The overall goal of this grant is to inform VA policy makers about the effect of offering concurrent care (disease modifying treatment -- Chemotherapy and XRT - while receiving hospice/palliative care) to Veterans dying of cancer. This goal and our project aims are highly consonant with those of the VA Palliative Care Initiative instituted in 2009. Unlike Medicare, for the last decade the Department of Veterans Affairs has emphasized the development of clinical services integrated with palliative care and community hospice services without having to forgo disease modifying treatments (i.e. chemotherapy and radiation). Conceptually, Veterans'ability to concurrently pursue disease modifying treatment and hospice care means that, unlike in Medicare, Veterans do not have to make the "terrible choice" of pursuing either disease modifying treatment or care focused on symptom control and completion of end-of-life tasks. Relying upon a combination of quantitative and qualitative methods, we seek to understand the costs and benefits of allowing Veterans with terminal cancer to continue to receive disease modifying treatment concurrent with hospice care. Based upon recent randomized trial evidence, we hypothesize that Veterans with newly diagnosed, metastatic, non-small cell lung cancer (NSCLC) in VAMC catchment areas that embrace the model of concurrent care for terminal cancer will use hospice and palliative care earlier than might otherwise be the case and will be exposed to fewer "heroic" measures such as ICU admissions, repeated hospitalizations and multiple cancer treatments in the last weeks and months of life. To address this hypothesis we will assemble a comprehensive longitudinal database merging VA health care utilization and Medicare enrollment and claims data for all Veterans who are cancer decedents between 2005 and 2013. Based upon these data, we will first characterize regional and VAMC level differences in the use of hospice/palliative care provided either by VA or Medicare and/or both. Next, we will characterize the rate of use of chemotherapy and XRT per month in the 6 months prior to death and, then, merging these sets of findings, estimate the national prevalence of concurrent treatment, how it varies regionally and how it has changed over the last decade. Then, using a difference in difference design, we will estimate the effect of a VAMC adopting a concurrent care model on the likelihood that Veterans who are newly diagnosed, metastatic, NSCLC patients will experience the following undesirable health services outcomes: 1) burdensome transitions in the last month of life;2) "late" (= 3 days before death) entry into hospice care;3) receipt of disease modifying chemotherapy or radiation therapy in the last month of life;4) an ICU stay in their last week of life and death in an ICU;and 5) survival. Based upon which VAMCs manifest higher levels vs. lower levels of concurrent care, we will undertake site visits to those medical centers, interviewing key informants in the oncology, palliative care and nursing services as well as speaking with the local community hospice leadership and other interested parties. Using next of kin surveys in the VA PROMISE database, we will calibrate the quality of life at the end of life assessments in PROMISE with claims based burdensome transitions at the end of life. Finally, using the quality of end of life valuations from the PROMISE data, we will undertake a formal cost-effectiveness analysis of the concurrent care model. If this observational study replicates the published small scale randomized trial, VHA not only has the evidence for more aggressively pursuing concurrent care but the VA is poised to influence national policy on this topic.
Improving the quality of end of life care is a major goal of the VHA. Recent evidence suggests access to cancer treatment together with hospice and palliative care improves patient outcomes and reduces costs. We propose to replicate this study since, unlike Medicare, the VA allows for such concurrent care. We will test the impact of concurrent care on veteran's quality of life, survival and costs and estimate its cost-effectiveness.