PROJECT SUMMARY/ABSTRACT In the United States (US) and Puerto Rico (PR), cancer is the second leading cause of death. Considering that the population is aging, an increase in the burden of cancer is expected in future years. Consequently, quality of care at the end of life (EoL) is critical for dying cancer patients and their families. Studies in the US have shown that EoL care is expensive for cancer patients and frequently has very low or no clinical value, and often is inconsistent with the patient's needs, values, and preferences. Conversely, timely receipt of palliative and hospice care can improve quality of life and decrease emergency department visits, hospitalizations, and the use of life-extending procedures. Even when progress is made towards improving the quality of EoL care, a growing body of evidence shows disparities among groups within the population. Understanding the different characteristics that affect the quality of care at the EoL is fundamental to implementing effective interventions to reduce inequity and improve the efficiency of the healthcare delivery system. There are few studies in the US and no study in PR evaluating the quality of EoL care from the multilevel perspective. This proposal would be the first of such a series of investigations. PR has a predominantly Hispanic, low-income, and aging population with high insurance coverage. Forty-eight percent of the PR population lives below the poverty level, and there is a shortage of medical specialists, including palliative medicine. This study provides an opportunity to investigate the variations in the use of recommended care, its cost, and the provider and patient's behaviors towards EoL decisions. This study seeks to advance our knowledge and understanding of EoL care among cancer patients by utilizing state-of- the-art claims data analytics coupled with surveys data collection. Towards this effort, we will merge the Puerto Rico Central Cancer Registry-Health Insurance Linkage Database with CMS hospice claims data and conduct physicians, patients, and caregivers’ surveys. Specific Aims are: 1) Assess the use of recommended care and the cost of care among cancer patients nearing EoL in PR; 2) Examine the perceived barriers to EoL discussion and decision-making process among oncologists in PR; and 3) Determine the knowledge, attitudes, and EoL planning among cancer patients and their caregivers in PR. This study will provide new evidence that will help us understand the complex process of EoL care and how different factors interact. Our most important contribution will be to document the potential multilevel influences on EoL care. This information will be essential for developing and implementing a multilevel intervention to improve the quality of care at the EoL. We expect our results can help improve health care delivery and enhance cancer patients' overall quality of life in a cost-effective way.