Modelling treatment and rehabilitation of stroke patients - Using simulation to evaluate the present and plan for the future

Stroke is a common cause of illness, disability and death. In Norway, it is estimated that approx. 12,000 people are affected annually. The interaction between hospitals, rehabilitation institutions and municipal services is extensive and complicated. In this project, these challenges for stroke patients in Norway are studied using statistical modeling and simulation. Data already collected for approx. 2,000 admissions to the stroke unit at Akershus University Hospital are combined with national patient databases. Statistical models examine how patient characteristics such as age, gender and municipality of residence are related to outcome measures such as duration of hospital stay, utilization of rehabilitation, anxiety, depression and health-related quality of life. Information on relatives of stroke patients has also been collected, among other things, to investigate whether their mental and physical health is affected by the stroke. In addition, time trends in the occurrence of stroke are studied. The analyses show that there are significant regional differences, but also a strong downward trend, which more than offsets the effect of the population aging. For younger people (under 45), the incidence is low, but more unchanged over time. On the other hand, the number of admissions to stroke units of patients who turned out not to have a stroke is on the rise, so there will not necessarily be a lower need for capacity in the country's stroke units in the future. The statistical models that have been developed, together with population models from Statistics Norway, can provide forecasts of patient flow through the treatment and rehabilitation services for stroke patients in Norway. The project has collaborated with a research community in Melbourne and investigated differences and similarities in the treatment of stroke in Norway and Australia.

Prosjektet har produsert viktig kunnskap om tidstrender for forekomst av slagsykdom og andre tilstander som medfører innleggelse på slagenheter ved sykehus i Norge. Gjennom kobling mot befolkningsframskrivinger for forskjellige deler av landet vil dette gi myndighetene mulighet til å planlegge kapasiteten for slagenheter bedre for de neste 20 årene.

This project will study the treatment and rehabilitation of stroke patients in Norway through statistical and simulation based modelling. We will use standard statistical modelling techniques to analyze and represent associations and causal effects among patient parameters like age, sex, socio-economic status and ethnicity and dependent variables like length of hospital stay, duration of rehabilitation, level of anxiety and depression and health related quality of life. We will develop discrete event simu lation models to describe the different paths that stroke patients follow through treatment and rehabilitation. We plan to make separate sub-models for the different phases of the disease: An Incidence model represents the random events of new stroke case s in the population. A Pre-hospital model represents the phase until hospitalization. A hospital?s stroke unit model represents internal queues, radiology and treatment processes. A rehabilitation model describes factors that govern the utilization of reh abilitation resources. A long-term care model represents duration and level of care, while a municipality model describes the local support that the patient receives. On top of all this we develop a demography model that represents time trends in the popu lation, which enables us to analyze how the treatment-rehabilitation machinery will respond to an ageing population. The analysis will uncover social and ethnic inequalities, as well as suboptimal current use of resources, which can be rectified. The pred ictions of future demands for treatment and rehabilitation will help us to identify bottlenecks before they become a problem. The Ahus project NOR-SPOT has collected detailed follow-up data from 1908 admissions to the stroke unit. In addition we will util ize national databases like the Norwegian Patient Registry. The combination of NOR-SPOT data and more coarse data for the entire Norwegian stroke population will enable us to analyze both in depth and width.