Ambulance demand: random events or predictable patterns?

Over the past 20 years there has been an increase in demand for emergency ambulance services across the developed world, placing significant strain on ambulance resources and the acute care sectors of the health care system and impacting on the quality of care delivered. It is estimated that over the next 20 years, ambulance demand will continue to increase due to the impact of an ageing population. Despite this, ambulance demand is an under-researched area of public health. Ambulance services provide the public with emergency health care and, if required, transport to hospital. To do so, ambulance services must ensure there are enough suitably trained staff and resources to cope with demand at any time of day or day of week. This requires an understanding not only of the types of clinical case types but the approximate temporal distribution of these cases. This thesis examines ambulance demand in terms of the types of cases that ambulance services attend and the temporal distribution patterns of those cases. This research used data obtained after paramedic assessment to examine the types of cases seen by ambulance services. This type of data is not commonly used to evaluate ambulance demand. The research highlighted the differences between ambulance clinical case type, point-of-call and hospital data. Clinical case type data generates more fine-grained, clinically relevant case definitions than the more commonly used point-of-call data obtained by the ambulance dispatch centre. Hospital data might not be an accurate representation of ambulance data as not all cases attended by ambulance are transported to hospital. This research found a significant proportion of cases would be missed if only hospital data was used. Hospital planners use clinical data and short-term temporal (time of day, day of week) demand patterns to monitor workforce and resource requirements. This thesis represents an examination of these patterns in large scale ambulance demand data. Individual case types have distinct time of day, day of week distribution patterns and knowledge of the frequency of clinical case types and their demand patterns can be used to understand day-to-day variation in overall ambulance demand. One of the biggest drivers of ambulance demand is ageing and this research modelled the relationships between age and clinical case type on demand and forecast the effect of increasing age on time of day and day of week patterns. This model highlighted important interactions between age and case types and showed how an ageing population will increase the relative frequency of cardiovascular disease and falls, increasing demand in the late morning when ambulance services are already at their busiest. This thesis describes the most comprehensive investigation of case types and associated temporal patterns in ambulance demand to date. It is apparent that analysing ambulance demand using clinical case type data combined with time of day and day of week patterns can provide information about current and future workforce and resourcing requirements. This knowledge is valuable for ambulance services and other health service providers as it can inform policy and practice strategies for ambulance demand management to maximise the pathways to care at a health system level and optimise scarce health system resources.