Interobserver variance in clinical scoring for cystic fibrosis

Area of Science:

• Pulmonology
• Clinical Research Methodology
• Medical Informatics

Background:

• Multicenter clinical research requires reliable methods for comparing patient status across diverse settings.
• Existing scoring systems may lack the simplicity or reliability needed for broad application in cystic fibrosis (CF) studies.

Purpose of the Study:

• To evaluate the interobserver reliability of a combined scoring system for assessing clinical status in cystic fibrosis patients.
• To determine if the proposed scoring system is suitable for use in multicenter clinical trials.

Main Methods:

• Five physicians independently scored 41 cystic fibrosis patients using the Doershuk modification of the Shwachman-Kulczycki system (history, physical, nutrition) and the Birmingham system (chest roentgenograms).
• Interobserver variance was calculated for individual categories and the total clinical score.
• The total consensus score was correlated with the NIH clinical score and pulmonary function measures (forced vital capacity, FEV1).

Main Results:

• Mean individual observer variance was low (1.6-2.9 points out of 25 for categories; 4.5-6.0 points out of 100 for total score).
• Coefficient of variance was approximately 10% for categories and 6.7% for the total score.
• The total consensus score showed high-confidence correlation with NIH scores, chest roentgenogram scores, and predicted FEV1/FVC values.

Conclusions:

• The developed scoring system exhibits acceptable interobserver variance, making it a viable tool for multicenter cystic fibrosis research.
• This scoring system provides a reliable method for standardizing clinical status assessment in multicenter studies.
• The system's strong correlation with established measures suggests its utility in evaluating treatment efficacy and disease progression.