- Research and Projects
- Get Involved
Neonatal services are a specialist service; services are commissioned and care is delivered across a network of hospitals. Capacity planning cannot be performed effectively at a single hospital since changes in capacity at one hospital affect demand at other hospitals in the network. The aim of this project was to develop a capacity/demand model to analyse provision of services across a network by; (i) looking at the ability of individual hospitals and the network to meet demand, (ii) the expected level of resource utilisation and, (iii) the distance of parents from the point of care of their infant.
A simulation model was built that allows alternative configurations of neonatal services to be evaluated. The model incorporates four levels of care, each requiring a different intensity of nursing. Mimicking the normal variation in the number births and care requirements of infants, the model also monitors the distance from home to point of care. In addition, it looks for the closest available cot, taking into account both infant needs and current neonatal workload (units become closed to new admissions when workloads exceed a given level).
One aspect of the model examines the tension between centralisation and localisation of services. As units are merged in the model, the variation in workload at the remaining units is reduced and less spare capacity is needed to cope with uneven workload. Resources may therefore be used more efficiently. However, this efficiency gain is accompanied by a significant impact on parents; travel distances and the number of parents living more than one hours drive from the point of care are significantly increased.
A simulation model was built, coupled to travel distances and times from home locations to points of care. Infants were categorised into seven groups depending on the level of maturity at birth, or the need for specialist services (cardiac and surgical care). The requirement for different types of care (intensive, high dependency, special, transitional) and the length of stay in each level of care was sampled from distributions for each infant category. The model searches for the closest cot to home where the unit is equipped to provide the care needed and where the unit is not closed due to excessive workload. Workload takes into account both the number of infants and the nurse requirements for different levels of care.
The model was used to examine a range of scenarios between full localisation (all services available in all hospitals) and full centralisation (services only available in one hospital). These extremes, whilst not viable options, helped provide the boundaries to the performance of the system. In order to meet the nurse staffing guidelines for neonatal care 90% of the time, it was predicted that nurse establishment (excluding supernummery and non-clinical staffing requirements) ranges from 200 at fully localised to 166 at fully centralised. However, this change is accompanied by a very significant impact on parents; when services are fully localised 98% of parents are within a one hour drive of the point of care, whilst when services are fully centralised only 40% of parents are within a one hour drive.
Using the key outputs of the model we were able to predict:
The model allows for the analysis of a complex system of care, where hospitals cannot plan independently of each other. Impact on parents was also evaluated allowing a balanced view of potential hospital efficiency against parental travel times.
There is a real tension between centralisation and localisation. In such cases it is important to model the impact to both hospital and patient (or parent).