Background
In high income countries, ~30% of pregnant women are provided with intrapartum antibiotic prophylaxis (IAP) for early‐onset group B streptococcal infection (EOGBSI). The infection rate is low, 0.2/1000 livebirths in our jurisdiction, and others. We hypothesised factors, other than IAP alone, were reasons for the low rate of EOGBSI.

Aims
Compliance with our local guideline, referred to here as ‘the guideline’.

Method
Compliance was defined as an initial dose of benzylpenicillin IAP followed by four‐hourly doses until birth. The study population was drawn from 4098 women who had 4100 pregnancies resulting in 4200 babies in an Australian birth setting from 1/1/2016 to 31/12/2016. Most, 93%, were eligible for universal GBS screening, 67% were reported as screened and 90% of these had a result documented; 23% were positive for GBS. A random sample (n = 223) was taken for further analysis.

Results
The adjusted odds of receiving benzylpenicillin IAP in accord with the guideline were three times higher among primiparous compared to multiparous women (P < 0.001, odds ratio (OR) = 3.4, 95% CI 1.7–6.7) and three times higher among women experiencing induction of labour compared to women who commenced labour spontaneously (P < 0.001, OR = 3.4, 95% CI 1.8–6.3). Of the 223 women, 188 received IAP: 176 received benzylpenicillin IAP, 31% (or 24% of the total sample) received this intervention in accord with the guideline, 24% received benzylpenicillin ≥4 h before birth but not in accord with the guideline and 44% received benzylpenicillin <4 h before birth. Conclusion We conclude that sub‐optimal compliance was largely a consequence of an unrealistic guideline.