Neonatal intensive care unit services - demand and capacity modelling: final report

4. Scenarios and assumptions to model impacts

The baseline period for the model is the twelve months from October 2022 to September 2023 (the most recent period at the outset of the project). The exception to this is the data collected from University Hospital Wishaw for which data from October 2021 to September 2022 was identified as being a more reflective baseline (see baseline section 3.1.1) for modelling.

4.1 Modelled scenarios and assumptions

Within the demand and capacity modelling, there are a series of scenarios which have been modelled. These include:

4.1.1 Demographic Projections

Population Change has been modelled based on the Population Projections for Scottish Areas (2018-based) for Age 0. The source of this is National Records Scotland. The population projections have been calculated at a Local Authority level and applied to the activity data at this level, based on the patient local authority of residence, to provide projected demographic change for each unit.

Incidence trends have been modelled to reflect expected trends in the incidence of a baby receiving neonatal care. This is to reflect the feedback that there is a trend towards providing active survival-focused care for extremely preterm infants (22-24 weeks of gestation) which brings high-intensity needs for longer periods. This has been assumed 1.5% growth per year for 22–24-week gestation. This effectively counters population change (reductions) for NICU sites.

4.1.2 Smallest Cohort

Gestation has been modelled on the basis of 100% of extremely preterm babies transferred to one of the three identified NICU sites, comprising all babies born at less than 27 weeks and all multiples at less than 28 weeks.

Birthweight: There is one principle modelled scenario for birthweight. 100% of babies with a birthweight of less than 800g will be transferred to the three identified NICU sites.

4.1.3 Sickest Cohort

We have worked with clinicians to identify a data definition for more mature babies considered the sickest informed by the Best Start Criteria and likely to require transfer. The data capture does not allow full identification of each episode of care that would align with criteria of complexity, such as “support of more than one organ in addition to respiratory support with an endotracheal tube (ETT)”, however criteria have been flagged within the activity data as potential proxies (in combination) to estimate the size of this cohort.

Recognising the challenge in retrospectively identifying this cohort, we have modelled a range of scenarios including a “minimum”, “likely” and “maximum” scenario to estimate impact and analyse the sensitivity of differing proportion of babies transferred to the three identified NICU sites. A summary of these three scenarios has been included below, it is worth noting that these proxy criteria are for the purposes of sizing the likely cohort only, and do not represent agreed guidelines (for example it may be seen as beneficial for all babies receiving therapeutic hypothermia to be transferred under the new model, as reflected in the ‘maximum’ scenario).

Prolonged IC stay babies that have not been identified in previous cohorts have also been included within the sickest cohort. To estimate the "ventilated at 48 hours and not improving" group in seen in the criteria, we have made an overall assumption over the proportion babies not already meeting other criteria, who receive intensive care for over 48 hours (LoS greater than two days). Similarly, we have modelled scenarios for prolonged IC stay babies based on a likely (25% of babies will be moved), minimum (10% of babies will be moved) and maximum (50% of babies will be moved) scenario.

4.1.4 Shift of days by level of care

Of the identified smallest and sickest cohort, the future modelling has assumed that the majority of IC days consolidate at the three NICU sites, as well as a proportion of the HD days prior to repatriation, and only a very small proportion of SC days:

• 95% of IC days of the identified cohort will shift to the nearest identified NICU
• 33% of HD days of the identified cohort will shift to the nearest identified NICU
• 5% of SC days of the identified cohort will shift to the nearest identified NICU

The 95% figure is based on feedback from sites that there will be some occasions where transfer is not possible or desirable, and marginal preterm cases where a baby is not deemed to require a transfer. Similarly, the 5% is a recognition that although almost all babies will have been repatriated to their local LNU prior to requiring this level of care, feedback has suggested that there will be a small proportion where this is not possible due to logistical or other reasons. 33% of HDU days is based on an average of 48 hrs in HDU against an overall average stay in HDU of 6 days. For modelling purposes for longer stays such as those seen for the extreme preterm and more complex cases, it is assumed that this overall proportion will hold.

4.1.5 Geographical flows

The core assumption is that future flows will follow the ethos of the Best Start Programme, to deliver care in the nearest appropriate centre, and modelling will assume that patterns of referrals follow the principle that activity from the five future LNU sites will be delivered in the nearest appropriate NICU centre in the future (based on postcode district where available, or the proportion of the local authority closest to each site). We have assumed that current flows to the three NICU sites would remain consistent.

To understand the sensitivity of this assumption, we have modelled a second scenario of historical flows, to understand the impact if neonatal activity follows historical ICU referral patterns.

4.1.6 Current and future occupancy

Future occupancy rates for neonatal cot requirements have been modelled based on an average of 80% across SC and HD cot days and 65% for IC cot days. This is to account for the additional variation in IC demand over time, as seen in the baseline analysis.

This will give a required operational (staffed) capacity in future requirements and has also been contrasted with the current available (physical) cot capacity for which the following rates are based.

4.1.7 Maternity capacity for in-utero transfers

We have sought to estimate the anticipated additional impact of in-utero transfers on maternity services in the three NICU sites. This has been based on the following assumptions:

• There are approximately 6 in-utero admissions for every 10 new neonatal admissions to IC or HD cots (including those who do deliver and those who do not)
• The average number of maternity bed days is 5.2 per in-utero transfer maternity admission.

These assumptions have been based on limited information and should be tested further as part of more focussed review of local maternity unit capacity, including the ability for maternity services at each of the NICU sites to manage additional demand.