Cancer risk evaluation and early detection are subject to serious limitatio
ns mainly related to human factors and to characteristics of the data invol
ved. To help overcome these problems, a computer-based system was designed
to provide the physician with a clearer clinical picture and aid in directi
ng patients to appropriate measures. Clinical and epidemiological data rela
ted to early cancer detection and to cancer risk factors was collected from
the literature and incorporated in a database, together with heuristic rul
es for evaluating this data. Individual data obtained from patients through
a questionnaire are input into CaDet, a computerized clinical decision sup
port system. A report summarizing patient data and cancer hypotheses, with
a scoring system that reflects degrees of alarm, is generated. The CaDet sy
stem, as well as some preliminary results of the clinical experience accumu
lated in its use, are described. These preliminary results suggest that the
approach may be useful in improving cancer risk assessment and screening i
n primary care setups.