Application of Hidden Markov Modeling to Objective Medical Skill Evaluation

The methodology for assessing medical skills is gradually shifting from subjective scoring of an expert which may be a variably biased opinion using vague criteria towards a more objective quantitative analysis. A methodology using Hidden Markov Modeling (HMM) and Markov Models (MM) were used to analyze database acquired the E-Pelvis (physical simulator) during a pelvic exam. The focus is on the method of selection of HMM parameters. K-Means is used to choose the alphabet size. Successful classification rates of 62% are observed with the HMM as opposed to 92% with the MM. Moreover, the MM provide an insight into the nature of the process while identifying typical sequences that are unique to each level of expertise, where the HM, given their nature as a black box model, do not.