74; 95% CI 0.67 to 0.82) faster, respectively, than women with low autonomy. Furthermore, women who resided in rural areas took 9% (AAF=1.09; 95% CI 1.01 to 1.19) more time to receive early ANC than women from urban areas. Lastly, women from Barishal, Chattogram, Rajshahi, Khulna and Rangpur division were 20% (AAF=1.20; 95% CI 1.03 to 1.40), 20% (AAF=1.20; 95% CI 1.09 to 1.33), 20% (AAF=1.20; 95% CI 1.06 to 1.35), 23% (AAF=1.23; 95% CI 1.09 to 1.39) and 36% (AAF=1.36; 95% CI 1.20 to 1.55) more likely to delay receiving early ANC, respectively, than women from Dhaka division.

Discussion

This research was designed to analyse the factors associated with early initiation of ANC among pregnant mothers in Bangladesh, employing a survival model as well as using the most recent nationally representative data. There has been an inadequate emphasis on the early initiation of ANC in Bangladesh, despite the growth in its coverage. However, various individual, household-level, socioeconomic, demographic and contextual factors significantly influence early initiation and adequate ANC coverage for ensuring the health and well-being of mothers and newborns.^{24 30 31}

The age of the mother significantly affects early ANC initiation, with middle-aged mothers (25–34 years) being more proactive in seeking early ANC than younger mothers (15–24 years). This finding aligns with previous research suggesting that older women may have greater awareness and experience regarding the importance of timely ANC visits.^{32 33} In general, middle-aged women are more likely to have had prior pregnancies, as increasing age might enhance their knowledge about potential pregnancy complications and the need for early care.^{34 35} Conversely, younger women may face barriers such as limited autonomy, lack of knowledge or hesitation to access healthcare services, as noted in other studies.^{36 37} However, cultural norms and societal expectations for younger mothers in some regions may also contribute to delayed ANC initiation, emphasising the need for targeted interventions to address these barriers.

The study found that women engaged in skilled work were more likely to access early ANC compared with housewives. This result is supported by literature indicating that employed women often have better financial resources, enabling them to prioritise health services, including early initiation of ANC.^{11 38} Based on the opposite view of early ANC initiation, a study found there is a significant association between maternal occupation and delayed initiation of ANC.³⁹ Working women have more income and education to make informed healthcare decisions, so they may seek more ANC services.⁴⁰ Employment may also increase opportunities for women to learn about ANC from healthcare providers. Due to their financial independence and ability to take medical leave, employed women may have more control over their health and pregnancy decisions.^{41 42} On the contrary, housewives, especially in patriarchal societies, might depend on their spouses or family members for financial support and decision-making, which could delay access to healthcare.⁴²

Furthermore, this research also revealed the results of previous studies that women with high autonomy in the household’s decision-making process significantly accelerate the early initiation of ANC service compared to women with low autonomy. This is consistent with research indicating that women who have control over financial and healthcare decisions are better able to prioritise their health needs.⁴³ Empowered women are also more likely to seek health information and make informed choices.⁴⁴ In contrast, women with limited autonomy may depend on their spouses or family members, who might not prioritise ANC.⁴⁵ Efforts to enhance women’s empowerment, such as education and economic independence, are crucial to improving maternal health outcomes.

This research revealed that the birth order of the child is also a significant predictor of receiving early ANC services in Bangladesh. One of the most important results is that women are more likely to initiate ANC services early during their first birth than the rest.⁴⁶ This might have happened because the mothers who did not face major complications during the first birth perceived a lower need for early initiation of ANC in subsequent pregnancies.

In the case of the husband’s profile, such as the husband’s age, educational qualification was found to be a significant factor. Women with older or more educated husbands were more likely to seek early ANC services. This finding is similar to a recent study in Bangladesh, which found that the women who had middle-aged husbands were more likely to receive early ANC service as compared with the younger husbands.⁴⁷ Also, a few previous studies found that women whose partners had completed secondary and higher education took significantly less time to receive early ANC compared with women whose husbands had no formal education.^{24 30} Educated husbands are more likely to understand the importance of early ANC and encourage their spouses to seek timely care.⁴⁸ Additionally, older husbands may have more life experience, facilitating access to healthcare services.

In accordance with the previous studies, the present results demonstrated that mass media exposure is one of the significant factors in accelerating the early initiation of ANC service in Bangladesh.^{49 50} Since the power of media is increasing day by day, the advertisements regarding maternal healthcare services and proper guidelines at different stages of pregnancy may help pregnant mothers be well-informed about suitable maternal health services. Media campaigns can reach a wide audience and influence health-seeking behaviour by addressing misconceptions and promoting positive practices. Conversely, women without media exposure may rely on traditional or informal sources of information, which might not emphasise the importance of early ANC.

The study revealed that women from wealthier households were more likely to access early ANC compared with those from poorer households. This finding aligns with the well-documented association between socioeconomic status and healthcare access.^{31 33 49} Wealthier households can afford transportation, healthcare costs and other related expenses, which are often barriers for poorer women.³³ Additionally, wealthier women may have greater exposure to health information and live closer to healthcare facilities.^{36 51} On the other hand, women from lower socioeconomic backgrounds often face compounded barriers, such as a lack of awareness, financial constraints and inadequate health infrastructure, highlighting the need for policies that provide financial and logistical support for disadvantaged groups.³⁶

Women who perceived distance to healthcare facilities as a minor issue were more likely to access early ANC compared with those for whom distance was a major barrier. This finding is supported by studies showing that proximity to healthcare facilities significantly influences service utilisation.³¹ Long travel distances and associated costs can deter women from seeking timely care, especially in rural areas where transportation infrastructure may be inadequate. Also, the public hospitals in Bangladesh provide health services at minimal cost, but the distance to the health facilities may create an economic as well as physical burden for them to seek maternal health services. Improving healthcare accessibility through decentralised services and community-based programmes can mitigate these barriers and promote timely ANC utilisation.

The study also found that rural women were less likely to access early ANC compared with their urban counterparts. This disparity aligns with existing literature highlighting urban-rural differences in healthcare access in LMICs.^{11 50 52} Urban areas typically have better healthcare infrastructure, greater availability of skilled providers and higher awareness of health services. In contrast, rural women often face challenges such as distance to healthcare facilities, lack of transportation and limited healthcare services, as discussed earlier.³⁷

These urban–rural differences are also evident across administrative divisions. Aligning with the regional disparities among early initiation of ANC in Papua New Guinea, this study found that the women from Barishal, Chattogram, Rajshahi, Khulna and Rangpur divisions were more likely to delay initiation of ANC compared with those from the Dhaka division.³⁸ Women in Dhaka may benefit from better health facilities and targeted interventions, while those in other divisions face systemic barriers. Tailored regional strategies, such as mobile clinics and community outreach programmes, are essential to address these inequities and promote timely ANC.

Limitations

The study has several limitations. Its cross-sectional design prevents causal inferences, and reliance on self-reported data introduces recall bias. Additionally, the proportionality assumption of the Cox proportional hazards model is difficult to assess robustly over the relatively short follow-up period (3 months). An alternative approach would be interval-censored time-to-event modelling using complementary log-log regression, which may better accommodate this structure. Future studies should adopt this framework to validate and refine our findings. Finally, while the study identifies key socioeconomic and geographic factors, it does not deeply explore health system constraints, such as healthcare provider attitudes and ANC service quality, that are critical to improving maternal healthcare.

Conclusions

This study highlights a critical gap in maternal healthcare utilisation in Bangladesh, revealing that only 4 out of 10 Bangladeshi women initiate ANC within the first trimester. Key determinants of early ANC initiation include maternal age, wealth status, husband’s education, women’s employment, media exposure and household decision-making autonomy. Women from wealthier households, urban areas and those with greater autonomy are more likely to seek early ANC, whereas regional disparities persist, with women in Dhaka initiating ANC earlier than those in Barishal, Chattogram, Rajshahi, Khulna and Rangpur.

In terms of improving early ANC initiation, targeted interventions should focus on reducing financial barriers, increasing awareness through mass media campaigns and expanding access to healthcare facilities, particularly in rural and disadvantaged regions. Strengthening women’s education, employment opportunities and decision-making autonomy can further enhance timely ANC utilisation. Addressing geographic barriers by improving healthcare infrastructure, introducing mobile clinics and enhancing community-based services is also essential. Additionally, improving health system factors, including the availability of skilled healthcare providers and the quality of ANC services, is necessary for sustainable improvements.

This study underscores the urgent need for comprehensive policy measures to promote early ANC initiation and reduce maternal and neonatal health risks in Bangladesh. Future research should explore longitudinal trends, healthcare system constraints and the effectiveness of tailored interventions to develop evidence-based strategies for improving maternal health outcomes.

Data availability statement

Data are available in a public, open access repository.