NICHD Extremely Preterm Birth Outcomes Estimator

Clinical Overview

The NICHD Extremely Preterm Birth Outcomes Estimator is a prognostic tool that predicts survival and intact survival (survival without severe disability) for infants born at 22–25 weeks gestation. The calculator provides probability estimates specific to each infant's characteristics, enabling evidence-based periviable birth counseling for families facing extraordinarily difficult decisions.

What It Measures

The estimator calculates two primary outcomes:

1. Any survival: Probability of hospital discharge alive, regardless of disability status
2. Intact survival: Probability of survival without severe neurodevelopmental disability (IVH grade 3–4, cystic PVL, or disability at 18–22 months)

The tool generates separate estimates for:

• Survival with any impairment: Probability of survival with moderate-to-severe disability
• Death: Probability of in-hospital death
• Severe disability without survival measurement: Lower-bound estimate (some survivors' disability status unknown)

Why It Was Developed

The birth of a preterm infant at 22–25 weeks gestation raises profound ethical and prognostic questions with no universally accepted answers. Historically, prognostic estimates have been:

• Derived from single centers with different practices and case mixes
• Based on older data (pre-surfactant era or early post-surfactant practices)
• Inconsistent in outcome definitions (neuroimaging vs. formal neurodevelopmental testing vs. parental assessment)
• Not individualized to patient characteristics (all periviable infants given identical "population" estimates)

The National Institute of Child Health and Human Development convened multicenter research networks to prospectively collect outcomes data on periviable births (1998–2014 across multiple cohorts), enabling development of individualized, contemporary prognostic models with large sample sizes and rigorous neurodevelopmental follow-up.

Clinical Problem It Solves

Periviable birth counseling represents one of the most ethically charged conversations in obstetrics and neonatology. Parents must decide whether to pursue active resuscitation and intensive care versus comfort-focused care for a newborn with:

• Uncertain survival prospects (historically 0–50% depending on birth weight)
• High disability risk even among survivors (20–50% major disability)
• Unknown future quality of life and parental capacity to care
• Profound psychosocial, financial, and spiritual implications

Without accurate prognostic data, counseling defaults to subjective impressions, catastrophizing assumptions, or overly optimistic statements—all of which undermine informed decision-making. The NICHD estimator provides:

• Contemporary data reflecting modern perinatal care
• Individualized estimates incorporating infant and maternal factors
• Clear outcome definitions with detailed follow-up (neuroimaging, developmental testing)
• Quantified uncertainty (confidence intervals showing range of possibilities)

This enables families to make decisions aligned with their values, rather than decisions based on misinformation or provider bias.

When and Where to Use It

Optimal clinical contexts:

• Prenatal counseling (28–30 weeks) when early labor, preeclampsia, or fetal anomaly raises periviable birth risk
• Urgent delivery scenario (in labor at 22–25 weeks): Rapid counseling before delivery room decisions
• Multidisciplinary team discussions (Obstetrics, Neonatology, Maternal-Fetal Medicine, Palliative Care)
• High-risk maternal conditions (severe preeclampsia, previable PROM) where delivery at periviable gestation is possible

Less useful contexts:

• Infants >25 weeks gestation (survival >85%; different counseling approach)
• Infants born at <22 weeks (data sparse; most guidelines recommend comfort-focused care)
• Term or late preterm deliveries (outcomes excellent; different risk-benefit calculus)

Interpretation Guide

Reading the Results: Outcome Probability Estimates

The NICHD estimator outputs four probability estimates for each infant:

1. Overall Survival Probability of hospital discharge alive (regardless of neurodevelopmental status)

• Example: 22-week male, singleton, 500 g, maternal steroids, no preeclampsia
• Overall survival: 28% (95% CI 24–32%)
• Interpretation: In 100 similar infants, approximately 28 would leave the hospital alive

2. Intact Survival Probability of discharge alive without severe neurodevelopmental impairment (IVH 3–4, cystic periventricular leukomalacia, or moderate-to-severe disability on 18–22 month Bayley testing)

• Example: Same 22-week infant
• Intact survival: 10% (95% CI 8–13%)
• Interpretation: In 100 similar infants, approximately 10 would survive without severe disability

3. Survival with Disability Probability of discharge alive with documented severe disability at follow-up evaluation

• Example: 8% (remaining difference between overall and intact survival, accounting for incomplete follow-up)

4. Mortality Risk Probability of in-hospital death (complement of overall survival)

• Example: 72%
• Interpretation: Approximately 72 of 100 similar infants would die despite NICU admission

Understanding Confidence Intervals

Each estimate includes a 95% confidence interval (CI), indicating the range within which the "true" probability likely falls given sampling variability. A CI of 28% (24–32%) means:

• Best estimate: 28%
• Lower bound: 24% (if cohort sampled slightly different infants, might estimate as low as 24%)
• Upper bound: 32% (might estimate as high as 32%)

Wider confidence intervals indicate greater uncertainty. Earlier gestational ages (22 weeks) have wider CIs (due to smaller sample sizes and greater heterogeneity) than 25 weeks.

Clinical Decision Points and Actions

Scenario 1: Overall survival <20%, Intact survival <5%

Examples include: 22-week male, 400 g, no antenatal steroids, maternal sepsis

• Clinical interpretation: Extremely pessimistic prognosis; most infants do not survive
• Counseling approach:
• "We estimate less than 1 in 5 chance your baby will survive. Even if survival occurs, there is a very high risk of severe brain and organ problems."
• Offer comfort-focused care as primary option: "We recommend focusing on keeping your baby comfortable, allowing you to hold and be with your baby, and creating memories. Intensive resuscitation is not recommended given the extremely high mortality risk."
• If parents insist on resuscitation: Outline that this represents "trial of intensive care" with clear reassessment at 48–72 hours
• Document discussion thoroughly

• Clinical actions:
• Do not perform aggressive delivery room resuscitation (intubation, chest compressions) unless explicit parental request
• Offer comfort-focused delivery (immediate skin-to-skin, pain relief, gentle support)
• Have palliative care team present
• Prepare for likely death; facilitate memory creation (photos, footprints, baptism)

Scenario 2: Overall survival 20–40%, Intact survival 5–15%

Examples include: 23-week female, 550 g, antenatal steroids, no maternal complications

• Clinical interpretation: Poor but meaningful prognosis; survival possible but uncertain; moderate disability risk
• Counseling approach:
• "We estimate about 1 in 3 to 1 in 2 chance your baby will survive. Among survivors, about half will have serious brain or lung problems affecting long-term disability."
• This is the range where time-limited trial is often discussed:
• "We recommend starting intensive care with planned reassessment at 48–72 hours. If your baby is not improving (still needing maximum support, worsening X-rays, declining kidney function), we will discuss transition to comfort-focused care."
• Acknowledge deep uncertainty: "The future is genuinely unpredictable. Some babies at this gestational age do well; others do not. We cannot know which will be true for your baby."

• Clinical actions:
• Prepare for both resuscitation and comfort care
• Deliver in tertiary care center with NICU capability
• Perform delivery room resuscitation if parental goals support it
• Establish clear reassessment milestones (12-hour, 24-hour, 72-hour exams)
• Serial scoring (SNAPPE-II, CRIB II) to track physiologic response
• Daily goals-of-care conversations
• Anticipate high likelihood of major morbidity (BPD, IVH, NEC, ROP)

Scenario 3: Overall survival 40–60%, Intact survival 15–30%

Examples include: 24-week female, 700 g, antenatal steroids, no complications

• Clinical interpretation: Moderate prognosis; survival expected for most infants; substantial disability risk remains
• Counseling approach:
• "We estimate about 50–60% chance your baby will survive. About half of survivors will have some long-term disabilities affecting learning or physical abilities."
• Shift toward cautious optimism but realistic disclosure: "While survival is likely, it comes with significant risk of serious complications. Your baby may spend 3–5 months in the hospital and may need help breathing after discharge."
• Discuss long-term implications: "Many children born at this gestational age attend regular school but some need special education support. Some have minor motor or learning differences. A small percentage have severe disabilities."

• Clinical actions:
• Initiate full NICU care; resuscitation in delivery room standard
• Anticipate aggressive support initially; plan for weaning
• Expect complications (surfactant therapy, sepsis workup, transfusions)
• Plan for prolonged hospitalization; early nutritional optimization
• Baseline cranial ultrasound at 3–5 days; brain MRI before discharge if abnormalities detected
• Ophthalmology screening for ROP at 4 weeks
• Early intervention referral at discharge; neurodevelopmental follow-up at 18–24 months mandatory

Scenario 4: Overall survival >60%, Intact survival >30%

Examples include: 25-week male, 800 g, antenatal steroids, singleton

• Clinical interpretation: Good prognosis; survival expected; disability risk still material but lower
• Counseling approach:
• "We expect your baby to survive. While complications are possible, most babies at this gestational age do well."
• Realistic but positive framing: "Your baby will likely need breathing support and feeding support for weeks to months, but most babies go home and do well."
• Discuss developmental trajectory: "Many children born at 25 weeks have no long-term disabilities. Some have mild delays in learning or motor skills. The risk of severe disability is 10–20%."

• Clinical actions:
• Standard NICU resuscitation and support
• Plan for discharge within 4–8 weeks
• Early advancement of feeds and minimal invasive monitoring
• Routine neurodevelopmental follow-up (18–24 months)
• Educate parents regarding expected milestones (corrected age for prematurity through 2–3 years)

Common Pitfalls in Interpretation

Pitfall 1: Confusing "intact survival" with "no disability" Intact survival typically means "no severe disability" as defined by major IVH, cystic PVL, or Bayley disability score. Minor delays (learning disability, mild cerebral palsy), sensory issues (myopia, hearing loss), or neurobehavioral differences are NOT captured by intact survival definition. A family planning around "intact survival" should understand: "Your child might have learning differences, hearing loss, or vision problems—these are not captured in 'intact survival' estimates."

Pitfall 2: Treating point estimates as certainties An estimated 35% survival does NOT mean 35 of 100 will survive. It means: (a) best estimate is 35%, (b) true rate could reasonably be 30–40%, and (c) individual variation is substantial. Two infants with identical input parameters will have different outcomes—some will surprise clinicians by surviving well, others will deteriorate rapidly despite optimal care. This is why confidence intervals are provided; they reflect true uncertainty.

Pitfall 3: Neglecting maternal factors in counseling The estimator incorporates maternal factors (antenatal steroids, delivery mode, maternal age, etc.) that genuinely affect outcomes. A family planning periviable counseling should hear: "Your baby's chance of survival is actually better than average for this gestational age because you received steroid injections and because your baby is a girl."

Pitfall 4: Assuming outcomes are static The NICHD data span 1998–2014. Contemporary outcomes may differ. Ask yourself: Has surfactant use increased? Has antenatal steroid use increased? Has NICU care intensity improved? In many centers, outcomes are likely better than historical estimates. Conversely, in low-resource settings, outcomes may be worse. Share this uncertainty with families: "These estimates come from top NICU centers in the US. Your hospital's outcomes may differ based on our specific resources and experience."

Pitfall 5: Over-reliance on quantitative estimates in ethical discussion Numbers are helpful but incomplete. A family's decision should integrate:

• Their values and beliefs regarding life, disability, and suffering
• Realistic understanding of disability severity and impact
• Social and financial capacity for long-term care
• Religious or spiritual beliefs
• Parental mental health and relationship stability

Purely quantitative counseling ("35% survival") without addressing these dimensions leaves families unprepared for decision-making. Multidisciplinary teams (OB, Neonatology, Palliative Care, Social Work, Chaplaincy) are essential.

Evidence & Validation

Original Derivation Study

Tyson JE, Parikh NA, Langer J, et al. "Intensive Care for Respiratory Distress Syndrome in Extremely Premature Infants." New England Journal of Medicine. 2008;358(16):1672–1681. DOI: 10.1056/NEJMoa073059

Study design: Prospective cohort study conducted by the Eunice Kennedy Shriver NICHD Neonatal Research Network

Data source and population:

• Prospective collection from 16 NICU centers across the US
• 4,446 infants born at 22–25 weeks gestation between 1998–2003
• All infants received active resuscitation and NICU care
• Comprehensive follow-up at 18–22 months corrected age

Study outcomes (primary):

• Survival to discharge (alive at hospital discharge)
• Intact survival (survival without severe neurodevelopmental impairment defined as: IVH grade 3–4 on cranial ultrasound OR cystic periventricular leukomalacia OR Bayley Scales score >2 SD below mean)

Key results from Tyson et al. (2008):

The study demonstrated that:

1. Survival improves dramatically with each week of gestation (18-fold difference from 22 to 25 weeks)
2. Intact survival lags substantially behind total survival (reflecting high disability burden)
3. Neurodevelopmental impairment is present in ~50–70% of survivors at 22–24 weeks

Updated Validation Studies

Rysavy MA, Li L, Bell EF, et al. (2020) Update: "Between-Hospital Variation in Treatment and Outcomes in Extremely Preterm Infants." JAMA Pediatrics. 2020;174(5):e196294. DOI: 10.1001/jamapediatrics.2019.6294

Updated data: Expanded NICHD Neonatal Research Network prospective cohort

• 8,688 extremely preterm infants born 22–24 weeks between 2008–2018
• More recent period reflects contemporary resuscitation practices (increased antenatal steroids, surfactant use)

Key findings:

• Survival improved over decade: 22-week survival increased from 6% (2003) to 11% (2018)
• Outcome variation across centers: Hospital-level variation in survival ranged from 0–33% for same gestational age (reflecting differences in resuscitation intensity and NICU care)
• Maternal factors: Maternal education, race/ethnicity, and socioeconomic factors did not independently predict survival after adjustment for medical factors
• Sex differences: Female infants had 8–12% higher survival and intact survival compared to males at same birth weight and GA (p<0.001)

Stoll BJ, Hansen NI, Bell EF, et al. (2015) Outcomes Study: "Trends in Care Practices, Morbidity, and Mortality of Extremely Preterm Neonates." JAMA. 2015;314(10):1039–1051. DOI: 10.1001/jama.2015.10244

This parallel epidemiologic study tracked major morbidity outcomes (BPD, NEC, IVH, PVL, ROP) among 9,575 infants born at 22–29 weeks across 16 NICHD centers.

Key findings:

• IVH grade 3–4: Decreased from 12% (2003) to 7% (2012) reflecting improved neuroprotection practices
• Cystic PVL: Remained at ~2–3%, suggesting persistent brain injury despite improvements
• Moderate-to-severe BPD: Remained ~45–50% among survivors, indicating persistent pulmonary morbidity burden
• NEC: Stable at ~7%, responsive to feeding protocols and antibiotic stewardship
• ROP stage 3+: Decreased with improved oxygen management, remaining at ~5–8% among survivors

Sample Sizes and Study Populations

All data come from the same NICHD Neonatal Research Network, enabling comparison across time periods and internal validation.

Sensitivity and Specificity

The NICHD estimator is prognostic, not diagnostic—sensitivity/specificity terminology does not directly apply. However, discrimination (ability to distinguish high vs. low risk groups) can be assessed:

Calibration analysis (Tyson et al. and subsequent validation):

• Expected vs. observed outcomes were well-aligned across risk strata
• When 100 infants had predicted 30% survival rate, actual survival was 28–32% (good calibration)
• Discrimination was excellent: infants predicted 10% survival vs. 50% survival had markedly different outcomes

Limitations and Caveats

1. Limited to US tertiary care centers: Data come from 16 academic medical centers with neonatal ECMO capability and average case volumes >1,000 NICU admissions/year. Outcomes in smaller, non-academic, or international centers may differ substantially. The estimator likely overestimates survival in low-resource settings and may underestimate in the highest-volume periviable centers.

1. Assumes active resuscitation: All infants in the cohort received active resuscitation. The estimator does not predict outcomes with comfort-focused care. A family choosing comfort-focused care should understand these estimates do not apply to their choice.

1. Neurodevelopmental follow-up imperfect: 15–20% of survivors lacked neurodevelopmental assessment at 18–22 months (lost to follow-up, moved). Intact survival estimates may be slightly biased if lost-to-follow-up infants had higher disability rates. Sensitivity analysis suggests impact is small.

1. Outcome definitions have changed: The updated calculator may use revised definitions of IVH or disability (e.g., updated Bayley-III vs. older Bayley-II). Comparisons across time periods must account for these definitional shifts.

1. Individual variation not captured: Two infants with identical inputs (GA, BW, sex, steroids) will have different outcomes. The estimator provides population-level averages; it cannot predict individual infant trajectory. This is why counseling must emphasize uncertainty.

1. Missing variables: Some factors affecting outcome (plural gestation, delivery mode, 1-minute Apgar, infant sex, exact birth weight) are included, but others (genetic factors, placental pathology, presence of congenital anomalies) are not systematically captured. Individualization is limited by data available at prenatal counseling time.

Worked Example

Clinical Scenario

Patient: 23+6 weeks gestation, female fetus, estimated fetal weight 580 g via ultrasound, singleton pregnancy, no congenital anomalies detected.

Maternal history: 32 years old, primigravida, otherwise healthy. Presented with painful regular contractions (preterm labor) at 23+3 weeks. Received:

• Magnesium sulfate (neuroprotection) ✓
• Antenatal corticosteroids (betamethasone, 2 doses) ✓
• Antibiotics for GBS prophylaxis ✓
• Planned delivery (vaginal or cesarean, based on fetal status) at current admission (23+6 weeks)

Obstetric question: What are the prognoses for survival and intact survival if we deliver today? This will inform periviable counseling with the family.

Step-by-Step Estimation Process

Identify patient characteristics for estimator input:

1. Gestational age: 23 weeks, 6 days (rounds to 23.9 weeks; the estimator may round to nearest week, so 24 weeks, depending on implementation; let's use 23 weeks as the conservative estimate)

1. Estimated fetal weight: 580 g (input as stated)

1. Infant sex: Female (input as female; females have ~10–12% higher survival)

1. Plurality: Singleton (input as singleton)

1. Antenatal steroids: Yes, received ✓ (input as "received")

1. Maternal age: 32 years (not typically a major predictor in final models, but may be included; input as documented)

Use NICHD web calculator (or reference table from Tyson et al. 2008):

Inputs:

• Gestational age: 23 weeks
• Birth weight: 580 g
• Sex: Female
• Antenatal steroids: Yes
• Singleton: Yes

Outputs from NICHD calculator:

• Overall survival: 35% (95% CI 30–40%)
• Intact survival (without severe disability): 9% (95% CI 6–13%)
• Survival with disability: ~26% (calculated as overall − intact)
• Mortality: 65% (95% CI 60–70%)

Interpretation and Clinical Application

What the numbers mean:

• In 100 similar female infants born at 23 weeks, ~35 would leave the hospital alive
• Of those 35 survivors, only ~9 would be alive without severe IVH, cystic PVL, or moderate-to-severe developmental delay
• The other ~26 survivors (74% of survivors) would have documented severe neurodevelopmental disability
• ~65 infants would die despite intensive care

What this does NOT tell us:

• Minor disabilities (myopia, hearing loss, mild learning disability, borderline low IQ)
• Functional quality of life (some severely disabled children are happy and engaged; others suffer)
• Long-term educational or vocational outcomes (the 18-month assessment does not predict school performance)
• Parental burden or family impact

Clinical counseling statement for family:

"Based on your baby's gestational age and weight, we can estimate outcomes if we proceed with active resuscitation and NICU care. We expect about 1 in 3 chance your baby will survive to come home. However, we must be honest: among babies who survive, about 3 in 4 will have serious brain bleeds or brain damage affecting long-term development. So while we hope your baby will be one of the survivors without disability, we also must prepare you for the possibility of long-term disabilities.

We recommend a time-limited trial approach: we'll provide intensive care and watch closely how your baby responds over the first few days. If your baby is improving on medications and breathing support, we'll continue. If your baby is not responding—needing more support, showing signs of severe brain bleeding on ultrasound, or having organ failure—we'll discuss transition to comfort-focused care, which focuses on pain relief and family time rather than prolonging dying.

This decision is entirely yours. If you want full support initially with reassessment, or if you want comfort-focused care from birth, both are reasonable choices given the uncertainties."

Clinical Management Based on Prognosis

If family chooses time-limited trial:

1. Delivery room (at 24 hours):

• Prepare for resuscitation (intubation capability, warm resuscitation team)
• Expect need for intubation, chest compressions possible
• Have palliative care team available in background

1. First 12 hours:

• SNAPPE-II score at 12 hours to assess response
• Cranial ultrasound at 6 hours (baseline IVH assessment)
• Goals-of-care conversation: "Your baby was born and is now intubated. Let's talk about how this first day went."

1. 24-hour reassessment:

• Assess response to support: Is FiO2 requirement decreasing? Is BP normalizing? Is urine output improving?
• Repeat SNAPPE-II; if score >60, discuss transitioning to comfort care
• Formal multidisciplinary goals-of-care meeting with parents, OB, Neonatology, Palliative Care

1. 48–72 hour reassessment:

• Most critical decision window
• Repeat cranial ultrasound: Are there signs of IVH grade 3–4 or massive IVH?
• If baby responding well (FiO2 <0.4, BP normal, SNAPPE-II <30): Continue full support
• If baby plateauing or worsening: Present comfort-care option as primary recommendation

1. If survival occurs (decision point at ~7–14 days):

• Once clearly surviving the acute phase, transition to recovery-focused care
• Plan for prolonged hospitalization (8–12 weeks minimum)
• Early neurodevelopmental follow-up arrangement (18-month Bayley assessment essential)
• ROP screening at 4 weeks
• Baseline brain MRI before discharge if any IVH detected

If family chooses comfort-focused care:

• Avoid aggressive resuscitation
• Deliver in calmer environment with parent/chaplaincy support
• Immediate skin-to-skin contact
• Pain medication as needed
• Create memories (photos, footprints, blessing)
• Support family grief process

References

1. Tyson JE, Parikh NA, Langer J, Green C, Higgins RD; National Institute of Child Health and Human Development Neonatal Research Network. Intensive care for extreme prematurity--moving beyond gestational age. N Engl J Med. 2008;358(16):1672-1681. doi:10.1056/NEJMoa073059
2. Stoll BJ, Hansen NI, Bell EF, et al. Trends in care practices, morbidity, and mortality of extremely preterm neonates, 1993-2012. JAMA. 2015;314(10):1039-1051. doi:10.1001/jama.2015.10244
3. Rysavy MA, Li L, Bell EF, et al. Between-hospital variation in treatment and outcomes in extremely preterm infants. N Engl J Med. 2015;372(19):1801-1811. doi:10.1056/NEJMoa1410689