Objective: To describe how decision making in the Cancer Subcommittee of the Ontario Committee to Evaluate Drugs (responsible for deciding which novel and costly cancer drugs will be funded in Ontario) is evolving along three scientific paradigms. Methods: We describe how these paradigms shape both criteria and process of decision making. We also systematically reviewed meeting transcripts to analyze decisions made in 2006. Results: Evidence Based Medicine (I) is part of decision making through rigorous evidence reviews and the implicit rule that drugs must pass the threshold of effectiveness to be funded. Although drugs must pass one evidence threshold to be licenced in Canada, higher standards are required for reimbursement (e.g. phase III controlled trial data, peer reviewed publication). Health economic criteria (II) are assuming greater weight in decision making, as the review process is standardized, committee members become more economically literate, and a cancer pharmacoeconomics unit is established. The process of decision making (versus decision criteria) is evolving using the ethical foundations of Accountability for Reasonableness (III), important tenets of which are transparency, accountability, and stakeholder involvement in the decision process. Review of the 2006 decisions showed that 16 of 37 drugs were funded (43%). Among negative funding decisions 86% were characterized by inadequate evidence (main reason in 43%), 71% were characterized by cost effectiveness concerns (main reason in 15% ), and 5% by ethical concerns (main reason in 5%). Forty-eight percent of decisions were multifactorial. Conclusion: Each paradigm used to make cancer drug funding decisions comes from a distinct intellectual tradition. Most decisions in 2006 were based on more than one paradigm. We believe that optimal decision making for cancer drugs involves integrating concepts from all traditions, involving both distinct decision criteria and decision processes. Integration requires judicious tradeoffs between both efficiency and equity, and evidence quality and efficiency/equity.