Traditionally, the systems used for statistical computing are in-house and present in the company’s own data centers. For smaller organizations, statistical computing may be simply having SAS on local workstations, but still in-house. With the emergence of cloud services, Biostatistics departments are increasingly opting to outsource their Statistical Computing Environments. In our experience, the following are the reasons we have seen for moving to a hosted model.
Quicker Startup Time: Hosting companies can deliver a full Statistical Programming environment in weeks or sometimes days. Internally it can be a struggle to line up the necessary IT project resources such as Project Managers, IT specialists and QA. Typically, the required IQ documentation will be provided by the vendor, relieving the organization of much of the time-consuming documentation required to meet regulatory requirements. In addition, we find the hosted model more efficiently keeps up with SAS upgrades.
Administration: Often resources of the internal IT organization may be scarce. Requesting new accounts or permissions can take days or weeks. SAS Upgrades and other server requests can get bogged down in the long queue of other projects the IT department supports. In our experience, a hosting vendor specializing in SAS is able to provide superior service or augment existing internal support. The business may also be able to take over more of the administration duties including the creation of accounts and user project access in order to speed up onboarding of new users. How much administration you can wrestle away from your internal IT varies by organization, but going with a hosted model can offer administration advantages.
Cost: It will be a large budget item to host your environment. However, the hosting costs are offset by not having to purchase and maintain internal servers and reduce the need for internal IT support resources. While quantifying the savings to the organization can be difficult, it is important to show that the economics make sense.
Hosting your SAS environment in the cloud comes with all the advantages of cloud computing. Without detailing the benefits of cloud computing, which is extensively covered elsewhere (Google), we like to think of how it simplifies or eliminates tasks that can take time away from the actual work of statistical programming and clinical reporting and allow focus on the core business. Who has time for planning and implementing performance enhancements, increasing SAS capacity, or disaster recovery, when these tasks can be passed to the SAS hosting vendor?