AbstractCancer occurrence was determined through 1988 among a cohort of 143,574 Kaiser Permanente Medical Care Program members whose prescription records were computer‐stored between 1969–1973. Age‐ and gender‐adjusted standardized morbidity ratios (SMRs) were determined by dividing the number of cancers observed among one category of drug users by the number expected based on the cancer experience of the cohort as a whole. This hypothesis‐generating screening programme examined 215 drugs and 56 cancer sites. Although a large number of drug‐cancer associations were nominally statistically significant, many of these have been previously reported and the primary focus of this report is on new findings. Among the new findings considered intriguing were associations of indomethacin with stomach cancer (SMR = 1.51, p < 0.05), prednisone with myeloid leukemia (SMR = 2.27, p < 0.05) and lymphosarcoma (SMR = 2.22, p < 0.002), sulfamethoxazole with lymphosarcoma (SMR = 2.90, p < 0.01), thiazides (SMR = 1.80, p < 0.05) and isoniazid (SMR = 8.46, p < 0.05) with gallbladder cancer.Given the long follow‐up of this cohort, drug‐cancer relationships that are manifested only after latencies approaching two decades would have sufficient time to become apparent. However, since these data are based on pharmaceuticals dispensed over a short time period and because limited data on potential confounders were available, more definite investigations should follow‐up these results.