PS1-33: CESR Proof-of-Concept Projects: Lessons Learned from the DCC Perspective

Abstract

Abstract Background/Aims Two multi-site proof-of-concept projects were initiated to demonstrate the viability and sustainability of the newly created Kaiser Permanente (KP) Center for Effectiveness and Safety Research (CESR). This presentation highlights lessons learned from the Data Coordinating Center (DCC) perspective as both projects conclude. Methods Each proof-of-concept project, charged with finishing within one year of start-up, assembled teams, including investigators and programmers from each KP site. CESR’s DCC supplied an overall project manager as well as consultation services on Institutional Review Board (IRB) issues, Data Use and Sharing Agreements(DUSA) and secure transfer of collected data. Other DCC services included SAS multi-site extraction programs for Virtual Data Warehouse (VDW) data and other non-VDW data. When non-VDW data was requested, the DCC also defined specifications for these tables. Cleaning of, Quality Assurance (QA) of and final combination of data from multiple sites was performed at the DCC. Results The proof-of-concept projects helped to refine CESR processes for future projects. Lessons learned include 1) consult a VDW specialist early in the planning process, including before IRB submission; 2) describe data to be extracted and outlined in IRB applications and DUSAs broadly enough to include related variables to minimize the number of time-consuming modifications (e.g., ask for newborn statistics, not head circumference or agpar5); 3) create and use templates for extraction programs, for programs which combine multi-site data and simple statistical processes; 4) date stamp programs and returning datasets and related output; 5) develop one location to allow investigators and programmers to access project related information and documents; 6) develop documents for tracking programs (including modifications) and datasets flowing between multiple sites and data coordinating center; and 7) invest time to create, update and communicate data flow diagrams (DFD) for each SAS program sent to sites. Conclusions Documenting and analyzing the processes involved in proof-of-concept projects can help inform and contribute to efficiencies in future projects.