Estimating the distribution of surgical durations based on eliminating the abnormal data

There are many manual errors about the records of surgical durations. Meanwhile, the extreme short and long records have an extremely small probability. The author provides a resample method to eliminating those data that are defined as abnormally small and big and test whether the distribution of surgical durations is log-normal distribution. The results proposed that doctors are still fit for log-normal distribution. Managers should urge the surgeons to complete surgeries without exceeding the surgical duration limit. By this way, productivity of the operation rooms will be improved to higher level.