Probability of fatigue failure and minimum sample size requirements for cyclically loaded bone.

Area of Science:

• Biomaterials Science
• Mechanical Engineering
• Orthopedic Research

Background:

• Bone fatigue-life data exhibit considerable scatter, necessitating probabilistic characterization.
• Existing fatigue testing literature lacks standardized sample size recommendations for bone.
• Accurate characterization of bone fatigue failure probability is essential for clinical and engineering applications.

Purpose of the Study:

• To determine the minimum sample size required for fitting a Weibull distribution to bone fatigue-life data.
• To evaluate the accuracy and precision of Weibull parameter estimation methods for bone fatigue.
• To provide a framework for sample size determination in future bone fatigue studies.

Main Methods:

• Utilized two existing experimental datasets of cortical and subchondral bone samples.
• Estimated Weibull parameters using maximum likelihood and rank regression methods.
• Employed Monte Carlo simulations and convergence analysis to ascertain minimum sample size.

Main Results:

• Suggests that up to 11 samples (n=11) may be required for adequate Weibull parameter quantification in bone fatigue tests.
• At the converged sample size, estimated Weibull parameters deviated 3%-25% from true population parameters, varying by estimation method.
• The maximum likelihood method demonstrated superior accuracy and precision in Weibull parameter estimation.

Conclusions:

• A minimum sample size of 11 is recommended for reliable Weibull analysis of bone fatigue data.
• The maximum likelihood method is preferred for estimating Weibull parameters in bone fatigue studies.
• These findings establish a critical guideline for future research on bone fatigue failure probability.