CDISC Standards for COAs using Item Response Theory (IRT)
December 18, 2023
The US Food and Drug Administration (FDA) recently finalized a guidance document outlining the agency’s thinking on Submitting Clinical Trial Datasets and Documentation for Clinical Outcome Assessments Using Item Response Theory. The guidance lays out technical specifications to consider when submitting clinical outcome assessment (COA) information that uses Item Response Theory (IRT) in a marketing application. The COA data can be from clinician-reported outcome (ClinRO), patient-reported outcome (PRO), observer-reported outcome (ObsRO), and performance outcome (PerfO) measures.
The COAs covered by this guidance could be static COAs developed and/or scored using IRT or COAs administered using Computerized Adaptive Testing (CAT). The FDA clarifies that IRTs are mathematical models describing the relationship between an item’s performance and characteristics with a patient’s status on the concept being measured. These COAs select items from an item bank representing a total set of items.
The guidance outlines specifics with respect to the following items:
- Documentation to be submitted to the FDA for the COA. This includes but is not limited to all potential items and responses, COA instructions received by responder, user manual, and scoring details.
- Timing of agency meetings to discuss use of the COA
- Dataset specifications for SDTM including:
- ZQ item dataset detailing all items within the item bank or items within a static COA
- Information to be included within the Trial Summary (TS) dataset
- Response data typically represented within the Questionnaires (QS), Functional Tests (FT), or Disease Response and Clin Classification (RS) dataset
- Dataset specifications for ADaM ADQS dataset
- How missing data should be handled within SDTM and ADaM datasets
Do you have questions on implementation of CDISC standards with respect to COAs? Contact us to talk to one of our Data Standards Specialists.
Charity Quick, Senior Director, Statistical Programming & Data Standards, has directed and supported statistical programming teams in more than 100 clinical studies across all development phases for more than 17 years. This includes analytical programming of ISS and/or ISE development in more than 15 marketing applications. Ms. Quick uses her deep industry experience and CDISC knowledge to develop strategic high-performing global programming and data standards teams to bring technical expertise and efficiencies to sponsor projects of all sizes. Her experience with data standards integration and reporting from study start to submission spans a variety of therapeutic areas with a focus on pain, CNS, and OA.
Heather Kopetskie, Senior Director, Biostatistics, has worked at Rho for more than 19 years. Ms. Kopetskie provides leadership and oversight for the biostatistics department and utilizes her experience to guide sponsors on the design and analysis of clinical trials. Her experience spans many therapeutic areas with an emphasis on solid organ and cell transplantation along with rare disease (orphan) products. Ms. Kopetskie has supported both federally funded projects and biotech/pharmaceutical companies. She has served as the biostatistical functional lead for the Immune Tolerance Network (ITN) Statistical and Data Coordinating Center and the Clinical Trials in Organ Transplantation (CTOT) Statistical and Clinical Coordinating Center overseeing more than 30 studies in the areas of allergy, immunology, and transplantation.