A statistical method to convert published response rates into marginal distributions with an example application in...

Area of Science:

• Medical research
• Biostatistics
• Clinical trial analysis

Background:

• Assessing chronic disease severity is crucial for effective management.
• Standard clinical trial metrics often use relative change scores (e.g., PASI 75 in psoriasis).
• There's growing interest in absolute disease severity levels, but comparing them across studies is challenging.

Purpose of the Study:

• To develop a statistical method for estimating absolute disease severity scores from aggregated relative improvement data.
• To enable more accurate comparisons of treatment efficacy using existing clinical trial data.
• To facilitate the assessment of patients achieving specific low levels of disease activity.

Main Methods:

• Developed a conversion method based on a latent 2-dimensional normal distribution model.
• The model accounts for baseline disease severity and allows for baseline inclusion criteria.
• Validated the model using simulated data and results from three Phase III psoriasis clinical trials.

Main Results:

• The developed method accurately predicted absolute Psoriasis Area and Severity Index (PASI) levels.
• Predictions closely matched actual absolute PASI proportions from clinical studies.
• Demonstrated the model's precision in converting relative improvement data to absolute severity estimates.

Conclusions:

• The novel conversion method reliably estimates absolute disease severity from relative changes.
• This approach can enhance the comparability of clinical trial data for psoriasis and potentially other chronic diseases like depression.
• Facilitates better understanding of treatment outcomes in terms of absolute disease burden.