Objective
The purpose of this study was to compare the effects of universal vs selective resuscitation on maternal utilities, perinatal costs, and outcomes of preterm delivery and termination of pregnancy at 20-23 weeks 6 days’ gestation.
Study Design
We used studies on medical practices, prematurity outcomes, costs, and maternal utilities to construct decision-analytic models for a cohort of annual US deliveries after preterm delivery or induced termination. Outcome measures were (1) the numbers of infants who survived intact or with mild, moderate, or severe sequelae; (2) maternal quality-adjusted life years (QALYs); and (3) incremental cost-effectiveness ratios.
Results
Universal resuscitation of spontaneously delivered infants between 20-23 weeks 6 days’ gestation increases costs by $313.1 million and decreases QALYs by 329.3 QALYs; after a termination, universal resuscitation increases costs by $15.6 million and decreases QALYs by 19.2 QALYs. With universal resuscitation, 153 more infants survive: 44 infants are intact or mildly affected; 36 infants are moderately impaired, and 73 infants are severely disabled.
Conclusion
Selective intervention constitutes the highest utility and least costly treatment for infants at the margin of viability.
Even with improved aggressive life-support measures, >75% of infants who are born at <24 weeks’ gestational age do not survive to hospital discharge. At least 50% of the survivors experience complications that are severe enough to affect their future quality of life. Because the establishment of an accurate clinical prognosis for survival in the immediate perinatal period is difficult, many centers routinely resuscitate extremely premature infants, despite emotional and financial burdens on parents and high risk of disabilities among surviving infants. Other perinatal providers hesitate to resuscitate fetuses and infants whom they judge to be nonviable or highly likely to experience life-long severe disability.
For Editors’ Commentary, see Table of Contents
The Born-Alive Infant Protection Act of 2002 (BAIPA) specified a legal definition of liveborn infants at any gestational age. Live birth is defined as “the complete expulsion or extraction from his or her mother of that member, at any stage of development, who after such expulsion or extraction breathes or has a beating heart, pulsation of the umbilical cord, or definite movement of voluntary muscles, regardless of whether the umbilical cord has been cut and regardless of whether the expulsion or extraction occurs as a natural or induced labor, cesarean section, or induced abortion.” In April 2005, Michael Leavitt, Secretary of the Department of Health and Human Services, issued federal guidance on BAIPA that required investigation of “all circumstances where individuals and entities are reported to be withholding medical care from an infant born alive in potential violation of federal statutes… .” If the delivery of an extremely premature infant is perceived as a medical emergency, then “the hospital and its medical staff would be required to perform a medical screening examination on that born-alive infant…, [and] there would then arise an obligation to admit the infant, or to comply with either the stabilization requirement or the transfer requirements, or risk a finding of an Emergency Medical Treatment and Labor Act or Child Abuse Prevention or Treatment Act violation.” BAIPA clarified the legal status of liveborn infants. It did not further specify what should be considered appropriate medical care for infants who are born at the margin of viability. To date, resuscitation practices currently vary across centers and may diverge from professional board guidelines. Providers therefore may believe that they could be subject to medicolegal suits or to federal enforcement of BAIPA legislative mandates.
We wished to characterize the theoretic effects of universal resuscitation of potentially viable infants on clinical outcomes (death, morbidity, and intact survival) of extreme prematurity. As a first analytic step, we designed decision-analytic models to test how universal resuscitation of all liveborn extremely premature infants would impact costs and maternal quality of life in comparison with current selective nonintervention practice.
Methods
Decision-analytic models were developed with TreeAge Pro 2009 software (TreeAge Software, Inc, Cambridge, MA). We used the entire cohort of US deliveries to model a cost-utility analysis that compared selective with universal resuscitation of infants who were born at 20-23 weeks 6 days’ gestational age. We restricted our analysis to this narrow, but controversial, range because infants who are born at <20 weeks’ gestational age have not been recorded to survive and because infants who are born at ≥24 weeks’ gestation routinely are resuscitated by most neonatologists. Two decision-analytic models were constructed to compare the outcomes of spontaneous preterm delivery and induced preterm deliveries/therapeutic abortions. We classified neonatal outcomes among survivors into 4 groups: (1) intact (no significant neurodevelopmental, visual, or auditory impairment), (2) mild sequelae (mild neurodevelopmental sequelae or correctable visual or auditory impairment), (3) moderate impairment (functionally independent with ambulant cerebral palsy [nonprogressive abnormality of movement and posture, with increased tone and stretch reflexes in ≥1 extremities], moderate cognitive deficit [Bayley Mental Developmental Index, <2 SDs below the mean], hearing loss that required amplification, or visual losses other than blindness, or (4) severe disability (dependent life with nonambulant cerebral palsy, profound cognitive deficit [Bayley Mental Developmental Index, >2 SDs below the mean], profound hearing loss, or blindness). The fifth neonatal outcome was death. Utilities were applied to each outcome status to generate quality-adjusted life years (QALYs). We examined costs from the societal perspective and utilities from the perspective of the mother. Sensitivity analyses were performed to test how robust both models are in response to changing outcome probabilities.
The decision model structures for preterm delivery and induced termination of pregnancy are shown in Figure 1 . The preterm delivery decision tree ( Figure 1 , A) begins with a decision node the branches of which represent 2 neonatal intensive care unit (NICU) management schemes: selective delivery room resuscitation vs universal delivery room resuscitation followed by intensive care, chance nodes whose branches represent live birth vs stillbirth, death in the delivery room vs survival to NICU admission, treatment of the infant, and finally infant outcomes. Figure 1 , B, shows the decision tree for induced termination of pregnancy, with the insertion of an earlier decision node for delivery by dilation and extraction (fatal) vs induction termination, with subsequent nodes similar to the preterm delivery model.
Probabilities were derived from a PubMed literature search of studies that had been published within 5 years of the 2005 guidelines for enforcement of BAIPA whenever possible ( Table 1 ). When no recent study was available, we extrapolated data from the most recent and robust of available relevant studies. We used population-based birth rates for prevalence estimates at each week of gestation, when available. We used geographically-based cohort studies of extreme prematurity, therapeutic abortion, and induced termination of pregnancy when population-based data were not available. Whenever available, we used data on outcomes of extreme prematurity from studies in the United States in preference to European population-based studies. When no other data were available, we extrapolated from clinical practice data from 2 university hospitals in San Francisco.
| Variable | Preterm delivery at 20-23 wk 6 d gestation | Input range | Therapeutic abortion at 20-23 wk 6 d gestation | Input range |
|---|---|---|---|---|
| Deliveries per y, n | 13,563 | 18,300 | ||
| Proportion of liveborn infants | 0.5544 | 0.35–0.76 | 0.0044 | 0.0–0.0144 |
| Proportion of liveborn infants resuscitated in delivery room | 0.4304 | 0–1.00 | 0.4304 | 0–1.00 |
| Proportion surviving resuscitation | 0.946 | 0.594–1.00 | 0.946 | 0–1.00 |
| Outcome probabilities of infants admitted to the neonatal intensive care unit | ||||
| Intact survival a | 0.013 | 0–0.29 | 0.013 | 0–0.29 |
| Mild sequelae a | 0.013 | 0–0.29 | 0.013 | 0–0.29 |
| Moderate morbidity | 0.021 | 0–0.09 | 0.021 | 0–0.09 |
| Severe morbidity | 0.032 | 0–0.13 | 0.032 | 0–0.13 |
| Death | 0.921 | 0.44–0.928 | 0.921 | 0.44–0.928 |
| Maternal utilities | ||||
| Intact survival | 1.00 | 0.90–1.00 | 0.90 | 0.80–1.00 |
| Mild sequelae | 0.90 | 0.80–1.00 | 0.81 | 0.71–0.91 |
| Moderate morbidity | 0.80 | 0.70–0.90 | 0.72 | 0.62–0.82 |
| Severe morbidity | 0.75 | 0.65–0.85 | 0.68 | 0.58–0.78 |
| Infant death | 0.90 | 0.80–1.00 | 1.00 | 0.90–1.00 |
| Maternal care costs, $ b , c | ||||
| Prenatal care | 10,376 | 5188–20,752 | 6917 | 3459–13,384 |
| Preterm delivery | 12,926 | 6463–25,852 | — | — |
| Termination | — | — | 1443 | 722–2886 |
| Induction termination | — | — | 12,926 | 6463–25,852 |
| Fetal and neonatal care costs, $ c | ||||
| No resuscitation | 1113 | 557–2226 | — | — |
| Delivery room resuscitation | 18,752 | 9376–37,504 | 6096 | 3048–12,192 |
| Neonatal intensive care unit for resuscitated infants | ||||
| Nonsurvivors-all | 91,033 | 45,517– 182,066 | 91,033 | 45,517–182,066 |
| Survivors | ||||
| 20-20 6/7 wk | 509,032 | 254,516–1,018,164 | 509,032 | 254,516–1,018,164 |
| 21-21 6/7 wk | 477,127 | 238,564–954,254 | 477,127 | 238,564–954,254 |
| 22-22 6/7 wk | 445,221 | 222,611–890,442 | 445,221 | 226,611–890,442 |
| 23-23 6/7 wk | 413,315 | 206,658–826,630 | 413,315 | 206,658–826,630 |
| Incremental cost of long-term care, $ c | ||||
| Survivor with mild sequelae | 52,056 | 26,028–104,112 | 52,056 | 26,028–104,112 |
| Survivor with moderate sequelae | 578,958 | 289,479–1,157,916 | 578,958 | 289,479–1,157,916 |
| Survivor with severe sequelae | 1,139,657 | 569.829–2,279,314 | 1,139,657 | 569,829–2,279,314 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
