Reinventing Evidence-Based Interventions?

To the Editor: Although I have long admired the contributions of Dr. Sullivan, "The Role of Services Researchers in Facilitating Intervention Research" (1) and the supporting commentary by Dr. Druss (2) in the May issue are potentially harmful. The authors argue that evidence-based interventions designed by "top-down" researchers are not sustained because frontline clinicians have not had an active role in their development and evaluation. Unless the clinicians participate in the design and testing of an intervention through "bottom-up" research, implementation ends when research funding ends. The authors advocate for bottom-up research initiated by clinicians, relegating researchers to a facilitative role.

If this approach were carried to its logical conclusion, each mental health facility would develop its own bottom-up interventions that, by definition, could not be independently replicated or disseminated without becoming a top-down approach. Aside from an unsupportable cost, each facility would reinvent its practice on a recurrent basis in the face of high staff turnover. Patients would continue to find themselves subject to idiosyncratic approaches of minimally established efficacy.

Dr. Druss, in turn, creates false dichotomies between top-down and bottom-up investigative approaches: rigor versus relevance, expertise versus practicality, and fidelity versus flexibility (2). In my experience, every top-down approach worth its salt is also relevant, practical, and flexible. Our own recent contributions are illustrative (3,4). Beyond an overly subjective definition of "evidence" and an epistemologic contortion regarding the production of evidence, Dr. Druss claims that evidence-based interventions need to be "reinvented" in order to accommodate different patients, providers, and organizations as well as financing and regulatory systems. "Invalidated" seems a better descriptor of the process. Where is the evidence that efficacious, top-down interventions don't work, as designed, in the real world? Even a cursory pass through the effectiveness literature on psychosocial treatment would suggest that effects most often equal or exceed those found in efficacy studies.

There are far more plausible causes of implementation failure (5) than clinicians' noninvolvement in the creation of evidence-based interventions, including an absence of dedicated funding and psychiatric leadership. In my experience, clinicians in public mental health facilities have had little or no influence on choice of interventions. Most hold a bachelor's degree or less. Typically, clinicians provide what administrators require—"services" that are able to be reimbursed, regardless of efficacy. If insurers recoil against reimbursing broadly replicated, evidence-based interventions, imagine how they would view the unique, unreplicated, and equivocal practices of individual agencies.

Finally, these articles do not accurately depict how top-down interventions are developed. Those known to me arose from a relevant, lengthy, and often in-depth, clinical experience. Theoretically grounded practice principles increasingly reflect the etiopathology of the disorder in question (3,4), an integration of theory and practice rarely seen in agency-initiated clinical research. The real question raised by Drs. Sullivan and Druss is whether the country will continue to inadequately treat the mentally ill by consigning them to the whimsy of idiosyncratic experimentation or ethically treat them with the evidence-based interventions of established efficacy and effectiveness that currently exist.