Project 27/28: Inquiry Into Quality of Neonatal Care and Its Effect on the Survival of Infants Who Were Born at 27 and 28 Weeks in England, Wales, and Northern Ireland
Objective. To identify variations in standards of neonatal care in the first week of life that might have contributed to deaths in infants who were born at 27 and 28 weeks' gestation.
Methods. A case-control study was conducted of infants who were born at 27 and 28 weeks' gestation in England, Wales, and Northern Ireland during a 2-year period. Cases were neonatal deaths; control subjects were randomly selected survivors at day 28. Main outcome measures were failures of prespecified standards of care or deficiencies in care reported by regional panels assessing anonymized medical records.
Results. Failures of standards of care relating to ventilatory support (adjusted odds ratio [OR]: 3.29; 95% confidence interval [CI]: 1.97–5.49), cardiovascular support (OR: 2.37; 95% CI :1.36–4.13), and thermal care (OR: 1.71; 95% CI: 1.21–2.43) were associated with neonatal death. Frequencies of unmet resuscitation standards (range: 3%–46%) and of delays in surfactant administration (range: 38%–40%) were similar in cases and control subjects. Panels identified significantly more deficiencies in all aspects of neonatal care in cases with the exception of the management of infection. Stratification by clinical condition of infants at birth showed a stronger association between overall standard of care and death when infants were in a good condition at birth.
Conclusions. Our findings suggest an association between quality of neonatal care and neonatal deaths, most marked for early thermal care and ventilatory and cardiovascular support. Poor overall quality of care was more strongly associated with deaths when the infant was in a good condition at birth.
Prematurity (birth before 37 weeks' gestation) accounts for 13% of registrable births in the United Kingdom1 and is the major cause (47%) of neonatal deaths.2 As such, it was chosen in 1997 to be the focus for the national Confidential Enquiry into Stillbirths and Deaths in Infancy (CESDI) program. CESDI, which is now the Confidential Enquiry into Maternal and Child Health, makes national recommendations from its findings aimed at improving the care of these immature infants.3
In the early to mid-1990s, premature infant survival rates in United Kingdom were 39% below 26 weeks' gestation and rose to nearly 80% by 28 weeks and 100% at 36 weeks.1,4–6 National data in 1998–2000 showed survival rates at 27 weeks' gestation of 86% and of 90% at 28 weeks.2 We hypothesized that, particularly at these gestations, standards of care may have an effect on survival. This project therefore focused on the maternity and neonatal care of infants who were born between 27 weeks and 0 days (27 + 0) and 28 weeks and 6 days (28 + 6) of gestation, hence the title “Project 27/28.”7 The aim of Project 27/28 was to identify variations in standards of antenatal care or neonatal care during the first week of life that might have contributed to death in preterm infants who were born at 27 and 28 weeks' gestation and to make recommendations for future practice on the basis of these findings. We report here on the neonatal findings of the project. Other details are available at www.cemach.org.uk/publications/p2728/mainreport.pdf.
During the 2-year period September 1, 1998, to August 31, 2000, all live-born infants who had a gestation between 26 + 0 and 29 + 0 weeks and were delivered in 294 maternity units in Northern Ireland, Wales, and England were notified to CESDI. From these, a minimum data set of information was recorded for 3522 infants who were born between 27 + 0 and 28 + 6 weeks.
A case-control design was then used. Detailed information for the enquiry was sought about all the neonatal deaths (cases) and a random sample of infants who survived up to 28 days. Reference numbers for all infants recorded in the logbooks were held in a central database and categorized by gestational age and plurality of pregnancy. A computer program then randomly selected from these categories to ensure appropriate frequency. Infants with lethal congenital anomalies were excluded. For the purposes of comparing characteristics and risk factors of mothers, 2 groups of mothers were also defined: (1) mothers of singleton or multiple births who all survived to day 28 and (2) mothers of singletons who died or of multiple births when 1 or more infants died before day 28.
The “exposure” was the departure from preset standards of care and panel assessments of deficiencies in care received by infants during the first week of life. Standards of care were preset by a group of 18 senior perinatal clinicians, academics, and epidemiologists (the Project 27/28 working group), who met in 1998 to define standards for the enquiry.8 Not all topics had well-established evidence-based standards. When strong evidence was absent, such as for resuscitation or ventilatory care, the working group set standards using national best practices or guidelines recommended by professional bodies and published either in national reports or in journals (see Appendix⇓, references 31–36). Panels reviewed the case notes and considered whether any aspects of neonatal care were substandard. The areas assessed were resuscitation at birth, early thermal care, surfactant therapy, ventilatory and cardiovascular support, management of early and late infection, overall organization of care, and ex utero transfer. Panels were asked each time they identified a deficiency in care to describe it by free text (available at www.cemach.org.uk). At the end of each enquiry, panels were asked to grade the overall obstetric and neonatal care received as good, acceptable, substandard minor, or substandard major. Crude odds ratios (OR) for neonatal death were calculated (with their 95% confidence intervals [CIs]) for different standards of care when poor care was defined as (1) departure from 1 of 10 preestablished standards of neonatal care and (2) as deficiencies in care as recognized by panels' opinion.
Multidisciplinary panels of independent experts (midwife, obstetrician, pediatrician, pathologist, neonatal nurse, anesthetist, and public health specialist) ranging from 3 to 18 participants (median: 8) were convened regularly in each health region to assess standards of care from the medical notes. These notes were anonymized, and neither the name of patients nor the name of the hospital was known to the panel members. After the enquiry, any form of identification was destroyed; therefore, no information about follow-up or morbidity outcomes was available.
Obstetric members of the panels were blinded to the outcome, but pediatric members were not, as blinding was not feasible for the assessment of care after the delivery of the infant. They nevertheless applied strictly defined criteria of assessment (available at www.cemach.org.uk).
Recording and collating “opinion” about the quality of patient care is complex. The panel judgments in these enquiries reflect individual health professionals' opinions; therefore, variation in responses is likely. This study built on the experience gained from previous exercises, which used a “second-pass” panel approach, whereby the same case was reviewed by 2 panels separately. These exercises had shown that there could be striking differences in opinion between panels.8,9 As a consequence, the method of Project 27/28 was modified: a structured approach was used with specific questions to the panels, standard documents, clearly predefined standards, and a grading system. A second-pass panel to assess the robustness of the Project 27/28 enquiry was set up for 92 cases and control subjects. Five to 6 second-pass panels were held in each region, and at least 1 neonatologist was available on each panel (range: 1–3).
Differences in categorical variables and trends were compared using the χ2 test and in continuous variables using the Student's t test or the Mann-Whitney U test as appropriate. The κ statistic was used to define agreement between panels 1 and 2. A κ >.6 was considered as good or very good agreement, .4 to .6 as moderate, .2 to .4 as fair, and <.2 as poor.10
Regardless of the quality of care, there are likely to be inherent differences between infants who do and do not survive: sicker infants are more at risk for a poor outcome regardless of the type of care that they receive. We therefore adjusted the analysis to take into account illness severity by using male gender, birth weight ≤5th centile, and the clinical condition of the infant at 5 minutes of life (defined as absent or ineffective respiration or a persistent heart rate below 100 beats per minute). Because no difference was found in terms of survival across different types of units where mothers and infants were treated,2 no adjustment by types of units was made. Adjusted ORs were obtained using multiple logistic regression, and significance was assessed using the likelihood ratio test.
Sicker infants may also need more medical intervention with the potential for more substandard care or errors. This hypothesis was explored by looking at the interaction between outcomes and standards of care for infants who were considered to be in poor or in good clinical condition at 5 minutes of life. There were varying amounts of missing data for the risk factors and standards, which are indicated in the tables as “not known.”
The overall neonatal mortality rate from the original complete data set was 12% (421 of 3522), and 421 control subjects were randomly selected. There were 81 exclusions (55 cases and 26 control subjects; the largest group excluded from the cases were those with lethal congenital malformations; Table 1). Enquiries were held on 761 infants (366 cases and 395 control subjects) and their 723 mothers (352 mothers of infants who died and 371 mothers of survivors).
There was no difference between mothers of infants who died and of those who survived in terms of age, ethnic group, civil status, relevant medical history before the pregnancy (eg, diabetes, hypertension, asthma, epilepsy, sickle cell disease), reported smoking habits, attendance at antenatal appointments, parity, history of previous preterm delivery, or obstetric history (Table 2). Regarding pregnancy-related clinical complications, no differences were detected between groups for preterm premature rupture of membrane (PPROM), duration of membrane rupture before delivery (although membrane rupture occurred earlier in the group of mothers of infants who died), pregnancy-induced hypertension, major placental bleeding, or chorioamnionitis. Bacteriuria was more frequent in the control group. Mothers of infants who survived presented more frequently with preterm labor. There was no evidence of differences between groups in the hospital settings where the mothers were delivered (number of deliveries, type of neonatal unit, and number of neonatal intensive care days). Considering antenatal management, a complete course of maternal steroids (>24 hours and <7 days before delivery) was just significantly more frequently given to mothers of infants who survived (crude OR: 0.74; 95% CI: 0.55–1.00), but this difference was no longer significant after adjustment for PPROM and preterm labor.
Antenatal Risk Factors and Early Neonatal Characteristics
Of the total 761 infants, 217 (29%) were from a multiple birth and 382 (50.2%) were born at 27 weeks. In the index cases, the median age at death was 3.3 days (interquartile range: 0.8–8.4 days). A total of 99 (27%) infants died within the first 24 hours of life, 51 (21%) in the first 12 hours, and 26 (7%) within the first hour.
There was more antenatal ultrasound evidence of fetal compromise and of poor fetal growth <5th centile and more urgent cesarean sections (for a clinical situation of impending fetal death) performed in the group of infants who died (Table 3). Infants who died had lower birth weight and were more likely to be male and to have non–life-threatening congenital malformations (Table 3). All of the recorded parameters reflecting the infants' initial condition (Apgar score, cord pH, temperature, and blood pressure in the first 12 hours; clinical risk index for babies [CRIB] score8) showed that infants who died had been in poorer condition at birth (Table 3).
Second-Pass Panel Validation (92 Cases and Control Subjects)
Concerning neonatal standards, agreement was “very good” or “good” in areas of care in which the responses requested were specific (resuscitation, surfactant administration) or the standards were based on quantitative measurements (temperature, the timing of the first dose of surfactant). There was only poor to moderate agreement when the responses were based on a subjective evaluation of care, such as an assessment of the quality of the effort to achieve a standard (for ventilation and cardiovascular support). There was fair agreement about the overall grade of care.
Standards of Neonatal Care
Nearly all infants were resuscitated at birth: 355 (97%) of infants who died and 365 (92%) control subjects. There were 5 resuscitation standards. One related to the pediatric staff present at delivery and the other 4 to resuscitation proficiency at delivery and on admission to the neonatal unit. Timely attendance by a full complement of skilled pediatric staff (a senior and a junior doctor or an advanced nurse practitioner) was not achieved in 46% of cases and 44% of control subjects. For cases and control subjects combined, staff attendance when standards were not achieved was consultant only (3%), middle-grade doctor only (21%), junior doctor or advanced nurse practitioner only (15%), midwife or neonatal nurse only (5%), and others (1%). Lack of tracheal intubation and positive pressure ventilation in infants who were in poor condition at 5 minutes were reported for one fifth of infants (18% of cases and 24% of control subjects). There was failure to treat a heart rate below 60 beats per minute at any stage by cardiac massage in 10% of cases and 20% of control subjects (not significantly different).
Although most (96%) infants received surfactant, administration within 1 hour after birth was not achieved in 38% of cases and 40% of control subjects. The timing of administration of the first dose of surfactant was similar for infants who died and those who survived at different time ranges (eg, 43% of cases and 40% of survivors received surfactant within half an hour; data not shown). However, exposure to poor care in early resuscitation and surfactant administration was not associated with a significantly increased risk for dying in the study.
Poor care in relation to 3 neonatal standards was associated with a significantly increased risk for dying: temperature control, ventilation, and cardiovascular support. Thus, a temperature ≤36.0°C on admission to the neonatal unit was reported in two thirds of all infants (73% of infants who died and 59% of control subjects) and was associated with a 70% increased risk for death. Poor ventilatory and cardiovascular standards were defined as failure to monitor or properly document blood gases or blood pressure and failure to make therapeutic adjustment to maintain blood gases or blood pressure within predefined limits (Appendix⇓). Standards of cardiovascular support failed to be achieved in 15% of infants who died and 7% of control subjects. Similarly, there was a failure to achieve adequate ventilatory support in 20% of infants who died and 7% of control subjects, and this was associated with a 2- to 3-fold increase in the risk for death.
Panels' Opinion of Deficiency of Neonatal Care
Except for the management of infection, deficiencies in the standards of care were identified for between 7% and 51% of infants, and these were consistently and significantly more frequent in cases than in control subjects (Table 5).
Overall Standards of Neonatal Care
Overall neonatal care was considered good or acceptable in 47% of infants who died and in 77% of control subjects. Major substandard care was significantly more frequent among infants who died (28%) than in those who survived (6%; Tables 6 and 7).
Considering condition at birth, there was always a nonsignificant tendency toward an increased risk for dying when infants who were in good clinical condition were exposed to poor care, compared with infants who were in poor clinical condition at birth. This interaction was highly significant (P < .001) when the overall grade of care was considered, ie, there was an increased risk for dying when infants who were in good condition at birth were exposed to a lower overall grade of care.
Project 27/28 is the first nationwide study to provide a detailed examination of the standards of neonatal care using panels of experts to conduct a structured assessment of anonymized obstetric and neonatal notes. Two regional UK studies have used the same enquiry method,11,12 and studies from the North American Vermont-Oxford Network have focused on specific clinical care areas using data collection from a large network of neonatal units.13
In keeping with the findings of others,14,15 mothers whose infants died were characterized by a lower incidence of preterm labor and earlier PPROM than control subjects. Bacteriuria was less frequent among mothers of infants who died, possibly because of its association with spontaneous preterm labor.16 The marked similarities between the maternal groups (Table 2) highlighted the difficulties of identifying the mother of an “at risk” infant before delivery. There were 3 notable differences between the mothers of infants who were born at 27 and 28 weeks' gestation compared with the general British maternity population: (1) there were more multiple births17; (2) there was a higher proportion of smokers17; and (3) there was a different distribution of ethnicity (not shown in Table 2): the study population comprised 79% white, 9% black, 8% Asian, and 5% others, compared with the general maternity population, which comprised 91% white, 2% black, 3% Asian, 1% mixed, and <1% others.18 In contrast to the maternal characteristics, there were multiple differences between the 2 groups of infants studied, which are shown in Table 3.
Population characteristics may affect outcome,19,20 and a variety of measures may be used to adjust for differences in risk-specific hospital performance. It was believed to be inappropriate to use the CRIB score,21 which is based on routine data recorded within 12 hours of life and can be used to describe risk-adjusted mortality, because it is partly dependent on early management.21–23 Furthermore, infants for whom the CRIB score was recorded were unrepresentative of the whole group in that they were older when they died and had higher Apgar scores and lower systolic blood pressure. Because other measures of illness severity may also be affected by inadequate care,21–23 we restricted adjustments in this study to the infant's gender, birth weight, and clinical condition within 5 minutes of birth. We considered the proportion of infants who were in poor clinical condition at 5 minutes of age in different neonatal settings where they were treated and found no significant differences (test for trend with tertiary units as reference: P = .26, data available but not shown).
There was a high proportion of cases overall for whom standards of neonatal care were not met (Table 4). However, exposure to poor care in early resuscitation and surfactant administration were not associated with a significantly increased risk for dying in the study. This might be a true finding or a consequence of a lack of power of the study, which had sufficient power to detect only differences between cases and control subjects of >7% in the proportion exposed to poor care at a 5% significance level.24 Our findings relating to delayed surfactant administration were similar to those from the Vermont-Oxford network during the same study period 1998–2000,25 and the conclusion of that study that there was a gap between the evidence from randomized trials and clinical practice may also apply to the United Kingdom.25 Overall, there was a high proportion of poor care in resuscitation and late surfactant administration, which, even if not directly causing death, might have had an impact on the quality of survival.
A significantly increased risk for dying was associated with poor-quality neonatal care in 3 areas (temperature control, ventilation, and cardiovascular support). Although hypothermia may indicate a preterminal condition rather than poor care, when the infants who died in the first 24 hours were removed from the analysis, a low admission temperature was still associated with death. The strength of the association (OR: 1.71) between hypothermia ≤36°C on admission to the neonatal unit and death in this study of infants who were born at 27 and 28 weeks' gestation was similar to that found by the EPICure study of infants <26 weeks' gestation when a cutoff point of ≤35°C was used4 and emphasizes the importance of good temperature control for these small infants.
The condition of an infant at 5 minutes of age is a marker of illness severity and of the amount of resuscitation required and hence of cold exposure. After adjustment for the condition of the infant 5 minutes after birth, there were only minor changes in ORs (from 1.91 to 1.71), but mortality was reduced when the incubator temperature was adjusted to maintain an anterior abdominal skin temperature above 36°C.26 This standard therefore remains an objective of good care in very preterm infants. Simple measures, such as occlusive wrapping in the delivery room, improve early temperature control in infants who are born at <28 weeks' gestation.27,28 An appropriate environmental temperature is also important, and the temperature of the delivery room should be at least 25°C.29
Three limitations of this study need to be considered. First, the proxy measure chosen for adjustment for illness severity might not properly reflect disease severity. Adjustment for illness severity resulted in only minor changes in ORs, and increased exposure of sicker infants to substandard care or errors may have had only a limited effect on the association between death and standards of care. In considering the possibility that substandard care and errors might be reduced by the presence of a senior clinician, we found that consultants (attending physicians) were more likely to be present at the resuscitation of sick infants (data available, not shown). This observation may explain the tendency for substandard care in resuscitation to be less frequently observed in infants who died than in those who survived (Table 4). The hypothesis that there was a bias because sicker infants were more likely to be exposed to potential errors was explored by looking at the interaction among outcome, clinical condition at birth, and standard of care. There was an increased risk for dying when infants who were in good condition were exposed to a lower overall grade of care compared with infants who were in poor clinical condition at birth (Tables 6 and 7).
A second potential limitation is that appraising the standard of care was difficult in the absence of a strong evidence base or easily measured criteria (Appendix⇓). For instance, evidence for recommending the timing of prophylactic surfactant at 27 and 28 weeks' gestation was lacking at the time the Project 27/28 working group convened before the study. The cutoff of 1 hour after birth was chosen to define delay because a UK neonatal staffing study30 had observed that surfactant was more likely to be administered within the first hour after delivery in well-staffed units of medium activity, where guidelines stated that prophylaxis should be given as soon as possible after birth. A standards document was compiled by the Project 27/28 Working Group2,8 to guide panel members and to improve the consistency of the assessment process. Analysis of the second-pass panel results suggested an overall robustness and consistency by the enquiry panels, especially when quantitative criteria were available.
A third potential limitation of the study was that pediatric members of the panels were not blind to the outcome. The knowledge that an infant had died may also have resulted in more stringent assessment of the quality of care that was given to infants who died. However, panels followed a strict protocol and had to justify their opinions using free text; most comments related to areas of care, such as ventilatory care, for which there was a strong association between poor care and death.
This study found an association between the quality of neonatal care and neonatal death. The association between poor clinical care and poor outcome was most marked for early thermal care and ventilatory and cardiovascular support. The association was strongest for infants who were born in good condition than for those who were already very sick. The high proportion of poor care that was found, even if not related to death, is likely to have had an impact on the quality of survivors.
This work involved the contributions of numerous people both outside and with the CESDI/Confidential Enquiry into Maternal and Child Health organization over several years. In particular, The Regional Coordinators for helping to conduct the study, the Regional Chairs for conducting the Panels, and the Project 27/28 Working Group for overseeing the setting up of the standards and the enquiry proforma; the team at the CESDI (now Confidential Enquiry into Maternal and Child Health) secretariat; Dr Joseph Kim for statistical advice in this article (Medical Statistics Unit, London School of Hygiene and Tropical Medicine); and, in support of the original Project 27/28, the members of the CESDI Executive Steering Group of the Maternal and Child Health Research Consortium, the members of the CESDI Interim Advisory Group and Recommendations Advisory Group, the support staff at the regional offices, the panelists, the many district coordinators, and others based throughout England, Wales, and Northern Ireland. We add a special thanks to the parents and their premature infants.
- Accepted March 7, 2005.
- Reprint requests to (D.A.) Confidential Enquiry into Maternal and Child Health (CEMACH) Central Office, 188 Baker St, London NW1 5SD. E-mail:
No conflict of interest declared.
- ↵All Wales Perinatal Survey and CESDI. Cardiff, Wales: Perinatal Survey Office, University of Wales College of Medicine; 1997 and 1999
- ↵CESDI. Confidential Enquiry into Stillbirths and Deaths in Infancy: 1st of January–31st December 1999. London, United Kingdom: Maternal and Child Health Consortium; 2001
- ↵Weindling AM. The Confidential Enquiry into Maternal and Child Health (CEMACH). Arch Dis Child.2003;88 :1034– 1037 [published erratum appears in Arch Dis Child. 2004; 89:293]
- ↵Costeloe K, Hennessy E, Gibson AT, Marlow N, Wilkinson AR. The EPICure study: outcomes to discharge from hospital for infants born at the threshold of viability. Pediatrics.2000;106 :659– 671
- Draper ES, Manktelow B, Field DJ, James D. Prediction of survival for preterm births by weight and gestational age: retrospective population based study. BMJ.1999;319 :1093– 1097
- ↵CESDI Project 27/28. Confidential Enquiry into Stillbirths and Deaths in Infancy. London, United Kingdom: Maternal and Child Health Consortium; 2003. Available at www.cemach.org.uk. Accessed October 24, 2005
- ↵CESDI. Confidential Enquiry Into Stillbirths and Deaths in Infancy: 8th Annual Report. London, United Kingdom: Maternal and Child Health Consortium; 1999. Available at: www.cemach.org.uk. Accessed October 24, 2005
- ↵CESDI. Confidential Enquiry Into Stillbirths and Deaths in Infancy: 6th Annual Report. London, United Kingdom: Maternal and Child Health Consortium; 1998. Available at: www.cemach.org.uk. Accessed October 24, 2005
- ↵Altman DG. Inter-rater agreement. In: Practical Statistics for Medical Research. London, United Kingdom: Chapman and Hall Publishers; 1991:403– 409
- ↵Draper ES, Kurinczuk JJ, Lamming CR, Clarke M, James D, Field D. A confidential enquiry into cases of neonatal encephalopathy. Arch Dis Child Fetal Neonatal Ed.2002;87 :F176– F180
- ↵Horbar JD. The Vermont Oxford Network: evidence-based quality improvement for neonatology. Pediatrics.1999;103 :350– 359
- ↵Chapple J. Perinatal mortality. In: Chamberlain G, Steer P, eds. Turnbull's Obstetrics. 3rd ed. Edinburgh, Scotland: Churchill Livingstone; 2001:729– 741
- ↵IFS. Infant Feeding Survey 2000. London, United Kingdom: The Stationary Office; 2002
- ↵Jain A, Fleming P. Project 27/28. An enquiry into the quality of care and its effect on the survival of babies born at 27–28 weeks. Arch Dis Child Fetal Neonatal Ed.2004;89 :F14– F16
- ↵Field D, Petersen S, Clarke M, Draper ES. Extreme prematurity in the UK and Denmark: population differences in viability. Arch Dis Child Fetal Neonatal Ed.2002;87 :F172– F175
- Baumer JH, Wright D, Mill T. Illness severity measured by CRIB score: a product of changes in perinatal care? Arch Dis Child Fetal Neonatal Ed.1997;77 :F211– F215
- ↵Hope P. CRIB, son of Apgar, brother to APACHE. Arch Dis Child Fetal Neonatal Ed.1995;72 :F81– F83
- ↵Fleiss JL. Statistical Methods for Rates and Proportions. New York, NY: John Wiley & Sons; 1981
- ↵Horbar JD, Carpenter JH, Buzas J, et al; Vermont Oxford Network. Timing of initial surfactant treatment for infants 23 to 29 weeks' gestation: is routine practice evidence based? Pediatrics.2004;113 :1593– 1602
- ↵Sinclair JC. Servo-control for maintaining abdominal skin temperature at 36°C in low birth weight infants. Cochrane Database Syst Rev.2002;(1):CD001074
- ↵Lyon AJ, Stenson B. Cold comfort for babies. Arch Dis Child Fetal Neonatal Ed.2004;89 :F93– F94
- ↵World Health Organization. Thermal Protection of the Newborn, WHO/RHT/MSM/97.2. Geneva, Switzerland: World Health Organization; 1997
- ↵British Paediatric Association. Neonatal Resuscitation. London, United Kingdom: The British Paediatric Association; 1993
- ↵Royal College of Paediatrics and Child Health. Resuscitation of Babies at Birth. London, United Kingdom: Royal College of Paediatrics and Child Health; 2000
- ↵Project 27/28 Working Group. Confidential Enquiry into Stillbirths and Deaths in Infancy. 1998. Available at: www.cemach.org.uk. Accessed October 24, 2005
- ↵Report of a Working Group of the British Association of Perinatal Medicine and the Research Unit of the Royal College of Physicians. Development of audit measures and guidelines for good practice in the management of neonatal respiratory distress syndrome. Arch Dis Child.1992;67 :1221– 1227
- ↵Soll RF, Morley CJ. Prophylactic Surfactant Vs Treatment With Surfactant (Cochrane Review). In: The Cochrane Library, Issue 2. Oxford, United Kingdom: Update Software; 1998
- ↵Royal College of Paediatrics and Child Health. Guidelines for Good Practice. Management of Neonatal Respiratory Distress Syndrome. London, United Kingdom: Royal College of Paediatrics and Child Health; 2000
- Copyright © 2005 by the American Academy of Pediatrics