A multiattribute-utility-function approach to weighing the risks and benefits of pharmaceutical agents.

Both the selection of doses of pharmaceutical agents and comparisons between pharmaceutical agents have long been based on the nonquantified concept of the risk-benefit ratio. Though useful, this concept implies a data comparison that is difficult to make: the toxicity versus the efficacy of a drug compound. This research demonstrates an approach for weighing risks and benefits by combining utility functions for human efficacy and toxicity with animal and laboratory toxicity information to develop an overall multiattribute utility function for an ophthalmic pharmaceutical agent, I-bunolol, intended for the treatment of glaucoma. With this multiattribute function and a small portion of the published data available for this drug, the expected utilities for six doses (including a control) could be compared and the value of this approach in drug-dosage selection demonstrated.