Psychiatrists' relationships with the pharmaceutical and device industries have been a growing focus of attention, with questions raised about the impact of those relationships on prescribing practices, diagnostic criteria, practice guidelines, continuing education, conduct and reporting of research, and patients' and public trust. Indeed, these concerns exist for the medical profession as a whole, with various remedial measures proposed. We suggest that such relationships can be understood as giving rise to a "principal-agent problem," which occurs when an agent (here, a physician) is engaged to advance the interests of another party, the principal (typically a patient), but also faces incentives to promote other interests. Studies suggest that at least some relationships--which include attending industry-sponsored presentations, meeting with marketing representatives, and accepting samples--can alter psychiatrists' and other physicians' behavior in ways that can compromise patients' interests, and that industry-funded research may create bias in the medical literature. These effects are difficult to detect in specific cases, however, because of asymmetries of information and may not be apparent even to physicians themselves. Principal-agent analysis suggests that the possible responses to such problems, including appeals to ethical principles, monitoring behavior, and managing risk-inducing situations, should include consideration of aligning agents' incentives with principals' interests. This type of analysis underscores the similarity of the issues raised by physicians' relationships with industry to problems that arise more generally in society, thus reducing physicians' potential affective responses to these issues and efforts to address them. Finally, such analysis directs attention to the benefits and costs of each alternative, thereby encouraging reliance on evidence as a basis for policy.