Tensor decomposition has recently been gaining attention in the machine learning community due to its versatility in processing large-scale data. In particular, it has become popular for the analysis of Electronic Health Records (EHR). However, this task becomes significantly more difficult when the data follows complex temporal patterns. This paper introduces the notion of a temporal phenotype as an arrangement of features over time. We propose a novel model integrating several constraints and regularizations to discover interpretable hidden temporal patterns. We validate our proposal using both synthetic and real patient data from the Greater Paris University Hospital. The results show that this technique outperforms the recent state-of-the-art tensor decomposition models.