High quality of care in neonatology implies providing an appropriate level of care to well newborn babies as well as more specialised care for the few babies who need it. Audit, surveillance and outcome studies may not always capture the complexity of quality of care and its contribution to outcome, and a more focused approach to standards of care evaluation may be required. Future progress in this field in the UK would benefit from a more coordinated approach from different organisations to bring together expertise in large database, management and analysis, audit and a national profile for feedback, evidence-based guidelines and guidelines development skills, expertise in the practice of changes together with the promotion by credible perinatal authorities of clinical practice.
Statistics from Altmetric.com
WHAT IS QUALITY IN NEONATAL CARE?
Quality of care has had various definitions in the literature over the years. The definitions have few points in common, although most of them assume that biomedical outcomes are important endpoints.1 With time, the definitions have also encompassed outcomes that relate to users’ or providers’ satisfaction and to financial aspects.1 2 More recently, a comprehensive definition of quality of maternity care was proposed by the Maternal Health Programme group of the London School of Hygiene and Tropical Medicine, UK1:
High quality care in maternity services involves providing a minimum level of care to all pregnant women and their newborn babies and a higher level of care to those who need it. This should be done while obtaining the best possible medical outcome, and while providing care that satisfies women and their families and their care providers. Such care should maintain sound managerial performance and develop existing services in order to raise the standards of care provided to all women.
This definition can easily be adapted to quality of care in neonatology. High quality care in neonatology implies providing an appropriate level of care to well newborn babies while providing a good quality of more specialised care to the few babies who need it. This should be done while obtaining the best medical neonatal outcome and satisfying parents and families as well as providers. This should be delivered in the context of sound managerial and financial performance and while developing existing perinatal services to further raise neonatal standards of care.1 Some studies are now including non-biomedical outcomes, such as staff well-being in the UK Neonatal Staffing Study3 or parents’ satisfaction.4
OUTCOME STUDIES IN RELATION TO QUALITY OF NEONATAL CARE
Outcome studies have evolved hand in hand with the development of neonatal care from its outset until today,5–7 and they represent a large percent-age of the publications in neonatology. These studies look at outcome in relation to populations of babies defined by their demographics or by medical conditions and related treatments. They constitute an important part of the information that neonatal clinicians need to support their own practice. Studies exploring the relationship between quality of care and outcome are less numerous and more recent. The UK neonatal staffing study3 described risk-adjusted mortality and morbidity outcomes in UK neonatal intensive care units in the context of differences in neonatal units’ volume, staffing level and compliance to national standards of care. It concluded that observed outcomes between units were not markedly different from those expected, given the degree of illness severity of their populations, although mortality did increase with increased staff workload and cot occupancy.3 Most quality of care attributes, such as audit, guidelines, training and leadership, were not independently related to outcomes although the rates of nosocomial infection decreased with quality improvement measures, such as the presence of an infection control nurse.
Marked variations in clinical practice and crude outcomes have been reported between units.8 9 These variations can be explained by differences in case mix or by chance, but even after adjustments, some units still have considerably fewer deaths than expected whereas others have more.8 Although some variations could be related to other risks factors not adjusted for, such as demographics of the population served by some hospitals, they could suggest possible differences in quality of care and therefore opportunities for quality improvement.8 Quality improvement is the main goal of the International Vermont-Oxford Neonatal Network (VON), a collaborative organisation of neonatal units that work to improve the quality of medical care by integrating research and clinical practice.10 Its database is the cornerstone of quality improvement efforts8 and is one of the largest anonymous databases on care and outcomes for infants 401–1500 g or <29 weeks’ gestational age. VON provides comprehensive, confidential, instantaneous online or quarterly and annual reports to participating neonatal units on performances, trends over time and benchmarking.8 In the UK, the Confidential Enquiry into Maternal and Child Health (CEMACH) is commissioned by the National Patient Safety Agency (NPSA) to carry out its Maternal and Child Health programme including perinatal mortality surveillance. CEMACH publishes perinatal mortality reports at a national level and for strategic health authorities, neonatal networks and National Health Service (NHS) trusts from the same data system.11
The crude neonatal mortality rates plotted against the number of live births, the average neonatal mortality rate and associated 95% confidence intervals are shown in fig 1. If a trust or neonatal network lies outside the 95% confidence limits, then it has a rate that is markedly different from the average rate. The further the point is outside the confidence limits, the less likely the rate is due to chance, and the more likely it is to be truly different from the national average. This figure, known as funnel plot, shows wide variations in the crude perinatal mortality rate among NHS hospital trusts.11 As previously discussed, these variations should not be interpreted as direct indicators of the quality of care. Other factors may influence rates such as sociodemographic characteristics of the population or the case mix or referral pattern of an individual hospital or differences in the way data are reported from individual trusts. Regarding the latter, CEMACH found, for example, that 11% of the notified neonatal deaths had a gestation between 17 and 22 weeks.11
Variations in outcome are influenced by the severity of illness of each individual baby10 and some form of adjustment using disease severity such as the CRIB score12 may be useful. Data needed for the CRIB score may be difficult to capture for surveillance purposes because of the amount of information required. CEMACH has successfully used a simple surrogate measure of illness severity for adjustment in a previous study by using the baby’s sex, birth weight and clinical condition within 5 min of birth.13 Along the lines of VON, CEMACH is now involved in a process of feedback, benchmarking and dialogue with each hospital. It will soon provide each unit with risk-adjusted outcomes so that published mortality data are more meaningful as quality measures. Such adjustment will include a unit’s level of care and referral patterns as well as illness severity measurement. To avoid pitfalls such as incomplete data and limited checks of data quality which can occur when using large database, CEMACH has a strong national data collection network of local health professionals in each maternity unit. The data are supplemented by reporting deaths using child health systems and local congenital anomaly registers, where data sharing arrangements have been in place for many years. This multiple source reporting leads to a high level of ascertainment of deaths. Data compiled centrally are also cross-validated with registration data on stillbirths and neonatal deaths from the Office for National Statistics.11
Similarly, a national project funded by the Healthcare Commission and supported by the Department of Health and the Royal College of Paediatrics and Child Health, the National Neonatal Audit Programme (NNAP), has been established with the aim of informing good clinical practice in aspects of neonatal care by auditing national standards.14 The British Association of Perinatal Medicine has established a minimum neonatal dataset with precise definitions of data items for standardised annual data reporting and eventually national data collection.15 Large regional datasets, such as the South of England Database (SEND), are now up and running and it is hoped that national neonatal reporting will be available in the future, allowing further exploration of quality of care and outcome.16
CONFIDENTIAL ENQUIRY METHODS, QUALITY OF NEONATAL CARE AND OUTCOME
Audit, surveillance and outcome studies may not always capture the complexity of quality of care assessment required and its contribution to outcome, and a more focused approach to standards of care evaluation may be necessary. For example, the patient safety imperative of neonatal encephalopathy led the NPSA to commission preliminary work from the National Perinatal Epidemiology Unit (NPEU). This report concluded that17:
given the complexity of data needed it may not be possible to collect the necessary level of detail required by current criteria to identify the contribution of intrapartum events and hypoxia in every neonatal unit.
Assessment of potentially avoidable cases of neonatal encephalopathy may therefore require a different approach. Confidential enquiry methods, developed initially by the Confidential Enquiry into Stillbirths and Deaths in Infancy (CESDI) and more recently by CEMACH, has shown over the years to be an appropriate tool for the detailed exploration of complex perinatal issues.18 19 The confidential enquiry methods comprise the evaluation of quality of care through established standards of care and the assessment of anonymised medical records by panels of independent experts blind to the outcome.20 CESDI and CEMACH have successfully collected large national case–control datasets on standards of care by submitting, for example, the care of individual babies to such expert external peer appraisal. This method has identified substandard care and deficiencies in the care delivered to premature babies born in England, Wales and Northern Ireland, and its association with mortality.13 18 The study found deficiencies in all aspects of neonatal care consistently more frequent in babies who died, even after adjusting for initial illness severity. A temperature ⩽36°C on admission to the neonatal unit was associated with a twofold increased risk of death. Delay in surfactant administration was found in more than 40% of cases, and the recommended composition of resuscitation team at birth was not in place in 45% of cases. The study also showed deficiencies in organisation of care and neonatal transport. CEMACH produced national recommendations for future practice based on these enquiry findings.18
In parallel, a new national position statement21 on the early care of premature babies has been developed by the UK’s leading national institution for clinical governance in neonatology, the British Association of Perinatal Medicine (BAPM). Further contribution to a recent national movement of quality improvement in neonatal care culminated in the setting up of clinically managed neonatal networks22 and the establishment of the National Neonatal Audit Programme.14 More recently CEMACH conducted a neonatal enquiry into the care of term babies of mothers with diabetes, which is expected to bring changes to practice nationally.23 It is also likely to inform the future diabetes in pregnancy national guideline developed by the National Collaborative Centre for Women’s and Children’s Health of the National Institute for Health and Clinical Excellence (NICE).24 Further neonatal enquiries will be developed in the future as part of the CEMACH programme.
RECOMMENDATIONS, QUALITY OF NEONATAL CARE AND OUTCOME
Moving from raw data collection and description to making recommendations brings challenges. Many clinical guidelines and recommendations are produced but with varying success with regard to their implementation.25 26 Various reasons have been put forward such as: “qualities of the guidelines, characteristics of the health care professionals, characteristics of the practice setting, incentives, regulation and patient factors”.27 Guidelines are more successful if:
the recommendations are based on low complexity procedures28;
the recommendations are compatible with existing norms and values, easy to follow and ideally not requiring new knowledge or skills29;
the recommendations are evidence based.30
It has been suggested that the quality of guidelines could be improved. The AGREE collaboration developed and validated an international instrument containing 23 items for the assessment of the quality of guidelines development.31 Guidelines adhering to these quality and attributes could improve implementation to some extent.32
ACTIVE DISSEMINATION OF INFORMATION, QUALITY OF NEONATAL CARE AND OUTCOME
Although neonatal audit and benchmarking of adjusted outcome information are considered as a platform for quality improvement efforts at local level, they are not sufficient to provide a basis for lasting improvement.8 Benchmarking has also its own methodological limitations and variations in outcome reflect a complex mix of different factors that are often difficult to disentangle.33 There are commonly gaps between research evidence, audit information, and policy and practice, thus making quality and outcome improvement difficult. Over 30% of care delivered in North America is not based on up-to-date information.34 35 These gaps need to be bridged to improve the quality of care. Systematic reviews of dissemination of information from the Cochrane Effective Practice and Organisation of Care Group—EPOC32 have shown that traditional, more passive ways used by health professionals to keep up to date with their practice, such as educational material, conferences and courses, have little impact on quality of care. Larger effects can be expected with more active methods of dissemination such as educational outreach by experts or trained facilitators and multifaceted interventions.36
CEMACH findings relating to delayed surfactant administration were similar to those from VON during the same study period 1998–2000.13 37 In addition, both studies concluded that there was a gap between evidence from randomised trials and clinical neonatal practice. A VON collaborative intervention study to promote evidence-based surfactant treatment for preterm infants using active dissemination techniques showed a large effect in surfactant uptake in labour wards and a marked reduction in the timing of surfactant administration from birth.37 Surfactant administration to infants less than 32 weeks’ gestation in Northern Ireland increased remarkably after a quality improvement programme was put in place.38 The BLISS cluster randomised controlled trial of the Effect of “Active Dissemination of Information” on standards of care of premature babies in England (BEADI) is a cluster randomised trial currently underway to assess the effect of active dissemination techniques on changes in policy and practice at neonatal units level in England. The trial is being done with the collaboration of the London School of Hygiene and Tropical Medicine, CEMACH and the EPICure2 study group.39 BEADI is testing the effects of focused interactive training workshops on neonatal outcomes.
In addition, clinicians may not respond to evidence by implementing changes in policies and practice. This may be so because of lack of knowledge or disagreement with recommendations but mainly because of local barriers to implementation.40 41 The reasons why clinicians do not respond to evidence is best addressed by qualitative studies that try to understand the process of knowledge transfer and local implementation barriers and facilitators.40–45 Very few studies have reported on these issues in the neonatal field.46 47 Given the relative paucity of such research evidence, the second objective of BEADI is to use qualitative methods to explore the process of implementation of recommendations derived from neonatal research findings and local barriers to changes in UK neonatal units.
An active component to promote changes in practice is therefore likely to be required to improve quality of care and outcome. Different active techniques have been described. Qualitative improvement collaborative programmes instituted within the VON neonatal network using behaviour changes theories have been associated, for example, with a reduction of nosocomial infections.48 But it is the combination of audit and surveillance using large neonatal database and quality improvement efforts that is more likely to bring changes in practice and better neonatal outcome. Taking for example policies at the edge of viability, the approach of the VON organisation has been twofold. Initially, an audit of outcome of babies with birth weights of 401–500 g within their neonatal networks was reported.49 This was followed by a collaborative quality improvement study that aimed at improving (i) the collaboration between obstetric and neonatal teams for the management of the pregnancy at the margin of viability and (ii) the team performance during high-risk deliveries using aviation-based communication techniques, and team training and video simulation.50
Fast changes in quality of care through patient safety improvements techniques such as the “care bundles” approach have been reported in adult intensive care settings.51 This approach allows a team of clinicians to assess how they deliver therapeutic interventions to patients. It is based on the development and implementation of a set of evidence-based interventions all of which should be delivered at the same time during a specific aspect of patient care, such as long line insertion.52 Such interventions still need to be tested for neonatal intensive care.
The assessment of quality of neonatal care and improvement of care techniques that will impact on outcome are complex and, compared with adult medicine, are still in their infancy. Future progress in this field in the UK would benefit from a more coordinated approach of different organisations to bring together expertise in large database management and analysis, audit and a national profile for feedback, evidence-based and guidelines development skills, expertise in practice of changes together with promotion by credible perinatal authorities on clinical practice. A “British Vermont-Oxford Network” approach could benefit from the organisation of an annual meeting on quality improvement in neonatology with representation from all the units in UK.
The author would like to thank D Elbourne and A Wilkinson for their critical review and comments of the manuscript and the team at the CEMACH central office for their support.
Competing interests: None.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.