Estimating procedure times for surgeries by determining location parameters for the lognormal model

Area of Science:

• Statistics
• Data Analysis
• Biostatistics

Background:

• Accurate estimation of distribution parameters is crucial for statistical modeling.
• The lognormal distribution is frequently used to model positively skewed data, such as surgical procedure times.
• Current methods for parameter estimation may not always yield optimal results.

Purpose of the Study:

• To empirically investigate methods for estimating the location parameter of the lognormal distribution.
• To identify the best order statistic for lognormal distribution parameter estimation.
• To compare the performance of different methods for identifying the optimal order statistic.

Main Methods:

• Empirical study using simulation data.
• Comparison of two models for identifying the best order statistic: conventional nonlinear regression and a data mining/machine learning technique.
• Evaluation of the impact of using the best order statistic versus the median on model performance.

Main Results:

• The study identifies a specific order statistic that is optimal for estimating the lognormal distribution's location parameter.
• Using the best order statistic, compared to the median, resulted in less frequent incorrect rejection of the lognormal model.
• The best order statistic led to more accurate critical value estimates and higher goodness-of-fit measures.

Conclusions:

• The selection of the appropriate order statistic significantly impacts the accuracy of lognormal distribution parameter estimation.
• Employing the identified best order statistic offers advantages in statistical model fitting and hypothesis testing.
• Improved estimation of surgical procedure times, facilitated by this method, can potentially enhance surgical operations.