Place of birth and postnatal transfers in infants with congenital diaphragmatic hernia in England and Wales: a descriptive observational cohort study

Objective To describe clinical pathways for infants with congenital diaphragmatic hernia (CDH) and short-term outcomes. Design Retrospective observational cohort study using the UK National Neonatal Research Database (NNRD). Patients Babies with a diagnosis of CDH admitted to a neonatal unit in England and Wales between 2012 and 2020. Main outcome measures Clinical pathways defined by place of birth (with or without colocated neonatal and surgical facilities), transfers, clinical interventions, length of hospital stay and discharge outcome. Results There were 1319 babies with a diagnosis of CDH cared for in four clinical pathways: born in maternity units with (1) colocated tertiary neonatal and surgical units (‘neonatal surgical units’), 50% (660/1319); (2) designated tertiary neonatal unit and transfer to stand-alone surgical centre (‘tertiary designated’), 25% (337/1319); (3) non-designated tertiary neonatal unit (‘tertiary non-designated’), 7% (89/1319); or (4) non-tertiary unit (‘non-tertiary’), 18% (233/1319)—the latter three needing postnatal transfers. Infant characteristics were similar for infants born in neonatal surgical and tertiary designated units. Excluding 149 infants with minimal data due to early transfer (median (IQR) 2.2 (0.4–4.5) days) to other settings, survival to neonatal discharge was 73% (851/1170), with a median (IQR) stay of 26 (16–44) days. Conclusions We found that half of the babies with CDH were born in hospitals that did not have on-site surgical services and required postnatal transfer. Similar characteristics between infants born in neonatal surgical units and tertiary designated units suggest that organisation rather than infant factors influence place of birth. Future work linking the NNRD to other datasets will enable comparisons between care pathways.


BACKGROUND
Congenital diaphragmatic hernia (CDH), a defect in the formation of the diaphragm, can result in abdominal organs herniating into the chest compromising lung development. 1 2This defect may manifest as poor lung function at birth, respiratory failure and death.Advancements in prenatal diagnosis, neonatal interventions and surgical techniques have improved outcomes of infants with CDH. 3 4 Antenatal screening aims to detect CDH early. 5In the UK, around 60% of CDH cases are antenatally diagnosed. 6This allows for parental counselling and shared decision-making on place of birth.However, evidence gaps on the optimal timing and place of birth can hinder informed decision-making. 7It is necessary to evaluate how neonatal services are organised to ensure infants with CDH are cared for in appropriately resourced settings. 7ational guidelines recommend that neonates requiring surgical care are born in maternity centres with colocated neonatal surgery. 8 9UK neonatal services are organised in networks. 10In several networks, neonatal surgery is provided

WHAT IS ALREADY KNOWN ON THIS TOPIC
⇒ National guidelines recommend that infants with known surgical anomalies are delivered in maternity units with colocated neonatal medical and surgical units.

WHAT THIS STUDY ADDS
⇒ Each year, around 147 babies with congenital diaphragmatic hernia (CDH) are admitted to neonatal units in England and Wales.Half of them are born in maternity units without an onsite surgical facility and consequently require postnatal transfer.⇒ Similar characteristics between those born in neonatal surgical units and tertiary designated units suggest that organisation rather than infant factors influence place of birth.⇒ Out of ten babies, around seven will survive neonatal discharge, half are discharged home and a quarter are discharged to other paediatric settings for ongoing care.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
⇒ We identified four different clinical pathways in England and Wales that may lead to variation in care and outcomes of babies with CDH.⇒ Data linkage between available health datasets is urgently needed to reliably evaluate factors that influence outcomes including the organisation of services.⇒ This would facilitate future research needed to determine whether variation in place of birth and care pathways observed in this study influence short and long-term outcomes of babies with CDH.

Original research
in 'stand-alone surgical units' without colocated maternity.Babies born in these networks are transferred postnatally.Previous research in England found that lack of colocation leads to avoidable postnatal transfers, with estimates of 31 cases of CDH undergoing an avoidable postnatal transfer annually. 7 11ere is mixed evidence on whether place of birth for CDH impacts outcomes. 12The role of place of birth in short-term outcomes has not been studied previously in the UK.Examining this relationship is complex, as both place of birth and outcomes are associated with an array of demographic, clinical and organisational factors.However,

Original research
routine data can potentially be used to examine the relationship between these factors and infant outcomes at a population level.
The National Neonatal Research Database (NNRD) comprises quality-assured data on all infants admitted to UK neonatal units. 13The NNRD enables us to study admissions of infants from birth to discharge from neonatal care.However, the NNRD does not capture care in other settings such as standalone paediatric surgical centres, paediatric intensive care or paediatric wards.
We aimed to describe current care pathways defined by place of birth for infants with CDH born in England and Wales and describe interventions and short-term outcomes.

Study design, setting and participants
We conducted the study in a two-stage process.
Stage 1: Discussions with experts in neonatal medicine, paediatric intensive care and neonatal surgery to explore care pathways for infants with CDH.
Stage 2: A retrospective observational study using routinely recorded data from the NNRD.Data were used to explore care pathways and short-term health outcomes.We report in line with Reporting of Studies Conducted Using Observational Routinely Collected Health Data guidelines. 14

Population
Data were extracted for all infants admitted to a neonatal unit in England and Wales between 1 January 2012 and 31 December 2020 with a diagnosis of CDH (as defined in online supplemental appendix).We excluded infants with inconsistent transfer patterns and those diagnosed with multiple congenital (non-cardiac) surgical abnormalities, for example, CDH and gastroschisis.
We used Office for National Statistics data on live births in England and Wales across the study period to estimate the prevalence of CDH. 15

Outcomes
We report primary outcome survival to discharge to home and neonatal discharge to other settings.Secondary outcome measures include postnatal management, transfer patterns and length of stay (table 1).
We define postnatal transfer as any transfer that requires ambulance transport.We sought clarification from the expert advisory panel on which units required postnatal transfer between neonatal and surgical units and sought consensus.Transfers were identified from the NNRD discharge destination code.

Statistical analysis
We report findings across clinical pathways and present descriptive statistics using median/IQR and percentage as appropriate.We report survival to neonatal discharge, discharge destination and length of hospital stay.This is a descriptive study and we have not undertaken any formal analysis to compare outcomes between pathways or adjust confounders.All analyses were performed using R V.3.6.

Research ethics and other approvals
We used deidentified data from the NNRD. 13 16All neonatal units agreed to the inclusion of their data in the study.

Stage 1: establish expert advisory board
Experts from across 6 out of 10 neonatal networks in England and Wales participated in the expert advisory board.Members included six neonatologists, eight neonatal surgeons and one paediatric cardiac intensivist.They worked in either tertiary neonatal units with colocated surgery or tertiary neonatal units without surgery on-site.The four pathways identified were based on birth in a maternity unit with: 1. Colocated tertiary neonatal and surgical units (neonatal surgical units).2. Designated tertiary neonatal unit for surgical conditions and transfer to a surgical centre (tertiary designated).

Tertiary neonatal unit not designated for surgical conditions
and transfer to a surgical centre (tertiary non-designated).4. Non-tertiary units without surgery and transfer to a surgical centre (non-tertiary).Birth in neonatal surgical units (1) is the only pathway that does not require postnatal transfer in an ambulance to a surgical unit.

Stage 2: routine data analysis using NNRD
Between 1 January 2012 and 31 December 2020, a total of 1335 babies with a diagnosis of CDH were admitted to a neonatal unit in England and Wales of which 16 were excluded from the study (figure 1).There were 6 108 030 live births during this period in England and Wales, resulting in an estimated live birth incidence of 2.2 per 10 000 (95% CI 2.1 to 2.3).The true live birth incidence is likely to be higher as our case number excludes babies born alive who died in the delivery room.

Clinical pathways
We identified 16 neonatal surgical units, 5 tertiary designated units and 5 stand-alone surgical centres without an on-site neonatal unit across England and Wales.During the study period 660/1319 (50%) infants were delivered in neonatal surgical units.The other half required postnatal transfer to a surgical unit (figure 1).We report yearly births across each clinical pathway during the study period (figure 2).We found that a median of 72 (67-76) of 146 cases were born in neonatal surgical units annually.

Infant and maternal characteristics
Table 2 summarises infant and maternal characteristics by pathway.Infants were most commonly male with a median gestational age of 38 weeks.Across the cohort, median (IQR) maternal age was 30.5 (26-35) years and half of the group were white British; between 10% and 20% of maternal ethnicity data were missing.

Organisational factors and postnatal transfers
More than 95% of infants born in surgical neonatal or tertiary designated units were admitted to the neonatal unit within the first 2 hours after birth compared with 74-78% in tertiary nondesignated or non-tertiary units.Transfer to a surgical centre occurred within the first 24 hours in 81% in non-tertiary centre and 60% in a tertiary non-designated centre.Infants in tertiary designated units were transferred later, commonly after day 3 from birth (median 4 (1.8-8) days).Infants in tertiary designated units were transferred to a stand-alone surgical unit without a neonatal unit in 71% (239/337) cases.One tertiary designated centre transferred infants to the surgical unit for surgery and then retrieved them after surgery.Infants born at this unit were not discharged from their electronic health system despite being transferred to a surgical unit as preoperative care and postoperative care occur within the tertiary designated neonatal unit.Across the cohort, 30% of infants (396/1319) were postnatally transferred to a stand-alone surgical unit.

Survival to a surgical centre
We found 659 infants born outside a surgical centre, of which 86% (565/659) survived to transfer to be admitted to a surgical

Survival to neonatal discharge and postnatal management during neonatal stay
In the first days after birth, 149 infants were transferred to a stand-alone surgical centre or other paediatric settings and did not return to an NNRD contributing unit.Transfers occurred on a median of day 2 (median (IQR) 2.2.(0.9-4.5) days) (online supplemental appendix) and these infants had minimal data.We therefore present in additional data reporting survival to neonatal discharge and postnatal management for 1170 infants (89% of the original cohort) excluding these 149 infants.We present postnatal management for 1170 infants that received the majority of care in a neonatal unit (table 2).Across all groups, infants commonly received a combination of multiple modes of ventilation during their neonatal stay.We found higher rates of inhaled nitric oxide and sildenafil use in neonatal surgical units and tertiary designated units.Surfactant was given outside the delivery room to a similar proportion of infants across all groups.

DISCUSSION
Over a 9-year period, 1319 infants with CDH were admitted to neonatal units in England and Wales.We identified four clinical pathways of care for neonates with a diagnosis of CDH.Half of the babies were born in maternity units with colocated neonatal surgical units and a quarter in tertiary designated units requiring postnatal transfer to stand-alone surgical centres.The transfer from tertiary designated centres to a surgical centre occurred at a median age of 4 days.The remaining quarter were born outside of these designated pathways.Infant characteristics and rates of intensive care support were similar for infants born in neonatal surgical units and in tertiary designated units.This suggests that organisational rather than infant factors influence place of birth and care pathway, particularly for babies with antenatally diagnosed CDH, who would be predominantly cared for across these two designated pathways.

Original research
We report a survival rate of 73%, consistent with previous data from England which estimated 1-year survival between 68% and 81%. 17 18hile comparison of survival outcomes between pathways is of interest, this was not undertaken formally in this descriptive study due to the unavailability of important confounders and mediators.These include information on fetoscopic endoluminal tracheal occlusion, a procedure which has been shown to improve survival to discharge in infants with severe left-sided CDH. 19Importantly, we lack information on whether CDH was antenatally or postnatally diagnosed, the laterality of the defect, defect type, lung-head ratio, antenatal treatment, timing of surgery and surgical complications.We speculate, for example, that the population of babies born in the tertiary non-designated and non-tertiary groups are likely to have been postnatally diagnosed due to smaller defects being undiagnosed antenatally and hence born outside a surgical centre.This would explain the more favourable survival to neonatal discharge and the shorter length of stay in the tertiary non-designated and non-tertiary groups.
Previous data from the USA have identified that being 'inborn' at the treatment centre is associated with mortality in CDH, 20 this is consistent with our findings.Whether there is a difference seen between the survival to neonatal discharge and length of stay between neonatal surgical unit and the tertiary designated unit groups warrants further exploration but requires data linkage in the UK to obtain additional information to enable case-mix adjustment.
Limitations to the study include missing data beyond the first few days of life for over one-third of babies born in tertiary designated centres transferred early to stand-alone surgical centres.To assess the impact of the missing data, we additionally reported outcomes for a subgroup of 1170 babies, excluding 149 babies (11% of the cohort), length of neonatal stay becomes longer and more consistent across the pathways in the subgroup.However, survival rate to neonatal discharge decreased for infants in tertiary designated units from 75% (260/337) to 64% (137/214).We speculate this is due to the disproportionate representation of deaths due to the inclusion of early mortality before transfer to a surgical centre, but exclusion of survivors transferred early to a surgical centre.The population in the subgroup may represent more severe CDH, particularly in the tertiary designated unit group.
A further limitation of the NNRD is that it captures data on neonatal unit admissions only and therefore while we found an estimated live birth prevalence of 2.2 per 10 000 (95% CI 2.1 to 2.3), this does not consider terminations of pregnancy or delivery room deaths. 17 21trengths of our study include the population-level coverage, including all babies with CDH admitted to neonatal units in England and Wales across a 9-year period.In England and Wales, babies with CDH will be admitted to a neonatal unit following birth as their first hospital episode, unless the antenatal plan is for palliative care on the postnatal ward, or the infant does not survive birth or the CDH is not detected prior to postnatal discharge.All other infants, even if they are transferred to a non-neonatal unit for ongoing care, are included thus reducing recruitment bias.
We have demonstrated the feasibility of using routinely collected data to identify the cohort of infants with CDH receiving care in the UK.Future research aimed at informing the configuration of care pathways and determining the optimal place of birth for babies with CDH must include outcome data from stand-alone surgical centres, as well as report on long-term health and education outcomes.To improve the accuracy and completeness of data and allow for more robust conclusions to be drawn, there are plans to link data from the NNRD with other sources of routine health data, such as the National Congenital Anomaly and Rare Disease Registration Service, Hospital Episodes Statistics and educational outcomes from the National Pupil Database for this population. 22This data linkage will enable future studies to explore the impact of birth location on outcomes of CDH while considering all confounders.To further enhance data quality, we recommend that centres carrying out neonatal surgery, including stand-alone centres, contribute to surgical datasets or registries to enable national audits and service evaluation.

Figure 1
Figure 1 Study population, clinical pathways defined by place of birth and survival to neonatal discharge.Number and proportion of infants admitted to each type of neonatal unit at birth displayed, together with the outcome of their neonatal stay: survival to neonatal discharge, death or discharge to other settings early not returning to an NNRD unit.CDH, congenital diaphragmatic hernia.

Figure 2
Figure 2 Number of cases per year born in maternity units with neonatal surgical units (green) and non-colocated units (stacked bar): designated tertiary units (blue), non-designated tertiary intensive care units (red) and non-tertiary units (yellow).

Figure 3
Figure 3 Sankey diagram with the final discharge destination recorded in the dataset for all infants across the four pathways (n=1319).

Table 1
Outcome measures of interest, definition and how they were derived

Table 2
Baseline characteristics and first transfer details by place of birth

of birth Neonatal surgical unit, n=660 Tertiary designated unit, n=337 Tertiary non-designated unit, n=89 Non-tertiary unit, n=233 Infant characteristics at birth
*Inhaled nitric oxide during NNU stayContinued

Table 4
Discharge destination, length of stay and survival to neonatal discharge excluding infants with minimal data transferred early to standalone units (n=1170) *Total population n=1170 (89% of whole cohort).CICU, cardiac intensive care unit; PICU, paediatric intensive care unit.

Table 3
Discharge destination, length of stay and survival to neonatal discharge for whole population (n=1319)