Accurate medical diagnosis is very important for patient treatment. The diagnostic framework is to connect the patient's information, X, such as clinical findings of pain, with the outcome of diagnosis, Y, such as cancer. The outcome Y sometimes includes both a decision regarding the disease and the classification of its grade. Moreover, some clinical indicators of X play important roles in the decision about disease and other indicators in the classification of disease. There is some uncertainty regarding information X and the outcome Y in the diagnosis and this is an aspect of "fuzziness" in medical diagnosis.As a new trial to clarify the fuzziness of diagnosis, we analyzed the clinical data on patients with Carpal Tunnel Syndrome (CTS) by the method of multivariate analysis.From the analytical results, we propose a new diagnostic system with two layers of diagnostic filters: A first Layer (for the decision about CTS) by the clinical indicators of Phalen's test and Ring sensory splitting and a second Layer (for the classification of CTS) by the indicators of Thenar muscle atrophy and Pinch disturbance.