Background: The successful implementation of maternal vaccination relies on results of clinical trials, considering the prenatal and postnatal attendance at selected healthcare institutions. This study evaluated factors influencing maternal/infant access to healthcare facilities to identify potential barriers to participation in future clinical trials on maternal vaccination. Methods: In this prospective, multi-centre, observational study, pregnant women (N = 3243) were enrolled at ten sites across Panama, the Dominican Republic, South Africa, and Mozambique between 2012 and 2014. They completed questionnaires at enrolment, delivery, and infant follow-up (90 days post-partum) visits, including questions on transportation, phone accessibility, alternative childcare, gestational age at enrolment, delivery location, and health status of their infant. Logistic regression was used to identify factors significantly associated with return to study site for delivery or infant follow-up visits. Results: Among 3229 enrolled women with delivery information, 63.6% (range across sites: 25.3-91.5%) returned to study site for delivery. Older women and those at later gestational age at enrolment were more likely to deliver at the study site. 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. Among 3145 women with live-born infants, 3077 (95.3%) provided 90-day follow-up information; of these, 68.9% (range across sites: 25.6-98.9%) returned to study site for follow-up visits. Women with other children and with lower transportation costs at delivery were more likely to return to study site for follow-up visits. Among 666 infants reported sick, 94.3% were taken to a healthcare facility, with only 41.9% (range across sites: 4.9-77.3%) to the study site. Conclusion: Although high retention was observed from enrolment through 90 days after delivery, post-partum surveillance should be broadened beyond the study sites and additional follow-up visits should be planned within the neonatal period. 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. Trial registration: The study was registered at ClinicalTrial.gov ( NCT01734434 ) on November 22, 2012.
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. Demographics and baseline characteristics of women included in the full analysis set N number of women per group, n (%) number (percentage) of women in each category, SD standard deviation 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 ({“type”:”clinical-trial”,”attrs”:{“text”:”NCT01734434″,”term_id”:”NCT01734434″}}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.