OBJECTIVES: To determine the quality of evidence required by the Pan-Canadian Oncology Drug Review in order to support a positive or conditional recommendation. METHODS: All 64 pCODR reviews from inception to December 31, 2015 were searched for relevant content. Factors of interest included: reliance on abstract-level data as primary clinical information; reliance on early-terminated trials; reliance on non-randomized information; reliance on unblinded trials; and the use of network meta-analyses to establish comparative efficacy. RESULTS: The use of abstract-level data appears to be significantly predictive of a negative recommendation, while the other factors have yet to demonstrate significant effects on the quality of recommendations by pCODR. CONCLUSIONS: Many factors contribute to success at pCODR, including the availability of other treatment options, (i.e. unmet need), and the burden of illness. The quality of the clinical evidence in support of reimbursement is only one factor among several, though it has clearly been shown to be the most important. Within the domain of clinical evidence, we have determined that certain limitations on the quality of the available evidence have demonstrated a meaningful effect on the likelihood of a poitive or conditional recommendation. Manufacturers would therefore be advised to consider these conclusions when determining the advisability of submitting earlier in their clinical development program. The implications of a negative recommendation may substantially outweigh the benefits of earlier access to markets.