Published online October 1, 2007
PEDIATRICS Vol. 120 No. 4 October 2007, pp. e815-e825 (doi:10.1542/peds.2006-3122)
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 CrossRef
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Van Reempts, P.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Van Reempts, P.
Related Collections
Right arrow Premature & Newborn

ARTICLE

Characteristics of Neonatal Units That Care for Very Preterm Infants in Europe: Results From the MOSAIC Study

Patrick Van Reempts, MD, PhDa, Ludwig Gortner, MD, PhDb, David Milligan, MDc, Marina Cuttini, MD, PhDd, Stavros Petrou, PhDe, Rocco Agostino, MDf, David Field, MDg, Lya den Ouden, MD, PhDh, Klaus Børch, MDi, Jan Mazela, MDj, Manuel Carrapato, MDk, Jennifer Zeitlin, DScl for the MOSAIC Research Group

a Department of Neonatology, Antwerp University Hospital, University of Antwerp and Study Centre for Perinatal Epidemiology, Flanders, Belgium
b Pediatric University Hospital, University of Saarland, Homburg, Germany
c Royal Victoria Infirmary, Newcastle upon Tyne, United Kingdom
d Unit of Epidemiology, Ospedale Pediatrico Bambino Gesù, Rome, Italy
e National Perinatal Epidemiology Unit, Oxford, United Kingdom
f Department for Mother's and Infant's Health, Hospital S. Giovanni Calibita–Fatebenefratelli, Rome, Italy
g Department of Health Science, University of Leicester, Neonatal Unit, Leicester Royal Infirmary, Leicester, United Kingdom
h Health Care Inspectorate, Ministry of Health, Welfare and Sports, the Hague, Netherlands
i Department of Pediatrics, Hvidovre University Hospital, Hvidore, Denmark
j Department of Neonatology, University of Medical Sciences, Poznan, Poland
k Department of Pediatrics, Hospital de São Sebastião, Sta Maria da Feira, Portugal
l Institut National de la Santé et de la Recherche Médicale, UMR S149, Epidemiological Research Unit on Perinatal and Women's Health, and Université Pierre et Marie Curie-Paris 6, Paris, France


   ABSTRACT
TOP
 ABSTRACT
METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
OBJECTIVES. We sought to compare guidelines for level III units in 10 European regions and analyze the characteristics of neonatal units that care for very preterm infants.

METHODS. The MOSAIC (Models of Organising Access to Intensive Care for Very Preterm Births) project combined a prospective cohort study on all births between 22 and 31 completed weeks of gestation in 10 European regions and a survey of neonatal unit characteristics. Units that admitted ≥5 infants at <32 weeks of gestation were included in the analysis (N = 111). Place of hospitalization of infants who were admitted to neonatal care was analyzed by using the cohort data (N = 4947). National or regional guidelines for level III units were reviewed.

RESULTS. Six of 9 guidelines for level III units included minimum size criteria, based on number of intensive care beds (6 guidelines), neonatal admissions (2), ventilated patients (1), obstetric intensive care beds (1), and deliveries (2). The characteristics of level III units varied, and many were small or unspecialized by recommended criteria: 36% had fewer than 50 very preterm annual admissions, 22% ventilated fewer than 50 infants annually, and 28% had fewer than 6 intensive care beds. Level II units were less specialized, but some provided mechanical ventilation (57%) or high-frequency ventilation (20%) or had neonatal surgery facilities (17%). Sixty-nine percent of level III and 36% of level I or II units had continuous medical coverage by a qualified pediatrician. Twenty-two percent of infants who were <28 weeks of gestation were treated in units that admitted fewer than 50 very preterm infants annually (range: 2%–54% across the study regions).

CONCLUSIONS. No consensus exists in Europe about size or other criteria for NICUs. A better understanding of the characteristics associated with high-quality neonatal care is needed, given the high proportion of very preterm infants who are cared for in units that are considered small or less specialized by many recommendations.

Key Words: very preterm infants • neonatal intensive care • level III unit • organization of care • regionalization

Abbreviations: IC—intensive care • MOSAIC—Models of Organising Access to Intensive Care for Very Preterm Births • CPAP—continuous positive airway pressure

It is generally agreed that very preterm infants should be delivered in maternity hospitals with an on-site neonatal unit that is capable of providing full intensive care (IC). This consensus reflects a large scientific literature showing that in most cases the prognosis for very preterm infants is better when they are born in these units, often labeled level III units.1–11 The American Academy of Pediatrics recommends that deliveries that occur before 32 weeks of gestation take place in such specialized units,12 and most European countries have passed laws or issued recommendations based on this premise.13,14 The proportion of very preterm infants who are delivered in maternity units with on-site IC has recently been proposed as a quality-of-care indicator for comparing perinatal health systems across Europe.15

There is less of a consensus, however, on the optimal structural characteristics of a perinatal unit, such as the minimum size or workload for achieving the best health outcomes for very preterm infants. Some studies find better outcomes in larger, more specialized units.11,16–18 In some settings, however, delivery and hospitalization in small units have led to similar outcomes for very preterm newborns.19–22 The question of whether optimal care can be provided in smaller units or outside tertiary units is relevant because there is a growing trend in many countries toward deregionalization of perinatal care.23,24

Understanding the reasons for these conflicting results is difficult because the characteristics of maternity and neonatal units are not often described in adequate detail. Instead, they are usually classified according to local definitions of levels of care. These classifications include 3 or 4 categories: one with no or very little neonatal care (usually labeled level I), a middle tier where some neonatal care is provided (level II), and a third tier corresponding to specialized care (level III). In many settings, a second middle-tier category exists25–27: neonatal units in this category can provide some IC but are not considered regional referral centers. For comparing results from different regions, it is important to determine how levels of care are defined in the various localities.

This analysis uses data from the MOSAIC (Models of Organising Access to Intensive Care for Very Preterm Births) project conducted in 2003 in 10 European regions to explore the organization of perinatal care for infants who are born before 32 weeks of gestation. The aims of this article are to describe and to compare critically existing guidelines for units that provide neonatal IC in the participating regions and to analyze the place of hospitalization of very preterm infants in light of the different criteria set out in these guidelines.


   METHODS
TOP
 ABSTRACT
 METHODS
RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Data on the characteristics of neonatal units were collected as part of the MOSAIC project in 10 European regions. MOSAIC included a prospective cohort study of all births that occurred between 22 weeks, 0 days and 31 weeks, 6 days of gestation in participating regions during 2003, as well as a survey of all maternity and neonatal units that collected information on the units’ characteristics in 2003 and their activity levels in the previous year, 2002. Participating regions, which represented distinct models of perinatal care as identified in a previous European collaborative project,14 were Flanders in Belgium, the eastern region of Denmark, 6 of 8 districts in the Ile-de-France region of France, Hesse in Germany, Lazio in Italy, the central and eastern regions of the Netherlands, Wielkopolska and Lubuskie in Poland, the Northern Region of Portugal, and the Northern and Trent Regions of the United Kingdom. In general, these regions are from European countries with comparable levels of infant health outcomes; infant mortality rates ranged between 4 to 5 per 1000 in 2003, with the exception of Poland, where infant mortality was 7 per 1000.28

Each regional MOSAIC team provided the governmental or professional guidelines in force in 2003 for levels of perinatal care.29–37 The criteria used to define level III units were abstracted from these documents and summarized. In France, Belgium, Italy, Netherlands, Poland, and Portugal, criteria for defining level III units were specified in statutes and applied directly to the units that participated in the MOSAIC study. In Germany, Denmark, and the United Kingdom, recommendations were issued by professional societies and were not legally binding. Maternity units were not labeled as level III units in these 3 countries, but these guidelines included criteria for units that deliver very preterm and other high-risk infants.

Data on the characteristics of maternity and neonatal units were collected by means of 2 structured questionnaires developed by the MOSAIC Study Group, which included neonatologists, obstetricians, and epidemiologists from each region. This analysis presents the results from questionnaires that were sent to neonatal units. This questionnaire was pretested with neonatologists outside the study regions, translated into local languages when necessary (in Italy, Germany, France, Poland, Belgium; administered in English in Portugal, Denmark, Netherlands, and the United Kingdom), and sent to the heads of all neonatal units in the MOSAIC regions. Nonresponders were contacted by telephone by the regional team coordinator. Response rates were 99% for all neonatal units and 100% for level III units.

The neonatal unit questionnaire collected information about activity levels in 2002 (number of admissions, number of infants ventilated, number of days of ventilation, number of infants receiving continuous positive airway pressure [CPAP], and number of days of CPAP). The number of IC beds and ventilators in use in 2003 was also recorded, as well as specific neonatal procedures such as assisted ventilation, parenteral nutrition by central catheter, inhaled nitric oxide therapy, high-frequency ventilation, extracorporal membrane oxygenation and renal dialysis (assessed by the question, "If an infant needs any of the following procedures, would you normally need to transfer the infant to another unit?"), and facilities available in the unit or in the hospital, including surgery. Information on staffing covered total numbers of pediatricians, pediatricians-in-training, and nurses and the presence of these personnel during the day, at night, and on Sundays. The question on the presence of personnel referred to a specific time of day and day of the week (Tuesday and Sunday at 11:00 AM, Tuesday at 1:00 AM). Finally, the existence of written protocols for the care of very preterm infants was investigated, including surfactant usage, postnatal corticosteroids, ultrasound screening, and follow-up programs. Some items included in the survey had a large number of missing values because many units do not seem to keep statistics for them (eg, number of patients receiving CPAP, number of days of CPAP). These items therefore could not be used. We also did not include data on total staffing because of the difficulty of defining each unit's total workload; this point is addressed in "Discussion."

Data from the MOSAIC cohort of very preterm infants were used to calculate the number of very preterm admissions for each neonatal unit. The number of "primary" admissions was determined by assigning each infant to the neonatal unit where he or she spent the first consecutive 48 hours of life or where he or she died when death occurred in the first 48 hours. Thus, each infant was counted only once. We used this definition because, in some regions, infants were admitted to neonatal care for a very short time after delivery before being transported to another, generally higher level unit, whereas in other regions, infants were transported directly from the maternity units without admission to a local neonatal unit. Our focus in this article is to analyze the characteristics of neonatal units where very preterm infants are hospitalized rather than the characteristics of the delivery unit. This point is addressed in "Discussion."

The MOSAIC cohort included 7222 infants, 4947 of whom were admitted to neonatal care in units in the study regions; the other 2275 were births that resulted from terminations of pregnancy, stillbirths, or deaths in labor ward. The inclusion period was all of 2003 except in the French region of Ile-de-France, where data were collected from February 1 to August 31, 2003. In the French region, data for these 7 months were extrapolated to a 1-year period to represent annual admissions. The regional teams also collected information on infants who were transferred to the units from outside the region, either by filling in MOSAIC questionnaires for these infants (in France, the United Kingdom, and Denmark) or by recording their number (in other regions). Very few infants were transferred into the regions in Italy, Poland, and Portugal. In Belgium and the Netherlands, several units received >5% of their total admissions from outside the region, but all of these units had a large number of admissions and adding 5% more to their total did not change their classification in these analyses. Adjustments to observed admissions therefore were not made for these regions.

Whenever possible, we used norms from recommendations to establish cutoff points for the variables in the analysis. When normative values were not available, we used the quartiles or quintiles of the observed distribution of the variable in MOSAIC level III units. Thus, quartiles were used for grouping the number of total admissions (<300, 300–449, 450–599, and ≥600) and that of ventilated infants (0–49, 50–99, 100–179, ≥180); quintiles were used for the number of very preterm (ie, before 32 weeks) admissions (<28, 29–49, 50–69, 70–104, and ≥105). The classification of number of both ventilated infants and very preterm admissions included the cutoff of 50, which was a minimum activity level mentioned in some of the recommendations. Recommendations for minimum number of admissions for a level III unit used a cutoff point for low birth weight infants (<1500 g), whereas MOSAIC used gestational age. In general, the proportion of infants who were born weighing <1500 g is equivalent to those born before 32 weeks.38 Cutoff points for IC beds were based on minimum criteria in the recommendations.

The analysis of unit characteristics covered only units that had at least 5 MOSAIC primary admissions in 2003, thus excluding units that only rarely cared for very preterm infants. In contrast, in the analysis of the place of hospitalization of infants in the MOSAIC cohort, all infants were included, even those who were admitted to units with fewer than 5 admissions per year.

We used {chi}2 tests to compare proportions across unit categories and regions and analysis of variance to compare means. Because of the large number of comparisons and because most of the cross-unit and cross-regional differences were significant at the P < .05 level, the results of significance tests are reported globally in the text. These analyses were conducted using Stata 8 for Windows (Stata Corp, College Station, TX).


   RESULTS
TOP
 ABSTRACT
 METHODS
 RESULTS
DISCUSSION
 CONCLUSIONS
 REFERENCES
 
The MOSAIC regions covered 545000 total births in 2003 and included 320 neonatal units, 65 of which were locally designated as level III units (Table 1). A total of 111 units admitted at least 5 very preterm infants as primary admissions in 2003. Of the remaining 209, 165 admitted no infants <32 weeks of gestation, 20 had 1 primary admission, 11 had 2, 8 had 3, and 5 had 4. In the MOSAIC cohort study, 4947 newborns were admitted to neonatal care. These infants had an average gestational age of 28.6 weeks (median: 29; SD: 2.2) and weighed 1202 g on average at birth (median: 1190; SD: 390). Units with fewer than 5 primary admissions provided care for 1.6% (n = 78) of the cohort admitted for neonatal care and for 0.6% (n = 8) of infants who were born before 28 weeks.


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

  		TABLE 1 Births and Neonatal Units in the MOSAIC Regions in 2003

 
Table 2 presents criteria for defining level III units on the basis of recommendations or laws in the MOSAIC regions in 2003. These recommendations were national, except in Lazio, where they were regional. Criteria concerning maternity units were included in 4 recommendations; 2 of them (Belgium and Netherlands) specified the number of maternal IC beds, and 2 (Portugal and Lazio) specified a minimum number of deliveries. In 3 countries, a minimum number of neonatal admissions for level III units was identified: 200 total admissions in the Netherlands and 50 very low birth weight admissions in Belgium and Germany. In Belgium, these units were also required to ventilate at least 50 infants per year. Six recommendations spelled out a minimum number of IC beds, ranging from 4 to 15.


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

  		TABLE 2 Criteria for Defining Level III Units in the MOSAIC Regions in 2003

 
Twenty-four-hour medical coverage was included in 6 of the recommendations (Belgium, France, Italy, Poland, Portugal, and United Kingdom); 4 of them (France, Italy, Poland, and Portugal) specified the presence of a neonatologist or qualified pediatrician (in France, no specialization in neonatology existed at the time of the study). The number of full-time-equivalent neonatologists per IC bed was defined in 2 recommendations (Belgium and Netherlands). Nursing requirements were specified in 2 ways: as the number of full-time-equivalent nurses per IC bed, ranging from 2.5 (Belgium) to 5 per bed (Netherlands), or as the number of nurses per IC bed required to be present 24 hours/day, which varied from 0.5 (France and Italy) to 1 (United Kingdom). These were recommended ratios for each country and do not necessarily represent real nurse-to-bed ratios.

Four recommendations stipulated the indications for transfer to level III units on the basis of gestational age and/or birth weight. In Denmark, level III care was recommended for extremely preterm infants (<28 weeks), whereas elsewhere the threshold was 32 weeks (Poland and Italy) and 30 weeks (Netherlands). On-site surgery facilities were mentioned in 3 recommendations (Germany, Netherlands, and Poland). All regions included additional requirements or criteria for defining level III units, such as responsibilities for transport, evaluation and audit, follow-up after discharge, teaching, and supervision of other maternity units with lower levels of care. In Belgium and France, neonatal units were required to be physically separate from pediatric wards and in Denmark, Germany, and Italy to be near to the delivery room.

Table 3 compares the characteristics of the neonatal units with at least 5 MOSAIC primary admissions by locally defined level of care (III versus I or II) and number of annual primary very preterm admissions (ie, <32 weeks of gestation). Differences in unit characteristics could be observed by level of care according to number of primary very preterm admissions, number of infants ventilated in 2002, IC beds, continuous medical staffing at night, on-site facilities, and written protocols. With the exception of total admissions, all items in this table differed significantly between level I or II and level III units at the .05 level (P values not shown). Nonetheless, some lower level units provided more specialized care, such as mechanical ventilation (57% of units) or high-frequency ventilation (20%), or had neonatal surgery facilities on-site (17%). Continuous medical coverage by a qualified pediatrician was available for 69% of level III units and 36% of the level I or II units.


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

  		TABLE 3 Characteristics of Units That Care for Very Preterm Infants in 10 European Regions According to Level of Care and Annual Admissions: All Units With ≥5 Primary Annual Admissions

 
The units’ characteristics also differed according to the number of annual very preterm primary admissions. All differences were significant at the .05 level, except for the proportion of units that provided ophthalmology and the proportion of units that did not have a qualified pediatrician on-site at night. Units with a larger number of admissions had more IC beds, ventilated more infants, and provided more services. Units with at least 50 very preterm admissions, however, were similar to larger volume units with respect to many services, and all were locally classified as level III. No significant differences were found among the 3 size categories above 50 with respect to high-frequency ventilation, neonatal surgery, pediatric cardiac surgery, neurosurgery, average number of services provided, and the presence of ≥2 written protocols. They differed significantly only with respect to inhaled nitric oxide therapy (P < .01) and the existence of a protocol for the administration of neonatal steroids (P = .04).

In contrast, there were significant differences between units that admitted between 29 and 49 infants and those with fewer than 29 admissions per year with respect to provision of parenteral nutrition (P < .01), high-frequency ventilation (P < .01), protocols for surfactant (P < .04), and protocols for ultrasound screening (P = .01). All units with ≥29 very preterm births provided ventilation for >24 hours and parenteral nutrition through a central catheter. Nonetheless, even among units with a low annual census of very preterm infants, some had ≥6 IC beds, ventilated ≥50 infants, and provided more specialized services such as high-frequency ventilation and neonatal surgery.

Table 4 presents the proportion of the MOSAIC prospective cohort of infants who were admitted to units that, on the basis of the criteria already discussed, could be considered as small-volume or less specialized for IC to very preterm infants. Twenty-eight percent of infants <32 weeks of gestational age were hospitalized in units with fewer than 50 such admissions, one quarter in units with fewer than 6 IC beds; 46% were in units without a qualified pediatrician on-site at night, and 41% did not have access to on-site surgery. Regions varied widely as to infants who were hospitalized in these small-volume and less specialized units. The percentage of infants in units with fewer than 50 very preterm admissions ranged from 10% in Ile-de-France to 58% in Lazio, Italy. High proportions of infants in smaller units were also found in Portugal (53%), Hesse (36%), and Trent (38%). The variability was equally large for the percentage admitted to units that ventilated fewer than 50 neonates per year (range: 2% in Flanders to 36% in Germany), had fewer than 6 IC beds (0% in Flanders to 74% in Italy), had no continuous pediatric/medical coverage on-site (0% in Italy and France to 100% in the United Kingdom and Netherlands), had no on-site neonatal surgery (0% in Flanders to 76% in Portugal), or had fewer than 2 neonatal protocols (6% in the United Kingdom north to 25% in Italy). Furthermore, regional rankings changed with the criteria adopted. Although Hesse had small units in terms of annual admissions, most infants were hospitalized in units with ≥6 IC beds. Similarly, in Ile-de-France, most infants were in large units but did not have access to on-site surgery.


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

  		TABLE 4 Percentage of Infants <32 Weeks’ Gestational Age and <28 Weeks’ Gestational Age Hospitalized in Smaller and Less Specialized Units as Defined by Multiple Criteria, All Infants Admitted to Neonatal Care in the MOSAIC Regions

 
Overall, infants who were born before 28 weeks of gestation were hospitalized in more specialized and larger volume units than the overall population of infants who were born before 32 weeks. Figure 1 illustrates this difference for unit size, measured by annual very preterm admissions for infants who were born before 28 weeks and at 28 to 31 weeks of gestation. In Denmark, for instance, 9% of infants who were born before 28 weeks were in smaller units, compared with 42% of those who were born between 28 and 31 weeks of gestation. In the other regions, the differences by gestational age were statistically significant, with the exception of Lazio, Hesse, and Flanders.


Figure 1
View larger version (20K):
[in this window]
[in a new window]

  		FIGURE 1 Percent of infants hospitalized in units with <50 very preterm admissions per year, according to gestational-age group, for all infants admitted to neonatal care in the MOSAIC regions ({chi}2 values for difference between gestational-age groups are shown in parentheses). NS indicates not significant; DK, Denmark; BE, Belgium; FR, France; DE, Germany; IT, Italy; NL, Netherlands; PL, Poland; PO, Portugal; UK, United Kingdom.

 

   DISCUSSION
TOP
 ABSTRACT
 METHODS
 RESULTS
 DISCUSSION
CONCLUSIONS
 REFERENCES
 
This study found that European regions differ widely in how they organized the care of preterm infants with respect to unit size (minimum admissions, number of IC beds), availability of services, and medical and nursing staffing and also in the underpinning recommendations. Minimum size criteria were specified in 6 of 10 recommendations in Europe for level III NICUs, but they varied considerably. More than one fifth of level III units in MOSAIC regions did not meet minimum cutoff points set for number of very preterm infants, infants ventilated, and IC beds. A similar proportion of infants who were born before 28 weeks of gestation received care in such units. The proportion of infants who were cared for in smaller and less specialized units varied greatly by region.

The diversity in both recommendations and the actual organization of neonatal units reflects, in part, the uncertainty that exists about the structural characteristics that are necessary for providing optimal neonatal IC. Most research in this area studies activity levels, probably because this information is easy to obtain and because there are clear hypotheses about the association among activity levels, experience, and quality, especially for volume of procedures performed.39 Nonetheless, the scientific literature offers contradictory results about the relation between volume and outcome for neonatal IC.40 Interpreting this literature is further complicated by investigators’ use of different cutoff points. Some studies find better outcomes at volumes higher than a daily census of 15 infants,16,17 whereas other studies use annual admissions and place the limit at 3641 or 50 very low birth weight admissions per year.18 Some studies have compared "smaller" with "larger" reference units and found an impact on outcome,42 but others reported that delivery and hospitalization in smaller units and level IIB units led to equivalent outcomes for very preterm newborns19–22; still others have found no relation between workload and outcome.43,44 In this study, we analyzed number of admissions in quintiles as observed in the 65 level III units in our sample; this categorization included the threshold of 50 very preterm admissions, which corresponds to the size criterion designated in recommendations of 2 MOSAIC countries and also used in previous studies.18 We found that units with between 50 and 69 very preterm admissions were similar in characteristics to larger units. In contrast, units with fewer than 28 admissions and those with 28 to 49 admissions differed significantly in the services provided and in the existence of protocols for the care of very preterm infants.

Our results suggest that 1 reason for the contradictory results of previous studies may be that smaller units differ with respect to other characteristics. Although the annual number of very preterm admissions correlated with a number of other characteristics from the minimum criteria recommendations, such as the number of patients ventilated, number of IC beds, and availability of surgery, some small units were more specialized than others. The specialization of small units may be associated with their ability to provide quality care. Similarly, there was considerable overlap between units that were labeled as level III versus level II. Some level II units provided specialized services, routinely cared for very preterm infants, and had 24-hour coverage by qualified pediatricians. One study previously reported that level II units with 24-hour neonatology coverage had outcomes equivalent to those of level III units,25 although this was not found in another study.45

Staffing is another element that has the potential to influence outcome. Low birth weight infants who were exposed to nursing staff-to-infant ratios of >1.71 had an adjusted risk for survival to hospital discharge 82% better than those who were exposed to lower staffing ratios in 1 study.46 Although the United Kingdom Neonatal Staffing Study did not show an association between staffing and outcome, mortality was linearly related to occupancy rates, and an increasing trend of mortality accompanied a lower nurse-to-infant ratio.44 Research conducted in adult ICUs has found that when intensivists are continually in charge of care, outcomes are better.47 In this study, we found differences both in recommended number of unit staff and in the qualification of medical professionals who should be present at all times in the unit. In several countries, the guidelines require on-site presence of a qualified pediatrician, whereas elsewhere, residents of various levels of pediatric training ensure night coverage. The regions varied highly in their practices with respect to this characteristic, a variation that reflected the differences in recommendations.

Our study could not measure actual staff-to-bed ratios because of the diversity of the organization of neonatal units in Europe. Although we collected data on the number of full-time staff in units, it was not possible to assess total unit workload accurately. As mentioned, the number of total admissions was only loosely related to the number of very preterm infants and ranged from <300 to >600 both in level III and in some level II units. Some units seemed to specialize in high-risk infants only, whereas others included high-dependency and special care nurseries. Recommendations most often set minimum criteria for nursing staff with respect to IC beds. However, some recommendations specified that each bed be covered by 1 nurse, whereas elsewhere, it was 2 beds per nurse, suggesting either that beds were used for infants who were less sick or that there were differences in the intensity of care provided to infants with similar health conditions. Our data were not detailed enough to pursue this analysis, but these differences made it difficult to be confident about using the number of IC beds for a comparative estimate of workload among the units.

Our analysis of the characteristics of the place of hospitalization of infants in the MOSAIC cohort revealed large differences among regions for the proportion of infants who were hospitalized in smaller volume and less specialized units. This analysis focused on units where infants were hospitalized for their first consecutive 48 hours of life, not on the neonatal unit at the place of birth, although this could also affect outcome. Our purpose was to assess norms with respect to where very preterm infants "should" be hospitalized. Infants are often delivered in hospitals that are not appropriate for their health status because of contraindications to in utero transfer, such as maternal hemorrhage or imminent delivery. In such cases, neonatal transfers are organized after delivery to a neonatal unit that is capable of providing appropriate care.

For the analysis of place of hospitalization, we used cutoff points that were derived from our descriptive analyses of the neonatal units with the aim of identifying units that were relatively less specialized and smaller volume than other units. These cutoff points do not represent accepted normative thresholds for defining less specialized or smaller volume units, because our review of the recommendations reveals that there is as yet no consensus on what these norms should be. This analysis does call into question the use of only 1 single structural characteristic for the classification of units.

We also analyzed place of hospitalization by gestational age group. Some authors claim that the benefits of higher workload or the provision of specialized care is restricted to gestational ages <29 weeks.48,49 Community and regional level III units may achieve comparable mortality outcomes for very low birth weight infants; however, they may also result in higher mortality for newborns who weigh <1000 g.50 Thus, optimal unit characteristics may differ according to infants’ level of risk. In several of the regions in this study, such as Denmark and the Northern Region of the United Kingdom, the marked difference in place of hospitalization of infants according to gestational age suggests that this type of consideration is taken into account in the organization of perinatal care; in other countries, this was not the case. The issue of whether specialized care should be reserved for only extremely preterm infants remains controversial because other research has found that very preterm infants who are cared for in smaller neonatal units have higher mortality regardless of gestational age.16,26,51


   CONCLUSIONS
TOP
 ABSTRACT
 METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
REFERENCES
 
The existence of minimum size and other structural criteria in recommendations for defining level of care reflects widespread beliefs that structural and organizational factors affect outcome for very preterm infants. However, the variations in the thresholds used and in factors selected for inclusion in these recommendations reveal the diversity of opinions about which structural factors really matter. This variability is not limited to Europe; other studies from the United States and Canada have also reported differences in the characteristics of NICUs, without, however, systematically describing such differences.43,52,53

Additional investigation is necessary to elucidate the impact of unit characteristics on health outcomes, and these analyses are planned by the MOSAIC group with its European cohort. However, our results point to a potential difficulty in analyzing differences in organizational structures when there is a large variability among countries. For instance, in countries where all or almost all neonatal units have low or moderate annual volume, it may not be possible to assess the impact of "large" units on outcome. The use of differing thresholds in studies from different countries may be 1 reason for their contradictory results. A similar problem exists for the analysis of staffing, such as the impact of having qualified pediatricians on-site at night, if policies are established at a regional or national level.

In the absence of clear evidence in support of policies to implement structural norms, actual practices and beliefs reflect many factors, including differences in the organization of maternity and pediatric care predating the development of IC, financing systems, and other health policies that do not necessarily relate to the care of very preterm infants. Achieving changes to existing systems will be constrained by these contextual factors. Nonetheless, given the large proportion of very preterm infants who are cared for in units that would be labeled small or unspecialized by many recommendations, our results underline the importance of gaining a better understanding of which structural criteria are associated with quality of care to provide guidelines for planners and clinicians who wish to improve health outcomes for this high-risk population.


   ACKNOWLEDGMENTS
 
This project was coordinated by Assistance Publique-Hôptiaux de Paris and cofinanced by the Research Directorate of the European Union (QLG4-CT-2001-01907).

The MOSAIC Research Group included locations in Flanders, Belgium (E. Martens, G. Martens, and P. Van Reempts); eastern Denmark, Denmark (K. Boerch, T. Weber, and B. Peitersen); Ile-de-France, France (G. Bréart, J.L. Chabernaud, D. Delmas, and E. Papiernik); Hesse, Germany (L. Gortner, W. Künzel, R. Maier, and B. Misselwitz); Lazio, Italy (R. Agostino, D. Di Lallo, and R. Paesano); eastern and central Netherlands (L. den Ouden, L. Kollée, G. Visser, J. Gerrits, and R. de Heus); Wielkopolska and Lubuskie, Poland (G. Breborowicz, J. Gadzinowski, and J. Mazela); Northern Region, Portugal (H. Barros, I. Campos, and M. Carrapato) Trent Region, United Kingdom (E. Draper, D. Field, and J. Konje); Northern Region, United Kingdom (A. Fenton, D. Milligan, and S. Sturgiss); Institut National de la Santé et de la Recherche Médicale, U149, Paris (G. Bréart, B. Blondel, and J. Zeitlin); external contributors (M. Cuttini and S. Petrou); steering committee (E. Papiernik, project leader, J. Zeitlin, research coordinator, G. Bréart, E. Draper, and L. Kollée).

We acknowledge the assistance of the personnel in the maternity and neonatal units in the regions that participated in the MOSAIC project.


   FOOTNOTES
 
Accepted Feb 23, 2007.

Address correspondence to Patrick Van Reempts, MD, PhD, Department of Neonatology, Antwerp University Hospital, Wilrijkstaat, 10, B-2650 Antwerp (Edegem), Belgium. E-mail: patrick.van.reempts{at}uza.be

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


   REFERENCES
TOP
 ABSTRACT
 METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 

  1. Kollée LA, Verloove-Vanhorick PP, Verwey RA, Brand R, Ruys JH. Maternal and neonatal transport: results of a national collaborative survey of preterm and very low birth weight infants in the Netherlands. Obstet Gynecol. 1988;72 :729 –732[Abstract/Free Full Text]
  2. Lubchenco LO, Butterfield LJ, Delaney-Black V, Goldson E, Koops BL, Lazotte DC. Outcome of very-low-birth-weight infants: does antepartum verus neonatal referral have a better impact on mortality, morbidity, or long-term outcome? Am J Obstet Gynecol. 1989;160 :539 –545[ISI][Medline]
  3. Modanlou HD, Dorchester W, Freeman R, Rommal C. Perinatal transport to a regional perinatal center in a metropolitan area: maternal versus neonatal transport. Am J Obstet Gynecol. 1980;138 :1157 –1164[ISI][Medline]
  4. Mondalou HD, Dorchester WL, Thorosian A, Freeman RK. Antenatal versus neonatal transport to a regional perinatal center: a comparison between matched pairs. Obstet Gynecol. 1979;53 :725 –729[Abstract/Free Full Text]
  5. Ozminkowski RJ, Wortman PM, Roloff DW. Inborn/outborn status and neonatal survival: a meta-analysis of non randomised studies. Stat Med. 1988;7 :1207 –1221[ISI][Medline]
  6. 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. N Engl J Med. 1982;307 :149 –155[Abstract]
  7. Truffert P, Goujard J, Dehan M, Vodovar M, Bréart G. Outborn status with a medical neonatal transport service and survival without disability at two years: a population-based cohort survey of newborns of less than 33 weeks of gestation. Obstet Gynecol. 1998;79 :13 –18
  8. Field D, Hodges S, Mason E, Burton P. Survival and place of treatment after premature delivery. Arch Dis Child. 1991;66 :408 –411[Abstract]
  9. 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]
  10. Shlossman PA, Manley JS, Sciscione AC, Colmorgen GH. An analysis of neonatal morbidity and mortality in maternal (in utero) and neonatal transports at 24–34 weeks’ gestation. Am J Perinatol. 1997;14 :449 –456[ISI][Medline]
  11. 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]
  12. Stark AR, American Academy of Pediatrics Committee on Fetus and Newborn. Levels of neonatal care [published correction appears in Pediatrics. 2005;115:1118]. Pediatrics. 2004;114 :1341 –1347[Abstract/Free Full Text]
  13. Debauche C, Van Reempts P, Kollée L, Chabernaud J, Zeitlin J, Velebil P. Maternal and neonatal transfer policies in Europe. Prenat Neonatal Med. 1999;4(suppl 1) :5 –14
  14. Zeitlin J, Papiernik E, Bréart G, Europet Group. Regionalisation of perinatal care in Europe. Semin Neonatol. 2004;9 :99 –110[CrossRef][Medline]
  15. Zeitlin J, Wildman K, Breart G, et al. PERISTAT: indicators for monitoring and evaluating perinatal health in Europe. Eur J Public Health. 2003;13(suppl) :29 –37[Free Full Text]
  16. 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]
  17. 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]
  18. Rogowski JA, Horbar JD, Staiger DO, Kenny M, Carpenter J, Geppert J. Indirect vs direct hospital quality indicators for very low-birth-weight infants. JAMA. 2004;291 :202 –209[Abstract/Free Full Text]
  19. Field D, Draper ES. Survival and place of delivery following preterm birth 1994–1996. Arch Dis Child Fetal Neonatal Ed. 1999;80 :F111 –F115[Abstract/Free Full Text]
  20. Arad I, Gofin R, Baras M, Bar-Oz B, Peleg O, Epstein L. Neonatal outcome of inborn and transported very-low-birth-weight infants: relevance of perinatal factors. Eur J Obstet Gynecol Reprod Biol. 1999;83 :151 –157[CrossRef][ISI][Medline]
  21. Rosenblatt RA, Macfarlane A, Dawson AJ, Cartlidge PH, Larson EH, Hart LG. The regionalization of perinatal care in Wales and Washington State. Am J Public Health. 1996;86 :1011 –1015[Abstract/Free Full Text]
  22. Meadow W, Mendez D, Makela J, Malin A, Gray C, Santos JD. Can and should level II nurseries care for newborns who require mechanical ventilation? Clin Perinatol. 1996;23 :551 –561[ISI][Medline]
  23. 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]
  24. 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]
  25. 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]
  26. Finnstrom O, Olausson PO, Sedin G, et al. The Swedish national prospective study on extremely low birthweight (ELBW) infants. Incidence, mortality, morbidity and survival in relation to level of care. Acta Paediatr. 1997;86 :503 –511[ISI][Medline]
  27. Empana JP, Subtil D, Truffert P. In-hospital mortality of newborn infants born before 33 weeks of gestation depends on the initial level of neonatal care: the EPIPAGE study. Acta Paediatr. 2003;92 :346 –351[CrossRef][ISI][Medline]
  28. EUROSTAT. Population Statistics; 2006. Available at: http://epp.eurostat.cec.eu.int/portal/page?_pageid=1996,45323734&_dad=portal&_schema=PORTAL&screen=welcomeref&open=/popul/popula/pop/demo/demo_mor&language=en&product=EU_MAIN_TREE&root=EU_MAIN_TREE&scrollto=0. Accessed August 1, 2007
  29. Berg D. The organization of perinatal centers. Geburtshilfe Frauenheilk. 1990;50 :76 –78[Medline]
  30. Ministère de l'emploi et de la solidarité. Décret no98–899 et no 98–9000 du 9 octobre 1998. Journal officiel de la république française. 1998:15343 –15343
  31. Parte I. Piano di ristrutturazione rete ospedaliera regionale. Classificazione delle unità di assistenza perinatale. Bollettino Regionale del Lazio. 1997; no. 25 del 10/09/1997—Allegato 2, sez. 1
  32. Rekomendacje postepowan w medycynie perinatalnej, wyd. III, red. J. Gadzinowski, G. H. Breborowicz, OWN, Poznan 2002
  33. Ministère des affaires sociales de la santé publique et de l'environnement. Moniteur Belge. 1996;20 :25264 –25281
  34. Health Council of the Netherlands. Perinatal Intensive Care. The Hague, Netherlands: Health Council of the Netherlands; 2000. Publication 2000/08E
  35. British Association of Perinatal Medicine. Standards for hospitals providing neonatal intensive and high dependency care (Second Edition) and categories of babies requiring neonatal care. Available at: www.bapm.org/publications. Accessed August 2, 2007
  36. Organisation of highly specialised functions in the Health Services. National Board of Health, Denmark 2001
  37. Direc

    Formula

    ão-Geral da Saúde Divisão de Saúde Materna, Infantil e dos Adolescentes. Saude Materno Infantil: Rede de Referenciacão Materno-Infantil: 2001:1 –72. Anex 6
  38. Buitendijk S, Zeitlin J, Cuttini M, Langhoff-Roos J, Bottu J. Indicators of fetal and infant health outcomes. Eur J Obstet Gynecol Reprod Biol. 2003;111(suppl 1) :S66 –S77
  39. Gandjour A, Bannenberg A, Lauterbach KW. Threshold volumes associated with higher survival in health care: a systematic review. Med Care. 2003;41 :1129 –1141[CrossRef][ISI][Medline]
  40. Sowden A, Aletras V, Place M, et al. Volume of clinical activity in hospitals and healthcare outcomes, costs, and patient access. Qual Health Care. 1997;6 :109 –114[ISI][Medline]
  41. Bartels DB, Kreienbrock L, Dammann O, Wenzlaff P, Poets CF. Population based study on the outcome of small for gestational age newborns. Arch Dis Child Fetal Neonatal Ed. 2005;90 :F53 –F59[Abstract/Free Full Text]
  42. Hummler HD, Poets C, Vochem M, Hentschel R, Linderkamp O. Mortality and morbidity of very premature infants in Baden-Wurttemberg depending on hospital size: is the current degree of regionalization adequate? [in German]. Z Geburtshilfe Neonatol. 2006;210 :6 –11[CrossRef][Medline]
  43. Horbar JD, Badger GJ, Lewit EM, Rogowski J, Shiono PH. Hospital and patient characteristics associated with variation in 28-day mortality rates for very low birth weight infants. Vermont Oxford Network. Pediatrics. 1997;99 :149 –156[Abstract/Free Full Text]
  44. United Kingdom Neonatal Staffing Study Group. Patient volume, staffing, and workload in relation to risk-adjusted outcomes in a random stratified sample of UK neonatal intensive care units: a prospective evaluation. Lancet. 2002;359 :99 –107[CrossRef][ISI][Medline]
  45. 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]
  46. Callaghan LA, Cartwright DW, O'Rourke P, Davies MW. Infant to staff ratios and risk of mortality in very low birthweight infants. Arch Dis Child Fetal Neonatal Ed. 2003;88 :F94 –F97[Abstract/Free Full Text]
  47. Pronovost PJ, Angus DC, Dorman T, Robinson KA, Dremsizov TT, Young TL. Physician staffing patterns and clinical outcomes in critically ill patients: a systematic review. JAMA. 2002;288 :2151 –2162[Abstract/Free Full Text]
  48. Johansson S, Montgomery SM, Ekbom A, et al. Preterm delivery, level of care, and infant death in Sweden: a population-based study. Pediatrics. 2004;113 :1230 –1235[Abstract/Free Full Text]
  49. Lee SK, McMillan DD, Ohlsson A, et al. The benefit of preterm birth at tertiary care centers is related to gestational age. Am J Obstet Gynecol. 2003;188 :617 –622[CrossRef][ISI][Medline]
  50. Gould JB, Marks AR, Chavez G. Expansion of community-based perinatal care in California. J Perinatol. 2002;22 :630 –640[CrossRef][Medline]
  51. Empana JP, Subtil D, Truffert P. In-hospital mortality of newborn infants born before 33 weeks of gestation depends on the initial level of neonatal care: the EPIPAGE study. Acta Paediatr. 2003;92 :346 –351[CrossRef][ISI][Medline]
  52. Canadian Paediatric Society. Levels of neonatal care. Paediatr Child Health. 2006;11 :303 –306
  53. Goodman DC, Fisher ES, Little GA, Stukel TA, Chang CH, Schoendorf KS. The relation between the availability of neonatal intensive care and neonatal mortality. N Engl J Med. 2002;346 :1538 –1544[Abstract/Free Full Text]
PEDIATRICS (ISSN 1098-4275). ©2007 by the American Academy of Pediatrics




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 CrossRef
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Van Reempts, P.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Van Reempts, P.
Related Collections
Right arrow Premature & Newborn