An analysis of the cost-effectiveness of selected protocols for the prevention of neonatal group B streptococcal infection.

Abstract

To determine the expected neonatal outcome in a hypothetical obstetric population with various screening and intrapartum management protocols for the prevention of neonatal group B streptococcal infections. We used decision analysis to investigate the performance of various antenatal and intrapartum group B streptococcal screening protocols combined with selective or universal intrapartum antibiotic prophylaxis in preventing neonatal early-onset group B streptococcal disease. Population characteristics, screening test performance, and treatment efficacy profiles were abstracted from previous investigations. In the absence of screening or treatment, the estimated neonatal attack rate and mortality rate were 3.6 and 0.60 cases per 1000 neonates, respectively. Universal antenatal screening at 26-28 weeks' gestation and selective intrapartum prophylaxis (treatment of colonized women with risk factors) or nonselective prophylaxis (treatment of all colonized women) had attack rates of 1.64 and 1.08 and mortality rates of 0.21 and 0.18 cases per 1000 neonates, respectively. Rapid intrapartum screening and selective prophylaxis had an estimated attack rate of 1.92-2.58 and a mortality rate of 0.26-0.40 cases per 1000 neonates, with results dependent upon the time required to attain test results. Similarly, rapid intrapartum screening and nonselective prophylaxis had an attack rate of 1.44-2.30 and a mortality rate of 0.24-0.38 cases per 1000 neonates. Empirical prophylaxis of all women delivering prematurely, regardless of culture status, had an attack rate of 2.40 and mortality rate of 0.36 cases per 1000 neonates. Under ideal circumstances, antenatal group B streptococcal screening at 26-28 weeks' gestation with vaginal-rectal cultures and subsequent intrapartum prophylaxis is associated with lower projected neonatal attack and mortality rates than protocols using rapid intrapartum screening methods. However, in a non-research population, difficulties frequently encountered with antenatal screening procedures may minimize these differences.