Purpose: To clarify the relevant criteria and optimal time for determining when stroke patients are ready to be discharged home or transferred to another hospital because the average length of a hospital stay for a patient discharged home from the hospital at which he/she first receives treatment is shorter (10.6 days) than that for a patient who is transferred to another hospital (20 days).

Introduction: Stroke patients (Japan Coma Scale (JCS) 1–3) are hospitalized for an average of 20 days when transferred from the hospital at which they first receive treatment. This is significantly longer than patients who are discharged home from the hospitals at which they initially receive treatment (10.6 days). Our hospital takes longer to decide when patients should be transferred because no criteria exist for the selection of patients and no clinical pathway for strokes has been implemented.

Method: To predict when patients can be discharged home after achieving a JCS score of 0, or “capable of a sitting position”, we investigated the number of days of hospitalization. We compared these patients with a similar group of patients who had changed hospitals. We assessed the sensitivity and specificity on each day of a patient’s hospitalization on which a significant difference existed between the two groups. A univariate logistic regression analysis was performed.

Results: Significantly more patients were discharged home after achieving a “JCS0”on days 3 to 7, day 9 or day 10 of their hospitalizations, but no significant difference between the two groups was detected using binomial logistic analysis and an ROC curve. However, significantly more patients were discharged home after achieving “capable of a sitting position” from days 2 to 10 of their hospitalizations. Binomial logistic analysis revealed a significant difference between the two groups in relation to “capable of a sitting position.” The cutoff point on the ROC curve was 2.5 hospitalization days.

In conclusion, it is possible that stroke patients with JCS scores of 1–3 who can assume a sitting position before day 3 of their hospitalizations can be discharged home.

　The purpose of this study was to construct a web-based visualized data analysis model with a business intelligence (BI) tool for the transurethral prostate surgery clinical pathway and then to make and use a common multicenter electronic clinical pathway. Various outcome data from 1,074 cases were compared with respect to surgical methods. The results were shared on the cloud via a commercially available spreadsheet with a BI tool and a mail option. Many types of inter- and intracenter variables were applied to the outcomes, including length of pre-operative stay, number of days of antibiotic dosage, number of days of drip infusion dosage, duration of catheterization, operative duration and length of post-operative stay. Policy differences were inferred between facilities and between doctors within facilities. The results in 24 cases in which multicenter common clinical pathways were used showed outcome improvement. While this study reveals the difficulty of developing and operating a multicenter common clinical pathway, it also points to the importance of doing so and of promoting standardization and improvement within facilities by comparing unique path analysis data on the web. The BI tool will be useful in comparing and sharing outcomes with the clinical pathway and in implementing improvements.