Background: Physicians who accept payments from a pharmaceutical company are more likely to prescribe that company’s cancer drug[s]. Whether long-term physician-industry relationships are more likely than intermittent relationships to result in practice changes is unknown. Methods: Open Payments data, containing industry-physician financial transactions, was linked to Medicare Part D prescription drug claims for 2013-2015. We identified physicians who treated each of 4 cancer types which had multiple orally-administered treatment options: renal cell (axitinib, everolimus, pazopanib, sorafenib, sunitinib), lung (erlotinib, afatinib), CML (dasatinib, nilotinib), and prostate (abiraterone, enzalutamide). We used modified Poisson regression to test if the number of years (either 1, 2, or 3) in which a physician received payments was associated with increased prescribing of the paying company’s cancer drug in 2015. We also tested whether physicians at NCI-designated cancer centers were more likely to receive industry payments. Results: The physician cohort sizes were: RCC, 674; lung, 966; CML, 367; prostate, 1,483. Controlling for physician characteristics including practice size, prescribing volume, and total dollar amount received, physicians who received payments in all 3 years (vs. 1 year) were more likely to use the paying company’s drug within RCC (RR:1.69, 95%CI 1.32-2.17) and lung cancer (RR: 1.42, 95%CI 1.15-1.77), but not CML (RR:1.13, 95%CI 0.94-1.35) or prostate cancer (RR: 0.93, 95%CI 0.85-1.01). Results were similar comparing physicians who received payments in 3 years vs. 2 years: RCC RR:1.53, 95%CI:1.21–1.94; lung RR:1.12, 95%CI:0.93–1.35; CML RR:1.13, 95%CI 0.94–1.35; prostate RR:0.93, 95%CI:0.85–1.01. Physicians at NCI institutions were less likely to have received any industry payments during the study period (RR:0.70, 95%CI:0.61-0.80). Conclusions: Longer-term industry relationships may be associated with greater changes in drug prescribing than time-limited ones. Conflict-of-interest policies and disclosures may be more informative if specifying duration of industry relationships. This study was limited by a short time range, and lack of payment records prior to 2013.