It may not be good. But change is inevitable. The traditional role of the clinical data management system (CDMS) is shifting and, in some cases, vanishing. In its place, new tools are becoming more accepted.
What about the most critical tool in all of clinical research? What about the database? Is there an optimal design for all sponsors? How can the needs of operational speed, on the one hand, be balanced with the demands of analysis and reporting about the operational realm?
Is the proper model for the database one large, unified centralized database that contains everything the company needs to know? Or is a federated approach, with linked subsidiary data reservoirs, the preferred solution?
Folks, we need help on this. What follows is our brief understanding of an unfolding debate in the industry. This question may have ramifications for the user experiences at clinical sites, and therefore for the rate of adoption of technology generally.
Reconciling the needs of operational teams (people who gather and clean data) and the needs of managerial teams (which analyze and report on data) will not be easy. A pure focus exclusively on the operational needs may ensure faster system performance--but require middleware that may or may not work. The unified data warehouse, on the other hand, may facilitate better analysis across multiple trials but impose draconian, unthinkable compromises on stakeholders (perhaps Phase I study teams) who need more specialized, custom-tailored solutions to get their work done. Some data warehouses, we dimly recall, have had lofty ambitions that have never quite been realized.
Here's the bottom line.
Electronic data capture is here to stay. Data warehouses are here to stay. What should connect them? How big and massive should clinical trial databases be? Will a federated approach be easier--or only appropriate for smaller companies that have not made major investments in Oracle databases? Readers are invited to write the correspondent with opinions, facts and real-world experiences.