OBJECTIVES: This study was performed to evaluate the occupational relationship on 190 cases of cancer selected out of 622 cases of cancer registered in two university hospitals from January 1, 1996 to December 31, 1997. METHODS: The selection criteria was for the patient to be more than 40 years old with lung, liver, urinary bladder, nasal cavity and skin cancer or leukemia. We reviewed the medical records to update the missing data and occupational histories. Telephone interviews were used to obtain complete occupational histories on the subjects. RESULTS: The sites of cancer in the order of relative frequency was lung (51.0%), followed by liver (32.9%), urinary bladder (14.1%) and skin (2.0%) in male, liver (41.5%), followed by lung (31.7%), skin (19.5%) and urinary bladder (7.3%) in female. The occupational histories of 190 cases with suspected cancer-causing occupations were recorded 5.8% on the doctor's medical records and 33.2% on the nursing records. The response rates of the telephone interviews were 87.4%. The distribution of occupation according to the telephone interviews was farmer (47.7%), office worker (16.1%), salesman (12.8%), production worker (6.7%), simple laborer (3.4%) and unknown (13.4%) in male, housewife (63.4%), farmer (17.1%), saleswoman (9.8%) and unknown (9.8%) in female. And there were two cases of suspected occupational relationships in the lung cancer cases. CONCLUSIONS: We could not discover definite cases of occupational cancer but found out two cases of suspected occupational relationships. Occupational cancer is likely to increase in the near future, so the efforts to detect occupational relationships with cancer should be continued.