INTRODUCTION

Prenatal care is a program of scheduled visits for pregnant women in order to prevent and detect complications of pregnancy, childbirth and puer-perium, and provide early treatment ¹. For this reason, it is part of universal policies and programs and is driven by guidelines, protocols and evaluation measurements. In order to accomplish the goals of early detection of risks and diseases that may affect mother and child, including infections, nutrition disorders, dental problems and biopsychosocial issues, pregnant women must join the program as soon as possible. The earlier the care is provided, the greater the opportunity to prevent, identify and address problems that might affect the health of the mother and that of her child. Some guidelines recommend joining the program before ten weeks of gestation ¹, while others recommend initiation before 14 weeks ². Early initiation of prenatal care is considered a predictor of adherence to the program ³. In Colombia, late initiation of prenatal care was reported in 60% of maternal deaths occurring during the period between 2004-2009 ⁴; consequently, evidence shows that late initiation of prenatal care entails a higher probability of negative outcomes for the mother and the baby ⁵.

Several international studies have identified mother-related factors associated with late initiation of prenatal care, including not being in a stable relationship, ⁶ low education level, unwanted pregnancy, lack of health insurance or a regular care institution, being younger than 20 or older than 35, multiparity, unemployment, living in socially depressed areas or having a low level of income ^6-14. In this regard, some authors recommend identifying, for each particular context, the specific issues affecting pregnant women in order to guide intervention strategies ¹⁵. Health authorities, decision-makers and healthcare service providers must be informed about factors associated with late initiation of antenatal care, so that strategies can be devised for attracting pregnant women to the program early on.

There is a paucity of studies in Colombia regarding access to prenatal control and associated factors ^16-25. In terms of the association between timely initiation and socioeconomic conditions, local information available has limitations in terms of the type of population included, consisting mainly of pregnant women from low socioeconomic brackets ^17,18, and the type of proxy variable used for assessing socioeconomic conditions, for example, the inability to pay for transportation to attend prenatal visits ¹⁹, the low income of the mother ¹⁶, and the economic condition of the household ²⁰. On the other hand, this association has not been shown consistently ^16-18, and this is compounded by the fact that the cutoff point to consider late initiation is widely variable ^17,18,21. Consequently, knowledge about the relationship between late initiation of prenatal care and economic factors is still limited.

Therefore, the primary objective of this study was to examine the prevalence of late initiation of prenatal care, and the secondary objective was to evaluate the association between late initiation and socioeconomic bracket.

MATERIALS AND METHODS

Design and population. Analytical cross-sectional study that included pregnant women over 18 years of age, living in the urban area of Bucaramanga, coming from all income brackets, who were seen in public and private healthcare institutions between June 2014 and December 2015 and who gave their consent to participate in the study. Sensory or communication impairment that prevented the participants from answering survey questions was considered the exclusion criterion.

Sample size and sampling. The study hypothesis was that pregnant women who live in areas classified as low socioeconomic brackets are at a higher risk of late initiation of prenatal care. In order to detect this association, a sample size was calculated in accordance with the recommendations by G. Liu ²⁶, using the Statcal routine of the Epi-info7 software ²⁷, with a figure of potential eligible cases of 13,521 (live births during 2014), a 95% reliability and a power of 80% to detect a 50% proportion of late initiation of prenatal care, and 1.1 due to design effect. With these conditions, the calculated sample size was 411 pregnant woman, allowing to find a difference in the proportion of late initiation of 30% between exposed (lower brackets) and non-exposed (high brackets) women to late initiation of prenatal care. Quota sampling was used, with proportional representation by socioeconomic brackets, from 1 to 6, in accordance with the classification provided by the Planning Bureau of Bucaramanga. In order to estimate the proportion of pregnant women to include by bracket, the proportion of deliveries in each of the brackets in 2013 was considered, based on the statistics of the Affiliation Registry System (RUAF). These statistics showed that 21.8% of the births came from bracket 1; 20.2% came from bracket 2; 25.6% from bracket 3; 18.6% from bracket 4; 8.5% from bracket 5; and 5.3% from bracket 6. Based on these proportions, the sample for collection was distributed as follows: bracket 1, 90 pregnant women; bracket 2, 83; bracket 3, 105; bracket 4, 76; bracket 5, 35; bracket 6, 22 women.

Procedure. Contact with the pregnant women was made in public and private healthcare institutions where they attended for laboratory tests, ultrasound scans or consultations with the health professionals in charge of prenatal care. The women were asked to sign an informed consent in order to participate. The data collection tool was administered to the women who agreed to participate. The tool consisted of a structured questionnaire prepared by the researchers, with closed, standardised questions designed to assess sociodemographic, gynaecological and obstetric, insurance, and prenatal care considerations. Before starting the research, a pilot test of the tool as well as of the data collection process was conducted with 30 pregnant women of all income brackets, and the necessary adjustments were made. A physician and a nurse experienced in population studies were entrusted with the data collection process. They received training and followed a standardised process to approach the pregnant women, give the informed consent and conduct interviews. They were supported in their-role by one of the researchers in charge of ensuring process and data quality. Enrolment of the women in the research was ended when the sample size for each bracket was reached.

Data management. Completed surveys were labeled with an identification code in order to ensure the confidentiality of the information. The collected data were stored in an Epidata® database and analysed using the Stata14® software package.

Measured variables. The measured variables were: a)the pregnant woman: age, education, marital status, having a paid job, gynaecological and obstetrical history, health services, and place of residence; b) healthcare services: type of affiliation to the General Social Security System (SGSSS) at the time of pregnancy, attendance to family planning programs, and preconception consultation; c) place of residence: housing bracket; d) the baby's father: whether he was living with the pregnant woman on the date of prenatal care initiation.

The dependent variable was late initiation of prenatal care, defined as initiation after 12 weeks of gestation. The independent variable of interest was the socioeconomic bracket reported by the woman according to the socioeconomic classification of her neighbourhood provided by the Municipal Planning Bureau ²⁸. Socioeconomic brackets are governed by the social stratification created in Colombia by Law 142 of 1994. According to the National Statistics Department (DANE), socioeconomic stratification is a classification of residential property that divides de population into six brackets of similar social and economic characteristics based on the physical characteristics of the dwelling and its surroundings. The purpose of such a classification is to focus public policy, mainly with the aim of charging differential rates for utilities and permit the allocation of subsidies ²⁹. Stratification in itself is considered a way to approach the hierarchical socioeconomic distinction between poverty and wealth which divides the population into six brackets from 1 to 6, where 1 is low-low, 2 is low, 3 is lower middle, 4 is middle, 5 is upper middle, and 6 is high. Brackets 1, 2 and 3 are beneficiaries of government subsidies: brackets 5 and 6 contribute; and bracket 4 is neither beneficiary nor contributor ³⁰.

Analysis. A description of the variables according to their measurement scale was initially done. Variables measured on a continuous ratio scale were described in terms of central trend and scatter, mean, standard deviation (SD) or median and interquartile range (IQR), depending on whether or not they showed a normal distribution on the Shapiro-Wilks test. Variables measured on ordinal and nominal scales were described using absolute and relative frequency tables and 95% confidence interval (CI). The proportion of late initiation of prenatal care was calculated by bracket and in two categories: low (1, 2 and 3) and high (4, 5 and 6). Late initiation of prenatal care was shown to behave similarly in brackets 1, 2 and 3, but showed a different behaviour in brackets 4, 5 and 6. Therefore, they were grouped in two brackets designated as low (1, 2 and 3) and high (4, 5 and 6).

A bivariate analysis was performed between late initiation and socioeconomic bracket, and the raw prevalence ratio (PR) was calculated with its 95% confidence interval (95% CI) and controlled for other potentially confounding variables. Those that modified raw PR by more than 10% and where the probability of type 1 error (p) was lower than 0.2 were introduced in a log-binomial regression model in order to estimate the adjusted prevalence ratio and their respective 95% CI ^31,32. All the calculations were done using the Stata14 software package, and differences were considered significant if confidence intervals did not include unit ¹ or p values were under 0.05.

Ethical considerations. Both the research protocol as well as the informed consent were endorsed by the Ethics Committees of the participating centres. All the pregnant women were given an informed consent before enrolment in the study.

RESULTS

Of 425 pregnant women asked to participate, 391 were included in the research, for a non-response proportion of 8% (34 women). Their reasons for not participating were time limitations for the interview or lack of interest in the study.

The median age of the participants was 25 years (range 18 to 43); the median level of schooling was 11 years (range 0 to 25 years); and the median gestational age at the time of the interview was 27 weeks (range 4 to 41 weeks).

In a total of 116 of the 391 participants (29.7 %) (95% CI: 25.2 %-34.5%) initiation of prenatal care was late. The following were the prevalences of late initiation by socioeconomic bracket: bracket 1: 32/84 (38.1%), bracket 2: 30/75 (40%), bracket 3: 35/96 (36.5%), bracket 4: 16/71 (22.5%), bracket 5: 3/47 (6.4%), and bracket 6:1/18 (5.6%). This shows that there was an inverse gradient between the prevalence of late initiation and the socioeconomic bracket. Thus, among pregnant women in the low brackets (1 to 3), there were 96 cases of late initiation among 256 women, for a prevalence = 37.5% (95% CI: 31.5-43.7), while there were 20 cases of late initiation among 135 women in the high brackets (4 to 6), for a prevalence = 14.8% (95% CI: 9.3-21.9), and a raw prevalence ratio (PR) of 2.53 (95% CI: 1.64-3.91). Other risk factors associated with late initiation in the bivariate analysis were age, less than 12 years of schooling, and non-affiliation to the SGSSS. In contrast, being in a stable relationship and remembering the date of the last menstruation were protective factors against late initiation of antenatal care (Table 1).

Table 1 Factors associated with late initiation of prenatal care. Bucaramanga, 2014-2015. Bivariate analysis. 

n: number of pregnant women.

PR: prevalence ratio.

95% CI: 95% confidence interval.

* Statistically significant association.

SGSSS: General Social Security System in Health.

The use of the multivariate log-binomial regression model for the analysis showed that late initiation of prenatal care was associated with low socioeconomic bracket, lack of affiliation to social security, low schooling, and age between18 and 24 years as risk factors. In contrast, remembering the date of the last menstruation and being in a stable relationship at the time of prenatal care behaved as protective factors (Table 2).

Table 2 Factors associated with late initiation of prenatal care. Bucaramanga, 2014-2015. Multivariate model. 

aPR: adjusted prevalence ratio.

95% CI: 95% confidence interval.

Adjustment goodness of the model: Pearson Ji2 99.03 and Prob > chi2 = 0.4802.

DISCUSSION

This study found a prevalence of late initiation of prenatal care of 29.7%. It also found an association between late initiation and living in a low socioeconomic bracket, lack of affiliation to social security at the time of pregnancy, low schooling, age between 18 and 24 years, remembering the date of the last menstruation, and being in a stable relationship during prenatal care, the latter as protective factors.

When comparing the prevalence of late initiation of prenatal care found in this study with the one reported by other authors in other countries, this prevalence was higher than the one found by Corbett in New Zealand (17%, with initiation at 18 weeks of gestation or later) ⁶, by Beeckman (10.8%, after 12 weeks) ⁹, and Fobelest (6.1%, after 14 weeks of pregnancy) in Belgium ⁸, by Heredia in Mexico (16.8% after 14 weeks of gestation) ¹³, and by Sinyange in Zambia (19%, after 14 weeks) ⁷. However, it is lower than the one reported in Myanmar by Aung (56.6%, 16 weeks of pregnancy) ¹⁴, in the United Kingdom by Creswell (37.5%, after 12 weeks of gestation) ¹¹, and in Ethiopia by Zegeye (73.8%, after 12 weeks of gestation) ¹⁰.

Compared to the findings of other studies conducted in Colombia, Miranda ¹⁶, Briceño ²² and Castillo ²⁴ reported the prevalence of late initiation using a cut-off point of 14 weeks.

The proportion of late initiation found in our study is higher than the one found by Miranda in Sincelejo (15.1%) ¹⁶, and by Briceño at a national level (25.3%) in users of the contributive insurance health regime ²². In contrast, it is lower than the one reported at a national level in users of the subsidised insurance health regime (52%) ²², and by Castillo et al., in 13 municipalities of the Department of Bolivar (36.7%)²⁴. Comparison of our results with the prevalence of late initiation found in prior studies conducted in the city, our results are lower than those reported by Ortiz et al. (50%, initiation after 12 weeks of gestation) ¹⁸ in teen-age pregnant girls in the public sector, and by Ruiz, (67.2%) in pregnant women in different income brackets and different types of insurance ²¹, and higher tan the 19% found by Moya for late initiation ¹⁷.

In terms of factors associated with late initiation of antenatal control, our findings regarding socioeconomic level are consistent with those of researchers outside Colombia, such as Corbett ⁶, Heredia ¹³, Beckam ⁹, and Fobeles ⁸, who also reported this association: the lower the socioeconomic conditions, the higher the probability of initiating perinatal care late in gestation. In Colombia, none of the prior studies that looked into the factors associated with late initiation of prenatal care included pregnant women in all socioeconomic brackets and only included low income population ^16,22,24, which could be considered as a surrogate of the low socioeconomic bracket.

It is also worth highlighting that the strongest association was found with the lack of health insurance. In this regard, our findings are consistent with the reports of different authors who have looked into prenatal care, such as Rodríguez ¹⁹, Briceño ²², Castillo ²⁴ and Vecino ²⁵ who found that affiliation to health insurance is associated with antenatal access and attendance. These results reinforce the knowledge available in the field of health systems in the sense that having insurance is an enabler of access and use of the services ^33,34.

On the other hand, being in a stable relationship reduces the probability of late initiation of prenatal care. Contrary to our findings, an association between a stable relationship and late initiation has been reported by other authors ^6,7.

Remembering the date of the last menstruation as a protective factor against late initiation might reflect the absence of irregularities that might prompt the woman to seek early medical assistance in order to rule out pregnancy

The strengths of this study include the use of a multivariate analysis model that allowed to reject the hypothesis of no association between low socioeconomic bracket and late initiation of prenatal care, adjusting for other factors. Moreover, standardisation of the tools and data collection processes reinforced the quality, validity and reproducibility of the data.

Some weaknesses of the study need to be considered when interpreting or generalising the results. To start with, weaknesses inherent to non-probabilistic sampling, which does not allow generalisation. On the other hand, living in areas classified in the low socioeconomic brackets does not necessarily reflect individual socioeconomic levels of all the residents clustered under the same socioeconomic bracket in Colombia ³⁵.

The results of this study highlight a challenge for the local health system because, within the framework of the Comprehensive Healthcare Policy (PAIS), the new Comprehensive Perinatal Maternal Care Route (RIA) includes the initiation of prenatal care before 10 weeks of gestation as one of the performance milestones.

CONCLUSIONS

This study showed late initiation of prenatal care in approximately one out of every three pregnant women. It also demonstrated the hypothesis that late initiation is associated with the low socioeconomic bracket. This finding suggests inequities in timely initiation of prenatal care among the poorer populations. The implications of these results point to the need to improve induced demand strategies in order to increase early attendance to prenatal care visits among women coming from the lower socioeconomic levels of the population.