Problem

Preterm birth (before 37+0 weeks of pregnancy) is the single biggest cause of neonatal mortality and morbidity in the UK, constituting 24% of child deaths in 2019–2020.1 Despite a 13% reduction in neonatal mortality in England and Wales between 2010 and 2019, there remains significant variation in these figures across the UK.2 3

Although survival rates are improving for preterm babies, rates of severe disability have not followed the same trajectory, and there is now a growing population of children living with neurodisabilities, such as cerebral palsy, due to their prematurity. Research suggests that approximately 1 in 10 ex-preterm babies are classified as living with a severe impairment and 1 in 10 with a moderate impairment, with greater impairment as gestational age decreases.4

The National Health Service (NHS) Long Term Plan, itself responding to the Better Births publication, has committed to realising a 50% reduction in stillbirth, maternal mortality, neonatal mortality and serious brain injury by 2025, with an increased focus on preterm mortality.5 6 In order to achieve this, the Maternity and Neonatal Safety Improvement Programme (MatNeoSIP) included ‘improvement in the optimisation and stabilization of the very preterm infant’ as one of their five primary drivers. Perinatal Excellence to Reduce Injury in Premature Birth (PERIPrem) felt that it was crucial to maximise this opportunity to improve outcomes and implement an additional four evidence-based elements in addition to the seven already stated, focusing also on optimal timing of interventions to drive down neonatal mortality and brain injury. The resultant bundle is displayed further (table 1).

There are 15 Academic Health Science Networks (AHSNs) across England, and they operate as the key innovation arm of the NHS. PERIPrem is a product of the merger of two successful, independent West of England AHSN (WEAHSN) ‘Evidence into Practice’ bids, put forward by Professor Karen Luyt (KL; Consultant Neonatologist and Professor of Neonatal Medicine at the University of Bristol) and Dr Sarah Bates (SB; Consultant Neonatologist at Great Western Hospitals NHS Foundation Trust) in 2019. The resultant project encompassed the whole South-West region, recognising the importance of regional collaboration to achieve implementation of the bundle elements.

Recognising bundle complexity, large geography, varied populations and reliance on ownership for success, PERIPrem invested in bringing together perinatal organisations caring for a population of 5 million comprising: 12 NHS Trusts (secondary care hospitals), South West AHSN and WEAHSN, South West Neonatal Operational Delivery Network (SWODN), parent partners, South Western Ambulance Service and clinical experts resulting in the UK’s largest perinatal quality improvement (QI) clinical collaborative.

The Trusts who signed up to PERIPrem were:

• Gloucestershire Hospitals NHS Foundation Trust.

• Great Western Hospitals NHS Foundation Trust.

• North Bristol NHS Trust.

• Northern Devon Healthcare NHS Trust.

• Royal Cornwall Hospitals NHS Trust.

• Royal Devon and Exeter NHS Trust.

• Royal United Hospitals Bath NHS Foundation Trust.

• Taunton and Somerset NHS Foundation Trust.

• Torbay and South Devon NHS Foundation Trust.

• University Hospitals Bristol and Weston NHS Foundation Trust.

• University Hospitals Plymouth NHS Trust.

• Yeovil District Hospital NHS Foundation Trust.

We also have industry partners Hologic & Vygon to thank for their contribution to facilitating PERIPrem.

The PERIPrem project aimed to:

1. Improve rates of care bundle implementation to mother–baby dyads who deliver at <34⁺⁰ weeks’ gestation using QI methodology.

2. Improve rates of individual element implementation to mother–baby dyads who deliver at <34⁺⁰ weeks’ gestation using QI methodology.

3. To understand the barriers and enablers associated with the PERIPrem approach on the implementation of a multiple-element standardised bundle.

4. To establish the impact of the PERIPrem approach on perinatal staff knowledge, skills and confidence in QI methodology, psychological safety and teamwork.

Longitudinal aims beyond the scope and time-frame of this current evaluation were to:

1. Reduce mortality and brain injury rates in infants born <34⁺⁰ weeks’ gestation in the South West of England.

2. Eliminate inequitable access to all elements of the care bundle for socially disadvantaged mothers and babies

Background

When tasked with providing a core outcome set for research involving premature infants, parents and patients demonstrated their priorities by placing survival, brain injury on imaging, general gross motor ability, general cognitive ability, quality of life, visual impairment/blindness and hearing impairment/deafness in their top 12 outcomes.7 A diagnosis of cerebral palsy can have an immeasurable impact on the life of a child and their family with one out of three unable to walk, one out of four unable to talk, one out of four having epilepsy, three out of four experiencing pain and one out of two having an intellectual impairment.8 The human importance of a project such as PERIPrem cannot be underestimated. The financial lifetime cost of cerebral palsy has been estimated at £700 000, with an increasing cost to NHS Resolution.9 10

Despite conclusive evidence of the effectiveness of many interventions, there remains marked variation in practice throughout the UK when one examines data collected by the National Neonatal Audit Programme via the BadgerNet neonatal clinical care dataset.11 Indeed, groups like the British Association of Perinatal Medicine (BAPM) and the Royal College of Paediatrics and Child Health call for QI methodology to be used to reduce variation in practice, which outcomes from projects like PReCePT have proven both feasible and sustainable.12

The evidence for each individual element of the PERIPrem programme is contained previously in table 1. It is not possible to quantify the impact of the bundle of interventions as a whole, but each intervention in itself carries significant opportunity to improve the mortality and risk of neurodevelopmental impairment for preterm infants.

PERIPrem aimed to build on the QI theory and the community of PReCePT to optimise rates of compliance to the bundle of evidence-based interventions and to examine whether such methodology could bring perinatal multidisciplinary teams (MDTs) together to aid challenging, complex objectives such as birth in the right place.

There have been multiple QI projects conducted in the past to improve rates of individual perinatal interventions, but to our knowledge, none have been as complex as PERIPrem in its ambition to approach eleven elements in 12 NHS Trusts. Empowering clinicians and parent partners to own the QI journey, supported by regional organisations, increases speed of roll out, maximises capacity and increases access to training. The translation of ideas and ambition into demonstrable improvement was made possible through the collaboration of all organisations offering particular skill sets to the ambition.

Measurement

The characteristics of the mother–baby dyads who were eligible to enter the PERIPrem project were collected via the PERIPrem Optimisation Tool, a smart Excel spreadsheet. The tool was submitted monthly by each unit to the evaluation team. The tool detailed all eligible babies born that month with information about their demographics to inform bundle eligibility, as well as completeness and timing of bundle elements as detailed in table 1. Teams were encouraged to use the PERIPrem Patient Passport to ease data collection (available here: https://www.weahsn.net/our-work/transforming-services-and-systems/periprem/periprem-quality-improvement-resources/).

In order to understand if adherence to the bundle had changed from preimplementation to postimplementation of PERIPrem two stratified samples were collected. The 12 Trusts were asked to complete a copy of the Optimisation Tool retrospectively for the first ‘x’ number of babies born after 1 January 2019 (18 months before implementation), a total of 120 babies across all 12 Trusts (special care units=5 babies, local neonatal units=10 babies, NICUs=15 babies). These 120 babies were compared with the last 120 babies (same number of babies from each Trust) born prior to 1 July 2021 (the close of implementation support).

Mixed-effects binary logistic regressions were conducted on this data using a two-level model, where the mother–baby dyad was the primary level of analysis and adjusted for clustering by Trust. Each of the bundle elements, as well as whether the dyad was fully optimised (received all the interventions), were entered into the model as the dependent variable. For the comparison of 2019 and 2021 data, time was added as a fixed factor. For the analysis of the data in the implementation phase, time was added as a continuous factor. For the optimisation score (a continuous variable), linear mixed models were used using the same hierarchical structure. It is important to note that linear mixed models can be limited by multicollinearity of predictors, and care was taken when deciding what predictors were included in the models, and final models were selected using ranking via Akaike’s Information Criterion.13

Predictors for each model included: baby characteristics (gestation, birth weight and single or multiple birth) collected at a patient level; maternity characteristics (ethnicity, age and deprivation) collected at a Trust level; and whether or not the unit was a NICU.

In order to understand the human factor mechanisms at play in the implementation of the PERIPrem approach, a range of measures and methods were adopted, including self-reported questionnaires and qualitative semistructured interviews. Quantitative self-reported surveys were administered in September–October 2020 and again in June 2021, via an online survey hosted by SurveyMonkey and included the following questionnaires:

• Psychological Safety using items taken from the Psychological Safety subscale in The Learning Organization Survey.14

• Staff Perceptions of Team Work using the TeamSTEPPS Teamwork Perceptions Questionnaire.15

• Barriers and Enablers to Implementation using qualitative semistructured interviews with staff and PERIPrem team members alongside an online survey based on the Theoretical Domains Framework for behaviour change, the barriers to QI identified in the published literature and expert opinion.16

Scores in the Psychological Safety subscale and the TeamSTEPPS Questionnaire were analysed using linear mixed models (with staff as primary level and Trust as secondary level) and responses to the questions around barriers and enablers to implementing QI methodology analysed using generalised estimating equations.15

Design

Using the QI implementation design of the AHSN team that delivered PReCePT under the leadership of KL and the local, multiple-element care bundle implementation experience of SB, the PERIPrem team assembled.12 The project team consisted of perinatal team members (midwives, obstetricians, neonatologists), the AHSN Patient Safety Collaborative team and parental coproduction.

Every Trust in the SWODN was invited to participate in the project. This invitation was extended to each Trust’s executive board, as well as to their perinatal teams. Every Trust accepted the invitation and the PERIPrem team proceeded by:

• Defining the 11 interventions, their evidence base and eligible cohort.

• Understanding the patient population through regional BadgerNet data and parent partner lived experience.

• Attempting to predict barriers to implementation between MDT members, local, regional and national systems.

• Designing resources to comprise data collection, parent and professional information sharing, QI processes and microsystem exploration.

• Supporting local teams with bespoke QI coaching.

• Monitoring progress in real-time, facilitating collaboration.

• Planning the evaluation, dissemination and communication of the findings locally, nationally and internationally.

Strategy

To structure the project, we employed the established QI Model for Improvement by the Institute for Healthcare Improvement (IHI) and elements of anthropological experience-based codesign.17 18

Implementation phase

The 12 Trusts joined the project over a staggered 3-month period dependent on local barriers. Team members varied in QI experience and project leadership. Teams met remotely with their QI coaches and the PERIPrem operational lead to establish an action plan and ensure that they were able to access the PERIPrem resources. They tended to choose to focus their efforts on elements where the most local improvement could be made and as this varied, so did the barriers encountered. Plan–Do–Study–Act structured cycles of change were encouraged, but teams were autonomous in their execution in response to their microsystem-specific demands, and in response so was the QI support that they received (figure 1.). The PERIPrem resources were iteratively adapted in response to feedback with a weekly newsletter that was emailed round to teams celebrating regional successes and innovations.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

The driver diagram designed on LifeQI by one of the local teams looking at approach to improving early breast milk.

In September 2020, additional funding of £500 per unit was made available via a grant from Hologic. With the help of senior staff PERIPrem leads used this money to bypass complex local procurement routes and create a range of interventions including: PERIPrem grab boxes, resources for tea trolley training, blankets, TransWarmer infant transport mattresses and Neohelp sterile polyethylene suits.

PERIPrem also supported the successful SWODN application to fund quantitative fetal fibronectin testing across all our units for free for a year to support place of birth, encouraging evidence-based QUIPP app calculation of risk of preterm birth, thus hoping to rationalise centralisation of mothers in extremely preterm labour.

Antenatal antibiotic administration to all mothers in established preterm labour was introduced in March 2021 alongside the introduction of optimal timing of the antenatal interventions to the Optimisation tool.

The implementation phase was extended from 6 to 9 months to take into consideration the multitude of barriers that COVID-19 wrought.

Lessons and limitations

As a change implementation construct, PERIPrem had many strengths both subjectively observed by the qualitative interviews with the local implementation teams and objectively measured through intervention implementation rates. The key strengths recognised were in response to the challenges faced:

COVID-19 placed constraints on staff place and pattern of work, reduced staffing numbers, resulted in significant additional mental load and prevented simulation and in-person collaboration.

Marrying this with the regional expanse of the project, the size of the MDT workforce involved and the scope of the interventions, teams struggled to find the time to implement change and report outcomes despite the funded hours, due at least in part to competing demands.

Our strengths:

• Innovation through a blended, remote approach to regional project design and implementation in a time of restricted travel.

• Comprehensive branding and packaging of support resources leading to regional continuity of care and use of a common language.

• Enabling transfer of up-to-date knowledge about the significance of these evidence-based interventions to a whole regional, perinatal workforce through a variety of approaches: infographics, videos and webinars.

• Delivery of real-time, visual data by local teams using the PERIPrem-supplied resources (Optimisation tool & LifeQI), and from the regional evaluation hub to provide teams with contemporaneous feedback on changes implemented.

• Being a realisation of translation of evidence into practice in a short time frame.

• The evidence of ongoing improvement in intervention delivery after the implementation phase and its associated support mechanisms has come to an end suggests successful embedding of change.

• The digital delivery of the project thereby reducing travel and the expected improvement in patient outcomes will result in a decreased carbon footprint.

PERIPrem was a challenging, multiple-element care bundle for Trusts to adopt, but the reality of the project ask was to reduce variation and ensure that every patient gets the best, pre-existing evidence-based care. While we were able to show significant improvements in element delivery between 2019 and 2021, the implementation period itself identified significant improvement in four elements only. The reasons behind this are complex as each intervention process is nuanced with its own barriers and enablers, but one also cannot underestimate the impact that the COVID-19 pandemic had across our region on staff capacity. Without a control group or data on other programmes Trusts were implementing, it is difficult to strictly infer whether the significant improvements observed between the 2019 and 2021 cohorts for several elements of the bundle were as a direct result of PERIPrem, but the contemporaneous relationship of the change is suggestive. National Neonatal Audit Programme data from 2019 to 2020 demonstrates insignificant change nationally in several PERIPrem elements including: birth in the right place for <27 week infants, antenatal steroid administration and normothermia, suggesting against our observed improvements being secondary to comparative practise change over the pandemic.11

Maternal characteristics were not available on a patient level, and therefore data obtained at Trust level were used. There were also differences in sample size dependent on the bundle element being analysed due to evidence-based gestation-related eligibility criteria. A further limitation specifically for the analyses of the 2019 and 2021 datasets was that these were stratified samples; however, distribution of the dyad characteristics was similar to the population used in the analysis of the implementation phase.

Patient and public involvement