During training in psychiatry, I learned that "splitting" is a primitive defense mechanism commonly used by patients with borderline personality disorder to manage intense contradictory feelings. I learned that the early phases of "splitting" could feel wonderful for the recipients, but then after the idealization was over, the situation would end in chaos and misery. I am very concerned that this notion of splitting is apt when we consider the larger implications of the article on cost savings from pill splitting by Weissman and Dellenbaugh in this issue.

These authors report on how instituting medication splitting as a cost-saving measure affected adherence and outcome in a cohort of patients with schizophrenia treated in the Veterans Affairs (VA) system. The findings imply that some medications, such as oral risperidone, can be split without serious adverse consequences. The results may actually suggest otherwise, but methodology issues are not my primary concern here.

My primary concern is that popularizing the use of pill splitting for cost-saving reasons may ultimately boomerang and not only increase pharmaceutical costs but also adversely affect medication adherence. How could this happen? On the basis of several years of consulting with the pharmaceutical industry, my guess is that this kind of study would steer a company away from a "flat pricing" policy in which the cost per pill is roughly the same regardless of dose. Such a study may also cause a company to reconsider whether it made sense to make the tablets easy to split. It is probably not a coincidence that risperidone has all of these patient-friendly dose-regimen features and has wound up being the choice for the VA pill-splitting project. Therefore, from a financial perspective, new medications may be "punished" for having patient-friendly regimen features. Then no one should be surprised when new medications are priced per milligram rather than what is effective for the individual patient, or when the medications are not easy to split or are inconvenient in other ways.

The overriding concern for the mental health field is the moral implication of legitimizing pill splitting—on a system level—as a way to lower medication costs. What if there were a VA study on the impact on adherence of ordering antipsychotic medications from Canada? Wouldn't the issue be whether it is appropriate for a federal agency to lower costs in this way?The real irony here is that the VA is studying potential costsavings from a pill-splitting project, whereas Medicare Part D does not permit price negotiations between the seller (drug company) and the purchaser (the federalgovernment). This is the real splitting that is going on, in my opinion.