Although it may be simple to evaluate some elements of clinical genetics, it is difficult to evaluate genetic counselling. We review previous studies of the outcomes of genetic counselling; although the methods used may be valid in research studies, there are practical and ethical difficulties in applying them to the measurement of clinical effectiveness in standard practice. No simple measures of outcomes would be suitable. Research evidence will be helpful in deciding what services it is appropriate to offer, and the quality of a service can then be assured by assessing the quality of the clinical process in three ways: 1) adherence to agreed protocols and standards of care; 2) peer review and audit of clinical activity; and 3) ongoing review of the satisfaction of clients and referring physicians with the service. The assessment of client satisfaction will need to be a sophisticated form of retrospective satisfaction with the service provided, and such a scheme has yet to be fully developed.