Principal findings
Impact of place of birth on mortality and key neonatal morbidities
In workstream 1, we found no difference in mortality up to the first year of life between birth and early care in a NICU (tertiary) compared with LNU (non-tertiary) setting. There was a significant risk of SBI in those born in LNU; the highest proportion of this was in babies born at 27 weeks, who were transferred out of LNU in the first 72 hours of life. When we removed babies transferred out in the first 72 hours of life, and then babies born at 27 weeks, from the analysis, our data lost significance.
We calculated that for babies born at 27 weeks gestation, birth in maternity services colocated with NICU reduced the risk of SBI from 11.9% to 7.7%, a reduction of 4.2%, with a NNT of 25. This means that 25 babies would need to be delivered in a NICU, rather than LNU, to prevent one SBI for babies born at 27 weeks gestation.
When we examined the association of SBI in more detail, analysing high- versus low-volume units (based on quartiles for IC days) we found that there was additional benefit for reduced SBI in high-volume units where the threshold for separation of high- versus low-volume units was 1614 IC days per year (upper quartile of IC days; NNT of 4). This was lost when the threshold for high- versus low-volume days was 440 IC days per year (lowest quartile of IC days).
Overall, 1 in 3 babies were transferred between units in this cohort, and 1 in 2 at 27 weeks gestation.
Does quality of care, independent of designation as neonatal intensive care unit or local neonatal unit, matter?
In workstream 2, we identified a positive association between units that comply with selected evidence- or consensus-based clinical practice, and a reduction in LOS. Units in high areas of social deprivation, and those with less staff, were less likely to meet best practice. These findings may have significant implications for cot capacity in neonatal units, parent and family engagement and taking baby home earlier, allocation of resources, and an overall cost saving for the NHS.
Cost-effectiveness of care in local neonatal unit compared with neonatal intensive care unit
In workstream 3, we found that the NHS costs for babies born at 27–31 weeks gestation was ~0.26 billion GBP annually, and that the largest contribution to this was the costs of daily care in neonatal units. These costs were highest for those born at 27 weeks gestation as they utilised the greatest number of neonatal bed-days. Within our cohort, their costs were more than three times those for babies born at 31 weeks gestation in England. In our cost-effectiveness analysis, we found that the cost of care was similar for babies born at 27, 28 and 29 weeks gestation in NICU and LNU, but that it was significantly lower in NICU for babies born at 30 and 31 weeks gestation. This was driven by the differences in the durations of the different levels of care provided in NIC compared with LNU, with more ICU/HDU days in LNU and more special care days in NICU.
Parent and staff perspectives on place of care
Our findings in workstream 4 suggest that for parents, continuity of care mattered most. Place of care discussions should include assessment of the burden placed on parents of various socioeconomic backgrounds and their consequent ability to maintain continuity in the face of disruptions. Supporting continuity133,134 is not only a vital element in improving parents’ subjective experience of neonatal services, but there is also strong evidence suggesting that continuity reduces stress on babies and improves both short- and long-term developmental outcomes.138
Engagement with stakeholders
In workstream 5, we demonstrate that there is an appetite to understand this better, to contribute ideas, and hopefully drive changes in decision-making that improve outcomes for very preterm babies, their families and the NHS.
Potential implications for health service delivery, families and society
The ‘safe’ threshold at which care can be delivered close to home is 28 weeks gestation
Care in either LNU and NICU for these babies is safe, based on current practice, and using OPTI-PREM outcome measures. Our data indicates no difference in outcomes between the babies born in NICU compared with LNU, even when early postnatal transfers were excluded, and after robustly adjusting for unmeasured confounders.
A higher severe brain injury rate occurs in preterm babies born at 27 weeks gestation in local neonatal unit
We cannot assign causality to postnatal transfers although it was an associated observation. This is because there may be other factors that have caused a baby to have a SBI, such as degree of illness in the baby and/or mother, and the finer details of care provided initially. However, to reduce the risk of SBI for babies born at 27 weeks gestation, antenatal triage of expected preterm births at this gestation, to maternity services colocated with NICU for care, would be indicated. This is because we would not always be able to predict which baby born at 27 weeks in LNU would be sick enough and/or need postnatal transfer for support in NICU.
Practically, a review of neonatal unit capacity may be required to safely implement this triage. This is because against current resources NICU teams do not have capacity to deal with existing caseloads and pathways of care. We studied 777 preterm babies born at 27 weeks gestation in LNU, and 18,847/29,842 (63.5%) of the cohort born in the 5-year period. Estimating from this, 1220 births (777 × 100/63.5) over the 5-year period means that approximately 240 additional preterm births per year in LNU would need to be antenatally triaged to NICU settings. To optimise care for preterm babies born at 27 weeks gestation, capacity to accommodate these babies in NICU as defined by OPTI-PREM findings is required. This capacity would need to include a review of training and rotation of staff to ensure skills are maintained especially to cater for the unexpected ill very preterm birth in LNU, or to provide support for existing LNU especially if geographically distant.
It would also require later review of care for babies born at 28 weeks gestation in LNU, to assess whether outcomes for these babies would be the same if born and cared for in either LNU or NICU setting, if LNU were not caring for babies born at 27 weeks gestation.
Our findings could guide individual ODNs to redirect obstetric, emergency care and ambulance services so that a mother with an expected preterm birth at 27 and at 28–31 weeks gestation has her baby delivered and cared for at a hospital with the most appropriate facilities for neonatal care for baby, while at the same time supporting parents and families better. Redirecting a mother to the correct maternity facility would be cost-effective, as opposed to transporting a preterm baby after birth to the most appropriate neonatal unit. Our work highlights the importance of joint working between maternity and neonatal service providers, rather than as separate ODNs.
Risk assessment of postnatal transfers of inadvertent births in local neonatal unit is required
For inadvertent births at 27 weeks in LNU, clinical judgement, balancing the benefits of clinical uplift against risk of SBI should be made, in determining need for early postnatal transfers to NICU neonatal units for ongoing care.
It is accepted that there will always be inadvertent births in LNU settings due to the unpredictability of preterm birth, and where mothers are too ill or too advanced in labour to facilitate safe in utero transfer. Optimal health service delivery might mean retaining care of babies born at 27 weeks gestation who are stable in LNU/low-volume neonatal units, to prevent any unnecessary potential SBI associated with postnatal transfer. However, this cannot be determined in advance. Clinical assessment is required, weighing risks of SBI associated with postnatal transfer to NICU for uplift care against risks of retaining care of the preterm baby and considering the expertise and resources to manage this need. Where care is retained, support for local teams (such as telemedicine, staff rotation, NICU oversight) need to be in place so that care and outcomes for very preterm babies are not compromised.
Capacity transfers in babies born inadvertently in local neonatal unit, at 27 weeks gestation, should be approached cautiously
OPTI-PREM was not designed to differentiate between reasons for postnatal transfer at 27 weeks. These could have been transfers for (a) ‘capacity’ reasons or for (b) protocol-driven reasons of stable preterm babies born at 27 weeks gestation or (c) based on clinical need. In most cases in LNU the distinction between a baby being transferred for either protocol (i.e. stable and ventilated at 48 hours) or progressive deterioration may be difficult to make. Clinical teams would need to judge this. Limiting capacity transfers in stable babies at 27 weeks gestation inadvertently born in maternity units colocated with LNU is the most practical and immediate intervention.
Capacity transfers at 28–31 weeks gestation should be approached cautiously
As care is equitable for babies born at 28–31 weeks gestation in NICU and LNU, one temptation may therefore be to recommend triage of care for these mothers and babies to LNU, so that NICU can care for the babies born at ≤ 27 weeks gestation. However, moving these babies postnatally or in utero as capacity transfers would constitute major disruptions for families and parents’ mental health. Factors around obstetric resources, the options for antenatal triage, and considering patient and family experiences by delivering care closer to home may need to be factored into the decision-making. We believe that caution needs to be adopted here. The OPTI-PREM study was not designed to distinguish between SBI that may be primarily related directly to the process of postnatal transfer and, alternatively, SBI associated with transfers primarily related to the degree of illness of the baby. To shift postnatal transfers from 27 weeks to 28–31 weeks gestation for capacity reasons may shift the burden of SBI to these gestational ages.
Costs of care
The costs of care for these babies, at each gestational age in weeks and overall, provides for the NHS a framework to facilitate adequate allocation of resources for very preterm babies.
That cost of care is lower at higher gestations in NICU, driven by fewer ICU days for babies born at 30 and 31 weeks, and potentially implies that efforts towards streamlining efficient care packages between NICU and LNU are required, rather than a review of place of birth for these babies.
That there is no difference in costs of care for babies born at 27 weeks gestation in LNU compared with NICU supports that recommending in utero transfer of all anticipated births at 27 weeks will not have a significant impact on the cost of care per baby while in neonatal care. However, the health economic cost of reducing a SBI for each baby, for short- and long-term outcomes, for their families, the NHS and society is likely to be substantial.
Evidence from workstream 2 (see Chapter 5), and our observations regarding TPN and surfactant use (see Chapter 7), supports future work examining costs and effects of care by quality of care provided. These will provide valuable insights that could be used in future efforts to optimise neonatal service delivery.
OPTI-PREM provides an analytic framework for future cost and effects analysis for key morbidities in very preterm babies. This scientific data could drive redistribution and allocation of resources for cost-effective interventions, that are evidence- and consensus-based. For example, if we understood the cost of disease in England, as is emerging from Canada,139 and the cost effects of evidence- and consensus-based practices such as 1 : 1 nurse provision, early BMF, and optimised delivery room non-invasive ventilatory support, we may be more effectively able to channel resources for cost-effective neonatal health care in England.
Implications for further research around making postnatal transfers safe
More scientific research is required around how better to optimise the process of postnatal transfers in the longer term, as these are unlikely to be prevented entirely. One in three babies born at 27–31 weeks were transferred between units in the postnatal period. This is accepted as a key area in which research needs to be optimised.
In addition, work on how better to optimise stabilisation and ongoing clinical care in a LNU setting for both mother and baby may be required. As we cannot completely exclude babies being transferred being more likely to be ill and at greater risk for SBI, research on how/whether resources should be directed (additionally) to optimising clinical training and support for teams in LNU as opposed to, or in support of, safer postnatal transfer is also required. This could include a review of thinking around provision of rotational medical and nursing care between LNU and NICU to limit deskilling of teams and the impact of postnatal transfers.
Implications for neonatal clinical teams and service provision
OPTI-PREM evidence supports adherence to national prespecified targets and aiming for the top end of performance in benchmarking exercises and accepted best practice. These, if resources are allocated appropriately (units that were understaffed and in areas of high indices of social deprivation were less likely to be higher performing), could have beneficial effects for families, babies, staff and the NHS in relation to reduced number of days in hospital. However, conceptualising what is realistic and achievable for units in areas of high indices of deprivation, and those not adhering to best practice models as defined within OPTI-PREM, is complex. It may require more than just a redistribution of funds and staffing. For example, an in-depth qualitative review of how better to recruit, retain, train and upskill staff may be required. This may include developing innovative models of rotation of staff between units of differing designations and possibly performance, to learn, export and maintain the skills necessary for optimal service delivery, even in impoverished settings where maternal morbidity, and therefore neonatal morbidity, is generally higher.
The observations in workstream 4 (see Chapter 8) highlight the importance of decision-makers and managers providing appropriate support for neonatal teams on the ground managing capacity and undertaking clinical care. Institutional processes and systems around capacity management and demand often drive internal conflict in clinical teams dealing with babies and families. More effort should be channelled into providing a layer of administrative support to clinical teams on the ground, allowing them to focus on delivering excellence in clinical care, with as little moral distress as possible, in trying to manage the needs of the individual baby versus available resources in units and networks.
Implications for babies
SBI may be associated with neurological findings ranging from mild developmental delay to severe cerebral palsy.140,141 From the EPICure studies, we understand better the educational and behavioural phenotype of preterm babies, and the cost to society is being increasingly reported in the literature .80,84,142–144 If we can prevent one preterm baby born at 27 weeks gestation having SBI, the cost savings to the NHS and society would be substantial, considering the lifetime healthcare costs of cerebral palsy,85,86 independent of the costs of other associations with prematurity.139 Supporting and including parents and families during their baby’s time in neonatal care will likely contribute to their better bonding with their preterm baby.
Implications for parents and families
Identifying that continuity of care has the most agency for parents is an important finding which should promote streamlining of services across neonatal units. Keeping families close to home will reduce stress and anxiety and support the mental and physical well-being of families. For those where transfer is necessary, strategies to better support parents social, economic and mental well-being need to be pursued. Working in partnership with parents will allow them to feel included, and to understand and accept the care pathways most appropriate for their baby.
Conclusion
Advances in neonatal care and improvements in outcomes, especially for the very preterm baby,145 has necessitated in England meta-management of neonatal health service delivery. Managed clinical networks comprising a mix of non-tertiary and tertiary services in different regions of the country, have evolved over the past 20 years,146–149 its philosophy being the provision of maternity (and neonatal) care close to home. However, with this came uncertainties of what this would mean for postnatal transfers, NHS resources and capacity, for very preterm babies, their families and society. Here we show evidence that place of birth matters, in all these aspects. We show that there is an appetite to understand this better, to contribute ideas, and hopefully drive changes that optimise outcomes for very preterm babies, their families and the NHS.
We found that the safe gestational age cut-off for babies to be born between 27+0 and 31+6 weeks and early care at either location was 28 weeks. We found no effect on mortality while in neonatal care or in infancy, including after sensitivity analyses. We found a significantly greater proportion of babies in LNU had SBI which appeared to be driven by those born at 27 weeks gestation, and those transferred in the first 72 hours. We identified that 25 babies at 27 weeks would have to be delivered in a NICU to prevent one SBI.
When we explored this difference in SBI between units further, we found that the significance remained in high-volume units providing the upper quartile of neonatal IC days (i.e. there was a significantly lower risk of SBI for babies cared for in high-volume neonatal units, which were all NICU), but not for low-volume units providing the lowest quartile of IC days (which included LNU). These observations support refinement of neonatal and maternity health service delivery for very preterm births in England.
Capacity for expectant mothers at maternity centres colocated with NICU is a major issue in England at present, so, simply suggesting to emergency teams, GP practices and obstetric teams that in utero transfers of all anticipated very preterm births at 27 weeks gestation should be redirected to these services, without addressing the wider capacity issues, will not succeed. At present there are neonatal operational delivery networks that co-ordinate provision of care and movement of newborn babies between units, and separate, less-developed maternity operational delivery networks, for mothers. To effect change based on our findings, obstetric and neonatal services need to co-configure or superimpose on each other, and effectively function as joint ‘perinatal’ networks. This is because capacity implications for one impact on the other. At the same time, there should be adequate information sharing to empower parents and expectant parents regarding decision-making on presentation to a maternity service colocated with a NICU for birth and early care, should they be expecting a birth at or before 27 weeks gestation6,7,14,132 in the UK.
Our study provides in-depth understanding of the costs of care, and the implications for services in caring for babies at these gestational ages. The inverse relationship between cost of care and degree of prematurity, and the risks of complications such as SBI, argue for more efforts to promote prevention of preterm births. That the burden of care is cost neutral within the neonatal period for the most immature of this cohort should argue towards NICU triage for anticipated births at 27 weeks gestation, coupled with the potential health economic benefits of preventing SBI. We have detailed information on the pressures and priorities harboured by staff and parents, and what potential avenues to pursue in working towards optimising delivery and quality of care of very preterm babies in England. Altogether, the findings from OPTI-PREM provide evidence supporting optimising health service delivery for very preterm babies in England through multiple avenues.