The effects of long-term fasting on the prognosis and hospital economy of hospitalised patient have not been established. To clarify the effects of long-term fasting on the prognosis and hospital economy of hospitalised patients, we conducted a prospective observational study on the length of hospital stay of patients hospitalised at thrity-one private university hospitals in Japan. We conducted a prospective observational study on the effects of fasting period length on the length of hospital stay and outcome of patients hospitalised for 3 months in those hospitals. Of the 14 172 cases of hospitalised patients during the target period on the reference day, 770 cases (median 71 years old) were eligible to fast for the study. The length of hospital stay for fasting patients was 33 (4–387) days, which was about 2·4 times longer than the average length of hospital stay for all patients. A comparative study showed the length of hospital stay was significantly longer in the long-term-fasting (fasting period > 10 d; n 386) group than in the medium-term-fasting (< 10 d; n 384) group (median 21 v. 50; P < 0·0001). Although the discharge to home rate was significantly higher in the medium-term-fasting group (71·4 % v. 36·5 %; P < 0·0001), the mortality rate was significantly higher in the long-term fasting group (10·8 % v. 25·8 %; P < 0·0001). It was verified that the longer the fasting period during hospitalisation, the longer the length of hospital stay and lower home discharge rate, thus indicating that patient quality of life and hospital economy may be seriously dameged.