Algorithms, Best Practices, and Expert Clinical Consensus

Abstract

September, 2000. Not long ago I participated in a discussion with a group of mental health administrators about practice guidelines and “best practices.” One member of the group, who represented the recipient constituency in a large state agency, spoke adamantly against any official endorsement of practice guidelines, arguing that they deprived recipients of the element of choice in their own treatment. This concern is quite a legitimate one, of course, but I disagreed with this assumption that practice guidelines deprive patients of choice; in contrast, I replied, good practice guidelines should strengthen the partnership between patient and provider, allowing betterinformed joint decisionmaking or, if you will, a process of mutual choice among options derived from evidence and expert knowledge. Many problems, however, permeate efforts to develop and maintain the algorithms and guidelines that define best practice. The relative merits of published treatment studies must be sorted out, along with their generalizability in the real clinical world. Decisions to develop algorithms based only on published randomized controlled trials, for example, can be persuasively defended, but there will often be a “disconnect” between these algorithms and the use of the newest treatments, such as recently introduced, promising new medications. I have been chairing the Work Group of the American Psychiatric Association that is developing a Practice Guideline for Borderline Personality Disorder (BPD), and we dealt with this problem by adding newer medications, such as the newer antipsychotics, to our proposed medication algorithms, with a special notation that these recommendations derive from clinical expertise, awaiting future randomized controlled studies. Obtaining the best consensus of the expert clinicians with the most experience treating patients with a particular disorder is then the next important challenge in developing guidelines, and this takes on special importance when published studies on the disorder are sparse. Explanatory models for patterns of symptomatic behavior can be enormously helpful in consolidating this clinical consensus, exemplified by the review in this issue by Links et al. of the role of impulsivity in patients with BPD. In this Journal, we have regularly presented reviews of clinically relevant algorithms and guidelines, as well as critiques and analyses of the guidelines themselves. In this issue, Trivedi et al. review the merits and drawbacks of computerized algorithms in clinical practice. An algorithm or guideline is, after all, no more than a shorthand distillation of the best evidence and clinical experience, to assist the busy practitioner in today's clinical world of information overload. In a way, this reflects the primary goal of this Journal, to keep the reader abreast of the most pertinent clinical wisdom. Also in this issue, Pies presents a “primer” of the cytochrome system, to clarify one important emerging database, and Caton et al. elucidate the nature of psychosis produced by psychotomimetic drugs, compared to that typical of the major psychotic illnesses such as schizophrenia.

Ähnliche Arbeiten