How good is this medical device? Bayesian mixed models for agreement measures

This project develops **Bayesian mixed models** for assessing agreement metrics between medical devices, focusing on chronic obstructive pulmonary disease (COPD). These models incorporate expert knowledge and allow robust inference by accommodating outliers using heavy-tailed distributions, such as the t-distribution. The project also provides a software package to facilitate the use of these methods, offering user-friendly tools for diagnostics and analysis. Applications include evaluating agreement between new diagnostic devices and existing gold-standard methods for COPD patients. By improving statistical rigor, the research aims to benefit clinical decision-making and prevent the misclassification of inferior or superior devices due to inadequate analyses. The project combines advanced statistical techniques with practical applications, contributing to medical statistics and clinical practice advancements.