Cancer patients are thought to be at high risk for venous thromboembolic events (DVT/PE). Beginning in October 2007, our tertiary cancer center instituted "mandated risk assessment" computerized DVT prophylaxis order entry, for all hospital admissions with an option for active opt out by the physician with a stated reason. Retrospective review of all DVT/PE events within 30 days of a hospital admission [any inpatient admission (IA) and outpatient surgery (OPS)] in comparable "optional (O)" (January 2005-September 2007) vs "mandated risk assessment (M)" (October 2007-May 2010) DVT prophylaxis order eras. Patient demographics, admission details, type of prophylaxis, treatment, and outcome were also analyzed. There were 16,363 for the O (11,944 IA/4,419 OPS) and 17,757 for the M (12,957 IA/4,800 OPS) DVT prophylaxis order eras. The number of DVT/PE events in the O era was 67 (prevalence 0.41%) versus 102 for the M era (prevalence 0.57%), P = 0.037. In the DVT/PE patients, DVT prophylaxis had been ordered during the index admission in 66 per cent for O versus 83 per cent for M (P = 0.008). Low-molecular-weight heparin was increasingly used in M era (33% vs 16%, P = 0.009). There was also no difference between O vs M era for status at DVT/PE diagnosis (outpatient 36% vs 24%) or associated symptoms. There were no deaths attributable to DVT/PE in the O era versus 3 deaths in the M era. Although DVT prophylaxis use improved with "mandated risk assessment" ordering, the DVT/PE incidence did not decrease. It may be difficult to overcome the surprisingly low baseline prevalence and multiple risk factors in this population.