Factors that Influence Teenage Antenatal Care Utilization in John Taolo Gaetsewe (JTG) District of Northern Cape Province, South Africa: Underscoring the Need for Tackling Social Determinants of Health

2.1. Study design

A cross-sectional study was conducted at all public health facilities in JTG district between September and November 2014 using mixed method approach. Both qualitative and quantitative data collection techniques were used for collecting primary data. Quantitative data was obtained using a standardized anonymous questionnaire adapted from validated tools[19,20], which included information on maternal reproductive knowledge, attitude and practice, socio-economic status, utilization of ANC services, delivery and postnatal visits (six-week), facility performances and supply side constraints in provision of maternal and child health services. Health facilities sampled include 2 District Hospitals, 5 Community Health Center (CHCs) and 37 Clinics. Questionnaires were translated from English into two local languages Afrikaans and Setswana languages and back translated into English to ensure accuracy

Qualitative data was collected through semi-structured interviews and Focus Group Discussion (FGD) from purposely-selected participants that included high reproductive health risk groups such as teenage mothers, mother who do not utilise ANC services at health services and traditional births attendants. The FGDs were captured using a digital recorder. The information then was transcribed for analysis where similar emergent teams were differently categorised. Lay counsellors who were trained for two days on data collection tools and research ethics collected the data. Lay counsellors signed a confidentiality agreement prior to data collection. Observational data and field notes were collected during field visits. Use of ANC services was determined by the participants self-report and was confirmed from clinical records.

2.2. Sampling techniques and sample size for facility based survey

Mothers attending six-week immunization for their infants in all public health facilities in JTG district were the study population from which the sample was drawn. By means of the 2011 census, the study population was estimated to consist of 72,000 women. The six-week immunization and reproductive health services utilization rate data from the 2012/13 District Health Information System (DHIS) was used to quantify the number of mothers expected to attend six-week immunization visit for their infants per facility over the study period. For the facility-based survey, a precision based sample size was calculated taking into account the expected annual utilization of ANC services, a precision level of 2-3%; and 95% confidence level. This provided a sample size of 383 respondent mothers who will be attending six-week immunization services for their infants in all public health facilities within the district. Sample size was allocated for each facility proportionate to their size as determined by the number of six-week immunizations coverage for 2013/14.

2.3. Data analysis

Data analysis was conducted using STATA version 13. Univariate, and multivariate logistic regression models were fitted to identify factors influencing the total number of ANC visits and late ANC booking. Results were expressed using descriptive statistics and where necessary, data was disaggregated by different sociodemographic and economic status of respondents. We used a dichotomous dependent or outcome variable of the study on pregnant women’s use of ANC visit, categorized into: - those who attend at least 4 ANC visits, and those who do not during the time of pregnancy. The literature shows that the relationship between a binary outcome variable of study such as attending at least 4 ANC visits during the time of pregnancy or failure and predictor variables that affect the outcome variable can successfully be explained using the binary logistic regression model.[21,22] Binary logistic regression analysis was performed to identify predictor variables that influence the number of ANC service utilization. The econometric measure of effect in binary logistic regression analysis is the odds ratio. Odds ratios arising from binary logistic regression analysis are estimated based on the maximum likelihood estimation technique.[23,24] The logit model is used quite extensively for the estimation of odds ratios in economic studies involving a dichotomous outcome.[25,26]

Consider a dependent variable Y with 2 possible outcomes (1, 0). The expression Y=1 represents the event that a mother has failed to attend 4 ANC visits. The expression Y=0 represents the event that a pregnant mother has attended at least 4 ANC visits successfully.

Previous studies revealed that utilization of ANC services is highly influenced by different sociodemographic, economic factors as well as specific individual-level characteristics both at the client and service providers’ side. Continuous predictable variables had to be dichotomised in order to perform Pearson’s Chi-square tests of association. For instance, the Pearson Chi-square test of association is used to test the null hypothesis that two factors such the dependent variable and independent variable are independent of each other, against the alternative hypothesis that the two factors are significantly associated with each other. The Pearson chi-square test of association is used only for the screening of variables.

Accordingly, we included several theoretically pertinent sociodemographic and economic independent variables (indicators) that influence uptake of pregnant women’s ANC services. Respondent women’s age was classified into age group of respondents less than 20 years (adolescent group), and age group older than or equal to 20 years of age (middle and older women). Women’s education level was defined in terms of the formal education system of South Africa: School not attended, elementary (grade 1-7), high school (8-12), college or above. Marital status of mothers categorized as married, living together, widowed/separated, and single. Place of residence was categorized as rural versus urban. Distance to health facility for ANC services categorized as more than or equal to one hour or less. Other independent variables include income, level of a mother, individual health seeking behavior and service provider’s attitude. Household assets, including ownership of durable goods (such as car, fridge, televisions, stove and washing machine). Dwelling characteristics included were source of drinking water, sanitation facilities, and construction materials and others used as predictor variables for binary logistic regression analysis. The identification of influential predictor variables was done based on odds ratios. In binary logistic regression analysis, influential predictor variables are characterized by odds ratios that are significantly different from 1, 95% confidence intervals of odds ratios that do not contain 1, and P-values that are smaller than 0.05, at the 5% level of significance.

2.4. Ethical consideration

The study was approved by the Northern Cape Provincial Health Research and Ethics Committee (PHREC), reference number NC-PHREC-2014/021.

3.1. Descriptive statistics

Of 383 sample mothers approached for study participation, 272 agreed to participate in the study and completed the survey questionnaire which resulted in a 72% of response rate. Out of 272 mother-infant pairs interviewed at 6 weeks immunization visits, 69.37% had secondary education. Only 8.9% have completed college or above. The majority (84.79%) respondent mothers were unemployed or housewives. Regarding the type of employment sector, government and company employment jointly account for only 15.27%. Most of the unemployed mothers depend for their source of income on child support grant, disability grant, and donations. However, for 54.76% respondent mothers, the total monthly income was less than or equal to one thousand Rand (100 US Dollars).

About one-fifth (18.68%) of the respondents were teenagers younger than 20 years of age, 71.96% were single mothers, Close to half (44.69 %) of mothers reported they had their first baby before the age of 20 years. Out of the 18.68% respondent mothers who were teens in this study, 94%, and 82.35% of them were unemployed and single mothers respectively. Regarding ages of the respondents, 15 years was the youngest age and 45 years was the oldest age. About 16.91% of the respondents were married; 9.56% were living together; 1.84% were separated/divorced; and 1.11% were widowed. About two-third (71.96%) respondents were single mothers.

In this study, only 28.62% respondents own houses, while 6.32% live in rented houses; 57.52% live with their relatives, and 9.67% lived with a spouse/partner. With regards to the main source of water used for drinking, 93.38 % respondents were using tap water either inside the house (16.54%); or yards (15.44%) or (61.76%) public taps. Close to ninety percent (88.97%) of the houses were connected to electricity, and for more than 80.44% of the respondents, electricity was the main type of fuel for cooking in the house. 21.72% households were using flush toilets, 43.82% pit latrines private, and 31.46% ventilated pit latrines. Table 1 presents a summary of the socioeconomic and demographic characteristics of the study respondents (Table 1).

Close to half (49 %) of the teenage pregnant women attended at least four ANC visit during their pregnancy. However, the majority (64.71%) of teenage mothers booked for the first ANC visit after 12 weeks into pregnancy. Only (35.29%) of teenage pregnant women booked for an early ANC first visit before 12 weeks of pregnancy as recommended by the South African Department of Health and WHO guidelines.

More than ninety percent (92%) teenage pregnant mothers attended their first ANC visits at public health facilities, and a few mothers attended private health facilities (6%) or both public and private facilities (2%). Regarding how pregnant women felt about their pregnancy, 25.49% of respondent teenage mother responded that the pregnancy was wanted, while the majority (74.51%) reported that their pregnancies were unwanted at that time. About 2 percent (1.96%) of teenage pregnant mothers completed college or above whilst the majority (86.27%) were in secondary education. 82.35% of them were single mothers, while 5.88% were married, 9.80% were living together, and 1.96% were separated/divorced. The bulk of them (94%) were unemployed, and for the majority of (58.7 %) respondent teenage mothers their total monthly income was less than or equal to one thousand Rand (or about $100).

3.2. Logistic regression analysis

A number of predictor variables influencing the use of ANC services were studied for two outcome variables using the multiple binary logistic regression models: (1) number of ANC visits (≥4 vs. < 4); and (2) time of attendance of ANC (early vs. late). We estimated the odds ratios (ORs) to assess the strength of the associations for the 95% of confidence intervals (CIs). Regarding the number of ANC visits, in multivariable regression after adjusting for employment status, marital status, relationship with the father of the child, waiting time, fear of HIV test, partner support, and education level, we found that mother’s age (OR=2.11; 95%CI = 1.04 - 4.27); distance to the nearest health facility (OR=3.38; 95%CI = 1.45-7.87); and client service satisfaction (OR=8.58; 95%CI =2.10-34.95) are significantly associated with poor uptake of ANC services (Table 2). This suggests that the probability greater number of ANC visits is 2.1 times more likely higher with age group older than or equal to 20 years compared to adolescents. Increased distance (more than or equal to one hour) to ANC services, is 3.3 times more likely to negatively impact uptake of ANC visits. Satisfied mothers with ANC services were 8.58 time more likely to use ANC visits. None of the predictor variable were strongly associated with the outcome variable time of first ANC (early or late) attendance (Table 3). A comparison of Table 2 with Table 3 shows that the independent variables did not affect both number of ANC visits and ANC booking the same way.

3.3. Focus group discussions

In addition to the surveys questionnaires, we conducted a focus group discussion to reveal a wealth of detailed information and deep insights on the utilization of ANC services and provide recommendations on strategies to tackle the challenges and achieve better health outcomes. For each group 10-15 key participants were selected from the following groups. (i) women who fully or partially missed reproductive health services; (ii) high risk reproductive health groups which include adolescents, mothers who plan to give birth at home, late presenters to reproductive health, and those women who have unplanned pregnancies; (iii) traditional birth attendants, and (iv) healthcare service providers.

We found that teenage pregnancy was more common in the JTG district and most of it was unplanned pregnancy. According to adolescents’ focus group informants, the main reason for unwanted pregnancy among teens was peer pressure from friends and economic dependency. Some of the teens had not completed school, were not married and the support they were getting from the child’s father were not consistent. According to one of the teenage pregnant women:

“At the earlier months of my pregnancy, he was there. But recently, I have no communication with the father of my child. Most of the time his cell phone is off and he is not responding to my calls. Because the child’s father is no longer able to support us, I will apply for the government child social grant. I will also request my parents to look after my child so that I will be able to look for work.”

Another teenage participant responded, “For me everything is still fine with this child’s father and he is still supporting me, although I had one more child from different father.”

Regarding knowledge about benefits of early and adequate number of ANC visits, the majority of respondents were aware of the benefits. They mentioned the role played by community healthcare workers. According to one of the respondents:

“The community healthcare workers are helping us to use health care services, remind patients about the date to go to the clinic visits. Sometimes they assist patients to the clinic, they encourage patients to take their pills as well as educate them on government’s policy with regard to their health. I think they are supporting the healthcare system in reaching the communities. However, some other participants suggested, “There is a need to have adequate nurses, and adequate space at the facility as well as priority should be given for pregnant women. The problem is that the clinics use a supermarket approach in providing health services, which means first come first served. This leads to a long waiting time for pregnant women and can be seen by other patients from the community which teen pregnant mother do not want.” In our district there are many facilities operating with only a nurse who has to perform all the services which include antenatal, family planning, chronic and also emergencies. On average one nurse has to see 35 patients a day but they often see more than that number daily”.

4.1. Limitation of the study

The study designed with a particular focus in JTG district. The findings of the study are limited to this district and cannot be generalized to the Northern Cape Province or the country. The study findings may not be representative of all pregnant mothers who do not attend six-week immunization due to child illness or death.