Division of Research Data-on-Demand: The Development Process for a Research Data Warehouse Rapid-Querying Tool

Authors

Daniel Ng, Kaiser Permanente Northern CaliforniaFollow
Dimitri Buzkov, Kaiser Permanente Northern CaliforniaFollow
Srivardhan Chimmula, Kaiser Permanente Northern California
Jamila Gul, Kaiser Permanente, Northern California

Publication Date

8-15-2016

Keywords

Virtual Data Warehouse, analytics

Abstract

Background/Aims: Even simple Virtual Data Warehouse (VDW) queries require writing a program in SAS (or other software). While this is a necessary part of a study’s overall analytical effort, investigators often need answers to basic questions prior to undertaking a formal study to develop hypotheses, estimate the population for a potential study, and ascertain other “what if” scenarios. Engaging an analyst can slow down the researcher’s initial explorations and distract the analyst from more complex duties. A means for researchers to run their own simple VDW queries can enhance the efficiency of both researchers and analysts and contribute to faster, higher-quality research.

Methods: The DOR Strategic Programming Group engaged in a process that took over a year to develop a VDW rapid-querying tool for researchers. This involved a process of gathering user requirements, development of technical requirements, assessment of technology alternatives, creation of conceptual solution architecture, logical data model and physical solution architecture, interface design, security design, user training, user beta testing and enterprise security assessment.

Results: The result of this development process was the successful implementation of DOR Data-on-Demand (D3) within the Division of Research. D3 has nearly three dozen active users. Based on user testing and feedback, D3 has been enhanced with additional data domains beyond HCSRN’s VDW data model, including custom metadata (e.g. text description search for diagnoses and procedures, a custom medication list), medication orders, vaccine data and disease registry data. D3 also has been adopted by medical group analysts to provide rapid response to urgent research questions impacting clinical operations and quality of care. We continue to explore other ways that such a rapid-querying tool can be used to improve the health of Kaiser Permanente Northern California members.

Conclusion: The creation of D3 has given researchers the means to create and run their own rapid VDW queries. In addition, research analysts and data managers have found it useful for running more complex queries, validating data and improving data quality. Beyond research, it can enhance the analytic capabilities of quality and operations support groups. Finally, it can potentially assist clinicians in providing better evidence-based care.

Recommended Citation

Ng D, Buzkov D, Chimmula S, Gul J. Division of Research Data-on-Demand: the development process for a research data warehouse rapid-querying tool. J Patient Cent Res Rev. 2016;3:203.