Modalities for diagnosis and prediction are used widely and account for a significant fraction of health care expenditures. However, the evaluation of the impact of tests on outcomes has proved to be a difficult undertaking. A key challenge stems from the ubiquitous role of tests in medical care. Insofar as a test provides information, which is subsequently incorporated into further diagnostic and therapeutic decision making, the test’s impact is strongly affected by the intervening decisions about medical care. Prospective, randomized studies of tests are relatively rare in comparison to studies of therapy, and tend to require large sample sizes. Thus, the study of the impact of diagnostic tests on subsequent process of care and outcomes is increasingly turning to the use of registries and large administrative and clinical databases. In this presentation we will begin with a quick review of the experience from a large national registry to evaluate PET in cancer. The focus of the presentation will be on the experience from ongoing research in the IDEAS Study, which is a national study conducted to evaluate the impact amyloid plaque imaging for Alzheimer’s Dementia on subsequent health care management and utilization. The study combines information from cohort of dementia or mild cognitive impairment patients who undergo amyloid plaque testing and are enrolled in a prospective registry, with information from CMS claims data on the registry cohort and a matched control group.