In late-stage elderly with decreased ADL and dementia, accurate nutritional history is difficult and nutrition screening that requires subjective information is often difficult. In addition, although Controlling Nutritional Status (CONUT) is recommended as an objective nutritional evaluation, the current situation is that it is difficult to put it into practical use due to the large number of missing values in blood data. Therefore, we focused on the modified CONUT method and investigated the relationship between the length of hospital stay, BMI, visceral adipose tissue index (VATI), subcutaneous adipose tissue index (SATI), and the survival at the end of 730-day observation period. In this study, 187 patients (78 males, 109 females, average age ± SD = 83.8 ± 5.9 years) were selected from 1,645 late-stage elderly patients aged 75 years or older who were admitted to our hospital from April 1, 2016 to March 31, 2017, and observed for 3 years. As a result, the modified CONUT method showed a high correlation with both the nutrition index of CONUT (p = 0.927) and the geriatric nutritional risk index (GNRI) (p = -0.718), and also correlated with the length of hospital stay (P = 0.002), BMI (P = 0.011), VATI (P<0.001), and SATI (P = 0.002). From the results of logistic regression analysis, the modified CONUT method was selected as one of the significant prognostic factors. These results suggest that the modified CONUT method is effective as a nutritional evaluation method and may be a predictor of prognosis.