Improving this country's very expensive and poorly performing health care system requires critical analysis of all health care expenditures and their attendant benefits. In a study reported in this issue, Valenstein and colleagues help frame the debate about costs and benefits of psychiatric care. These authors capitalized on the availability of comprehensive data from the VA health system to estimate the cost of implementing monitoring procedures that have been recommended by the FDA for reducing suicide risk in depression treatment.

The authors estimate that the protocol would cost the VA system up to $270 million annually. Although this represents somewhat less than a 1% increase in the VA's health care budget, it is nonetheless a substantial sum. For example, it represents nearly half of annual federal expenditures for the Mental Health Block Grant, which helps to provide safety-net mental health services to the civilian population. The magnitude of the expense underlines the importance of understanding the validity of the FDA guidelines for preventing suicide and, perhaps, of the FDA's considering the costs of its recommended protocols in relation to anticipated benefits. Clearly, this admonition is relevant to the full spectrum of health care interventions, with treatments for psychiatric conditions being no more and no less important than other therapies. Some would criticize any consideration of cost in the face of a life-saving and clinically sensible intervention or would appropriately worry about including the cost of nonintervention. However, maximizing the efficient utilization of scarce health care resources requires that we understand all the costs and benefits.

For these reasons, and given the elevated suicide risk for veterans and the expense of the FDA protocol, further research is needed. The availability of this large VA data set might provide the statistical power for a longitudinal, experimental analysis in which differing intensities of follow-up could be implemented with several samples and evaluated in relation to completed suicides or suicide attempts. Longitudinal research might also help to better identify individuals for whom this level of vigilance is merited and perhaps suggest a stepped approach to assessing suicidality with graded follow-up. Studies could investigate the use of automated methods to perform follow-up screening. Veterans would report their thoughts and feelings, which in some cases would trigger personal follow-up.

Analysis of both costs and benefits must be part of the contemporary health care debate. Although recent congressional interest in comparative effectiveness research does not explicitly consider costs, it certainly anticipates such consideration. The cost analysis presented by Valenstein and colleagues highlights the importance of understanding differential effectiveness and of conducting research to develop evidence-based protocols that consider comprehensive costs as well as benefits.