"Clinical Global Impressions" (ECDEU): some critical comments

Area of Science:

• Psychometrics
• Clinical Trial Methodology
• Pharmacological Research

Background:

• The Clinical Global Impressions Scale (CGI) is widely used in clinical trials.
• Previous experience highlighted potential limitations in the CGI's application.

Purpose of the Study:

• To investigate the semantic, logical, and statistical properties of the CGI.
• To identify and address inconsistencies and limitations in the CGI's current version.

Main Methods:

• Analysis of semantic, logical, and statistical properties of CGI items.
• Evaluation of item distribution and construction based on clinical trial data.

Main Results:

• The CGI exhibits inconsistent rating behavior and contains redundant information.
• Several CGI items show abnormal distribution properties and questionable measurement significance.
• Item #3 (pure drug effect) lacks a basis for judgment.
• Comparing current condition to admission status is not recommended.

Conclusions:

• The current CGI version requires careful application and interpretation.
• Specific items should be used judiciously or omitted.
• Parametric statistics are not recommended due to poor distribution properties and limited significance of change ratings.