This document provides technical specifications for the submission of clinical outcome assessment (COA) data that use Item Response Theory (IRT) and supplements the FDA Center for Drug Evaluation and Research (CDER) Patient-Focused Drug Development (PFDD) Methodological Guidance Series.
As described in the Biomarkers, EndpointS, and other Tools (BEST) Resource glossary, a COA refers to the assessment of a clinical outcome made through a report by a clinician, a patient, a non-clinician observer, or through a performance-based assessment; thus, there are four types of COAs:
1. clinician-reported outcome (ClinRO),
2. patient- reported outcome (PRO),
3. observer-reported outcome (ObsRO),
4. performance outcome (PerfO) measures.
IRT is a family of mathematical models that describes the functional relationship between item performance, item characteristics, and the patient’s status on the construct being measured.
COAs that use IRT include static, fixed-form COAs that are developed and/or scored using IRT, or COAs that are administered using IRT-based ComputerizedAdaptiveTesting (CAT).
CAT is a sequential form of individual testing administered by a computer in which successive items in the COA measure are selected for administration based primarily on the item’s psychometric properties and content in relation to the patient’s responses to previous items.
COAs that use IRT can leverage an item bank for item selection.
An item bank represents the total set of items from which a subset is selected during COA measure development or selected for the patient during adaptive testing.
COAs covered in this guidance (referred to as ‘covered COAs’ throughout this document) include:
(1) fixed-form (i.e., static, fixed-length) COAs that are developed and/or scored using IRT and
(2) COAs administered using IRT-based CAT.
This document provides specifications for the submission of Clinical Data Interchange Standards Consortium (CDISC) Study Data Tabulation Model (SDTM) and Analysis Data Model (ADaM) datasets.
These technical specifications aim to provide general guidelines for standardized dataset content and structure, to facilitate FDA review of the covered COA within the marketing application that the submitted data are intended to support.
The SDTM and ADaM specifications outlined in section 4.0 Overview of Dataset Specifications are not prescriptive and do not include an exhaustive list of all datasets, variables, and controlled terminologies to be submitted for FDA review.
The specifications outlined in section 3.0 Documentation Specifications and section 4.0 Overview of Dataset Specifications are pursuant to discussions with FDA and may vary by clinical drug development program and clinical trial therein.
Read also:
