🔸 Understanding the difference:
The FDA‘s latest guidance on patient-focused drug development emphasizes the importance of clinical outcomes assessments (COA) as endpoints for regulatory decision-making. One section of the guidance discusses the use of “Change from Baseline” versus “Percent Change from Baseline” as endpoints, specifically focusing on COA scores. However, the key discussion points apply to other continuous variable endpoints as well.
🔹 Change from Baseline considerations:
When using the change from baseline as an endpoint, it’s important to consider the following:
1️⃣ Ordinal scores make it difficult to interpret the change from baseline because the difference between two ordinal scores may not correspond to the same degree of change in the underlying health state.
2️⃣ If expressing treatment effects in terms of change from baseline aids interpretation, it can be achieved by using statistical models, such as ANCOVA, to derive predicted follow-up scores on treatment based on a given baseline score.
3️⃣ In situations where randomized controlled trials are not feasible, such as evaluating certain devices, change from baseline endpoints may be the best available option.
🔸 Challenges with Percent Change (PC) from Baseline:
While percent change from baseline may offer easier interpretability, it comes with several important challenges:
1️⃣ PC treats baseline and follow-up scores differently, leading to varying interpretations. For example, positive and negative percent changes may indicate improvement and deterioration, respectively, despite having the same absolute change.
2️⃣ PC from baseline is undefined when the baseline score is zero, requiring imputation techniques for analysis.
3️⃣ PC from baseline scores often exhibit highly non-normal distributions, making them challenging to model statistically.
🔹 Expert recommendations:
If the treatment effect is expected to be multiplicative rather than additive, applying a logarithmic or similar transformation to COA scores before comparing groups may be beneficial.
🔸 FDA and EMA guidelines:
Both the FDA and the European Medicines Agency (EMA) have provided guidelines on adjusting for baseline values in clinical trials. They suggest including baseline values as covariates in the statistical model, which allows for comparing change from baseline or raw outcome variables. This adjustment improves precision and provides flexibility in endpoint choice.
While percent change from baseline may be used as the primary efficacy endpoint in certain clinical trials, it is crucial to consider the limitations associated with this approach.
🔗 For more details, refer to the article: https://lnkd.in/ghhkqWB6