Factors influencing access of pregnant women and their infants to their local healthcare system: a prospective, multi-centre, observational study

Between 1990 and 2015, the global under-five annual mortality rate has declined by 53%, from 91 to 43 deaths per 1000 live births [1]. Moreover, an acceleration in the improvement of child survival has been observed during this period, with annual rates of reduction in mortality increasing from 1.8% in 1990–2000 to 3.9% in 2000–2015 [1]. The decline in mortality has been slower in neonates than in post-neonatal under-five children (47% versus 58% globally), with neonatal mortality rates that fell from 36 to 19 deaths per 1000 live births between 1990 and 2015. Almost 45% of the 5.9 million deaths reported worldwide among children younger than 5 years in 2015 occurred in neonates, and the neonatal period remains the most vulnerable time for a child’s survival [1‐4]. The vast majority of newborn deaths and almost all maternal deaths (99%) take place in developing countries [3, 5]. As lack of access to quality healthcare at critical time points prior to, during, and after delivery may influence mortality rates [4, 5], identifying barriers to healthcare utilisation could reduce infant and maternal deaths.

Immunisation of pregnant women can protect neonates from diseases which may present within the first days of life: tetanus vaccination is recommended by the World Health Organisation [6], some countries have introduced routine seasonal influenza vaccination [7], and others recommend primary or booster vaccination against pertussis [8, 9]. Maternal vaccination against other diseases, such as Group B Streptococcus (GBS), which can kill infants in the first hours of life, is currently under investigation [10, 11]. GBS is a major cause of neonatal sepsis and meningitis, and is classified as early-onset disease when presenting between 0 to 6 days of age and late-onset disease when presenting between 7 to 89 days of age [12]. In previous clinical trials, a single dose of an investigational GBS vaccine was administered to women at 24–35 weeks of gestation, but the optimal vaccination schedule has not been determined [10, 11].

The successful implementation of maternal vaccination will rely on results of clinical trials, taking into account the prenatal and postnatal attendance at the selected healthcare institutions serving as study sites. This study identified factors influencing whether pregnant women would enrol in clinical trials and return to study site for delivery, infant follow-up visits, or medical assistance in case of infant disease. The aim was to identify potential barriers to participation in future large-scale efficacy trials of investigational GBS vaccines administered during pregnancy.

This prospective, multi-centre, observational study was conducted between November 2012 and January 2015 in potential sites for a phase 3 maternal immunisation clinical efficacy study in which pregnant women from countries with reported high incidence of GBS disease in infants could be enrolled. Although 13 sites originally qualified for participation, three sites, one each in Malawi, South Africa and Panama, did not recruit any participants. The study was therefore conducted at ten sites: three in Panama, two in the Dominican Republic, four in South Africa, and one in Mozambique (Table 1). The study sites were urban, except the semirural site in Mozambique.

The study was conducted in accordance with the principles of Good Clinical Practice and the Declaration of Helsinki. The study protocol and related documents were approved by the institutional review boards or ethics committees in each study country: Comite de Docencia e Investigación, Hospital Santo Tomás, Panama; Consejo Nacional de Bioetica en Salud, the Dominican Republic; University of the Witwatersrand Human Research Ethics Committee (Medical), South Africa; Health Research Ethics Committee Research Development and Support Division, Stellenbosh University, South Africa; and National Bioethics Health Committee, Mozambique. Written informed consent was obtained from all women, or their parent/legal representative if minors, prior to enrolment in the study. Informed assent was also obtained from women who were classified as minors. The study was registered at www.​clinicaltrials.​gov (NCT01734434).

Women who sought antenatal care at the study site were eligible for participation if they were at ≥24 weeks (Panama, the Dominican Republic, and site SA_1 in South Africa) or 28–34 weeks (sites SA_2, SA_3, and SA_4 in South Africa and the site from Mozambique) of gestation at enrolment, and if they provided written informed consent. Following a protocol amendment, the gestational age eligibility window was modified to be consistent with recommendations from the GBS scientific committee specifying that the investigational GBS vaccine should be administered in pregnancy between 28 and 34 weeks of gestation. The new criterion for gestational age was only implemented at three sites from South Africa and the site in Mozambique (SA_2, SA_3, SA_4, and MO_1) because the others had already completed enrolment or were nearing completion at the time of the protocol amendment.

Data were collected from women at each of the three study visits (enrolment, delivery, and infant follow-up visit) by means of questionnaires (Additional file 1). The follow-up visit occurred when the infants reached 90 days of age, which is the upper age limit used to define late-onset GBS infection [12]. The questions were intended to identify the barriers that may hinder access to healthcare or participation in a future clinical trial for pregnant women in each of the settings. Questions asked at all three visits were: duration, cost and type of transportation to the study site, access to a telephone, and availability of care for other children. Specific questions at the enrolment visit (prenatal) were gestational age and its method of estimation, estimated delivery date, and details of the clinic for the woman’s well-child visit (immunisation). At the delivery visit, the time and place of delivery, labour onset, pregnancy outcome, health status of infant, and day of discharge were recorded. Women not delivering at the study site were asked to provide this information telephonically or by completing the non-study site delivery questionnaire. If needed, a home visit was undertaken as per local practice. Additional time points for data collection were the well-child visit (within 21 days after the estimated delivery date) or the 90-day follow-up visit. At the infant follow-up visit (90 days after delivery), the following information was recorded: the health status of the infant, the well-child visit attendance, and whether the infant was sick during the first 90 days of life (in which case further information was recorded on whether medical care was sought or not, where medical care was sought, or the reasons for not seeking medical care). Education about signs and symptoms of potential sepsis in infants was provided to nearly all women (97.3%) at enrolment and at delivery. Women were counselled to seek urgent medical care if their infant developed any symptom of sepsis. Potential GBS infections could be detected if enrolled women identified these specific sepsis symptoms in their child. Investigators recorded all study data on electronic case report forms stored in a secure electronic data capture system for validation.

The study sample size was driven by the enrolment capacity of the sites. Summary statistics were calculated for all recorded questionnaire data, and were presented as frequencies and percentages. Means and standard deviations were presented for continuous variables (e.g. age, gestational age). Analyses were performed on overall data, by region (Latin America or Africa), by country, and by study site. Statistical testing was used only for descriptive purposes, and no multiplicity adjustments were performed.

Stepwise logistic regressions were performed on all enrolled women and on those who provided data after enrolment. Formula and details on odds ratios (ORs) calculation are given in Supplementary Methods (Additional file 2).

To evaluate factors that might influence whether pregnant women would return to study site for delivery, pregnant women were categorised into two groups: those who returned to a study site to deliver and those who did not. Pregnant women whose delivery status was unknown were grouped with those who delivered at a non-study site. To explore the seeking of medical attention, only women with a sick newborn or infant were included for the first occurrence of illness. The predictor variables included in the logistic regressions were: country of study site; age of pregnant woman; gestational age of pregnancy at enrolment; mode, duration, and cost of transportation to the study site; and alternative childcare (no children, children but with no alternative childcare, children with alternative childcare). The stepwise method was used with p-values to enter factors set at 0.05 and p-values to remove factors set at 0.06. All 95% confidence intervals (CIs) were calculated using the Clopper-Pearson method. All statistical analyses were performed using Statistical Analysis System version 9.1 or a later version.

This study identified factors influencing whether pregnant women would enrol in clinical trials, and would return to study sites for delivery, infant follow-up visits, or to seek medical assistance for their sick infant. By identifying restrictions to maternal/infant healthcare access in potential clinical sites, this study provided a potential platform for improving compliance and retention for future phase 3 clinical efficacy studies of investigational GBS vaccines. Apart from greater efficiency in clinical trials, there is evidence that ante- and post-partum visits improve maternal and infant health, and increase perinatal infant survival rates by offering opportunities for education and medical interventions. A greater understanding of predictors of pregnant women access to healthcare services may therefore also have programmatic implications for the improvement of healthcare utilisation.

Out of all pregnant women who were approached, 10% were not eligible for the study and could not participate. The reasons for not participating were estimated with the most common being lower than required gestational age. Proportions of ineligible women and reasons for ineligibility will be useful in the operational planning of clinical trials enrolling pregnant women, e.g. women with too low gestational age at screening could potentially participate if they return later during pregnancy.

Women at later gestational age at enrolment were more likely to deliver at the study site. In clinical trials, the gestational age is estimated to identify the appropriate time for maternal vaccination to ensure optimum placental antibody transfer to the infant [13]. Here, methods of estimating gestational age varied among sites, the most common being ultrasound or LMP in Latin America and South Africa, and fundal height in Mozambique. Ultrasound is considered the most accurate method, but is less available in resource-limited settings [13‐15]. The LMP method can be universally applied, but relies on self-assessment and reporting, and is potentially unreliable [14]. The fundal height method is routinely used in many resource-limited countries [16].

Older women were also more likely to return to study sites for delivery. This observation is in line with previous studies suggesting that women in their 30s attended antenatal healthcare more frequently than younger or older women, but contrasts with other studies suggesting that age had no effect on antenatal healthcare utilisation [17]. Another finding was that women with other children were more likely to return to study site for follow-up visits. A potential explanation for this unexpected finding could be that multiparous women often return to the site of their previous delivery for antenatal visits and are therefore more likely to return to the same site for delivery and follow-up visits.

Additionally, we identified factors that may hinder access to healthcare facilities for pregnant women and their infants. Mode, duration, and cost of transportation had a significant influence on whether women would return to study site for delivery and follow-up visits. Transportation was previously identified as a barrier for access to obstetric care in low- and middle-income countries, with increasing cost, distance, and travel time reducing the likelihood of returning [17‐20]. Surprisingly, we found that increased transportation costs at enrolment, which were lower than transportation costs at delivery, were associated with increased likelihood of women returning to study sites for delivery. This observation suggests that women who can afford to pay for transportation to antenatal visits are more likely to return to the same site for delivery. Another potential explanation could be that women living further from the study site may stay nearby when their due date was coming closer, which would then decrease the transportation costs at delivery. In contrast, lower transportation costs at delivery were associated with an increased likelihood of women returning to the study site for follow-up visits. Our results suggest that transportation cost is a barrier for routine follow-up visits, but not for hospital delivery, which is an important finding to consider when preparing clinical trials on maternal immunisation. In Africa, this finding should be interpreted with caution considering the differences in minimum salary in the respective countries [21].

Over 90% of women sought medical care at the study site or another healthcare facility for their sick infant, suggesting that the probability to detect potential GBS infections was high. However, more widespread efforts are needed to collect detailed follow-up information since only two thirds of live born infants were brought to the study site for the 90-day follow-up visits. Each site from a clinical study should establish surveillance in surrounding facilities where sick infants may be seen.

The study had a number of limitations. Although education about signs and symptoms of sepsis was provided to nearly all women, the study failed to clearly define what “sick” or “illness” meant, allowing thresholds for considering a child as ill that could differ by country. In addition, the incidence and causes of disease in infants were not collected as the majority of sick children were not assessed at the study site. When choosing sites for potential phase 3 trials on investigational maternal GBS vaccines, it is important to understand the neonatal invasive GBS disease incidence in the area; however, these estimates are lacking for many low- and middle-income countries [22]. The low proportion of women younger than 18 years of age (0.2%) was another potential limitation. Indeed, 16% of pregnant women in sub-Saharan Africa are under 19 years [23]; these young women are often at higher risk for GBS due to a lack of sexual health education [24] and a lower level of natural immunity against colonisation. Young pregnant women are an important group to study because they are often underrepresented in clinical trials due to legal or logistical reasons. This study was also limited by the fact that only healthcare sites large enough to conduct phase 3 clinical trials were used. These were in large urban areas, except the semirural site in Mozambique. Therefore, factors influencing whether women would return to healthcare facilities may vary for rural areas and smaller healthcare sites. The absence of data on the following variables was a further limitation of this study: the mode of delivery, the occurrence of complications during delivery, and the quality of the antenatal and birthing experience, which may all impact the likelihood of returning for a follow-up visit. Finally, failure to collect information on the impact of other important factors, such as education, parity, and income, was another limitation [25, 26].

In this prospective, multi-centre, observational study, older pregnant women and those at later gestational age at enrolment were more likely to deliver at the study site. Women with other children were more likely to return to study site for follow-up visits. While heterogeneities were observed at site level, shorter travel time at delivery and increased transportation costs at enrolment were associated with increased likelihood of women returning to study site for delivery. Lower transportation costs at delivery were associated with increased likelihood of women returning to study site for follow-up visits. Although high retention was observed from enrolment through 90 days after delivery, post-partum surveillance should be broadened beyond the study sites and an additional follow-up visit should be planned within the first month of life, the most vulnerable time for a child’s survival. The factors influencing maternal/infant access to healthcare facilities and the issues identified in this study should be taken into consideration in planning future clinical studies on maternal immunisation in low- and middle-income countries.