The first workshop will be held in the winter of 2020/2021 and be focused on analysis of population data and demonstrating how studies using established administrative data resources such as Medicare claims databases combined with complex well-established and innovative analytic approaches (such as partitioning analyses, time-series based methods of projection and forecasting, and stochastic process models) can be used to uncover previously overlooked or understudied aspects in this area of research.
The second workshop will be in the winter of 2021/2022 and extend this focus to include the analysis of clinical datasets routinely collected in Medical centers (such as Duke Warehouse data) and demonstrate how new and established analytic methods can be rigorously applied to such data to contribute to identifying some of the causes of persistent health disparities between specific groups of the U.S. population and narrowly defined patient strata. Analyses of such increasingly available large health datasets provide an opportunity to obtain nationally representative race/ethnicity-related and geography/area-specific results based on individual-level measures that reflect the real care-related and epidemiological processes ongoing in the U.S. healthcare system.
Two specific aims are envisioned:
Aim 1. Increase collaboration across the scientific community; promote the diffusion of new and established analytic methods across the multiple academic disciplines engaged in the study of: i) disparities in AD/ADRD risks and outcomes as well as other age-related diseases, ii) forecasting approaches for prevalence and mortality of AD/ADRD and other diseases, iii) genetic effects on disparities in health outcomes, and iv) analysis of Medicare and other administrative claim-based data. Bridge the gap between the older and younger generations of health researchers and expose all participants to the different ways even established statistical methods are used across different academic fields.
Aim 2. Introduce the opportunities offered by clinical datasets routinely collected by individual Medical centers, the common pitfalls encountered when analyzing site-specific data, and ways to overcome them. In addition to the subject area of Aim 1, the following topics will be introduced: i) sources of clinic-related disparities in AD/ADRD and other age-related diseases, ii) comorbidity, multimorbidity, therapeutic and surgical treatments, social and genetic factors as sources of disparities in health outcomes of AD/ADRD and other diseases, and iii) forecasting of clinic outcomes and approaches for analyses of health interventions.
Recent advances in machine learning and well-established or recently developed advanced statistical approaches will be discussed at both Workshops. The focus of these workshops, however, is not on the development of methodology but rather on demonstrating their use in practical cases of analyses of disparities in AD/ADRD and other age-related diseases; to demonstrate how the capabilities and level of detail of any given methodology scale with the type of data it is applied to and provide some practical and potentially actionable results.