Bayesian method for modeling male breast cancer survival data.

Area of Science:

• Biostatistics
• Survival Analysis
• Computational Statistics

Background:

• Vast amounts of health data necessitate advanced statistical methods for valid conclusions.
• Male breast cancer survival data from 1973-2009 requires robust analytical models.
• Health science administration generates large datasets requiring sophisticated analysis.

Purpose of the Study:

• To develop a statistical probability model for predicting male breast cancer patient survival.
• To analyze survival times for male breast cancer patients diagnosed in the USA.
• To provide predictive insights into future survival for male breast cancer cases.

Main Methods:

• Utilized a random sample of 500 male patients from the Surveillance Epidemiology and End Results (SEER) database.
• Employed the exponentiated Weibull model with a novel Bayesian method and Markov chain Monte Carlo (MCMC) for parameter inference.
• Assessed model goodness-of-fit using Akaike Information Criteria (AIC), Bayesian Information Criteria (BIC), and Deviance Information Criteria (DIC).

Main Results:

• The exponentiated Weibull model demonstrated a good fit for male breast cancer survival data.
• Statistical inferences of posterior parameters were derived.
• Mean predictive survival times, 95% predictive intervals, predictive skewness, and kurtosis were calculated.

Conclusions:

• The developed model and findings can inform treatment planning and healthcare resource allocation for male breast cancer.
• Results may stimulate further research into male breast cancer survival.
• The study provides valuable statistical insights into male breast cancer prognosis.