Screening patients admitted to a rehabilitation center has become important. Actually, co-existing diseases are not very often evaluated and their importance is underestimated. At our department 166 consecutive patients were enrolled. The Cumulative Illness Rating Scale (CIRS) appeared to be the most suitable for these cases. In this series, stroke patients presented with higher severity and higher comorbidity than the hip fracture patients. The hip fracture cases were older but the stroke cases had higher severity and comorbidity. Comorbidity, also, showed a significant negative correlation with FIM in the stroke patients. These data show that severe comorbid conditions influence the functional autonomy. Severity and comorbidity were correlated both in the hip fracture and stroke cases. A review of geriatric literature demonstrates lower values in patients in rehabilitation. The difference is due to our accurate selection of patients at admission, where general health conditions are considered. In conclusion, the CIRS should be used as a method for selecting patients at admission and as a prognostic index for improvement at discharge. The CIRS, however, has some inconveniences and amelioration is necessary, such as the inclusion of a double testing (admission-discharge), psychiatric disturbances and a new item for skin alone. The Severity Index was higher in women, who were older than men, whereas, comorbidity was the same. In the patients suffering from hip fracture, the age was higher in women, but dependence, severity and comorbidity did not statistically differ between the groups.