A comparison of statistical methods for combining relative bioactivities from parallel line bioassays

Area of Science:

• Pharmacology and Toxicology
• Biostatistics
• Bioassay Methodology

Background:

• Estimating common bioactivity from multiple parallel line bioassays is crucial for drug development and toxicological studies.
• Existing methods include ordinary unweighted average, weighted average, and maximum likelihood approaches, with similar techniques applied in meta-analysis.
• Assessing the performance of these statistical methods is essential for reliable bioactivity estimation.

Purpose of the Study:

• To compare the performance of ordinary unweighted average, weighted average, and maximum likelihood methods for estimating common bioactivity.
• To evaluate these methods based on coverage probabilities and confidence interval lengths through a simulation study.
• To identify the most suitable methods for analyzing multiple parallel line bioassays.

Main Methods:

• A simulation study was conducted to assess different statistical methods.
• Methods compared include ordinary unweighted average, various weighted average approaches, and maximum likelihood estimation.
• Performance was evaluated by comparing coverage probabilities of true relative bioactivity and confidence interval lengths.

Main Results:

• The ordinary unweighted average method demonstrated the best coverage probability, albeit with slightly wider confidence intervals.
• Weighted average methods provided good coverage and narrower intervals for homogeneous bioactivities.
• For heterogeneous bioactivities, weighted average methods performed well with a liberal significance level for homogeneity testing; maximum likelihood methods showed good coverage for homogeneous bioactivities.

Conclusions:

• The ordinary unweighted average method is a preferred approach due to its consistent coverage probability.
• Two specific weighted average methods developed for bioassays are also recommended, particularly for homogeneous data or with appropriate heterogeneity testing.
• The choice of method depends on the homogeneity of the bioactivities being combined.