GBS colonization rates of 14.96% were based on weighted ethnic specific colonization rates (n= 1,329) from two recent Israeli studies [1, 2]. Based on data from the Israel Center for Disease Control and the Ministry of Health, the incidence rate of early onset GBS in 2005-2006 in Israel was 0.315 per 1000 births and 0.253 per 1000 births in 2008-9.
Assuming sensitivity and specificities for culture screening of 0.95 and 0.97 respectively , we back-calculated, the underlying natural incidence (null scenario) of early onset GBS, under the baseline assumption in 2008-2009, that 15% of the mothers-to-be nationwide received culture screening on an opportunistic basis (see Appendix 1 for full details).
Finally we estimated the incidence in a future intervention scenario where 85% of mothers-to-be would receive culture screening, a figure currently attained in the USA (3). Of the 15% who will not be screened, around half were unscreened due to pre-term births .
We assumed that mothers who were given a scheduled C-section birth would not receive IAP. Currently 18.9% of Israeli births [4, 5] are delivered by C-section of which 30% are scheduled in advance and so would not receive IAP.
Based on evidence from a Cochrane meta-analysis , we assumed that IAP with penicillin would prevent 83% of the true positive identified cases from giving birth to a child infected with GBS. We assumed that 0.022% of the mothers who received IAP would develop an anaphylactic reaction , with an attendant case fatality rate (CFR) of 10% .
In 2010, there were 166,184 live births (including 3,784 twins and 77 other multiple births) and 501 still births, resulting from 162,740 deliveries in Israel . By multiplying the birth rate by the GBS incidence rate and adjusting for the efficacy of the intervention, we calculated the number of infants born with GBS in the following three scenarios: the null scenario, the present scenario and the future intervention scenario, where 85% of mothers to be would be culture-screened at 35–37 weeks.
Ninety percent of the children with EON GBS, were assumed to suffer from only sepsis, whilst the remaining 10% had meningitis in addition to the sepsis, based on the past ten years experience of the Shaare Zedek Hospital in Jerusalem.
The reported range of the percentage of meningitis cases that develop long term neurological sequelae increased from 28.4% (n=218) in studies from 1973-1985 [10–17] to 47.5% (n=141) in studies reported from 2000-2012 [18, 19]. As no studies were identified during the period 1986-1999, the baseline estimates used in our model were based on the 47.5% figure from the two recent studies from the 21st century [18, 19]. In addition, we conducted a sensitivity analysis over a range of values suggested by the literature.
The neurological sequelae consisted of severe (31.3% of all neurological sequellae), moderate (33.4%) and mild (35.3%) neurological disorders [18, 19]. In addition a further 4.6% of meningitis cases suffered from severe or profound bilateral hearing loss .
Costs of intervention
No additional marginal costs were imputed for the taking of the culture as this was considered to be part of the routine examination of the mother-to-be. Laboratory costs of analyzing the culture test (including the culture tube) were 25.62 NIS per test. No provision was made for physician labour time as this was considered to be part of the routine pre-natal check-up in Israel.
Penicillin costs were based on an average of two doses (given four hours apart) per birth (average delivery time 6–8 hours) of 5 million units of benzyl penicillin costing 11 NIS per 10mu vial (Personal Communication: Alan Greenberg, Chief Pharmacist, Shaare Zedek Hospital, Jerusalem).
No extra marginal cost was incurred for nurses time in setting up the IV drip.
Costs of anaphylactic shock, GBS cases and sequelae
Each fatal anaphylactic shock case was assumed to cost one day's stay in an Intensive Care Unit (5,142 NIS). Non-Fatal anaphylactic cases were assumed to spend an extra two days in hospital  costing 4,208 NIS.
Hospital costs (in relation to non-GBS births) of 29.458 NIS and 37,872 NIS for GBS sepsis and sepsis with meningitis cases were based on 14 and 18 days stay respectively (obtained from an analysis of the case-notes of 24 infants hospitalized for GBS in Shaare Zedek Hospital, Jerusalem) in a quasi- Intensive Care Unit (ICU) at 2,106 NIS per day (Ministry of Health, Price List 2010). Fatal cases were assumed to cost the equivalent of one days stay (5,142 NIS) in an ICU.
Post hospital care (e.g. outpatient follow up visits) was assumed to amount to 36.0% of in-hospital care costs . Both in- and post-hospital costs for GBS children were adjusted by factors of 38.8% and 21.0% respectively in order to estimate the additional costs for GBS children compared with births to children without GBS .
Costs of acute care were calculated by multiplying the expected number of cases by the unit costs of health services received. For chronic sequelae (deafness, and brain damage), the expected age-specific number of cases was multiplied by the age-specific discounted lifetime costs.
Costs of treating long-term sequelae were based on lifetime costs discounted at 3% per annum. Data on costs of deafness were retrieved from a recent cost-utility analyis of cochlear inplantsa. These costs included cost of aids or implants, hearing tests, ear moulds, special therapy, school visits, acoustic classrooms, amplification for teachers, remedial education for children with and without cognitive complications.
Lifetime costs were estimated to be around 2,640,519 NIS, 1,281,790 NIS and 316,180 NIS respectively for severe, moderate and mild neurological brain damageb. These costs included initial diagnostics and care costs, care costs in subsequent years, additional medical costs and special education costs.
Age-specific lifetime costs of caring for brain damage (diagnosis, medical care, special education, rehabilitative day care, residential care) were obtained from the ministry of healtha,b and the ministry of social affairsc.
Around 75% and 5% of severe and moderate-mild brain damaged persons were assumed to be cared for in residential care centers, reflecting the trend of de-institutionalisation  at a cost of around 105,000 NIS per year.
Since the analysis was from a social perspective, estimates of lost productivity due to sequelae were included and based on the assumption that persons with moderate or severe sequelae did not participate in the workforce. Hearing challenged persons and persons with mild sequelae were assumed to participate fully in the labor force. Average discounted lifetime employment costs of 1,102,287 NIS per person were calculated from national average gross wages  increased by a factor of 15% to include employers pension, national insurance and educational fund contributions and by age-specific unemployment rates .
QALY losses due to morbidity of GBS caused sepsis and sepsis-meningitis cases were calculated by multiplying the expected number of cases by the expected duration of the illness (14 days and 18 days respectively), and the disability weights of 0.31  and 0.62  respectively attached to acute illness and the age-specific average QALY weight of an infant not affected by GBS sepsis or meningitisa.
QALY losses due to morbidity from chronic sequelae from meningitis were calculated by multiplying the expected number of sequelae by the average lifelong disability weight of 0.964 [26–28] attached to the sequelae and the age-specific average QALY weight of a person not affected by GBS meningitis . The results were discounted at a rate of 3% per annum over the persons remaining life expectancy .
QALY losses due to morbidity from anaphylactic reactions were based on an assumed disability weight for meningitis (0.61) and the average duration of acute reaction was assumed to be one day.
QALY losses due to mortality were estimated from the product of the following:
the number of incident cases
case-fatality rate of 2.8% representing the excess mortality in GBS as opposed to non-GBS cases [29, 30] or the 10% case fatality rate from anaphalactic reactions .
age and gender specific life expectancies at birth of 80.2 for males and 82.1 for females  in Israel,
age-specific QALY weights of a healthy persona.
A spreadsheet model was built incorporating technical, epidemiological, health service utilization and costs, demographic and economic data described above.
The cost utility ratio calculated the net costs per Quality Adjusted Life Year (QALY) added of the intervention of universal screening followed by IAP prophylactics, using the formula.
All costs are at presented in mid-2010 price levels, at the average annual exchange rate of 3.588 shekels to the US dollar . Costs are viewed from a societal perspective (i.e. including estimates of lost productivity in addition to health and welfare services costs).
Estimates of QALYs added by the intervention do not include those arising from reduced caregiver burden for sequelae since such data is not available. All future costs and QALYS were discounted at an annual rate of 3%.
Taking into account the resources available in Israel, an intervention is defined as being very cost-effective and cost-effective if the cost per QALY is less than the per capita GNP of 106,548 NIS in 2010  or between 1–3 times the per capita GNP (106,548-319,644 NIS) respectively. If the cost per QALY is more than three times the GNP per capita (319,644 NIS) then the intervention is regarded as not being cost-effective .
Averted QALY losses are calculated by summing the mortality and morbidity gains from decreased incidence of GBS as a result of the universal screening intervention.