The demand for services for predicting, diagnosing, and managing genetic diseases or diseases with a genetic component is likely to increase faster than the availability of services from medical geneticists and genetic counselors. Health care systems may also impose limitations on referrals to these specialists. If genetic problems are not to be missed and excessive referrals are to be avoided, non-geneticist practitioners will have to recognize when genetic problems should be considered, and initiate diagnosis and even management. Primary-care-centred systems offer the greatest potential for maximizing overall cost-effectiveness, by reducing the demand for specialty services not essential for improving health. But primary-care-centred systems may pose a risk of underdetection and undermanagement of genetic problems if practitioners are not actively supported by information and other educational networks. Several models for dealing with these challenges are presented, including algorithms that aid in recognizing genetic problems.