This study was carried out in Bule Hora Town, which is located West Guji Zone of the Oromia region in the lowlands of southern Ethiopia. According to the West Guji Zone health and demographic surveillance department dataset, which is derived from the 2007 census by 2.4% inflation. ¹⁴ The town, which is located 365 kilometers from the capital city of Addis Ababa along the main connecting road, is under zonal administration and has a total population of 157,148. The town is characterized by an urban setting, a mostly agrarian economy, and a population that is mostly from the Oromo ethnic group, namely the Guji agro-pastoral community. A facility-based unmatched case-control study design was employed. To address the determinants of early antenatal care utilization, the study followed specific procedures conducted between June and July 2025.

Quantitative data were collected using structured questionnaire. The study population consisted of pregnant women who attended ANC at Bule Hora University Teaching Hospital during the study period and were randomly selected. Cases were defined as women who initiated ANC before 12 weeks of pregnancy (early initiators), while controls were those who started ANC at or after 12 weeks (late initiators). ¹⁵ Eligible participants included were women aged 18 and older, as well as mature minors under 18, who had resided in the study area for more than six months. Women with unknown gestational ages at their first visit, hearing or speech impairments, serious pregnancy complications, or mental illness were excluded from the study. The study participants were approached by the principal investigator and interviewers with extensive experience in quantitative research.

Dependent/Outcome variable: Early ANC initiation, defined as initiation before 12 weeks of pregnancy (early initiators/cases = 1), versus initiation at or after 12 weeks (late initiators/controls = 0).

Independent/Exposure variables: Sociodemographic factors (age, residence, marital status, personal and husband’s education and occupation, income, and distance to a health facility); knowledge-related factors (awareness of ANC importance, timing, number of visits, services, danger signs, and information sources); individual and obstetric factors (parity, gravidity, previous and current ANC timing, delivery location, history of abortion/stillbirth, past ANC use, family planning, pregnancy planning, complications, and transportation access); and finally, health facility and cultural factors (facility proximity, service costs, drug availability, waiting times, provider attitude, husband accompaniment, community support and norms, and cultural barriers).

The sample size was calculated using a double population proportion formula in Epi-Info version 7.2.2.6, based on a key determinant from prior literature. ¹⁶ The first calculation produced 416 participants with a power of 90%, a 95% confidence level, a case-to-control ratio of 1:3, an exposure proportion of 35.7% among controls, and an adjusted odds ratio of 2.15. The final sample size was 456 (114 cases and 342 controls) after accounting for a 10% non-response contingency. Simple random sampling was used to choose study participants from the hospital’s antenatal care register. The Epi-Info random number generator was used to classify eligible pregnant women at random.

Data quality was ensured through pretested and refined questionnaire before deployment on KoboToolbox. Data collectors received two-day training on protocols and digital data collection using KoboToolbox to ensure tool use, data accuracy, and ethical procedures for two days. Built-in validation checks (such as skip logic, range limits, and required fields) were applied to minimize entry errors. Supervisors review submitted forms daily for completeness and consistency, whereas the principal investigator performs random cross-checks and downloads the data periodically to identify anomalies or missing values. After collection, all data was cleaned, verified for completeness, and exported to statistical software for final consistency checks before analysis.

The data was collected using face-to-face interviews from pregnant women who can provide informed consent using Kobocollect app. The questionnaire was adapted from Ethiopian Demographic and Health Survey and different related literature. ^{17, 18} Informed consent was obtained before conducting the interviews. The data collectors are professionally Nurses or health officer, or midwives have experience of working in the ANC unit from other facilities were participated and one supervisor was recruited from zonal health office.

All complete questionnaires were checked for accuracy, consistency, and completeness before being entered into KoboToolbox and exported to SPSS version 27 for analysis. To find factors linked to early ANC utilization, descriptive statistics, bivariable and multivariable logistic regression analyses were carried out; variables with p ≤ 0.25 in the bivariable analysis were included in the multivariable model to assess confounding effects; statistical significance was set at p < 0.05; the Hosmer-Lemeshow test was used to verify model fitness before final analysis.

Early ANC: Pregnant women who visit the ANC unit before 12 weeks of pregnancy (i.e., early initiators of ANC). ¹⁵

Knowledge of ANC: Assessed by questions on key ANC topics (early booking, danger signs, complications, timing, importance, services, gestational age). Scoring ≥50% is classified as good knowledge; <50% as poor knowledge. ¹⁹

Long Waiting Time: The total time a pregnant woman waits at registration, triage, and diagnostic units before seeing a provider. A cumulative waiting time of ≥120 minutes is considered long. ²⁰

Ethical clearance was obtained from Bule Hora University following the Declaration of Helsinki and was approved by Institutional Research Ethics Review Committee (BHU/IRERC) with protocol number of BHU/IRERC/030/17. A support letter was obtained from the Bule Hora Town health department to the hospital. All participants were informed about the purpose and the procedures of the study. All responses were kept confidential and anonymous. The study protocol was approved by the Bule Hora University Institutional Review Board (IRB). Prior to enrollment, informed consent was obtained from all participants: written consent was obtained from those with formal literacy, while witnessed verbal consent was obtained from participants unable to read or write.

A total of 448 pregnant women were included, with 112 cases and 336 controls, yielding a response rate of 98.2%. Most participants were urban residents, 394 (87.9%). The majority were married, 441 (98.4%), with slightly fewer in controls, 108 (96.4%) vs. 333 (99.1%) in cases. Secondary education was the most common level, 123 (27.5%), and housewife was the dominant occupation, particularly among cases, 208 (61.9%) vs. 37 (33.0%) in controls. Husbands of 142 (31.7%) participants had higher education, slightly more among controls, 41 (36.6%). Low household income (<8,319 ETB) was reported by 273 (60.9%) participants, more so in cases, 213 (63.4%) vs. 60 (53.6%) in controls ( Table 1).

Most participants knew the importance of ANC, 436 (97.3%), and its benefits, 433 (96.7%). Knowledge of early ANC booking was higher among controls, 106 (94.6%), than cases, 154 (45.8%). Fewer cases, 67 (19.9%), knew the correct time to start ANC (<12 weeks), compared to 71 (63.4%) controls. Monitoring maternal and fetal health was reported by 309 (71.4%), while early detection of complications was noted by 295 (68.1%). Only 138 (30.8%) correctly identified the first trimester as the appropriate time to begin ANC. Regarding necessary ANC visits, most participants believed fewer than 4 visits were sufficient 276 (61.6%). Knowledge of pregnancy danger signs was 189 (42.2%), with vaginal bleeding being the most recognized sign, 125 (66.1%). The main sources of ANC information were community health workers, 170 (37.9%), and health workers, 116 (25.9%).

As shown in the figure below, overall knowledge of ANC was higher among early ANC utilizers (cases) with 92 (82.1%) demonstrating good knowledge, compared to 164 (48.8%) of late utilizers (controls). Overall, 256 (57.1%) of participants had good knowledge, while 192 (42.9%) had poor knowledge of ANC ( Figure 1).

Out of 448 participants, 135 (30.1%) were nulliparous, 126 (28.1%) primiparous, and 187 (41.7%) multiparous. Regarding gravidity, 139 (31.0%) were primigravida and 309 (69.0%) multigravidas. In their last pregnancy, only 49 (17.2%) initiated ANC in the first trimester, while 142 (49.8%) started in the second trimester, and 29 (10.2%) in the third. Place of previous delivery showed that 245 (54.7%) gave birth at a health institution, 69 (15.4%) at home, and 9 (2.0%) on the way. A total of 70 (15.6%) had a history of abortion and 33 (7.4%) experienced stillbirths. Family planning use before pregnancy was reported by 250 (55.8%), and ANC attendance in a previous pregnancy by 274 (64.6%). Pregnancy was planned by 331 (73.9%), and 117 (26.1%) experienced complications.

Out of the participants, 309 (69.0%) of participants reported proximity to a health facility, with a higher proportion among cases 101 (90.2%) than controls 208 (61.9%). Long waiting times (>120 minutes) were experienced by 178 (39.7%), predominantly among controls 172 (51.2%) versus only 6 (5.4%) cases. Most participants rated provider attitudes as good or very good: 421 (94.0%). In terms of socio-cultural and support factors, husband accompaniment to ANC was reported by 221 (52.1%), and 321 (71.7%) were exposed to maternal health messages through media. Perceived community support for ANC was reported by 359 (80.1%), and 397 (88.6%) believed ANC attendance was a community norm. Decision-making regarding ANC was self or jointly made by 403 (90.0%) participants and 13 (54.2%) reported husband accompaniment.

In the bivariate analysis, variables with early ANC (p < 0.25) and were included in the multivariable logistic regression. After adjusting for potential confounders, the multivariate logistic regression analysis showed that younger pregnant women aged 15–24 were nearly three times more likely to delay ANC initiation compared to those aged 35–44 (AOR = 2.79, 95% CI: 1.12, 6.96). Women with poor knowledge of ANC were about four times more likely to delay ANC than those with good knowledge (AOR = 4.19, 95% CI: 2.24, 7.82). Those who experienced long waiting times before seeing a provider were four times more likely to delay ANC (AOR = 4.02, 95% CI: 2.23, 7.24). Similarly, women in polygamous marriages had higher odds of delayed ANC initiation (AOR = 3.74, 95% CI: 1.53, 9.16), and those perceiving poor community support for ANC were over three times more likely to delay seeking care (AOR = 3.63, 95% CI: 1.44, 9.17) ( Table 2).

This institutional-based case-control study has several limitations. Firstly, it cannot establish causal relationships for all significant associations. Secondly, recall and social desirability biases may affect the accuracy of self-reported data, particularly regarding missed doses and socioeconomic status.