Background <p>Hazardous drugs such as antineoplastic agents pose occupational risks to healthcare workers, making appropriate chemotherapy spill management essential for workplace safety. Hands-on training opportunities in pharmacy education are often limited by safety concerns and resource constraints. Virtual reality (VR) simulation may provide immersive, guided rehearsal without hazardous-drug exposure, but comparative evidence remains limited.</p> Objective <p>This study examined changes in pharmacy interns’ self-perceived knowledge, understanding, and ability following VR simulation–based training and traditional instructor-led instruction and explored whether short-term changes differed between the two cohort-defined instructional orders.</p> Methods <p>This cohort-allocated crossover quasi-experimental study included 37 pharmacy interns at a tertiary medical center in Taiwan. Instructional order was assigned by training cohort rather than by individual randomization and was completely confounded with cohort and calendar period. Cohort 1 (<i>n</i> = 17) received instructor-led instruction followed by VR training; Cohort 2 (<i>n</i> = 20) received the reverse order. Four single-item domains—knowledge of spill-kit contents, understanding of spill-management procedures, ability to independently don personal protective equipment (PPE), and ability to independently manage a spill—were assessed using a 5-point response scale at baseline, after the first intervention, and after crossover. Non-parametric tests were used.</p> Results <p>After the first session, self-perceived scores increased significantly across all four items in both cohorts (all <i>p</i> &lt; 0.001). The instructor-led-first cohort showed greater first-phase changes than the VR-first cohort in self-perceived knowledge of spill-kit contents (<i>p</i> = 0.006) and ability to independently don PPE (<i>p</i> = 0.047); differences in the other two items were not statistically significant. These findings could not be attributed to instructional modality alone. Following crossover, second-phase change scores did not differ significantly between the cohort-defined orders (<i>p</i> = 0.221–0.956). All participants met the prespecified satisfaction criterion for the VR training (37/37, 100.0%; rating ≥ 4/5); 27 (73.0%) identified instructor-led explanation combined with VR as the most helpful option, and 28 (75.7%) indicated that VR could not replace instructor-led explanation.</p> Conclusions <p>Both first-session instructional approaches were followed by short-term increases in pharmacy interns’ self-perceived knowledge, understanding, and ability. Because instructional order was completely confounded with cohort and calendar period, the absence of significant second-phase differences neither establishes equivalence between the orders nor rules out an instructional-order effect. VR may serve a complementary rather than competing instructional role and provide a safe approach for rehearsing hazardous-drug spill-management procedures. These conclusions are limited to short-term self-perceived outcomes and post-training learner perceptions.</p>