The objective was to assess the extent of inappropriate hospital admis
sion and stay in an adult patient population and identify potential co
rrelates of such inappropriateness. Design: a cross-sectional survey o
n a sample of 1,082 in-patients using a modified version of the approp
riateness evaluation protocol (AEP). Setting: adult acute departments
in a 950 bed teaching hospital in the greater Milan area. The patient
sample consisted of patients cared for at 1 of the participating depar
tments on any of 3 index days between October 1989 and June 1990. The
results show that overall 27% of the admissions and 40% of the hospita
l days were inappropriate. The rate of inappropriate admissions was hi
gher for patients admitted during week days and was also significantly
associated with the admitting ward and the age of the patient. Inappr
opriate hospital-day stays were related to the type of ward and to the
location during the stay, with those sampled in the later part of the
ir stay having the highest rate of inappropriateness. Most of the hosp
ital days rated as not appropriate did not require any further stay (7
5%). Conclusions: a substantial proportion of hospital use was found t
o be medically unnecessary and, for the most part, due to hospital fun
ctioning or the behaviour of the doctors. Some properties of the AEP (
high inter-rater reliability, predictivity of expected associations) w
ere also confirmed. The basic features of this revised version of the
AEP make it a good candidate for becoming of great importance in monit
oring the effect of the changes the Italian National Health Service is
currently undergoing.