Validation of the Prolonged Length of Stay Score in the Dijon Stroke Registry

Our results suggest that the PLOS score was neither a good predictor of LoS 6 7 days nor of LoS 6 30 days, although it worked better for survivors. This contrasts with previous results [2, 3] . One possible explanation may be the greater median LoS in our population, compared with previous studies. The proportion of patients with LoS 6 7 days ranged from 60% for patients with a PLOS score = 0 to approximately 80% in other categories except for a PLOS score 6 6 (only 3 patients, interpretation therefore limited). The PLOS score seems valid in health systems with a short median LoS for stroke patients, but is less useful in those with long LoS. As LoS depends on several factors, including accessibility of diagnostic resources and availability of postacute facilities [5] , our findings indicate that the PLOS score requires validation in other populations before generalization.