Characteristics of sites and participants

Parts of this chapter have been reproduced from Daniels et al.1 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: https://creativecommons.org/licenses/by/4.0/. The text below includes minor additions and formatting changes to the original text.

Accrual period

The first site opened to recruitment on 26 July 2017, and the trial closed to recruitment on 30 April 2019. Ten usual-care sites and eight rapid test sites met or exceeded their recruitment target of 86 women. As the sample size requirements were met (i.e. at least eight sites in each arm recruited 86 women per site), the Trial Steering Committee requested that recruitment ended.

Site-level randomisation

Overall, 22 maternity units agreed to participate and were randomised to usual-care or rapid test pathways. Initially, we randomised 20 sites (see Appendix 1). The allocations were revealed to the units at the same time, after all units had commited to participation. Following randomisation, two sites, one allocated to each strategy, requested withdrawal from the study. One usual-care site withdrew as it lost a key staff member and one rapid test site was not able to implement the rapid test using the no-consent model. We replaced these with two other sites that were randomised to usual-care and rapid test strategies, respectively. The anonymised maternity unit characteristics that were used within the minimisation algorithm are shown in Table 3. The units had around 2000–6000 births per year that were not planned elective caesareans. There was variation in the reported rates of IAP to prevent early-onset GBS infection.

Individual-level accrual

The cumulative rate of accrual of women into the GBS2 trial, by screening pathway, is shown in Appendix 2. The date on which the woman was identified as having GBS risk factors was used as the date of study entry. The weekly total number of intended vaginal deliveries was requested from each unit to enable monitoring of the accrual rate. In the early phase of the trial, usual-care units were allowed to exceed their target of 86 women and continue to 100 women per unit. This was curtailed after the second unit reached a total of 100 women. At the end of the recruitment period, when sample size requirements were met, recruitment at two rapid test units was stopped: in one when 83 women had been recruited and in one after three women had been recruited. An estimate of the proportion of women recruited into the study of the total from the population giving birth vaginally or by emergency caesarean (using data from Table 3, pro rata for the duration of participation) is shown in Table 4.

Characteristics of the included sites and participants

Maternity sites

Following randomisation, the maternity units were balanced according to the minimisation characteristics (Table 5).

Participants

Overall, 1687 women were identified and added to the trial database during the recruitment window for the 20 randomised sites. There were 52 women from two usual-care maternity units who were added to the database but who were ineligible for the study, as the database eligibility criteria initially did not reflect the protocol. A further three women in each group who, on closer examination, were added to the database in error, as they were not eligible and one woman’s record was duplicated. These women’s data were excluded from the analysis, leaving 1628 women in the data set.

The overall numbers of women recruited were 722 in rapid test sites and 906 in usual-care sites. There were 67 pairs of twins and three trios of triplets and, therefore, the number of babies available for assessment was 749 in the rapid test group and 951 in the usual-care units (Figure 7). Similar proportions of women in the two study groups were recruited from the two regions, and the women were similar in their parity and type of birth. A higher proportion of women in the rapid test sites entered labour after being induced than in the usual-care units (Table 6).

FIGURE 7 The CONSORT flow diagram of sites and participants in the GBS2 study. (Figure not reproduced here; see the original article.)

In 93% of women there was one risk factor for neonatal GBS disease, and two risk factors in 6–7% of women. The most common risk factor was preterm labour. The distribution of risk factors among the study population is shown in Tables 7 and 8.

Among those women with only one risk factor, more women in rapid test units had GBS detected in the current pregnancy (41%, 293/722) than those in the usual-care units (31%, 278/906). In addition, 15% (139/906) of women in usual-care units had a raised maternal temperature compared with 8% (55/722) of women in rapid test units. GBS was detected in the vaginal or rectal swab commonly in both groups, followed by diagnosis in the midstream sample (Table 9).

Compliance to group allocation and test strategy

At a site level, complete compliance was achieved because the rapid test machines were supplied only to those sites allocated to the rapid test. In 53 women (7.35%), a swab was taken but the test failed to yield a result because of equipment problems; half of these problems arose in two maternity units where the GeneXpert machine developed an intermittent fault during the recruitment period. A further 23 tests were considered invalid, as > 15 minutes had passed between swabbing and initiating the test, with other reasons or no documented reason for the test failing in 24 attempts. There were 56 women who should have had a second test because their labours failed to progress within 48 hours, but who were missed. Among individual participants, for three women in the rapid test strategy group it is unknown whether or not a swab was taken and no test results are available. We were unable to calculate the duration between the test becoming available, the midwife collecting the result and the IAP being administered.

Of the 241 women in rapid test units who tested positive, 79% (190/241) received IAP to prevent GBS vertical transmission and a further 8% received antibiotics for other reasons. Of the 316 women who tested negative for GBS, 17% received IAP to prevent GBS and a further 39% did so for other indications. Table 10 shows the adherence to the test result.

Table 11 provides the various indications for IAP for positive and negative rapid tests and women with failed or missed test.