Data-driven knowledge discovery of intervention patterns for older adults with and without end-of-life care interventions using visualization techniques

Abstract

The majority of hospice care in the United States (also known as end-of-life care [EOLC]) is home-based, provided by public health and home care agencies. Worldwide, palliative and EOLC care are often combined and can be provided years before death. In the United States, however, as most reimbursement for EOLC is limited to six months before death, palliative care services are often separate from EOLC. A systematic review of home-based palliative care outcomes in the United States found strong evidence for lower hospitalization rates and lower costs and limited evidence for high patient satisfaction, increased dying at home, and quality of life improvement. To study home-based EOLC, data from 1167 clients with and EOLC intervention were matched 1:1 with older adult health care clients by gender and age using the Omaha System. Those with an EOLC intervention had 41.6% more total interventions, 59.0% more total visits, and 25.6% fewer problems than those without an EOLC intervention. Data visualization techniques from exploratory data analysis were then used to compare this data to standardized guidelines. Some overlap between guidelines and data was found, but granularity increased when terms were combined, showing the ability of the Omaha System terminology to adapt to the level of granularity needed, making it ideal for intervention dataset analysis. This study leveraged Omaha System data from practice settings to discover novel EOLC intervention patterns for older adults. These methods may be used to generate new practice-based evidence for other populations, settings, and practices.