: The candidate's goal is to become an independent health services researcher with a focus on studying health disparities via large datasets. The proposed research project will assess the relationship between race and repeat and follow-up tests (RAFTs) after colonoscopy using Medicare administrative data. This award has three broad objectives that will enable Dr. Richards' progression to an independent researcher. First, the candidate will obtain a deeper understanding of advanced statistical methods, such as survival analysis and hierarchical linear modeling. Second, the candidate will learn how to use and gain practical experience with large datasets. Third, the candidate will develop a deep understanding of issues related to health disparities and health services research. ? ? Prior research by the candidate demonstrated that African Americans receive less colonoscopy and sigmoidoscopy than whites. However, further study by the candidate showed that blacks have a higher use of repeat and follow-up tests (RAFTs) after colonoscopy. With this in mind, the candidate proposes to study why Black Americans have an increased rate of RAFTs after colonoscopy, in the elderly Medicare population. The proposed study will utilize and link several large Medicare and administrative datasets. We will assess the relationship between race and facility and/or provider characteristics using advanced statistical techniques. The effects of race and provider most likely take place at several different levels (e.g., micro and macro). At the micro (individual) level, traditional statistical modeling may detect a relationship between provider specialty and RAFTs; however, at the macro (aggregate) level a stronger relationship may become apparent when neighborhood effects and provider volume are also taken into account. To do this, advanced statistical techniques, such as Hierarchical Linear Modeling (HLM), must be used to take into account the complex nature of these relationships. The proposed project for this award is intended to complement the career development plan since both are aimed at having the candidate learn and apply advanced statistical methods such as HLM and survival analysis, as well as having him become facile in using longitudinal data and complex administrative datasets. Although the results of this proposed study will have particular relevance for Black Americans undergoing colonoscopy, the methodology learned will be useful for establishing a career in health services research. ? ?
