Can experts agree when to hospitalize adolescents?

Rates of psychiatric hospitalization and lengths of stay for adolescents have been a focus of recent controversy. With the advent of managed care, hospital systems and third-party payers are looking for ways to decide when hospitalization is indicated. The authors sought to determine whether experts could agree on the appropriateness of putative indicators for hospitalization of adolescents for conduct disorder or substance abuse.

Using a methodology developed at the Rand Corporation and previously applied to procedures in medicine and surgery, the authors developed a list of possible indications for hospitalization of adolescents with conduct disorder and/or substance abuse. A nine-member panel of experts in these areas, balanced by geography, academics/clinical practice, and whether the expert was in charge of a hospital unit, then rated the appropriateness of each indication twice under a modified Delphi procedure.

Using prespecified definitions for agreement, after the initial rating the panel had low levels of disagreement (11%) and moderate levels of agreement (28%) on more than 1,900 possible indications for hospitalization. Despite an expanded number of indications, the panel reduced disagreement to less than 5% and increased agreement to more than 55% after the second round of ratings.

The consensus achieved compared favorably with the results of similar panels judging the appropriateness of procedures in medicine and surgery. The methodology is applicable to studies of the appropriateness of pharmacological or psychotherapeutic interventions in both child and adult psychiatry. The results of such studies can form the basis for rational utilization review and treatment authorization decisions.