OBJECTIVES: The interpretability of clinical outcome assessments is centred around the concept of meaningful change and evidence-based approaches for determining the magnitude of change that is relevant to patients. Differentiating within-patient change from between-group difference in change is important for valid inference. The former serves to define individual-level responders while the latter helps guide interpretation of aggregate-level statistical contrasts between treatment arms. The publication of the draft FDA Patient-Focused Drug Development Guidance 4 has renewed discussion of methods for derivation of between-group thresholds. This presentation will provide an applied illustration of derivation processes by which both threshold types can be established.

METHODS: This study used an interim data cut a 52-week randomized controlled trial assessing the efficacy of Depemokimab every 26 weeks versus placebo. The Asthma Daytime Symptom Diary (ADSD) and Asthma Nighttime Symptom Diary (ANSD) are two separate six-item patient reported outcome measures completed by participants in study 206713 to evaluate the core symptoms of asthma during the day and night, respectively. Thresholds were established using anchor-based quantitative assessment methods: within-patient thresholds denoted a difference between timepoints while between-group thresholds reflected inter-group differences in that change over time.

RESULTS: 335 patients, 58.2% female, average age 54 years and three-quarters non-smokers across a diverse geography, were included in the analysis. Within-patient change thresholds denoting improvement on the ADSD ranged from 1.2 to 2.2. For the ANSD, within-patient change thresholds ranged from 1.4 to 1.8. Evaluation of thresholds for interpreting meaningful between-group difference suggested a range of 0.8 to 1.7 for the ADSD. Equivalent estimation for the ANSD suggested a threshold range of 1.0 to 1.8.

CONCLUSIONS: Thresholds for meaningful within-patient change and between-group difference for the ADSD and ANSD have been successfully estimated using anchor-based strategies, which will be presented and discussed in full.