A belief in the validity of informed consent is one of the most import
ant consequences of the doctrine of autonomy in medical ethics. Truly
informed consent requires full disclosure of all relevant information
by the doctor, competence of the patient to appreciate what the inform
ation signifies, understanding of the facts and issues by the patient,
a voluntary choice by the patient and an autonomous authorisation for
treatment or entry into a trial. Each of these conditions is hard to
fulfil. In particular, full autonomy cannot exist in illness, and this
is acknowledged by the act of consultation. The issue is further comp
licated by the stochastic nature of biological systems and the respons
es to illness and treatment. Disclosure of likely outcomes is, of nece
ssity, unsatisfactory when a patient seeks surety when entering the cl
inical process with a serious and life threatening disease. Neither st
rict logic nor the law provide answers to this problem. The legal and
moral issues have become confused. This is unfortunate, because the le
gal concern often centres on avoidance of actions in law rather than o
n the more fundamental issue of benefit to the patient. There is a nee
d to teach doctors that discussion is a necessary part of the doctor-p
atient relationship, that fully informed consent is seldom-if ever-pos
sible, and that skill in understanding what the patient is seeking is
more important than the development of rigid and legally ''complete''
consent forms.