Related Experiment Videos
Assessing the reliability of two toxicity scales: implications for interpreting toxicity data
M D Brundage1, J L Pater, B Zee
1Department of Community Health and Epidemiology, Queen's University, Kingston, Ontario, Canada.
Journal of the National Cancer Institute
|July 21, 1993
Summary
Toxicity scales used in cancer trials show limited reliability. Experienced data managers disagreed on toxicity grading, indicating a need for improved methods in cancer clinical trials.
Area of Science:
- Oncology
- Clinical Trials
- Biostatistics
Background:
- Cancer therapy toxicity is a critical endpoint in clinical trials.
- Current toxicity grading scales lack established reliability and validity.
Purpose of the Study:
- To assess the reliability of the National Cancer Institute of Canada Clinical Trials Group (NCIC-CTG) expanded toxicity scale.
- To evaluate the reliability of the World Health Organization (WHO) standard toxicity scale using simulated patient data.
Main Methods:
- Seven experienced data managers interviewed 12 simulated patients each.
- Inter-rater and intra-rater reliability were calculated using kappa statistics and proportions of agreement.
- Analyses were repeated for condensed toxicity grading categories (none, low-grade, high-grade).
Main Results:
- Modest inter-rater reliability was observed, with kappa values ranging from -0.04 to 0.82 for clinical categories.
- Proportions of agreement for clinical categories ranged from 0.52 to 0.98.
- Condensing toxicity grades improved agreement statistics, but significant discrepancies persisted.
Conclusions:
- Experienced data managers exhibit variability in assessing patient-reported toxic effects.
- Neither the NCIC-CTG nor the WHO toxicity scale demonstrated superior reliability.
- Differences in the breadth of recorded toxic effects were noted between the scales.