By Steve Light
DataCeutics offers 23 years of programming experience to the pharmaceutical and biotech industry. These groups of highly experienced programmers are well-versed on CDISC, foundational standards, domain, and data mapping, and use of specific tools to assist during this period of data transition and standardization.
Why STDM andADaM?
One of the motivations for the development of CDISC, SDTM, and ADaM standards stems from the need of regulatory reviewers to reduce the time and effort required to locate and interpret data for each submission they face. As of late 2016, the FDA requires clinical data to be submitted in SDTM format. While SDTM has been encouraged by the agency for several years, sponsors will lose the option of providing tabulations without submitting SDTM. Standardization has been good for the industry, making structure and content of clinical data predictable and not dependent on conventions unique to a single sponsor. Use of common standards reduces the effort required for technical staff to become oriented to new projects, whether within a single company or across companies. It drives efficiency and quality of data preparation and reporting and enables the development of standardized tools to screen data for quality issues and to automate presentation and analysis. As well, SDTM encourages the re-use of code and software for data preparation and reporting.
The value of standardization is being internalized by sponsors. SDTM has become a standard component of clinical data systems. Standardization simplifies routine tasks and speeds the process of data summaries and reports. It also allows programmers to spend time on more challenging projects rather than routine tasks and to develop metadata-driven systems to control development and to monitor the quality of the outputs.
Many companies are using in-house tools for mapping or converting clinical data into SDTM and ADAM. DataCeutics uses Pinnacle 21 Community or Enterprise, a commercial product widely used for performing QC of SDTM or ADaM deliverables and ensures consistency with defined standards and terminology. Also used are DataCeutics own metadata-driven tools and a well-defined and proven process for SDTM implementation and verification.
The conversion of ‘raw’ data to SDTM can be a straight forward or complex process, depending on many factors. DataCeutics has experience working with legacy data as well as CDASH-based source data to prepare SDTM domains. As well, experience is also available for less predictable additional sources of data such as externally-sourced labs, ECGs or event/assessment adjudications. Integrating data from these various sources requires the sort of programming and analytical experience that DataCeutics offers.
When Can We Start?
Until recently, SDTM was often addressed retrospectively and as a final step before finalization of a submission. As CDISC standards have become more tightly integrated into the project lifecycle in research organizations, availability of SDTM data has been introduced, in time, to serve as the foundation of ADaM analysis data, also used as part of clinical study reports. The path to SDTM and ADaM for any research is dependent on (as always) priorities and timelines. For new or recently implemented studies, SDTM design and mapping can closely follow CRF and clinical database finalization. Development of conversion programs can be initiated with test data extracts and will gain strength as actual collected data becomes available.
ADaM data must maintain traceability to SDTM sources, and development of ADaM data requires a finalized SAP. The design of ADaM can be started once SDTM specs are developed. ADaM program development is dependent on the availability of SDTM datasets from which to program.
Each of these efforts naturally requires an accompanying QC/verification process. At DataCeutics our procedures, processes, and tools support development and verification. Our goal is to maximize use of metadata-driven tools to provide high-quality results and speed of production for our clients.
Staffing to support SDTM and ADaM is not easily predictable without understanding particulars of the study and systems in use. Many factors come into play but most important is the timeline of the project. Compressed schedules will call for more staffing to allow concurrent development efforts. Extended timelines (often dictated by study duration) will allow for lower staffing. At a minimum to do SDTM, two programmers plus a manager are required. ADaM can be more involved; the specifications for efficacy datasets, in particular, are often (usually) developed by statisticians who will perform the statistical analysis.
Important factors to consider and what your conversation with DataCeutics regarding CDISC will include:
- Project Timeline
- Complexity of the Study
- Number of Sources of Data
The ideal process for CDISC, STDM, and ADaM:
- Number of Sources of Data
- CDASH (we hope!) Database and a CRF Finalization
- Exporting of Test Data
- SDTM Design/Mapping
- SDTM CRF Annotation
- SDTM Programming and Verification
- ADaM Design (could start earlier)
- ADaM Development and Verification
We will also consider Data Management standards you currently have in play:
- Study metadata (define.xml) and reviewer guide development and verification
- Standardized Lab Data
- Standard Terminology Consistent with CDISC
- System in your environment
- Software tools available for use
- Repository of Code for Reuse
It is best to get started for SDTM and ADaM during CRF review of annotations of your collected data. SDTM design can start as soon as the raw database is finalized. Test data can be used for testing before patients are even enrolled in the study.
Custom Software and copyrights are transferred to Client. DataCeutics will reserve to itself only a license to keep a copy of the Custom Software, which may be used to support and maintain the Custom Software, and create derivative works.