Cancelled operations are not infrequent, can annoy patients and waste
hospital resources. They were monitored prospectively to determine the
ir frequency, causes and any impact of surveillance-by-objective plus
in-house education. The study involved nine departments and was done i
n three phases over 47 consecutive months. During the study period 19
661 operations were performed and 1976 (9.1%) were cancelled, the rate
s in Phases 1, 2, and 3 being 11.8%, 6.7% and 6.2%. The difference bet
ween Phase 1 and Phase 2 was significant (p<0.001) but that between Ph
ase 2 and Phase 3 was not. The study was not designed to influence pat
ients, and the frequency of cancellations they had caused remained unc
hanged at 2.6%, 1.8% and 1.8%. Nurses contributed least (1.4%) to the
burden of cancellations at a frequency of 0.14%, 0.12% and only 0.04%
respectively. Medical staff contributed most (over 70%) and their thre
e most frequent reasons, which also proved resistant to change, were c
o-morbid slates, overbooking and blood not available. Surveillance ach
ieved the desired impact in five other reasons for cancellation which
became virtually eliminated, viz. anaesthetist or surgeon unavailable,
the difficult intubation, medical students' examination, no consent f
orms and the arbitrary cancellation. We conclude that the frequency of
cancelled operations can be reduced significantly by surveillance but
it should include targeting surgeons as individuals by name. However,
there were two limitations. No impact occurred when physicians' perce
ption was that medico-legal issues were involved, and once the interve
ntion stops any improvements achieved can disappear.