Published online November 20, 2006
PEDIATRICS Vol. 118 No. 6 December 2006, pp. e1667-e1679 (doi:10.1542/peds.2006-0612)
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow P3Rs: Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when P3Rs are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow E-mail this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My File Cabinet
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via CrossRef
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Haberland, C. A.
Right arrow Articles by Baker, L. C.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Haberland, C. A.
Right arrow Articles by Baker, L. C.
Related Collections
Right arrow Premature & Newborn

ARTICLE

Effect of Opening Midlevel Neonatal Intensive Care Units on the Location of Low Birth Weight Births in California

Corinna A. Haberland, MD, MSa, Ciaran S. Phibbs, PhDa,b,c,d and Laurence C. Baker, PhDa,b,e

a Center for Health Policy/Center for Primary Care and Outcomes Research
b Health Research and Policy
c Pediatrics, Stanford University School of Medicine, Stanford, California
d Health Economics Resource Center, Veterans Affairs Palo Alto Health Care System, Palo Alto, California
e National Bureau of Economic Research, Cambridge, Massachusetts


   ABSTRACT
TOP
 ABSTRACT
DATA AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
OBJECTIVE. Despite evidence and recommendations encouraging the delivery of high-risk newborns in hospitals with subspecialty or high-level NICUs, increasing numbers are being delivered in other facilities. Causes for this are unknown. We sought to explore the impact of diffusion of specialty or midlevel NICUs on the types of hospitals in which low birth weight newborns are born.

DESIGN. We used birth certificate, death certificate, and hospital discharge data for essentially all low birth weight, singleton California newborns born between 1993 and 2000. We identified areas likely to have been affected by the opening of a new nearby midlevel unit, analyzed changes over time in the share of births that took place in midlevel NICU hospitals, and compared patterns in areas that were and were not likely affected by the opening of a new midlevel unit. We also tracked the corresponding changes in the share of births in high-level hospitals and in those without NICU facilities (low-level).

RESULTS. The probability of a 500- to 1499-g infant being born in a midlevel unit increased by 17 percentage points after the opening of a new nearby unit. More than three quarters of this increase was accounted for by reductions in the probability of birth in a hospital with a high-level unit (–15 points), and the other portion was resulting from reductions in the share of newborns delivered in hospitals with low-level centers (–2 points). Similar patterns were observed in 1500- to 2499-g newborns.

CONCLUSIONS. The introduction of new midlevel units was associated with significant shifts of births from both high-level and low-level hospitals to midlevel hospitals. In areas in which new midlevel units opened, the majority of the increase in midlevel deliveries was attributable to shifts from high-level unit births. Continued proliferation of midlevel units should be carefully assessed.

Key Words: health services research • low birth weight • NICU • neonatal care • premature infants

Abbreviations: LBW—low birth weight • HMO—health maintenance organization • VLBW—very low birth weight

Evidence suggests that high-risk newborns born in hospitals with subspecialty NICUs have better outcomes than those born in hospitals with no NICU or only midlevel (specialty) facilities,1–13 and current recommendations call for delivery of high-risk (<1500 g or <32 weeks’ gestation) newborns in subspecialty facilities.14–16 Implementation of regionalized systems during the 1970s and early 1980s seemed to have some success in increasing the share of high-risk births taking place in hospitals with appropriate NICU facilities.17–20 However, these trends slowed, or even reversed, during the 1990s.10–12,21,22 This shift has raised concerns about the potential for worse newborn health outcomes10–12,23 and questions about the causes of deregionalization and potential policy options.

One potential cause is the proliferation of midlevel NICUs. In particular, the number of small, community NICUs increased significantly during the 1990s. In California, the location of our study, there were 48 new midlevel units established or upgraded between 1990 and 2001. This occurred primarily in urban and suburban areas that were already being served by tertiary centers.12,24,25 Although a range of factors such as insurance status, maternal education, and obstetric risk factors can influence the choice of hospital for birth, increased availability of, and easier access to, midlevel units in some areas may have led to increased use.26–31

The literature on the effects of NICU availability on use of different levels of care is sparse. Although recent studies have shown that more low birth weight (LBW) infants are being born and cared for at midlevel unit hospitals, none have specifically addressed how the entrance of a new midlevel NICU into an area affects where LBW infants are born in that area. One previous study of the impacts of managed care on newborn care showed some evidence suggesting that the existence of midlevel units could move newborns away from subspecialty centers.32 Given the evidence for potentially worse outcomes among high-risk newborns born in midlevel facilities, the opening of new midlevel units could, therefore, have a significant impact on these infants.

We investigated how the locations of births were affected by the addition of new midlevel NICUs. We studied patterns of hospital use for deliveries in areas likely to have been affected by the opening of a midlevel unit over the period 1993–2001 and compared them with patterns in areas less likely to have been affected by new units. If the introduction of a new unit was important in determining patterns of hospital use for mothers of high-risk newborns, we expected to see patterns of births change more in areas affected by the opening of a new midlevel unit than in other areas.


   DATA AND METHODS
TOP
 ABSTRACT
 DATA AND METHODS
RESULTS
 DISCUSSION
 REFERENCES
 
Newborn Data
We used data from California birth certificate records for 1993–2000 linked to mortality records and hospital discharge records (from the Office of Statewide Health Planning and Development). For essentially all newborns in California, the data identify a range of clinical characteristics, demographics, mortality, and the hospital at which the infants were born and other hospitals used during the first year of life.33,34

For analysis, we extracted singleton births with birth weights between 500 and 2499 g that took place in California between 1993 and 2000. Because the determinants of hospital choices can vary between rural and urban areas, we used the zip code of mother’s residence to select births from zip codes that (1) have their center within 1 of the 25 most populous counties in California (these areas contain ~90% of the state population), and (2) had at least 1 LBW birth in each of the years 1993–2000. We excluded infants with a diagnosis of trisomy 13, trisomy 18, or anencephaly. We included data on 177277 births.

NICU Levels
The previous literature on California NICUs commonly groups hospitals into 4 categories12,32,35,36: level I (cared only for healthy neonates and those with minor medical problems), level II (cared for moderately sick infants but did not provide assisted ventilation for >4 hours), level II+ (community NICU, an expanded level II unit that provided long-term ventilatory support, and some, but not all, of the other services normally provided by level III units), and level III (provided a full range of specialized neonatal intensive care, including pediatric subspecialty consultants and surgery). Previous studies have also identified volume as an important factor in NICU outcomes.11,12

We present results based on a simplified grouping that collapsed these categories to place hospitals into 1 of 3 categories for analysis: "low-level" facilities are those with level I facilities, "midlevel" facilities are all level II units and level II+ units with average daily censuses under 15, and "high-level" facilities are level II+ facilities with average daily census of 15 or more and level III facilities. This grouping eases the presentation of results while capturing important patterns in the data. We placed the smaller level II+ units in the midlevel group and the larger level II+ units in the high-level group based on previous studies reporting that high-volume level II+ NICUs had LBW outcomes similar to level III units.11 The actual number of level II+ units assigned to the high-level group is small. In the year 2000, there were a total of 82 level II+ units in California: 7 of those were placed in the high-level NICU category, and 75 were placed in the midlevel category. Often, these high-level level II+ facilities function as tertiary facilities, except that they do not perform surgeries on site. Some of these units are closely affiliated with nearby children’s hospitals and patients requiring surgery are transferred to those facilities as needed. We verified that our results are robust to assigning all the level II+ units to the midlevel group.

For each birth, we coded the category of NICU care available in the hospital of birth on the basis of these groupings.

Identifying New NICU Areas
We categorized each analysis zip code on the basis of whether or not it was likely to have been influenced by the opening of a new midlevel NICU between 1993 and 2000. To do this, we computed the distance from the center of each zip code to the address of the nearest midlevel unit in each year and examined patterns in the observed distances for each zip code over time. Those zip codes that had a decrease in the distance to the nearest midlevel unit at some point between 1993 and 2000 were identified as areas in which the choices available to mothers giving birth were likely to have been affected by the opening of a new midlevel unit. We termed these areas "new-NICU areas."

In zip codes that continued with the same distance to the nearest midlevel unit throughout, we expected that choices available to mothers were less likely to have been affected by the opening of a new NICU. We termed these areas "no–new-NICU areas." Note, however, that it is possible that new midlevel NICUs opened near enough to these zip codes to affect choices to at least some extent, although not near enough to result in classification of the area as a new-NICU area. Thus, it remains possible that some births in even these areas were affected by new midlevel unit openings, although presumably to a lesser degree than births in the new-NICU areas. Our analysis approach requires only that births to mothers residing in the new-NICU areas were more likely to have been affected than births in these other areas. Based on these measures, we coded each birth according to whether or not the mother’s residence was in a new-NICU area. Births to mothers who live in new-NICU areas are termed "new-NICU area births."

Thirteen zip codes had increases in the distance to the nearest midlevel unit during the study period because of hospital closures or hospitals upgrading a midlevel unit to a high-level unit. The results presented below exclude 1444 births in these zip codes, but our results are robust to including them.

Analysis of NICU Use Patterns
To investigate the effects of opening new midlevel NICUs, we examined changes in the distribution of births across hospitals with different NICU capabilities over time in new-NICU areas compared with changes over time in other areas. If the introduction of a new unit is an important driver of changes in patterns of hospital use for high-risk newborns, we expected to see patterns of births change more in the new-NICU areas where the distance to the nearest midlevel unit declined than in other areas where it did not.

We began by plotting the share of births in low-level, midlevel, and high-level facilities by year, comparing patterns in new-NICU areas and other areas. We then focused on just births in 1993 and 2000 to provide clear before and after comparisons. We tabulated shares of births at different level hospitals in each year and examined changes in the distribution over time.

To control for baseline characteristics of the new-NICU areas and other areas, as well as for variations in the characteristics of the populations across areas, we used multinomial logistic regression. The dependent variable is the level of NICU facilities available in the birth hospital (3 categories: low-level, midlevel, and high-level). The independent variables include a dummy for births in 2000 to control for time trends common to both types of areas and a dummy for new-NICU areas to control for baseline characteristics of new-NICU areas as opposed to other areas. The key independent variable is an interaction term between these 2 measures, capturing the change over time in locations of birth in new-NICU areas, relative to the change in other areas. We controlled for birth weight (500–749, 750–999, 1000–1249, 1250–1499, 1500–1749, 1750–1999, and 2000–2499 g), the presence of anomalies previously identified as being associated with a higher risk of mortality12 (2 categories: any anomalies or no anomalies), maternal age (3 categories: ≤19, 20–29, and ≥30 years), maternal education (3 categories: <12, 12, and >12 years), maternal race (5 categories: white, Hispanic, black, Asian, other), and mothers’ expected source of payment for the delivery hospitalization (5 categories: health maintenance organization [HMO], non-HMO private coverage, Medicaid, self-pay, or other). We used results from these models to compute predicted probabilities of being born in low-level, midlevel, and high-level centers before and after the adoption of the relevant midlevel unit, holding all control variables fixed at their sample means.

We used {chi}2 tests to examine the statistical significance of the changes in distributions and the statistical significance of coefficients in the multinomial logistic regressions.

We performed these analyses for all births in the sample and then separately for newborns 500 to 1499 g and 1500 to 2499 g.


   RESULTS
TOP
 ABSTRACT
 DATA AND METHODS
 RESULTS
DISCUSSION
 REFERENCES
 
Of the 892 zip codes represented in our analysis sample, 97 had a decrease in the distance to the nearest midlevel NICU, signaling the introduction of a new nearby midlevel unit. The other 795 zip codes had no change in distance to the nearest midlevel NICU during the study period. In the new-NICU zip codes, the average distance to a midlevel NICU declined from 12.0 miles in 1993 to 7.6 miles in 2000. Among the zip codes with no change in distance, average distance was 6.5 miles. That is, on average the areas where a new nearby NICU opened started with additional midlevel choices but came, over the study period, to more closely resemble the other areas in terms of distance to the nearest midlevel unit.

Of the 175833 LBW births we examine, 19058 were associated with new-NICU zip codes.

Table 1 summarizes information about the characteristics of births in the sample. Mothers who lived in new-NICU areas were slightly younger, less educated, and less likely to be black than those who lived in other areas. There were no statistically significant differences in insurance status, birth weight, or the presence of anomalies.


View this table:
[in this window]
[in a new window]

  		TABLE 1 Sample Characteristics

 
Figure 1 plots trends over time in the shares new-NICU area births that took place in low-level, midlevel, and high-level hospitals. Figure 2 plots these analogous trends among births in the areas with no new NICUs. There was a clear progression of change toward more births in midlevel NICUs over time in areas that had new nearby units open, whereas there was little evidence of changes in areas that did not.


Figure 1
View larger version (16K):
[in this window]
[in a new window]

  		FIGURE 1 Share of births in low-level, midlevel, and high-level NICU hospitals in areas with a new nearby midlevel NICU opening between 1993 and 2000.

 

Figure 2
View larger version (15K):
[in this window]
[in a new window]

  		FIGURE 2 Share of births in low-level, midlevel, and high-level NICU hospitals in areas with no new nearby midlevel NICU opening between 1993 and 2000.

 
Focusing on data from 1993 and 2000, Table 2 summarizes the distributions of new-NICU area births and births to mothers who lived in other areas. The new-NICU areas and the areas without new NICUs began in 1993 with different distributions of births across hospital types. Among births to mothers who lived in newNICU areas, 57% of 1993 very low birth weight (VLBW) births were initially at high-level centers, and 31% of VLBW births were at midlevel centers. In comparison, areas that did not see the opening of a new midlevel unit had 51% of VLBW births in a high-level hospital and 43% of VLBW births in a midlevel hospital.


View this table:
[in this window]
[in a new window]

  		TABLE 2 Distribution of Births According to NICU Level in Birth Hospital, 1993 and 2000

 
By 2000, in the areas where a new nearby midlevel unit opened, the share of VLBW births in hospitals with midlevel units had increased by 18 percentage points to 49%, the share taking place in high-level centers decreased by 19 points to 38%, and the share in low-level facilities increased by <1 point (P < .001).

Among births associated with areas that did not have a new unit open, the differences between the distributions in 1993 and 2000, although still statistically significant, were much smaller. There was a small increase in the share of births taking place at midlevel facilities and corresponding small declines in the share of births in low-level centers and high-level centers, which may be attributable to the fact that even areas that did not meet our definition for having a nearby new midlevel unit open may have been affected to some extent by the opening of some new units, perhaps a bit farther away.

To ensure that our results are not because of differences in the baseline characteristics of different areas or their populations, we estimated multinomial logistic regression analyses that controlled for underlying characteristics of areas and populations. Results for key variables are shown in Table 3 (see Appendix 4 for full results). For VLBW births, the regression coefficient capturing the 1993–2000 change in the likelihood of being born in a low-level hospital relative to a midlevel hospital in new-NICU areas is –0.653 (P < .05). The corresponding regression coefficient capturing the change in the likelihood of being born in a high-level hospital relative to a midlevel hospital in new-NICU areas is –0.695 (P < .01). The fact that both of these coefficients are negative suggests that the share of VLBW births in midlevel units increased in new-NICU areas relative to rates of births in both low- and high-level hospitals.


View this table:
[in this window]
[in a new window]

  		TABLE 3 Estimated Changes in the Probability of Birth in a Low-Level, Midlevel, and High-Level NICU After the Nearby Opening of a New Midlevel Unit, Adjusted

 

View this table:
[in this window]
[in a new window]

  		APPENDIX 1 Congenital Anomalies Used in the Regression Models

 
To evaluate the magnitudes of the effects, we used the regression results to compute the changes in the predicted probabilities of births in low-, mid-, and high-level hospitals in new-NICU areas, using changes in other areas to adjust for general trends in locations of birth. Results are shown in the bottom portion of Table 3. For VLBW newborns, results show a predicted 17-percentage-point increase in the probability of birth in a midlevel hospital in new-NICU areas, most of which is accounted for by a decline in the probability of birth in a high-level hospital. We estimated the models also for LBW and 1500- to 2499-g newborns and found similar patterns.


   DISCUSSION
TOP
 ABSTRACT
 DATA AND METHODS
 RESULTS
 DISCUSSION
REFERENCES
 
Proliferation of midlevel NICUs is associated with changes in the types of hospitals at which VLBW and LBW newborns are delivered. The patterns we observed associated with opening new midlevel NICUs suggest that the rapid growth in the number of midlevel units in California observed during the 1990s contributed significantly to the movement of LBW and VLBW infants from high-level NICUs to midlevel facilities.

Specifically, when we examined VLBW and LBW births to mothers who resided in areas where a new midlevel NICU opened between 1993 and 2000, we found significant increases in the share of births taking place at midlevel units after the new unit opened. We found that most of the increase in the share of VLBW and LBW newborns taking place in midlevel hospitals was accounted for by a decrease in the share of these newborns who were delivered in hospitals with high-level NICU facilities.

By contrast, there were much smaller changes in the locations of VLBW and LBW births to mothers who resided in areas that were not close to newly opened midlevel NICUs. It is interesting to note that there was some deregionalization evident even in births to mothers who lived outside of the new-NICU areas. This may reflect generalized trends toward deregionalization driven by a range of forces. It may also reflect the diluted but still potentially nonnegligible effects of opening new midlevel units. It may be, for example, that new units opened near enough to some areas to have a small effect on hospital choice for mothers in those areas, but not near enough to trigger our mechanism for defining new-NICU areas.

One valuable way to interpret the results is to take the trends for the no–new-NICU areas as an indication of a baseline trend toward deregionalization, and to interpret the difference between the changes seen in the new-NICU areas and those in the no–new-NICU areas as an indication of the portion of deregionalization driven by the introduction of new midlevel units. Because there is some chance that even the baseline trends observed in the no–new-NICU areas were caused by the opening of new midlevel units, looking at the difference might understate the true effects of midlevel unit openings, but the difference would still provide a conservative estimate of the size of the effect. In this light, the fact that the changes we observe in the locations of births to mothers in new-NICU areas are much larger than changes for births to mothers in other areas is consistent with the view that opening new midlevel NICUs was an important factor in deregionalization.

The size of the shift we observed is quite large. Opening a new midlevel unit is associated with an increase of 17 percentage points in the probability that a VLBW birth to a nearby mother will take place in a midlevel unit, coupled with a corresponding decline of ~15 percentage points in the probability that the birth will take place in a high-level facility, and a 2-percentage-point decline in the probability the birth will take place in a low-level hospital. For LBW births, opening a new midlevel unit is associated with an increase of ~21 percentage points in the probability that a birth to a nearby mother will take place in a midlevel unit, coupled with a corresponding decline of 14 percentage points in the probability that the birth will take place in a high-level facility and a 7-percentage-point decline in the probability the birth will take place in a low-level hospital.

Between 1993 and 2000, among all births <2500 g, there was a noticeable amount of deregionalization in California. Overall, the share of births in midlevel units increased by 3.8%, and the share in high-level facilities declined by 4%. For VLBW births, the share of births in midlevel units increased by 4.9%, and the share in high-level facilities declined by 7.2%. Our results suggest that nearly all of this shift can be traced back to the opening of new midlevel units. Simply stated, the majority of the VLBW births in these newly opened midlevel units seem to be births that would have occurred in the high-level units had the new midlevel unit not opened.

We believe a plausible interpretation of these results is that the opening of midlevel units played a causal role in deregionalization during the 1990s. However, as with most nonrandomized studies, we are unable to definitively rule out the possibility that other forces caused the results we observe. Our study design should minimize the possibility that other factors drove the changes in birth locations we saw in new-NICU areas. If there were other factors, they would have had to be present in just those areas that were most affected by new midlevel NICU openings during this time period. We are unaware of likely candidates, although it is possible that they exist.

One natural question is, are the shifts we saw caused by reductions in the capacity of high-level facilities? We believe this is unlikely. We are unaware of changes in high-level unit capacity that would have affected only our new-NICU areas during this time period. More generally, the number of LBW births in California actually declined between 1993 and 2000,37,38 and we are aware of no generalized reductions in the number of high-level NICU beds between 1993 and 2000.39

The role of patient or doctor preferences for deregionalization may warrant additional study. It is possible, for example, that patients in some areas had a strong preference for midlevel units. This preference would provide reason for nearby hospitals to add midlevel units, and then reason for people to use them once they were in place. We do not have the data that would be needed to provide empirical evidence about the role of preferences. However, although we cannot rule out variations in preferences as a potential factor driving our findings, we would not expect them to play a strong role. It is not clear that one would naturally expect there to be highly localized variations in patient preferences. It is also the case that many of the areas that received new midlevel NICUs did have access to midlevel units even in 1993, and we still saw changes associated with the opening of new units. This suggests to us that, whatever the underlying reasons for the opening of new midlevel units, their opening did play an important role in shifting the location of births.

It is possible, even likely, that other factors besides the opening of new midlevel units can contribute to deregionalization. Factors such as the location and amenities of other hospitals and the preferences of patients may play important roles. Our results should not be interpreted as suggesting that midlevel NICU diffusion is the only important factor in deregionalization but rather as providing support for the view that midlevel diffusion is an important factor.

That the opening new midlevel NICUs can change hospital use patterns is perhaps not that surprising, particularly in retrospect. Other studies have shown that travel time and convenience can be significant drivers of hospital-use patterns.26,27,29,30,40 However, significant shifts of VLBW and LBW births from high-level hospitals to midlevel hospitals are concerning. Current guidelines call for VLBW births and/or births of infants <32 weeks’ gestation to take place in high-level facilities.14–16 In addition, evidence indicates that outcomes for newborns weighing <2000 g delivered in high-level centers are better, and these shifts may be associated with changes in mortality rates. Studies in the last 15 years have found that such infants in high-level units have much better outcomes than those born in midlevel and low-level units.7,10,11,41,42 For infants weighing 2000 to 2499 g the data are mixed. Powell et al10 found that outcomes for these larger LBW infants were better in nontertiary settings; however, they also found that higher risk deliveries of infants in this birth weight range were concentrated at tertiary hospitals. There is also evidence that for all infants born in hospitals with high-level NICUs, the outcomes are better.12 Thus, it is very likely that disproportionately moving LBW and VLBW births from high-level units into midlevel units would tend to raise mortality rates higher than they otherwise would have been.

In addition, as overall survival increases (mortality rates for VLBW infants in the United States have dropped [31.8% in 1990 to 24.7% in 2000]), complications like intraventricular hemorrhage become more of a concern in these patients. Synnes et al43 found rates of intraventricular hemorrhage to also be lower in bigger tertiary care centers. Thus, one can also infer that shifts in birth patterns resulting in proportionately less VLBW and LBW infants being born in high-level centers could raise complication rates higher than they might have been.

At the same time, a full evaluation of the impact of opening new midlevel units would have to include other factors besides morbidity and mortality of newborns. In some cases, it may be that new midlevel units increase convenience for families. We would expect that the most significant benefits of these types would be found in rural areas where new midlevel NICUs could greatly reduce travel time to a hospital with at least some NICU capabilities. Because our study only included newborns in nonrural areas, it seems likely that this would be unlikely to be a strong offsetting factor, but it may be important in some cases. More generally, the majority of the new midlevel units introduced in California were added in suburban or urban areas that were already serviced by high-level and other midlevel units.11,12,24,44

In some areas, better access to midlevel units may improve access for higher birth weight children who benefit from the more advanced services available in midlevel units but would otherwise not receive them. Although mortality rates for higher birth weight infants are quite low, other aspects of morbidity could be affected.

There may also be cost advantages if midlevel NICU stays are less expensive. Cost differences between midlevel units and high-level units remain unclear. Phibbs et al12 found no difference in the risk-adjusted costs in California, whereas Sanderson et al8 found no difference between the costs for infants weighing 500 to 1499 g in level II and level III units but found "enhanced level II" units to have significantly less costs (approximately $9000 in Medicaid reimbursements per admission) than level III units in South Carolina. Additional research will be needed to assess whether there are costs or cost savings involved in referring all high-risk infants to high-level units.

LBW neonates are a small segment of the overall population. In 2004, only 8.1% of newborns were LBW and 1.47% was VLBW. Thus, the affected proportion of the newborn population is relatively small.45 However, poor outcomes in this group are expensive,46 both during acute and long-term care and from both a societal and personal perspective. Much progress has been made over the last 4 decades to help improve the outcomes in these patients, but diffusion of these neonatal care technologies and knowledge has not been shown to improve infant outcomes and can worsen them as high-risk deliveries are drawn away from high-level units. Weighing the potential for higher mortality rates among VLBW and LBW newborns with the possibility of benefits like added convenience for parents, improved access for higher birth weight newborns, cost changes, is inherently subjective and is complicated by a lack of complete data. Nonetheless, we believe that these results suggest the importance of renewed scrutiny and discussion of efforts to further expand the supply of new midlevel NICUs.


View this table:
[in this window]
[in a new window]

  		APPENDIX 2 Complete Regression Results

 


   ACKNOWLEDGMENTS
 
Dr Haberland was supported by Agency for Healthcare Research and Quality institutional training grant HS000028, the Lucille Salter Packard Pediatric Research Fund, and the National Institutes of Health Loan-Repayment Program. Dr Phibbs was supported by National Institutes of Health grant RO1 HD36914 from the National Institute of Child Health and Human Development and Agency for Health Research and Quality.

We thank Drs Pete Groeneveld and Susan Schmitt for help with SAS and Stata programming and Dr Paul Wise for insight and advice on the manuscript.


   FOOTNOTES
 
Accepted Jun 19, 2006.

Address correspondence to Corinna A. Haberland, MD, MS, Stanford University, Center for Health Policy/Center for Primary Care and Outcomes Research, 117 Encina Commons, Stanford, CA 94305. E-mail: corinnah{at}stanford.edu

The authors have indicated they have no financial relationships relevant to this article to disclose.


   REFERENCES
TOP
 ABSTRACT
 DATA AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 

  1. Paneth N, Kiely JL, Wallenstein S, Marcus M, Pakter J, Susser M. Newborn intensive care and neonatal mortality in low-birth-weight infants: a population study. New Engl J Med. 1982;307 :149 –155[Abstract]
  2. Gortmaker S, Sobol A, Clark C, Walker DK, Geronimus A. The survival of very low-birth weight infants by level of hospital of birth: a population study of perinatal systems in four states. Am J Obstet Gynecol. 1985;152 :517 –524[ISI][Medline]
  3. Paneth N, Kiely JL, Wallenstein S, Susser M. The choice of place of delivery. Effect of hospital level on mortality in all singleton births in New York City. Am J Dis Child. 1987;141 :60 –64[Abstract]
  4. Mayfield JA, Rosenblatt RA, Baldwin LM, Chu J, Logerfo JP. The relation of obstetrical volume and nursery level to perinatal mortality. Am J Public Health. 1990;80 :819 –823[Abstract/Free Full Text]
  5. LeFevre M, Sanner L, Anderson S, Tsutakawa R. The relationship between neonatal mortality and hospital level. J Fam Pract. 1992;35 :259 –264[ISI][Medline]
  6. Kirby RS. Perinatal mortality: the role of hospital of birth. J Perinatol. 1996;16 :43 –49[Medline]
  7. Menard MK, Liu Q, Holgren EA, Sappenfield WM. Neonatal mortality for very low birth weight deliveries in South Carolina by level of hospital perinatal service. Am J Obstet Gynecol. 1998;179 :374 –381[CrossRef][ISI][Medline]
  8. Sanderson M, Sappenfield WM, Jespersen KM, Liu Q, Baker SL. Association between level of delivery hospital and neonatal outcomes among South Carolina Medicaid recipients. Am J Obstet Gynecol. 2000;183 :1504 –1511[CrossRef][ISI][Medline]
  9. Samuelson JL, Buehler JW, Norris D, Sadek R. Maternal characteristics associated with place of delivery and neonatal mortality rates among very-low-birthweight infants, Georgia. Paediatr Perinat Epidemiol. 2002;16 :305 –313[CrossRef][ISI][Medline]
  10. Powell SL, Holt VL, Hickok DE, Easterling T, Connell FA. Recent changes in delivery site of low-birth-weight infants in Washington: impact on birth weight-specific mortality. Am J Obstet Gynecol. 1995;173 :1585 –1592[CrossRef][ISI][Medline]
  11. Cifuentes J, Bronstein J, Phibbs CS, Phibbs RH, Schmitt SK, Carlo WA. Mortality in low birth weight infants according to level of neonatal care at hospital of birth. Pediatrics. 2002;109 :745 –751[Abstract/Free Full Text]
  12. Phibbs CS, Bronstein JM, Buxton E, Phibbs RH. The effects of patient volume and level of care at the hospital of birth on neonatal mortality. JAMA. 1996;276 :1054 –1059[Abstract]
  13. Yeast JD, Poskin M, Stockbauer JW, Shaffer S. Changing patterns in regionalization of perinatal care and the impact on neonatal mortality. Am J Obstet Gynecol. 1998;178 :131 –135[CrossRef][ISI][Medline]
  14. American Academy of Pediatrics; American College of Obstetricians and Gynecologists. Guidelines for Perinatal Care. 5th ed. Elk Grove Village, IL: American Academy of Pediatrics; 2002
  15. Stark AR; American Academy of Pediatrics, Committee on Fetus and Newborn. Levels of neonatal care. Pediatrics. 2004;114 :1341 –1347[Abstract/Free Full Text]
  16. Committee on Perinatal Health. Toward Improving the Outcome of Pregnancy: The 90’s and Beyond. White Plains, NY: March of Dimes Birth Defects Foundation; 1993
  17. Committee on Perinatal Health. Toward Improving the Outcome of Pregnancy. White Plains, NY: March of Dimes; 1976
  18. McCormick MC, Shapiro S, Starfield BH. The regionalization of perinatal services: summary of the evaluation of a national demonstration program. JAMA. 1985;253 :799 –804[Abstract]
  19. Bowes WA. A review of perinatal mortality in Colorado, 1971 to 1978, and its relationship to the regionalization of perinatal services. Am J Obstet Gynecol. 1981;141 :1045 –1052[ISI][Medline]
  20. Cordero L, Backes CR, Zuspan FP. Very low-birth weight infant: I. Influence of place of birth on survival. A J Obstet Gynecol. 1982;143 :533 –537
  21. Mehta S, Atherton HD, Schoettker PJ, Hornung RW, Perlstein PH, Kotagal UR. Differential markers for regionalization. J Perinatol. 2000;20 :366 –372[CrossRef][Medline]
  22. Gould JB, Marks AR, Chavez G. Expansion of community-based perinatal care in California. J Perinatol. 2002;22 :630 –640[CrossRef][Medline]
  23. Warner B, Musial MJ, Chenier T, Donovan E. The effect of birth hospital type on the outcome of very low birth weight infants. Pediatrics. 2004;113 :35 –41[Abstract/Free Full Text]
  24. Howell EM, Richardson D, Ginsburg P, Foot B. Deregionalization of neonatal intensive care in urban areas. Am J Public Health. 2002;92 :119 –124[Abstract/Free Full Text]
  25. Phibbs C, Baker L, Schmitt S, Danielsen B, Phibbs R. The effects of NICU patient volume and NICU level at the hospital of birth on neonatal mortality over time for infants with a birth weight <2000g: California 1991–1999 [abstract]. Pediatr Res. 2003;53 :442A . Paper presented at: Pediatric Academic Societies Meetings; May 3–6, 2006; Seattle, WA
  26. McGuirk MA, Porell FW. Spatial patterns of hospital utilization: the impact of distance and time. Inquiry. 1984;21 :84 –95[ISI][Medline]
  27. Phibbs CS, Mark DH, Luft HS, et al. Choice of hospital for delivery: a comparison of high-risk and low-risk women. Health Serv Res. 1993;28 :201 –222[ISI][Medline]
  28. Bronstein JM, Capilouto E, Carlo WA, Haywood JL, Goldenberg RL. Access to neonatal intensive care for low-birthweight infants: the role of maternal characteristics. Am J Public Health. 1995;85 :357 –361[Abstract/Free Full Text]
  29. Piette JD, Moos RH. The influence of distance on ambulatory care use, death, and readmission following a myocardial infarction. Health Serv Res. 1996;31 :573 –591[ISI][Medline]
  30. Mooney C, Zwanziger J, Phibbs CS, Schmitt S. Is travel distance a barrier to veterans’ use of VA hospitals for medical surgical care? Soc Sci Med. 2000;50 :1743 –1755[CrossRef][ISI][Medline]
  31. Wall SN, Handler AS, Park CG. Hospital factors and nontransfer of small babies: a marker of deregionalized perinatal care? J Perinatol. 2004;24 :351 –359[CrossRef][Medline]
  32. Baker LC, Phibbs CS. Managed care, technology adoption, and health care: the adoption of neonatal intensive care. Rand J Econ. 2002;33 :524 –548[CrossRef][ISI][Medline]
  33. Danielsen B. Probabilistic Record Linkages for Generating a Comprehensive Epidemiological Research File on Maternal and Infant Health. Rocklin, CA: Health Information Solutions; November 4, 2000
  34. Herrchen B, Gould JB, Nesbitt TS. Vital statistics linked birth/infant death and hospital discharge record linkage for epidemiological studies. Comput Biomed Res. 1997;30 :290 –305[CrossRef][ISI][Medline]
  35. Bhatt DR, Stillwell J. California Directory of NICUs, Neonatologists, Perinatologists, Pediatric Surgeons and Congenital Heart Surgeons. Elk Grove Village, IL: American Academy of Pediatrics, Section on Perinatal Pediatrics; March 1, 2002
  36. California Children’s Services. Manual of Procedures, Standards for Neonatal Intensive Care Units. Sacramento, CA: Department of Health Services, State of California; January 1, 1999
  37. Martin JA, Hamilton BE, Ventura SJ, Menacker F, Park MM, Sutton PD. Births: final data for 2000. Natl Vital Stat Rep. 2002;50 :1 –101[Medline]
  38. Ventura SJ, Martin JA, Taffel SM, Mathews TJ, Clarke SC. Advance report of final natality statistics, 1993. Mon Vital Stat Rep. 1995;44(6 suppl) :1 –88
  39. California Office of Statewide Health Planning and Development. California Office of Statewide Health Planning and Development Documentation of Hospital Financial Reporting Data. Sacramento, CA: California Office of Statewide Health Planning and Development; 2004
  40. Luft HS, Garnick DW, Mark DH, et al. Does quality influence choice of hospital? JAMA. 1990;263 :2899 –2906[Abstract]
  41. Bode MM, O’Shea T M, Metzguer KR, Stiles AD. Perinatal regionalization and neonatal mortality in North Carolina, 1968–1994. Am J Obstet Gynecol. 2001;184 :1302 –1307[CrossRef][ISI][Medline]
  42. Chien LY, Whyte R, Aziz K, et al. Improved outcome of preterm infants when delivered in tertiary care centers. Obstet Gynecol. 2001;98 :247 –252[Abstract/Free Full Text]
  43. Synnes A, MacNab Y, Qiu ZG, et al. Neonatal intensive care unit characteristics affect the incidence of severe intraventricular hemorrhage. Med Care. 2006;44 :754 –759[CrossRef][ISI][Medline]
  44. Phibbs CS. Are bigger units better? Paper presented at: Hot Topics in Neonatology; December 9, 2003; Washington, DC
  45. Hamilton BE, Martin JA, Ventura SJ, Sutton PD, Menacker F. Births: preliminary data for 2004. Natl Vital Stat Rep. 2005;54(8) :1 –17
  46. Schmitt SK, Sneed L, Phibbs CS. Costs of newborn care in California: a population-based study. Pediatrics. 2006;117 :154 –160[CrossRef][ISI][Medline]
PEDIATRICS (ISSN 1098-4275). ©2006 by the American Academy of Pediatrics



This article has been cited by other articles:


Home page
 NEJMHome page
C. S. Phibbs, L. C. Baker, A. B. Caughey, B. Danielsen, S. K. Schmitt, and R. H. Phibbs
Level and Volume of Neonatal Intensive Care and Mortality in Very-Low-Birth-Weight Infants
N. Engl. J. Med., May 24, 2007; 356(21): 2165 - 2175.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow P3Rs: Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when P3Rs are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow E-mail this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My File Cabinet
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via CrossRef
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Haberland, C. A.
Right arrow Articles by Baker, L. C.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Haberland, C. A.
Right arrow Articles by Baker, L. C.
Related Collections
Right arrow Premature & Newborn