Characterising end of life health care expenditure

Date of ISAC Approval: 
Lay Summary: 
The increase in healthcare spending observed in England over the past few decades has placed a greater burden on government funding, particularly at a time when financial constraints are tight. This has led to an increasing interest in understanding the drivers of healthcare costs and identifying more cost-effective ways to provide care to patients. An area that has attracted particular attention is end-of-life (EOL) care. The hospital expenditure an individual gives rise to over a lifespan is increasingly focused on the last few years prior to death. This suggests that a greater understanding of the drivers of healthcare costs at EOL is critical to enable policy makers to develop better strategies to reduce costs without compromising EOL care quality. For example, policies to achieve cost savings in the treatment of long-term conditions associated with persistent use of care over a prolonged period may free resources that can be used to support high quality care the last few months of life. Such strategies are likely to represent an efficient use of resources and be ethically acceptable. Our research will identify patterns of EOL spending in both general practice and hospital care settings across a diverse range of patients and health conditions. This will allow us to investigate healthcare costs over time in the years prior to death and how these vary with both patient characteristics and health conditions. In turn this will help inform policies towards a more effective use of healthcare funds without damaging the quality of care patients receive.
Technical Summary: 
Proposals aiming to reduce health care costs often target the high cost of care during the last year of life. However, focusing on such a short period is myopic. The majority of the highest cost group patients are not in their last year of life. At the population level, the proportion of the annual spending that is due to caring for individuals in their last year of life is considerably smaller than the proportion of the three years spending that is due to caring for individuals in the last three years of life. To gain a complete understanding of EOL health care expenditure attention needs to be paid to patterns or trajectories of spending during longer periods before death. This project will characterise EOL care into profiles of expenditure during the final three years of life and will determine the drivers and characteristics that underlay such profiles. Our sample are individuals who died during the calendar years 2012-2014. We will use CPRD data to identify primary care activity and linked HES data to identify inpatient, outpatient and A&E activity over a period of three years prior to their death. For each decedent, we will calculate aggregate health care costs in 36 monthly intervals. We will estimate group-based trajectory models (GBTM) to identify distinct trajectories of health care spending in the data such as persistent high users, progressive users, and late rise users of EOL care. Once the expenditure trajectories have been defined, the probability of membership to a cluster will be modelled as a function of patient-specific covariates such as age, diagnostic groups; chronic conditions, and number of co-morbidities. More detailed information about the profile of spending near EOL can shed light on potential strategies to mitigate costs while preserving high-quality care at EOL.
Health Outcomes to be Measured: 
The outcomes will be patient's health care expenditures measured in each of the 36 months prior to death. They will comprise primary care, inpatient, outpatient, and A&E costs

Nigel Rice - Chief Investigator - University Of York
Nigel Rice - Corresponding Applicant - University Of York
Panagiotis Kasteridis - Collaborator - University Of York
Dr Rita Santos - Collaborator - University Of York

HES A&E;HES Admitted;HES Outpatient;ONS;Patient IMD