What is my baby’s Kaiser sepsis risk?
Enter the birth details and the mother’s risk factors to estimate a newborn’s early-onset sepsis (EOS) risk per 1,000 births with the Kaiser sepsis calculator models (2024 or the original 2017), at birth and after the baby’s exam. For information only. This is not medical advice or a diagnosis. Do not use it to make treatment decisions without a clinician. No licensed clinician has reviewed this page.
- Risk at birth
- 0.29
The early-onset sepsis risk at birth is 0.29 per 1,000 births (0.1 per 1,000 if the baby appears well).
- If the baby appears well
- 0.1
- If the exam is equivocal
- 1.05
- If the baby shows clinical illness
- 4.17
Risk at birth: 0.29. The early-onset sepsis risk at birth is 0.29 per 1,000 births (0.1 per 1,000 if the baby appears well).
This calculator gives general information only. It is not medical advice and does not replace a doctor or other health care provider. Talk with a professional about your results. Terms of use
How to calculate
Estimates a newborn’s risk of early-onset sepsis per 1,000 births from maternal factors with the Kaiser Permanente models (2024 update for universal GBS screening, or the original 2017 model), and after a well, equivocal, or ill clinical exam.
Example with the default inputs (Kaiser model 2024: universal GBS screening, EOS incidence (per 1,000 births) 0.5 per 1,000, Gestational age: weeks 40, and days 0, Highest maternal temperature 98.6 °F, Rupture of membranes (hours) 10, Mother’s GBS status Negative, Antibiotics in labor None, or any given less than 2 hours before birth): The early-onset sepsis risk at birth is 0.29 per 1,000 births (0.1 per 1,000 if the baby appears well).
Method: Risk at birth = 1 ÷ (1 + e^−x), x = intercept + b1 × temperature (°F) + b2 × GA + b3 × GA² + b4 × (ROM hours + 0.05)^0.2 + b5 × (GBS prophylaxis ≥ 2 h or broad-spectrum 2 to 3.9 h) + b6 × (broad-spectrum ≥ 4 h) + b7 × GBS positive + b8 × GBS unknown. 2024 model: 0.85194656, −7.72247124, 0.09842383, 0.86770862, −2.13142945, −2.33985917, 1.02265353, 1.13710111, exam LRs 0.36, 3.65, 14.5. 2017 model: 0.868, −6.9325, 0.0877, 1.2256, −1.0488, −1.1861, 0.5771, 0.0427, exam LRs 0.41, 5, 21.2. After the exam: odds × LR.
- The 2024 model is for babies born at 35 weeks or later where GBS screening is universal; the 2017 model is for 34 weeks or later where it is not (Kaiser 2024 model FAQ).
- Risks are per 1,000 live births, rounded to 2 decimals, halves up.
- The page shows risk only. It does not show the care recommendations of the Kaiser tool.
Worked examples
Each example is checked against the calculator on every build.
- Kaiser model 2024: universal GBS screening, EOS incidence (per 1,000 births) 0.5 per 1,000, Gestational age: weeks 40, and days 0, Highest maternal temperature 98.6 °F, Rupture of membranes (hours) 10, Mother’s GBS status Negative, Antibiotics in labor None, or any given less than 2 hours before birth gives Risk at birth 0.29, If the baby appears well 0.1, If the exam is equivocal 1.05, If the baby shows clinical illness 4.17.Source: Kaiser 2024 model FAQ: intercept 57.29929499 − ln[(0.9995 ÷ 0.0005) × (0.0002763 ÷ 0.9997237)] = 57.8926
- Kaiser model 2024: universal GBS screening, EOS incidence (per 1,000 births) 0.5 per 1,000, Gestational age: weeks 39, and days 3, Highest maternal temperature 101.5 °F, Rupture of membranes (hours) 20, Mother’s GBS status Positive, Antibiotics in labor GBS prophylaxis, 2 hours or more before birth gives Risk at birth 1.31, If the baby appears well 0.47, If the exam is equivocal 4.76, If the baby shows clinical illness 18.63.Source: Kaiser 2024 model: 39 weeks 3 days, 101.5 °F, 20 hours, GBS positive, GBS prophylaxis: x = −6.6382, 1.3077 per 1,000
- Kaiser model 2024: universal GBS screening, EOS incidence (per 1,000 births) 0.3 per 1,000, Gestational age: weeks 37, and days 0, Highest maternal temperature 100.2 °F, Rupture of membranes (hours) 30, Mother’s GBS status Unknown, Antibiotics in labor Broad-spectrum, 4 hours or more before birth gives Risk at birth 0.44, If the baby appears well 0.16, If the exam is equivocal 1.6, If the baby shows clinical illness 6.32.Source: Kaiser 2024 model, 0.3 per 1,000 (intercept 57.3816, the FAQ table gives 57.4): x = −7.7316, 0.4386 per 1,000
- Kaiser model 2017: no universal screening, EOS incidence (per 1,000 births) 0.5 per 1,000, Gestational age: weeks 40, and days 0, Highest maternal temperature 98.6 °F, Rupture of membranes (hours) 10, Mother’s GBS status Negative, Antibiotics in labor None, or any given less than 2 hours before birth gives Risk at birth 0.14, If the baby appears well 0.06, If the exam is equivocal 0.69, If the baby shows clinical illness 2.93.Source: Puopolo 2011 model with the coefficients in Nguyen 2024 (Cureus) and the Kaiser 0.5 per 1,000 intercept 40.5656: 0.1384 per 1,000
- Kaiser model 2017: no universal screening, EOS incidence (per 1,000 births) 0.5 per 1,000, Gestational age: weeks 39, and days 3, Highest maternal temperature 101.5 °F, Rupture of membranes (hours) 20, Mother’s GBS status Positive, Antibiotics in labor GBS prophylaxis, 2 hours or more before birth gives Risk at birth 1.4, If the baby appears well 0.57, If the exam is equivocal 6.95, If the baby shows clinical illness 28.84.
- Kaiser model 2017: no universal screening, EOS incidence (per 1,000 births) 0.3 per 1,000, Gestational age: weeks 35, and days 0, Highest maternal temperature 100.2 °F, Rupture of membranes (hours) 30, Mother’s GBS status Unknown, Antibiotics in labor Broad-spectrum, 4 hours or more before birth gives Risk at birth 1.01, If the baby appears well 0.41, If the exam is equivocal 5.01, If the baby shows clinical illness 20.89.
- Kaiser model 2017: no universal screening, EOS incidence (per 1,000 births) 0.6 per 1,000, Gestational age: weeks 41, and days 2, Highest maternal temperature 101.3 °F, Rupture of membranes (hours) 5, Mother’s GBS status Negative, Antibiotics in labor None, or any given less than 2 hours before birth gives Risk at birth 1.73, If the baby appears well 0.71, If the exam is equivocal 8.61, If the baby shows clinical illness 35.5.
Medical disclaimer
This calculator is for information and education only. It is not medical advice, and its result is not a diagnosis. Do not use it to start, stop, or change a treatment without a clinician who knows your health history. No licensed clinician has reviewed this page. This site's editors checked the formulas and worked examples against published sources. If you think you have a medical emergency, call your local emergency number.
When this does not apply. The models are for newborns in the first days of life: the 2024 model from 35 weeks, the 2017 model from 34 weeks (Kaiser 2024 model FAQ; Puopolo 2011). They do not apply to babies born earlier or to later infections. Choose the 2024 model where pregnant people have universal group B streptococcus (GBS) screening, as in the US, and the 2017 model where they do not; the wrong model can misstate the risk when the GBS status is unknown. The page gives no care recommendations.
How the risk is worked out
The page has both Kaiser Permanente models. Both use the same inputs and the same form of equation; the coefficients, intercepts, and exam likelihood ratios differ.
- 2024 model (the default). Kaiser's update, built from 412,543 births in 2010 to 2020, after universal GBS screening began (Kuzniewicz 2024; Kaiser 2024 model FAQ). Kaiser suggests it where GBS screening is universal, as in the US. It is for babies born at 35 weeks or later, and only ampicillin, penicillin, or cefazolin count as GBS prophylaxis.
- 2017 model (the original). Puopolo 2011 and Escobar 2014, for babies born at 34 weeks or later. Kaiser suggests it where GBS screening is not universal.
Risk at birth, 2024 model (Kaiser 2024 model FAQ)
x = intercept + 0.85194656 × T − 7.72247124 × GA + 0.09842383 × GA² + 0.86770862 × (ROM + 0.05)^0.2 − 2.13142945 × A1 − 2.33985917 × A2 + 1.02265353 × G+ + 1.13710111 × G?
risk at birth = 1 ÷ (1 + e^−x), with T, GA, ROM, A1, A2, G+, and G? as defined for the 2017 model below. A1 counts GBS prophylaxis with ampicillin, penicillin, or cefazolin only; GBS prophylaxis given with an aminoglycoside, aztreonam, azithromycin, or metronidazole counts as broad-spectrum.
The intercept is 57.29929499 at the Kaiser Northern California rate ȳ = 0.0002763 (0.2763 per 1,000). For another incidence τ (per birth), Kaiser gives intercept = 57.29929499 − ln[((1 − τ) ÷ τ) × (ȳ ÷ (1 − ȳ))]. This gives, for example, 57.3816 at 0.3 per 1,000 and 57.8926 at 0.5 per 1,000; Kaiser's table prints the same values to one decimal (57.4 and 57.9) for all 14 rates it lists.
Exam likelihood ratios (the point estimates in Kaiser's 2024 model update): well appearing 0.36, equivocal 3.65, clinical illness 14.5.
Risk at birth, 2017 model (Puopolo 2011)
x = intercept + 0.8680 × T − 6.9325 × GA + 0.0877 × GA² + 1.2256 × (ROM + 0.05)^0.2 − 1.0488 × A1 − 1.1861 × A2 + 0.5771 × G+ + 0.0427 × G?
risk at birth = 1 ÷ (1 + e^−x)
- T: the mother’s highest temperature in labor in °F. A temperature in °C is converted: T = °C × 9/5 + 32.
- GA: gestational age in weeks, weeks + days ÷ 7 (39 weeks 3 days = 39.4286).
- ROM: hours from rupture of membranes to birth.
- A1 = 1 for GBS-specific antibiotics 2 hours or more before birth, or broad-spectrum antibiotics 2 to 3.9 hours before birth; otherwise 0.
- A2 = 1 for broad-spectrum antibiotics 4 hours or more before birth; otherwise 0. With no antibiotics, or any given less than 2 hours before birth, A1 = A2 = 0.
- G+ = 1 if the mother is GBS positive; G? = 1 if her GBS status is unknown.
- Intercept by the baseline EOS incidence per 1,000 live births, as in the Kaiser calculator:
| Incidence per 1,000 | Intercept |
|---|---|
| 0.1 | 38.952265 |
| 0.2 | 39.646367 |
| 0.3 | 40.0528 |
| 0.4 | 40.3415 |
| 0.5 | 40.5656 |
| 0.6 | 40.7489 |
| 0.7 | 40.903919 |
| 0.8 | 41.0384 |
| 0.9 | 41.1571 |
| 1 | 41.263432 |
| 2 | 41.965852 |
| 4 | 42.676976 |
The coefficients are those of the Kaiser model as printed by Nguyen and others (2024) and in the Kaiser site's EMR implementation FAQ, which also gives the (ROM + 0.05)^0.2 term, the temperature in °F, and the likelihood ratios. They match the adjusted odds ratios in Puopolo 2011, Table 5 (for example e^0.868 = 2.38 per °F and e^1.2256 = 3.41 for the ROM term).
Risk after the exam (Escobar 2014), 2017 model
posterior odds = odds at birth × LR, where odds at birth = p ÷ (1 − p); posterior risk = posterior odds ÷ (1 + posterior odds).
| Exam | Likelihood ratio (LR) |
|---|---|
| Well appearing | 0.41 |
| Equivocal | 5 |
| Clinical illness | 21.2 |
These are the upper limits of the 95% confidence intervals in Escobar 2014, the values the original calculator uses. All risks are shown per 1,000 live births (risk × 1,000), rounded to 2 decimal places, halves up, read from the decimal value (within 1e-9, relative, of a half counts as the half).
The Kaiser tool also gives care recommendations for clinicians. This page does not.
Inputs the calculator accepts
Gestational age 34 weeks 0 days to 43 weeks 6 days; temperature 35 to 42 °C (95 to 107.6 °F); ROM 0 to 240 hours.
Worked examples by hand
2024 model, 40 weeks, 98.6 °F, ROM 10 hours, GBS negative, no antibiotics, incidence 0.5 (the page default). Intercept 57.8926. x = 57.8926 + 84.0019 − 308.8988 + 157.4781 + 0.86770862 × 1.5865 (1.3766) = −8.1496 (from the unrounded terms). Risk = 1 ÷ (1 + e^8.1496) = 0.29 per 1,000; well appearing (odds × 0.36) 0.10; equivocal (× 3.65) 1.05; clinical illness (× 14.5) 4.17.
2024 model, 39 weeks 3 days, 101.5 °F, ROM 20 hours, GBS positive, GBS prophylaxis 2 hours or more, incidence 0.5. 1.31 per 1,000; 0.47 if well, 4.76 if equivocal, 18.63 if ill. The same baby scores 1.40 per 1,000 with the 2017 model.
2024 model, 37 weeks, 100.2 °F, ROM 30 hours, GBS unknown, broad-spectrum antibiotics 4 hours or more, incidence 0.3. 0.44 per 1,000; 0.16, 1.60, 6.32.
The next four use the 2017 model.
40 weeks, 98.6 °F, ROM 10 hours, GBS negative, no antibiotics, incidence 0.5. x = 40.5656 + 85.5848 − 277.3 + 140.32 + 1.2256 × 10.05^0.2 (1.5864 × 1.2256 = 1.9443) = −8.8852. Risk = 1 ÷ (1 + e^8.8852) = 0.0001384 = 0.14 per 1,000. Well appearing: odds 0.0001384 × 0.41 → 0.06; equivocal × 5 → 0.69; clinical illness × 21.2 → 2.93.
39 weeks 3 days, 101.5 °F, ROM 20 hours, GBS positive, GBS-specific antibiotics 2 hours or more, incidence 0.5. 1.40 per 1,000 at birth; 0.57 if well, 6.95 if equivocal, 28.84 if ill.
35 weeks, 100.2 °F, ROM 30 hours, GBS unknown, broad-spectrum antibiotics 4 hours or more, incidence 0.3. 1.01 per 1,000; 0.41, 5.01, 20.89.
41 weeks 2 days, 38.5 °C (101.3 °F), ROM 5 hours, GBS negative, no antibiotics, incidence 0.6. 1.73 per 1,000; 0.71, 8.61, 35.50.
Other questions people ask
What is the Kaiser sepsis calculator?
It is a model from Kaiser Permanente Northern California that estimates the chance that a baby born at 34 or 35 weeks or later has early-onset sepsis, a bacterial blood infection in the first days of life. Puopolo and others (2011) built the risk at birth from maternal factors; Escobar and others (2014) added the baby’s condition. When Kaiser hospitals used it, antibiotic use in the first 24 hours fell from 5.0% to 2.6% of newborns (Kuzniewicz 2017).
What goes into the risk at birth?
Five facts known at delivery: gestational age, the mother’s highest temperature in labor, how long the membranes (waters) were ruptured, the mother’s group B streptococcus (GBS) status, and the type and timing of antibiotics in labor. The model also needs the baseline rate of EOS where the baby is born.
How does the baby’s exam change the risk?
The risk at birth is multiplied, as odds, by a likelihood ratio for the exam. The 2024 model uses the point estimates from the 2024 Kaiser model update: 0.36 if the baby appears well, 3.65 if the exam is equivocal, and 14.5 if the baby shows clinical illness. The original 2017 model used the upper 95% limits: 0.41, 5, and 21.2.
Does this page say whether a baby needs antibiotics?
No. The Kaiser tool pairs each risk with care recommendations for clinicians. This page shows the risk numbers only. Decisions about tests, monitoring, and antibiotics belong to the baby’s care team.
Which model should I use?
Kaiser suggests the 2024 model where pregnant people have universal group B streptococcus (GBS) screening, as in the US, and the original 2017 model where they do not. In the 2024 model an unknown GBS status triples the odds (odds ratio 3.11), because few mothers are untested where screening is universal; where screening is not universal, that would overstate the risk. The 2024 model is for 35 weeks or later; for a baby born at 34 weeks, use the 2017 model.
What incidence should I choose?
The incidence is the baseline rate of early-onset sepsis per 1,000 live births at the hospital or in the region. The Kaiser tool names 0.3 per 1,000 for Kaiser Northern California and 0.5 per 1,000 as the US national rate.