Utilization of health care services at the end of life

The project is organised in three parts. Part I: Where do people die? Trends in place of death (PoD). In this part we have investigated how PoD is influenced by the characteristics of the dead, such as age, gender and cause of death, and by supply side factors. We have decomposed time trends in the distribution of PoD. The analyses are based on data from the Cause of Death Register from 1987-2011. More than half of the increase in the proportion of deaths in nursing homes (from 29.5% in 1987 to 45.5% in 2011) can be predicted from changes in gender, cause of death and age among the deceased. Trend analysis of standardized PoD proportions shows that there has been a clear trend towards nursing homes as a place of death for cancer patients throughout the period. The impact of system variables has been analysed using multinomial multi-level regression analysis (data for the years 2003-2011) and the results support the hypothesis that PoD can be linked to service capacities (including bed capacity in hospitals and nursing homes, share of the home service) and travel time to hospitals. Part II: Patterns of health care utilization and costs at the end of life. In this part we have identified the use of health and care services in last years of life and calculated associated costs. We have highlighted the importance of proximity to death versus age of service costs by comparing ulisation and costs in thelast year of life with one year's consumption for "survivors" (alive at least 2 years after the observation year). The analyses are based on data from the Cause of death register linked to data for various service registers: NPR (specialist health services), KUHR (primary care), IPLOS (care services) and the Norwegian Prescription Database. We have investigated consumption patterns before death for those who died in 2011 and examined the importance of individual characteristics, interaction between different service components and differences by place of residence and place of death. Less than 1 percent of the population dies every year. Costs in the last year of life (365 days) for those who die make up more than 10 percent of total estimated annual health and care costs. The substantial increase in cost in high age among both elderly decedents and survivors relates to higher long-term care costs. Health care cost decreases among elderly, peaking in the late 50s among decedents and early 80s among survivors. Cost per decedent in the last year of life are much higher than average cost of survivors at all ages, but the ratio decreases with age. For specialist somatic care, primary care physicians and prescription drugs, differences in cost per user contribute more to differences in average cost between decedents and survivors than differences in user rates. The opposite was found for specialist mental health care and long-term care. In the last year of life, the proportion with service use increases with proximity to death. This applies to most of the services that has been investigated. An increased share of long-term stay in nursing homes is reflected in a decline in the proportion of home help. 1/3 was long-term nursing home resident and more than 20% had short-term nursing home stay the last month before death. The number with hospital admissions increases much especially the last months of life, and more than 50% had at least one day in hospital last month. Nearly 20 percent of the deceased changed their residence during the last year of life. If we count the number of transfers between different places of residence and institutional stays, then more than half had more than 3 transfers the last year of life. The most common transfer is between home and hospital. More than 30% had more than 3 hospital admissions. The last 7 days there are more transfers from hospital to short-term stays in nursing homes than to home. Of those living in ordinary homes at the time of death, almost half died in hospitals, more than ¼ in nursing homes and 1/5 at home. The proportion being hospitalised the last year of life is much lower for those who live in nursing homes throughout the period (above 1/3) than those who live in ordinary homes (more than 80%) and sheltered care homes (70%). The share with short-term stays in nursing homes the last year is higher among those who change residence than among those who do not change their place of residence. These service utilisation patterns vary with place of residence, PoD, and characteristics of the deceased such as age, gender, cause of death and marital status. Part III: Terminal care in municipal care settings. In this part we have examined terminal care in the municipalities through a survey among employees in nursing homes and home care services, among other things about experiences regarding relatives' ability to accept the imminence of death and relatives? ability to reach agreement when deciding on behalf of patients unable to consent.

Resultatene fra prosjektet forventes å ha nytteverdi i første rekke for helsemyndigheter og tjenestetilbydere på lokalt, regionalt og nasjonalt nivå, og derigjennom kan resultatene også påvirke tjenestetilbudet til mennesker i livets sluttfase. Prosjektet vil også kunne ha betydning utover landets grenser ved at det bidrar til den internasjonale forskningen på dette feltet. Det har også bidratt til nettverksbygging som har resultert i nye prosjektsamarbeid.

The overall objective of our project is to describe the patterns of care utilization and costs at the end of life and to identify factors associated with utilization and place of death. The project will be organized in three parts: Part I: Where do peop le die? Trends in the place of death. Part II: Patterns of health care utilization and costs at the end of life. Part III: Terminal care in municipal care settings assessed by patients' relatives and health care professionals The project will use a com bination of registry data, surveys and qualitative interviews to answer these questions.