The objective of the proposed study is to quantify the burden of psychological morbidity as well as health care costs and utilization among people (probands) whose spouses do or do not have cancer. The parent grant for the study is a National Institute on Aging (NIA) grant that uses one million married couple's Medicare data from 1991-2001 to determine the inter-relationship of morbidity and mortality between elderly spouses (R01 AG017548). We will take advantage of this existing longitudinal, nationally representative dataset to develop a new line of inquiry into the burden of caregiving for spouses with cancer. Our analysis will use advanced and complementary statistical techniques to quantify how being a caregiver specifically influences proband health and well-being. We propose to do two types of comparisons: (1) between individuals, comparing probands with spouses who have cancer to probands with spouses who do not have cancer, and (2) within individuals, comparing probands to themselves at times when their spouse does have cancer and does not have cancer. For our first analysis, we will use propensity scores to create matched comparison groups of couples where the """"""""case"""""""" couples include a spouse with cancer and the """"""""control"""""""" couples include a spouse without cancer. The couples will be matched on all observable characteristics, allowing us to isolate and measure the impact of cancer caregiving on proband mental health, health care utilization, and health care costs. Because we recognize that using propensity scores will not allow us to control for all possible characteristics among couples, we will conduct an additional analysis that controls for all observable and unobservable characteristics within couples. For our second analysis, we will use fixed effects Cox models to examine how the onset of cancer in our """"""""case"""""""" couples influences the frequency and timing of proband mental and physical health events and costs. Using the fixed effects Cox models, we will be able to examine how spouse cancer hospitalization influences proband health; these models will also control for any effects on proband outcomes due to spouse death. The combination of a large, comprehensive data source and appropriate statistical methodology will allow us to make a unique contribution to the research on how caregiving for cancer affects proband psychological, physical, and financial health. Findings from this research will lead to the development of several possible new lines of research: (1) it may lead to an R01 proposal that creates an intervention targeted at those probands most likely to suffer from high psychological, physical, and financial burdens associated with caregiving for spouse cancer patients; and (2), it may provide the foundation for similar research conducted both before and after the implementation of the Medicare Part D prescription drug benefit, to examine how the implementation of the benefit influences both spouse and proband health outcomes. This application represents an important developmental and empirical step towards a larger agenda which examines the inter-relationship of illness in spouses and family members. ? ? ? ? ?
