Aims Determining which babies should receive antibiotics for possible early onset sepsis (EOS) is challenging. Kaiser Permanente have developed an electronic calculator providing an individualised estimation of EOS risk.

We performed a meta-analysis quantifying how many culture positive EOS cases might be ‘missed’ using the calculator, in addition to cases missed using National Institute for Health and Care Excellence (NICE) guidelines (CG149 2012).

Methods A systematic literature search using a modified cluster technique, snowballing from studies citing the article in which the calculator was widely publicised (Kuzniewicz 2017) on Ovid MEDLINE, Embase, Maternity & Infant Care Database and Google scholar. Reference lists were reviewed.

Studies were eligible if they presented data evaluating the calculator, either by retrospective case review or prospective cohort study and identified at least one episode of EOS.

The primary outcome measure was numbers of culture positive EOS cases where the calculator did not recommend empirical antibiotics. If the NICE guidelines would not have recommended treatment either this was not classified as a ‘miss’. Risk of bias was assessed using QUADAS-2.

Data were pooled using a random effect meta-analysis, quantifying heterogeneity using I². A subgroup analysis was performed using data from studies of babies exposed to chorioamnionitis.

Results Eleven studies were eligible. There were 75 EOS cases and a minimum of 14, and a maximum of 22 cases where use of the calculator would have resulted in delayed or missed treatment, compared to if NICE guidelines were followed.

The probability of ‘calculator’ delayed or missed treatment for an EOS case (additional to cases missed by following NICE guidelines) were best case 0.19 [95% confidence intervals 0.11 – 0.29,I² 0%], worst case 0.31 [95% CI 0.17 – 0.49, I² 37%].

The probability of missing cases was significantly (p=0.03) more in babies exposed to chorioamnionitis, up to 0.56 [95% CI: [0.25, 0.82], I² 0%].

All included studies had a low/moderate probability of bias.

Conclusion A substantial proportion of EOS cases were missed by the calculator. Further evaluation of the calculator is recommended before it could be safely introduced into UK clinical practice.