Our hospital is responsible for acute care as part of a regional-based integral medical care center and uses Diagnosis Procedure Combination (DPC). We attempt to increase patient discharge rates within period II in the DPC / Per-Diem Payment System (DPC/PDPS). However, hospitalization of cerebrovascular disease patients tends to be prolonged by reduced activities of daily living caused by disease and patient background. Thus, we surveyed factors related to extended hospitalization of patients with cerebral infarction, who are among the largest number of patients with cerebrovascular disease. In this survey, we retrospectively analyzed 109 hospitalized patients who underwent medical treatment from May 2016 to March 2017. Patients were divided into two groups (within period II, and period III and above). We performed univariate and multivariate analysis on factors contributing to prolonged hospitalization. Multivariate logistic regression analysis revealed that place of residence (pre- and posthospitalization) affected the hospitalization period. Among them, patients who were admitted from their homes and discharged to facilities other than their homes had the longest hospital stay. Interestingly, univariate analysis showed a significant difference (P<0.001) between the two groups in the National Institute of Health Stroke Scale (NIHSS) score on admission, while there was no significant difference (P = 0.65) with multivariate analysis. It implied that NIHSS score affected the decision about recuperation location after discharge, but not the hospitalization period. Based on our study, we identified key issues that need to be addressed: 1) the time to determine appropriate medical care after the acute phase, and 2) the waiting period for proper transfer. Therefore, to promptly discharge patients from hospital, we believe it is important to provide early intervention by ward nurses at the time of hospitalization, and to review our ongoing efforts to strengthen the system and enhance collaboration within our medical center.