Because population-based surveillance of cancer incidence across a wide variety of industries and occupations is quite rare and because the vast literature in occupational cancer epidemiology concentrates to a great extent on cancer risks among white males, new methods are needed to generate hypotheses about occupational cancer risks. The results of a series of studies conducted during the past six years suggested that the telephone interview could be utilized as an effective method for occupational cancer surveillance. A 10-minute telephone interview that was developed to collect occupational histories, smoking histories, and other related data and some methodologic issues that were tested during a pilot study are described. The authors found the telephone interview to be an effective instrument for obtaining these data. This study suggests that it is important to obtain as high a proportion as possible of the responses from the study patients rather than proxy respondents; that when one cannot interview the patient, proxy respondents can provide much of the data requested; and that supplementing population-based cancer surveillance system data with these interview data is useful in routine monitoring.