DDReg Pharma

DDReg Pharma

The Use of Item Response Theory in Clinical Outcome Assessments

Clinical outcome assessments (COAs) are a critical component of clinical trials and regulatory submissions for regulatory agencies like the US FDA. The US FDA defines a COA as a “measure that describes or reflects how a patient feels, functions, or survives”; COAs provide important information to the US FDA on the effectiveness of therapies so that they can make informed decisions on the approval or clearance of these medical therapies. Indeed, as healthcare shifts towards patient-centred care, there is a need for implementing more systematic methods and a more patient-focused approach towards product development and regulatory decision making. The US FDA’s patient-focused drug development (PFDD) guidance documents addresses how stakeholders can systematically collect and submit “real” patient data and other information gathered directly from patients and caregivers to enhance product development and regulatory decision making.

Patient-Focused Drug Development Program

Currently, under the PFDD program, the US FDA has issued 4 guidance documents that address critical aspects of collecting and submitting robust “real” patient data:

  • Guidance 1: Collecting comprehensive and representative input
  • Guidance 2: Methods to identify what is important to patients
  • Guidance 3: Selecting, developing, or modifying fit-for-purpose clinical outcomes assessments
  • Guidance 4: incorporating clinical outcome assessments into endpoints for regulatory decision making.

Two new guidance documents have recently been issued to supplement the PFDD guidance series, which communicate the US FDA’s thinking on COA data that can be incorporated in premarket applications. The first guidance, Submitting Clinical Trial Datasets and Documentation for Clinical Outcome Assessments Using Item Response Theory, outlines technical considerations for sponsors with respect to submission of COA data that uses Item Response Theory (IRT). There are different types of COA data that can come from various sources including patient-reported outcome (PRO) measures, observer-reported outcome (ObsRO) measures, clinician-reported outcome (ClinRO) measures, and performance outcome (PerfO) measures.

When it comes to regulatory decision making, the US FDA evaluates how COA-based endpoint results correspond to a treatment benefit that is meaningful to patients. There are many approaches for collecting evidence in order to support the interpretability of COA-based endpoints. COA scores can be translated to their corresponding patient experiences through many approaches. For measures developed using IRT, the meaning of different scores can be enhanced by using IRT item parameters, as highlighted in guidance 4 of the PFDD Guidance Series: Incorporating Clinical Outcome Assessments Into Endpoints for Regulatory Decision-Making

What is the Item Response Theory?

The IRT is a group of mathematical models which describe the “functional relationship between item performance, item characteristics, and the patient’s status on the construct being measured”. COAs that incorporate IRT include static, fixed-form COAs that are developed using IRT, or COAs that are administered using IRT-based Computerized Adaptive Testing (CAT).

The guidance document provides specifications for sponsors that want to submit COA data which use IRT which allow sponsors and the FDA to communicate on issues related to clinical trial design or conducts that could affect the content of various datasets used.

Key considerations for document and clinical trial dataset submission

The guidance document covers important aspects that sponsors must consider when submitting clinical trial datasets and documentation for COAs using IRT. It addresses the content requirement and timing for covered COA documentation where sponsors should consult the FDA during the IND phase to help determined the extend of evidence requirement to be submitted for covered COA and obtaining agreement with respect to the proposed fixed/static COA measure. It also outlines the dataset specifications with respect to Study Data Tabulation Model (SDTM) and Analysis Data Model (ADaM). 


The FDA’s guidance on Clinical Outcome Assessments utilising Item Response Theory aims to assist sponsors in meeting expectations, offering technical specifications, definitions, and practical guidance on submission requirements. By embracing IRT, sponsors can enhance precision and efficiency in measuring PROs, ultimately advancing patient-centred care in clinical research.

As leading regulatory consultant, DDReg has experience in navigating complex regulatory requirements including those for clinical phases of product development with respect to clinical data submission. Read more about navigating nuances for clinical trials: UK MHRA Navigating Nuances in Clinical Trial Applications.