You are on page 1of 11


n e w e ng l a n d j o u r na l


m e dic i n e

special article

Level and Volume of Neonatal Intensive Care and Mortality in Very-Low-Birth-Weight Infants
Ciaran S. Phibbs, Ph.D., Laurence C. Baker, Ph.D., Aaron B. Caughey, M.D., Ph.D., Beate Danielsen, Ph.D., Susan K. Schmitt, Ph.D., and Roderic H. Phibbs, M.D.


There has been a large increase in both the number of neonatal intensive care units (NICUs) in community hospitals and the complexity of the cases treated in these units. We examined differences in neonatal mortality among infants with very low birth weight (below 1500 g) among NICUs with various levels of care and different volumes of very-low-birth-weight infants.

We linked birth certificates, hospital discharge abstracts (including interhospital transfers), and fetal and infant death certificates to assess neonatal mortality rates among 48,237 very-low-birth-weight infants who were born in California hospitals between 1991 and 2000.

Mortality rates among very-low-birth-weight infants varied according to both the volume of patients and the level of care at the delivery hospital. The effect of volume also varied according to the level of care. As compared with a high level of care and a high volume of very-low-birth-weight infants (more than 100 per year), lower levels of care and lower volumes (except for those of two small groups of hospitals) were associated with significantly higher odds ratios for death, ranging from 1.19 (95% confidence interval [CI], 1.04 to 1.37) to 2.72 (95% CI, 2.37 to 3.12). Less than one quarter of very-low-birth-weight deliveries occurred in facilities with NICUs that offered a high level of care and had a high volume, but 92% of very-low-birthweight deliveries occurred in urban areas with more than 100 such deliveries.

From the Health Economics Resource Center and the Center for Health Care Evaluation, Veterans Affairs Palo Alto Health Care System, Menlo Park, CA (C.S.P., S.K.S.); the Department of Pediatrics (C.S.P.) and the Department of Health Research and Policy and the Center for Primary Care Outcomes Research (C.S.P., L.C.B.), Stanford University School of Medicine, Stanford, CA; the National Bureau for Economic Research, Cambridge, MA (L.C.B.); the Department of Obstetrics and Gynecology (A.B.C.) and the Department of Pediatrics and the Cardiovascular Research Institute (R.H.P.), University of California, San Francisco; and Health Information Solutions, Rocklin, CA (B.D.). Address reprint requests to Dr. C. Phibbs at the Health Economics Resource Center (152), VA Palo Alto Health Care System, 795 Willow Rd., Menlo Park, CA 94025, or at N Engl J Med 2007;356:2165-75.
Copyright © 2007 Massachusetts Medical Society.

Mortality among very-low-birth-weight infants was lowest for deliveries that occurred in hospitals with NICUs that had both a high level of care and a high volume of such patients. Our results suggest that increased use of such facilities might reduce mortality among very-low-birth-weight infants.

n engl j med 356;21  may 24, 2007


The New England Journal of Medicine Downloaded from on April 1, 2013. For personal use only. No other uses without permission. Copyright © 2007 Massachusetts Medical Society. All rights reserved.

Copyright © 2007 Massachusetts Medical Society. including our own. most studies were based on data collected before the routine use of surfactant-replacement therapy. We used the draft rather than the final version because the draft was a more accurate reflection of how hospitals in California were actually operating. level 2. Each neonatal intensive care unit (NICU) that offers a lower level of care must have a formal contractual relationship with a NICU that provides tertiary care.6 as well as other investigators. deregionalization) has resulted in increasing numbers of high-risk newborns receiving care in low-volume units offering midlevel care. 2013.32 The birth-certificate data were also successfully linked to infant discharge abstracts from the receiving hospital for 99% of the infants who were transferred to another hospital. 2007 The New England Journal of Medicine Downloaded from nejm. The study was approved by the human subjects committees at Stanford University and the California Office of Statewide Health Planning and Development. These infants are a vulnerable group and thus particularly likely to be affected by hospital services. For personal use only. lower-level NICUs are associated with worse outcomes. a NICU that provides mechanical ventilation without restrictions but does not provide major surgery.1-3 many studies of neonatal care have shown a lower mortality rate in hospitals with higher volumes of patients than in those with lower volumes. at each level of care.838 infants). involved relatively small samples or narrowly defined networks and thus could not adequately assess interaction between volume and level of care.9-17. More than 99% of the maternal and infant discharge abstracts were successfully linked with infant birth certificates. which has substantially improved mortality rates among very-low-birthweight infants. In addition. a NICU that provides mechanical ventilation with restrictions (e. and cared for.4% of births but 51% of infant deaths. a NICU that provides major neonatal surgery but neither open-heart sur­ gery nor extracorporeal membrane oxygenation (ECMO). level 3C. and the requirement for obtaining informed consent was waived.11. after the increased use of antenatal corticosteroid therapy that occurred after the publication of the results of the National Institute of Child Health and Human Development consensus conference in 1994. No other uses without permission.22-25 have previously demonstrated a relationship between the level of NICU care and neonatal outcome.9-17 Growth in the number of NICUs in community hospitals throughout the United States over the past two decades (i. we used data collected from all hospitals in California from 1991 to 2000 to examine the effects of NICU level of care and patient volume on mortality among very-low-birth-weight infants. with assigned levels of care and specific guidelines that define the characteristics of infants who should be delivered.g. only for infants with a birth weight above 1000 g). for the most part. California birth and death certificates were linked to hospital-discharge abstracts for both mothers and infants.nejm. One of the complexities in addressing this question is that the units with the highest level of care are also typically those with the highest volume. in 2000.18-21 It is uncertain whether lower-volume.4. level 3B.31.. particularly among infants with very low birth weight (below 1500 g). 1991.5.8 Higher levels of care are associated with lower neonatal  may 24.27-30 Me thods Study Design We obtained data on very-low-birth-weight infants born in California hospitals and on in-hospital infant and fetal deaths for the period from January 1. most previous studies.. they accounted for only 1.e. level 3A.26 In the current study. 2000 (66. These data reflect outcomes reported after the reduction in mortality associated with the introduction of surfactant-replacement therapy in 1990 2166 and. However. The death certificates included both infant and fetal death certificates.33 to differentiate NICUs in community hospitals from true tertiary or regional perinatal centers (level 3C or 3D). .6. no NICU. The final version does not allow for NICUs that provide mechanical ventilation without re- n engl j med 356.21  www.4-6. a NICU that provides cardiac surgery requiring cardiopulmonary bypass or ECMO. to December 31. Levels of Care We defined levels of care as follows: level 1. and level on April 1. We.4-7 Other studies have examined the association between the level of neonatal care and outcomes. All rights reserved. a NICU that provides care for mildly ill infants but does not provide mechanical ventilation. These definitions are based on the draft version of the American Academy of Pediatrics report on NICU levels of care8. making it difficult to ascertain whether both volume and level are independent predictors of neonatal outcome.The n e w e ng l a n d j o u r na l of m e dic i n e P aralleling the literature on adult care. Neonatal care is formally regionalized.

86).nejm. we created categorical variables for the volume for each level of care.28 Regressions run separately for each level of care showed that the effects of the volume of verylow-birth-weight infants varied according to the NICU level. We tested several different nonlinear function­ al forms using birth weight and gestational age but used categorical variables for the final model because they produced a better fit. empirically from our data (see Section A-1 of the Supplementary Appendix. All rights reserved. in-hospital fetal deaths account­ ed for 22. available with the full text of this article at www. No other uses without permission.27. We developed the model using a random 50% sample and then validated it by applying the estimated coefficients to the remaining data. The dependent variable was inhospital neonatal-related or fetal death. female singletons.34 We controlled for the year to offset the decline in neonatal mortality over the course of the study period. For gestational When added to the neonatal deaths. Most n engl j med 356. version 9. We assigned levels of NICU care to each hospital. nejm. and there was a noticeable shift that we could not confirm had occurred after upward in the levels of care provided (Fig. . and we considered only those variables that were present at birth (see Table A-4 of the Supplementary Appendix).org on April 1. In 2000. Copyright © 2007 Massachusetts Medical  may 24.237 infants. resulting in a final sample of 48. 2007 We used logistic regression to estimate odds ratios for mortality associated with the NICU level of care and annual volume of very-low-birthweight infants. To arrive at a conservative estimate of the number of fetal deaths that occurred after the mother was admitted to the hospital. We also excluded fetal deaths 1991 and 2000. For ease of presentation. We used separate birth-weight functions for male singletons. We tested a wide range of clinical and demographic variables from the birth certificate and discharge data to control for risk factors. the standard definition of a “neonatal death” — death within 28 days after birth — may be biased by the exclusion of continuously hospitalized infants who die after 28 days.35 When applied to the entire data set. we use the term “neonatal-related deaths” to refer to all neonatal deaths plus any deaths that occurred between 29 days and 1 year after delivery if the infant was continuously hospitalized. many California NICUs were actually providing this level of care during the study period. Thus. Neonatal and Fetal Deaths hospital admission (5777 fetal deaths).21  www. A Hosmer–Lemeshow test revealed an acceptable fit (P = 0. we also counted the number of very-low-birth-weight infants who received care at each hospital (both those born in the hospital and those born elsewhere). A total of 6892 infants were transferred between hospitals. The standard errors for the hospital-level independent variables were corrected for within-hospital clustering with the use of the “cluster” option in Stata software. 1). especially the ability to perform very rapid cesarean deliveries.Deregionalization and Very-Low-Birth-Weight Mortality strictions but that do not provide major surgery (level 3B in the draft version). For personal use only. the model again fit well (P  =  0. deaths after 28 days accounted for 7.8% of total deaths.5% of all neonatal deaths. can result in the live birth of infants who would otherwise die in utero. 2013. and deaths among them were attributed to the birth hospital. these infants remained in the sample. the exclusion of in-hospital fetal deaths would introduce a systematic bias against hospitals with large or high-risk obstetrical services. we used 2-week intervals through 33 weeks. and multiple births with 100-g intervals up to 1000 g and 250-g intervals from 1000 to 1500 g. we used ICD-9-CM codes R e sult s to identify and exclude infants with such anomalies (7667 infants) (see Table A-3 of the Supple. Thus. Statistical Analysis Because of improvements in neonatal care. Clinical Mod­ ification (ICD-9-CM) codes (see Table A-2 of the Supplementary Appendix) from the mother’s discharge abstract for procedures that are performed only if the fetus is still alive. area under the receiver-operatingcharacteristic curve = 0. 9th Revision.The number of NICUs increased slightly between mentary Appendix).34 by the Hosmer– Lemeshow test. For each year. Because some congenital anomalies can increase the risk of death among infants with very low birth weight. we identified in-hospital fetal deaths using the Internation­ al Classification of Diseases. Data on infants with a birth weight below 500 g (5157 infants) were excluded to be consistent with previous studies and because of the variability in decisions about whether to treat such infants. Differences among hospitals in the level of ob­ stetrical care. 2167 The New England Journal of Medicine Downloaded from nejm.13). for each year.

or 3D NICU that treated at least same level of care. ECMO.5% in 2000 (Table 1). as were most of ies at hospitals with lower-level and lower-volume Table 2.or moderate-volume units. Copyright © 2007 Massachusetts Medical Society.) The from 35.. deliverH/T 33p9 Enon Combo low. PLEASE NOTE: the NICUs that upgraded their Figure level has of been service NICUs were associated with an increased risk of redrawn and type has been reset. level 3B denotes no restrictions on mechanical ventilation but no major surgery. level of care. The percentage of very-low-birth-weight NICUs that treated 50 or fewer very-low-birthdeliveries in NICUs that treated 51 to 100 of these weight infants per year than in units with larger infants was constant over this time.exception of two very small groups of hospitals: tality rates among NICU level-of-care and volume those with level 2 NICUs that treated more than groups (Table 2).The n e w e ng l a n d j o u r na l of m e dic i n e 200 180 160 15 140 20 16 19 15 18 16 18 15 18 16 18 16 19 16 21 15 26 15 31 Level 3D Level 3C Level 3B Level 3A Level 2 No. death (Table 3). Mortality decreased as patient 25 very-low-birth-weight infants (four hospitals volume increased within each level of care and in 2000) and level 3A NICUs that treated more with higher levels of care within each volume than 50 such infants (three hospitals in 2000). A NICU that treats 50 very- Line 4-C Revised 2168 n engl j med 356. . According to Level of Care. only for infants whose birth weight is greater than 1000 g).nejm. Table 2 also shows the distribution of though the number of NICUs in these two groups several risk factors for death according to the was smaller than normal for categorical variables. No other uses without permission. with the There was a wide range in the unadjusted mor.6% in 1991 to 21. which shows odds ratios for death associthat treated more than 100 such infants decreased ated with the other covariates in the model. 3C. The risk of that treated 26 to 50 very-low-birth-weight infants death was significantly higher in level 3B and 3C annually. Number of NICUs.21  www. (Table A-1 of the Supplementary Appendix). and there volumes. with no mechanical ventilation. model specification tests showed that they should As compared with deliveries at hospitals with not be combined with smaller NICUs with the a level 3B. For personal use only. adjusted for the risk factors shown Please check carefully. volumes were significantly increased. or both. and level 3D denotes cardiac on April 1. in California. Levels of care were empirically determined by the authors on the basis of a modified version of American Academy of Pediatrics definitions. 8 Level 2 denotes an intermediate-care NICU. 3C. 1 of 1 REG F CASE EMail 3rd SIZE ARTIST: of the new NICUs in California in the 1990stswere H/T 100 very-low-birth-weight infants per year. level 3C denotes RETAKE major 1st neonatal surgery but no cardiac AUTHOR: Phibbs ICM 2nd surgery and no extracorporeal membrane FIGURE: oxygenation (ECMO). (Table A-5 of the Supplementary ApJOB: 35615 ISSUE: 04-12-07 eries in hospitals with level 3B. of NICUs 120 100 80 19 60 40 62 20 0 55 57 27 28 40 42 45 44 48 51 51 51 50 46 30 33 33 33 33 32 27 60 55 54 53 50 49 49 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 Figure 1. The risk of death for NICUs with various were only minor changes at NICUs with other combinations of lower levels of care and patient levels of care. 2007 The New England Journal of Medicine Downloaded from nejm. level 3A denotes mechanical ventilation with restrictions (e. The percentage of very-low-birth-weight deliv. and 3D NICUs pendix. All rights reserved. Al­  may 24. 1991–2000.g. odds ratios decreased as volume increased within with most of this decline offset by the increase each level of care and as the NICU level of care in deliveries at hospitals with level 3B or 3C NICUs increased within each volume group. 2013.

org  may 24.21  www. level 3A mechanical ventilation with restrictions (e. level 3B no restrictions on mechanical ventilation but no major surgery.0 1.8 25.6 1. including infants transferred to or received from other facilities.9 4. and the proportion of such deliveries occurring in these hospitals has declined over time. However. . We also performed an analysis stratified according to birth weight. to address the possibility that obstetrical volume accounted for our results.0 2.4 3. Our results were robust when obstetrical volume 2169 n engl j med 356. this restriction excluded only 8% of such deliveries. 1991 and 2000.nejm. Although the associations between mortality and NICU level and volume were great­ er for the smallest infants (below 1000 g). most of the increase in the risk of death was accounted for by hospitals with small to moderatesize NICUs. Some limitations of our study should be noted. we estimated that 21% of the deaths of very-low-birth-weight infants in the year 2000 were potentially preventable (see Table A-7 of the Supplementary Appendix). or both. The ability to provide rapid emergency cesarean sections not only prevents some fetal deaths but also results in the delivery of many infants in healthier condition. Dis cus sion Our study shows strong associations between both NICU level and volume at the delivery hospital and mortality. of Infants Level 1 ≤10 >10 Level 2 ≤10 11–25 >25 Level 3A ≤25 26–50 >50 Level 3B or 3C ≤25 26–50 Level 3B. Table 1.3 4. Because of the observational design.3 3. Assuming that only 90% of the deliveries of very-low-birth-weight infants in the large urban areas could be shifted to hospitals with tertiarylevel NICUs that care for at least 100 such infants annually.4 7. 2013.8 21. Results were materially unchanged when we included infants with congenital anomalies and when we excluded in-hospital fetal deaths. this analysis was restrict­ ed to geographic areas with at least 100 very-lowbirth-weight deliveries in 2000.* Level of Care and No.9 35. level 3C major neonatal surgery but no cardiac surgery and no extracorporeal membrane oxygenation (ECMO).4 0. high-level NICUs may also have better obstetrical care. we created a model based on estimates that included obstetrical volume. hospitals with large. as well as the lack of other data about obstetrical services.9 4. the effects on our results were minimal (see Table A-6 of the Supplementary Appendix).0 1.6 6.1 2.8 8. Copyright © 2007 Massachusetts Medical Society.5 19. or 3D 51–100 >100 1991 6. 2007 The New England Journal of Medicine Downloaded from nejm. Levels of care were empirically determined on the basis of a modified version of American Academy of Pediatrics definitions. The results of a model limited to infants born after 1995 were consistent with the overall results (data not shown). Factors other than NICU level and volume may explain the observed associations.5 3.6 percent 2000 4. For personal use only. ECMO. Thus. high-level NICUs represent the minority of NICUs in California. All rights reserved.g. limited our ability to investigate the role of obstetrical factors.8 Level 2 denotes an intermediatecare NICU. Fewer than 25% of very-low-birth-weight infants were delivered in such hospitals in 2000. Our analysis of data from a 10-year period strengthens the evidence from pre- vious studies that used data from a period of 1 or 2 years. and level 3D cardiac surgery. only for infants whose birth weight is greater than 1000 g). with associated improvement in survival. it was not possible to assess whether there was a causal relationship between NICU features and neonatal mortality.. High-volume. We estimated the number of potentially preventable deaths on the basis of the odds ratios from Table 3 and the distribution of very-lowbirth-weight deliveries across NICUs in 2000 from Table 1.5 * The numbers of very-low-birth-weight infants are the total numbers treated at each hospital.Deregionalization and Very-Low-Birth-Weight Mortality low-birth-weight infants per year corresponds to an average NICU census of about 15 patients.5 3. with no mechanical ventilation. Distribution of Very-Low-Birth-Weight Deliveries According to Annual Patient Volume and Level of NICU Care Available at the Delivery Hospital in California. 3C. Since most births occurred in the large urban areas of California. The very strong correlation between NICU volume and number of deliveries. Because the distance from the mother’s home to a hospital determines the feasibility of delivery at that hospital. they were still significant for the larger infants (see Table A-6 of the Supplementary Appendix).org on April 1.2 2. No other uses without permission. Mortality was lowest when very-low-birthweight deliveries occurred in hospitals with terti­ ary-level NICUs that treat more than 100 of these infants annually.1 25. For example.

1 3.7 3. of deaths (%) 909 (34.0 3.704 2536 (19.5 1.2) 342 (23.2 3.2) 166 (23.2 3.0 900–999 g 3.7 3.9 3.2 3.8 11.2 3.6 4.7) 1359 (20.0 4.9 9.5 7.6 2.2 3.3 10.7 1.0 3.1 4.1 3.2 8.3 3.8 1. 2007 1000–1249 g 10.6 12.2 8.8 9.8 11.3 6.9 3. All rights reserved.2 3.1 1.9 3.0 1000–1249 g 3.4 6.6 4.6 1.21  www.1 8.8 1.1 13.4 13.5 1.5 4.3 13.1 3.3 9.5 1.0 2.9 13.4 1.2 10.4 4. 700–799 g 1.3 3.7 3.6 2.8 1.7) 872 2270 4071 717 1450 1470 1357 6567 12.6 3.7 2.3 1.2 2.2 3.8 800–899 g 3.8 2. No other uses without permission.8 2.7 4.8 3.3 2.1) 4.7 10.0 1.1 3.1 800–899 g 1.4 3.4 4.5 5.744 1–10 Infants 11–25 Infants >25 Infants ≤25 Infants 26–50 Infants >50 Infants ≤25 Infants 26–50 Infants 51–100 Infants Level 2 Level 3A Levels 3B and 3C Levels 3B.3 1.5) 405 (29.5 11.4 1.7 2.3 2.1 3.0 The New England Journal of Medicine Downloaded from nejm.4 14.3 3.4 600–699 g 1. of infants 2636 No.4 14.0 4.9 9.7 1.1 9.3 3.7 5.9 10.9 1.4 7.5) 597 (26.8  may 24.5 14.0 4. 2013.* Characteristic 1–10 Infants No.9 1.4 12.3 3.1 1.3 1000–1249 g 7.7 5.0 2.5 3.5 3. Copyright © 2007 Massachusetts Medical Society.7 3.5 2. For personal use only.7 9.7 11.7 1.4 3.6 n e w e ng l a n d j o u r na l 700–799 g 4.9 5.9 9.1 2.5 m e dic i n e 1250–1499 g 14.4 1.0 3.9 3.7 1250–1499 g 12.3 3. and 3D† >100 Infants 12.6 3.4 3.8 5.nejm. 3C.3 1.1 2.6 3.2 2.3) 863 (21.2 Singleton male 500–599 g 5.2 3.5 1.5 1.1 14.2 2.5 1.3 7.2 2.6 0.7 1.7 1.4 1.7 2.9 12.1 11.2 2.0 3.0 3.8 n engl j med 356.3 9.5 8.1 16.1 11.6) 300 (20.2 9.9 3.0 4.2170 Level 1 >10 Infants 1379 275 (31.5 13.8 700–799 g 2.5 3.7 0.8 9.6 .0 1.5 4.2 0.4 3.2 2. Characteristics of Very-Low-Birth-Weight Infants According to NICU Annual Volume and Level of Care.3 2.8 1.1 3.5 1.0 3.3 3.5 1.5 2.1 The Table 2.5 4.9 4.5 1.5 2.4) 308 (22.0 2.2 12.0 2.7 0.1 4.6 3.5 4.4) Birth weight.1 900–999 g 4.8 5.2 3.5 3.0 5.2 2.9 3.6 1.5 8.0 4.3 600–699 g 3.2 9.9 9.0 1.9 7.6 2.7 2.5 3.3 3.9 1.5 3.8 600–699 g 4.3 1250–1499 g 6.0 900–999 g 1.4 2.4 3.2 4.0 3.0 1.1 3.0 10.9 3.3 10.4 2.9 11.5 5.2 13.0 of 800–899 g 3.9) 2299 (18.9 on April 1. sex.7 1.9 8. 1991–2000.9 5.9 2.7 4.0 7.1 6.8 4.0 3.6 3.6 3.3 11.7 1.9 3.1 2.5 1. and plurality (%) Singleton female 500–599 g 4.7 1.3 11.7 Multiple 500–599 g 1.4 3.1 0.5 3.3 2.0 3.

5 38.1 6.Gestational age (%) 12.2 10.8 16.4 12.1 15.2 9.8 17. Copyright © 2007 Massachusetts Medical Society.9 10.9 20.1 15.5 11.07 3.7 53.4 11.0 5.9 21.9 The New England Journal of Medicine Downloaded from nejm.2 12.9 34.1 16.15 2. 2007 Fetal or infant condition (%) 7.1 12.1 27.1 0.9 18.3 12.nejm.7 ≥34 wk 18.6 32–33 wk 7.6 9.7 20.00 0. For personal use only.07 2.3 35.3 17.5 1.5 0.5 0.0 5.05 2.3 11.9 17.3 9.5 14.5 33.7 0.9 8.5 Large for gestational age¶ 11.1 10.6 18.0 13. All rights reserved.9 19.1 16.7 11.9 0.8 24.2 10.1 19.6 23.6 2.0 14.1 7.7 8.7 Insurance (%) Medicaid 54.5 10.3 12.6 10.2 Small for gestational age¶ 8.5 60.6 7. Fetal distress 6.8 33.4 12.0 9.6 16.2 13.9 11.0 5.0 12.5 21.3 42.2 9.3 18.1 3.4 <24 wk 13.7 14.4 17.6 11.6 14.7 Black race (%)‡ 8.1 Uninsured 9.2 11.14 2.3 3.9 18.3 21.0 5.8 1.3 19.0 0.0 9–11 yr 25.9 41.6 16.3 31. No other uses without permission.7 12.5 17.2 12 14.8 13.1 17.6 21.8 20. 2013.0 11.0 15.8 17.1 33.9 7.2 15.7 14.7 9.6 15.5 13.00 0.1 17.2 46.5 28.8 26–27 wk 15.2 23.9 12.5 12.2 6.8 5.4 15.7 14.3 0.1 15.3 16.4 15.3 17.21  www.6 10.1 8.00 2.2  may 24.8 Hydrops due to isoimmunization‖ 0.5 30–31 wk 11.9 Maternal educational level (%)§ 16.6 15.3 ≤8 yr on April 1.2 9.8 20.9 10.5 12.1 3.8 4.7 10.4 25.6 9.9 13.5 15.8 5.8 39.34 Hemolytic disorder 1.11 3.3 19.1 12.4 11.8 19.4 80.8 14.17 2.2 2.3 12.4 0.6 10.0 5.6 21.4 17.4 0.2 5.5 22.6 16.9 3.7 Health maintenance organization 8.0 18.0 7.6 17.9 19.07 2.5 21.0 3.1 9.2 9.7 18.7 36.7 5.6 11.1 8.1 46.8 17.2 10.9 10.4 10.0 9.3 16.0 13.6 12.9 12.9 28–29 wk 17.5 8.8 24–25 wk 14.1 6.0 23.2 Deregionalization and Very-Low-Birth-Weight Mortality n engl j med 356.4 13.8 2171 .5 5.3 9.

0 0.3 and 761.9 8. † No NICU with a 3D level of care had fewer than 51 patients. Data on race or ethnic group were obtained from information recorded on the birth certificate.6 8.4 1.7 1. and 760.57 0.3 n e w e ng l a n d j o u r na l Other†† 0.0.7 9. along with infants who had other congenital anomalies.8 3.8 9. as reported by the mother.6 0.2 1.88 0. ** Data on maternal hypertensive disorders and noxious substances were based on ICD-9-CM codes 760. and 3D† 51–100 Infants >100 Infants The Table 2.65 1.72 Placental hemorrhage 6.5 3.72.0 8.5 5.4 6.7.6 Oligohydramnios 0. ‡ The Asian.66 1.4 1.74 1. ¶ Information on these conditions was based on ICD-9-CM codes for small for gestational age and large for gestational age (764 and 766. All rights reserved.5 0. .8 Prolapsed cord 2.9 3.3 1.2 11.2172 Level 1 >10 Infants 1–10 Infants 11–25 Infants >25 Infants ≤25 Infants 26–50 Infants >50 Infants ≤25 Infants Level 2 Level 3A Levels 3B and 3C 26–50 Infants Levels 3B.1 5.72 of n engl j med 356.5 1.4 4.1 3. For personal use only. 760.41 6.8 7.9 0.3 6.8 1.6 6.7 1.9 1.0.3 9.2 3.0 7. (Continued. respectively. * Data include all in-hospital.9 0. 2013. Hemolytic disease without a diagnosis of hydrops was included in the hemolytic-disorders variable. 760.6 Infant affected by maternal ­ condition** ­ 0. Data on maternal education were obtained from information recorded on the birth certificate.0 1. 3C.2 3. and infants with this condition were excluded.) Characteristic 1–10 Infants High-risk maternal condition (%) 6.3  may 24.1 8. Native American. Infants with a birth weight below 500 g and infants with major congenital anomalies were excluded.0 3.1 8.8 1.64 0. No other uses without permission.4 7.1 0.0 1. ‖ Nonimmune hydrops was classified as a congenital anomaly.56 on April 1.4 0.9 6.4 1.6 4. 2007 m e dic i n e The New England Journal of Medicine Downloaded from nejm. and Hispanic categories had no significant effect on mortality and were excluded from the final model.21  www. as reported by the mother.48 5.1 6.73.237). ††Data on chronic maternal circulatory and respiratory diseases and incompetent cervix were based on ICD-9-CM codes 760.2 5. § The other educational categories (12 years and <4 years of college) had no significant effect on mortality and were excluded from the final model.4 8.3 4.5 0.6 Premature rupture of membranes 3.22 0. Copyright © 2007 Massachusetts Medical Society. respectively).5 1. very-low-birth-weight deliveries and fetal deaths (a total of 48.4 0.

they should not explain our findings.04–1. All rights reserved. several obstetrical conditions (premature rupture of the membranes. our study showed that the NICU volume and level in the hospitals where very-low2173 n engl j med 356. No other uses without permission. The risk factors we assessed did not differ significantly among the level-of-care and volume groups. 2007 The New England Journal of Medicine Downloaded from nejm.37) 1.51 (1. Although our model controlled for many potential confounders.53 (2.02–3. volume explained less of the variance in mortality than it did in our study.6 Further.96–1. Odds Ratios for Mortality among Very-Low-Birth-Weight Infants. particularly community hospitals. These data are of high quality. sex.08 (0.4. polyhydramnios.86.01 1. but the role of obstetrical volume merits further investigation.* Level of Care and No. ­ in California.17–1.00) <0.98–1. Infants with major congenital anomalies or a birth weight below 500 g were excluded.Deregionalization and Very-Low-Birth-Weight Mortality was added to the model.001 <0.00 0. such as intraventricular hemorrhage and chronic lung disease.52) <0. which. Data from the Vermont Oxford Network also demonstrated considerable variation in outcomes across hospitals after taking the effects of NICU level and volume into account.95) 1. type of insurance. our results suggest that reductions in mortality could be achieved by moving from low to moderate volumes. multiple gestation. fetal distress. and fetal and neonatal conditions (small for gestational age.91–3.35–2. In conclusion. The reference group was hospitals with a level 3B. Although it would be more difficult to regionalize very-low-birth-weight deliveries in more sparsely populated areas of the United States. which may be a more feasible goal in these areas.39 On the basis of our model. and hemolytic disorders).001 Odds Ratio (95% CI) P Value * The area under the receiver-operating-characteristic curve was 0. but it is possible that unmeasured differences among the groups affected the results. and efforts to increase regionalization are likely to draw some opposition. hydrops. black race.nejm. we estimated that increased regionalization of NICU care may have the potential to prevent 21% of deaths among very-low-birth-weight infants. or 3D NICU that treat at least 100 very-low-birth-weight infants per year. Copyright © 2007 Massachusetts Medical Society. of Infants Level 1 ≤10 >10 Level 2 ≤10 11–25 >25 Level 3A ≤25 26–50 >50 Level 3B or 3C ≤25 26–50 Level 3B.34) 1. One potential explanation for these differences is that our data included a broad­ er sampling of hospitals. exceptionally large for gestational age.  may 24.18) 1.37–3. Given that some high-risk cases are selectively referred to large tertiary-care centers. A recent study showed that a higher NICU volume was associated with a lower risk of intraventricular hemorrhage. For personal use only. Our observation that 92% of the very-low-birth-rate deliveries in our study occurred in urban areas with more than 100 such deliveries suggests that it would be geographically feasible to regionalize the vast majority of these deliveries Table 3.72 (2. the relationship between volume and outcomes other than mortality requires additional study.001 <0.24) 1.13) 2. ideally. are also important.21  www.001 <0. consequently.39 (1.37 However.001 0. . To do so would probably require the addition of some large perinatal centers. There could be some disadvantages to closing facilities that are not included in our estimates. gestational age. we could address only those variables available from birth certificates and dis­ charge abstracts. The only outcome we assessed was mortality. and oligohydramnios). According to NICU Level of Care and Annual Patient Volume. year (2000 was the reference variable). including a causal relationship between large. Our exclusion of infants with life-threat­ ening congenital anomalies eliminated any bias due to referrals that were restricted to infants with treatable anomalies.12–1.78 (1. placental complications. in these analyses.19 (1. we would expect such factors to introduce a bias against the highest-level NICUs. 3C. The model included birth weight.22 2.001 on April 1.28–2.36 but they do not include information on all the potential differences in mortality. maternal educational level.56–2.30 (1. 2013.08 2. other outcomes.50) <0.001 1. Studies using data from the Vermont Oxford Network showed weaker relationships between NICU volume and mortality38.26) 1. would be strategically located to maximize geographic access and could be created through the mergers of existing smaller NICUs. This estimate relies on several assumptions.002 <0.38. or 3D 51–100 >100 1.39 than we observed. Standard errors were corrected for clustering of patients within hospitals. and a very high level of regionalization.22 (0.21) <0.69 (1. high-level NICUs and reduced mortality.88 (1.

Pulmonary surfactant therapy. 2013. Rubenfeld GD. 1968-1994. Less than a quarter of very-low-birth-weight infants are born in hospitals with such NICUs. 19911999. 6. Dr. Infant mortality statistics from the 2000 period linked birth/infant death data set. JAMA 1995.328:861-8. Wright LL. Paneth N. Phibbs CS. Holgren EA. 29. Am J Obstet Gynecol 1995. Liu Q. Very-low-birth-weight outcomes of the National Institute of Child Health and Human Development Neonatal Research Network. 3.16: 43-9. Rosenblatt RA. 32.109:745-51. Chassin MR. Maenacker F. 2000. Am J Obstet Gynecol 2001. Dr. Ekbom A. Little GA.108:426-31. Jobe AH. N Engl J Med 1982. Age at death used to assess the effect of interhospital transfer of newborns. Med Care 1979. Horbar JD. Halm EA. Badger GJ. from Blue Cross of California (Wellpoint). the mortality was lowest for deliveries that occurred in hospitals with highlevel and high-volume NICUs. Perinatal regionalization and neonatal mortality in North Carolina. Howell EM. Suser M. Fisher ES. Cifuentes J. Hospital volume and the outcomes of mechanical ventilation. 20. Kiely  may 24. Richardson D. Supported by a grant (HD-36914) from the National Institute of Child Health and Human Development and the Agency for Healthcare Research and Quality. 31. Bode MM. JAMA 1996. and low birth weight newborns. Stark AR. The relation of obstetrical volume and nursery level to perinatal mortality. 21. 184:1302-7. Luft HS. Wallenstein S. Trends in mortality and morbidity for very low birth weight infants.137:511-20. Stukel TA. Hart LG. Carpenter JH. Danielsen B. Pediatrics 2004. Ruys JH. 22. Lee C. Are neonatal intensive care resources located according to need? Regional variation in neonatologists. Brand R.178:131-5.276: 1054-9. [Erratum. 7. 24. Hickok DE. Logerfo JP. Pediatrics 2002. Whyte R. Pedi­ atrics 2004. Mayfield JA.141:60-4. Stevenson DK.] 9.110:143-51. Goss CH. Ann Intern Med 2002. 2174 n engl j med 356. May 1991 through December 1992. Heagerty PJ.114:1341-7. Baker LC. Kiely JL. 23. The effect of patient volume and level of care at the hospital of birth on neonatal mortality. Eas­ terling T. Phibbs RH. Pediatrics 2002. Rocklin. Phibbs CS. Baker SL. Our results suggest that increased regionalization of perinatal care might reduce mortality among very-low-birthweight infants. Nesbitt T. Pediatrics 2005. Liu Q.346:153844. 15. Marcus M. Herrchen B. Chang CH. Association between level of delivery hospital and neonatal outcomes among South Carolina Medicaid recipients. Am J Public Health 1990. Carlo WA. Baldwin LM. Vital statistics linked birth/infant death and hospital discharge record linkage for epidemiological studies. Copyright © 2007 Massachusetts Medical Society.50(12): 1-28. For personal use only. Comput Biomed Res 1997. Changing patterns in regionalization of perinatal care and the impact on neonatal mortality. References 1. Am J Dis Child 1987.173:1585-92. Outcomes of regionalized perinatal care in Washington State. 12.21  www. Johansson S. Obstet Gynecol 2001. Preterm delivery. Pediatrics 2006. Measuring the effectiveness of perinatal medical care. N Engl J Med 1993.307:149-55. Holt VL. Effect of corticosteroids for fetal maturation on perinatal outcomes. N Engl J Med 1979.92:119-24. Metzguer KR.81:404-11.179:37481. Foot B. Am J Obstet Gynecol 1998. 98:247-52. 2. 28. Neonatal mortality for very low birth weight deliveries in South Carolina by level of hospital perinatal service. Little GA. Schmitt SK. Sanderson M.118(6):e1667-e1679. Baldwin LM. Kahn JM. CA: Health Information Solutions. Paneth N. Poskin M. Chien LY.nejm. Verloove-Vanhorick SP. 11. Fisher ES. Fanaroff AA. 173:1423-31. West J Med 1988.149:98-102. Ginsburg P. Phibbs reports serving as a scientific advis­er for the Leapfrog Group’s evidence-based hospital referral pro­gram and as chairman of the neonatal subcommittee. Baker reports receiving consulting fees from a consortium of hospitals in the state of Georgia. Stiles AD. Lee SK. Stukel TA. Shaffer S. Is volume related to outcome in health care? A systematic review and methodologic critique of the literature. 26. Powell SL. Levels of neonatal care. 8.73:854-61. beds. Natl Vital Stat Rep 2002.273:413-8. Am J Public Health 2002. Kiely JL. Gould JB. Am J Obstet Gynecol 1998. Jes­ per­sen KM. Rosenblatt RA. and this percentage has been declining over time. Kramer AA. Menard MK. Connell FA. Effect of opening midlevel neonatal intensive care units on the location of low birth weight births in California. Aziz K. N Engl J Med 2006. et al. Sappenfield WM. J Perinatol 1996. O’Brien CR. Stockbauer JW. 5. Pediatrics 2001. 16. 355:41-50.115:1118. Williams RL.The n e w e ng l a n d j o u r na l of m e dic i n e birth-weight infants are born is strongly associated with mortality. 14. Should operations be regionalized? The empirical relation between surgical volume and mortality. 27. 25. Mathews TJ. Deregionalization of neonatal intensive care in urban areas. Sappenfield WM. Chang CH. and infant death in Sweden: a populationbased study. Mayfield JA. 80:819-23. Bronstein J. 10. No other potential conflict of interest relevant to this article was reported. 301:1364-9. Enthoven AC.113:1230-5. Kirby RS. Matthew D. Paneth N. Wallenstein S. et al. Thiessen P.17:95-110. 4. Pakter J. Buxton E.183:1504-11. Phibbs RH. No other uses without permission. Susser M. Perinatal mortality: the role of hospital of birth. Goodman DC. Schoendorf KS. Pediatrics 1988. 17. 30. Improved outcome of preterm infants when delivered in tertiary care centers. 19. Am J Obstet Gynecol 1995. 18. Susser M. NIH Consensus Development Panel on the Effect of Corticosteroids for Fetal Maturation on Perinatal Outcomes. Bronstein on April 1. and from Catholic Healthcare West and serving as an expert witness in legal matters related to certificate-of-need legislation. Pediatrics 1984. Verwey RA. Yeast JD. Phibbs CS. Ebeling MC. MacDorman MF. Montgomery SM. The choice of place of delivery: effect of hospital level on mortality in all singleton births in New York City. Mortality in very preterm and very low birth weight infants according to place of birth and level of care: results of a national collaborative survey of preterm and very low birth weight infants in the Netherlands. 2007 The New England Journal of Medicine Downloaded from nejm. Probabilistic record linkages for generating a comprehensive epidemiological research file on maternal and infant health. O’Shea TM. 13. Newborn intensive care and neonatal mortality in low-birth-weight infants: a population study. Mortality in low birth weight infants according to level of neonatal care at hospital of birth. level of care. All rights reserved. The relation between the availability of neonatal intensive care and neonatal mortality.30:290-305. . Bunker JP. N Engl J Med 2002. et al. Chu J. Goodman DC. Haberland CA. Recent changes in delivery site of low-birth-weight infants in Washington: impact on birth weightspecific mortality. Am J Obstet Gynecol 2000.

Am J Public Health 1991. Rogowski on April 1.Deregionalization and Very-Low-Birth-Weight Mortality 33. The importance of assessing the fit of logistic regression models: a case study.23(5): 88-97. Kenny M.21  www. Stata user’s guide. Neonatal intensive care unit characteristics affect the incidence of severe intraventricular hemorrhage. perinatologists. California direc- tory of NICUs. n engl j med 356. et al. Stith SA. Copyright © 2007 Massachusetts Medical Society. JAMA 2004. Copyright © 2007 Massachusetts Medical Society. 2013. Horbar JD. Bhatt DR. July through December 1988. California Department of Health Services. Meux EF. 37. Lemeshow S. 291:202-9. Stillwell J. Elk Grove Village. Synnes AR. Staiger DO. College Station. All rights reserved. Health Aff (Millwood) 2004. 38. Report of results from the OSHPD reabstracting project: an evaluation of the reliability of se­ lected patient discharge data. TX: Stata Press. Geppert J. Zach A. No other uses without permission. 35. Taber S.81:1630-5. 39. 1990. 2002. Horbar JD. 8th  may 24. 44:754-9. Hosmer DW. 2007 2175 The New England Journal of Medicine Downloaded from nejm. Sacramento: Office of Statewide Health Planning and Development.nejm. Qiu Z. Variations in the quality of care for verylow-birthweight infants: implications for policy. Indirect vs direct hospital quality indicators for very low-birth-weight infants. IL: American Academy of Pediatrics. neonatologists. Med Care 2006. For personal use only. pediatric surgeons & congenital heart surgeons. . Rogowski JA. Carpenter J. 34. 36. Macnab YC. Staiger DO. 2003.