Don’t Miss the Latest

When you subscribe to the d-wise blog, you’ll get the latest
industry trends, news and tips right in your inbox!

No, thanks. I don‘t need to stay current.

The Solution Landscape Part 3: SCE

The final part of our three-blog series on the solution landscape brings us to the Statistical Computing Environment (SCE). Perhaps somewhat inaccurately named, the SCE covers the umbrella of an overall clinical framework, extending well beyond statistical.SCE hosted?

The SCE is, in essence, a managed clinical data environment. It “provides a foundation for documenting rigor in the analysis and reporting of clinical trial results while increasing productivity.” Where a CDR is strictly about data, the SCE has a broader scope, often encompassing the CDR.

An increased need for transparency and traceability in light of heightened regulations (such as 21 CFR Part 11) and data standards (such as CDISC’s SDTM and ADaM clinical data standards), has elevated the importance of a structured SCE. In addition, outsourcing and data sharing up the ante for industry in terms of improving data quality and integrity.

The purpose of a structured SCE is to support good statistical and clinical data management practice, which can be achieved through a combination of a technology protocol together with well-defined operational processes. The end result, when thoroughly implemented, should be replicable research, guaranteed data integrity, and validated analysis and reports.

A 2011 Drug Information Journal paper prepared by industry biostatisticians outlined six elements integral to statistical programming environment: controlled access, version control to ensure traceability and manage processes, dependency management of statistical outputs and inputs, versatile analysis tools, the ability to organize and enforce process rules, and metadata management.  

An SCE helps to make processes far more transparent, which in turn makes it easier to track the progress of reports, submissions, and other projects. Other advantages include the ability to break activities in smaller, bite-size chunks and a more organized approach to the creation of documentation.

Implementation of an SCE requires a number of fundamental steps, including:

  • An understanding of the existing environment and how users of the system work

  • A repository and platform technology, which might be a commercial solution, such as SAS Drug Development, or, as might better suit small companies, a custom built environment that includes a validated analytics server implemented in a virtual machine (VM) environment

  • Development of new business processes to help ensure successful adoption of the SCE

  • Integration of the SCE with other clinical systems

  • Implementation of programming standards

  • System configuration and validation

As in any implementation, d-Wise recommends an iterative process, identifying critical pieces to undertake first, for example the establishment of a standard hierarchy and guidelines for moving data in and out of the system. It is also best to identify and define risks alongside a mitigation plan, and to ascertain resources and responsibilities. d-Wise also highly recommends beginning with a pilot implementation to test tools and processes.

For an SCE to fulfill its mission, which should be to improve statistical competence and efficiency by making it easier to leverage clinical trial data, understanding and defining the needs of the organization needs to be a first step, followed by a clear development of standard operating procedures and processes.

In reality, most solutions on the market as CDRs are first and foremost SCEs with CDR modules or capabilities. This includes SAS’ Drug Development (SDD), Oracle’s Life Sciences Hub, and Entimo’s entimICE DARE.

For smaller companies, the cost of acquiring, implementing and validating an SCE solution can be hard to justify. In that case, an alternative is to build a solution – with external help. d-Wise, for example, has worked with clients to build cost-effective and simple solutions, where possible using open source technologies  for system components and virtualization technology to limit validation overheads. This has been achieved by building project teams from both d-Wise and the client, with d-Wise providing the roles of system architect, system administrator, and project manager, and the client providing various roles to represent the users of the system.

A successful solution – whether purchasing off-the-shelf solutions or building an SCE – requires an evaluation of the current environment and users’ needs before determining requirements for the new environment. 

For more information on d-Wise's whitepaper on a hosted SCE please go to: http://www.d-wise.com/Statistical-Computing-Environment/

Share

About the Author

Leave a Comment