Adopting a continuous improvement model for future DSM revisions

The approach used for making changes to both the DSM and ICD has been up to now to revise the manuals in their entirety at fixed (albeit variable) intervals. Typically, these diagnostic revision efforts have been multi‐year affairs involving the appointment of committees of experts tasked with making changes with the goal of improving the validity, reliability, and clinical utility of the diagnostic systems1, 2.

While this approach has the advantage of facilitating standardized communication among users of the classifications by ensuring the uniformity and stability of the diagnostic definitions in the time interval during which that edition of the manual is in effect, it prevents the incorporation of new scientific knowledge into the manual as it emerges, a limitation that has become especially problematic given the extended intervals between revisions that have characterized the most recent editions of the DSM and ICD (19 years and 23 years, respectively).

Advances in digital publishing that allow instantaneous dissemination of changes at minimal cost have paved the way towards the American Psychiatric Association (APA) adopting a continuous improvement model for the DSM, in which revisions are pegged to specific scientific advances. Thus, rather than waiting until the next wholesale revision to implement a clinically useful change (such as incorporating a solidly validated biomarker into the definition of a disorder), such a change could be put into effect as soon as it has been determined that it is diagnostically advantageous to do so. Moreover, implementing a continuous data‐driven approach has the added advantage of discouraging changes that are not well supported by empirical evidence. As described by Kendler in his accounting of the history of the DSM‐5 Scientific Review Committee, there is an inherent trend towards making changes built into the DSM revision process: “for workgroup members, it is a natural source of pride to ‘make a difference’, to ‘put their mark’ on the document” (3).

A new DSM web portal (www.dsm5.org) has been set up by the APA to field proposals for changes on a continuous basis. Submissions will be web‐based, with proposers required to provide supportive information in a structured format, including the reasons for the change, the magnitude of change, data documenting improvements in validity across a range of validators, evidence of reliability and clinical utility, and a consideration of current or potential deleterious consequences associated with the proposed change. It is anticipated that most submissions will come from interested persons (e.g., psychiatric researchers, individual clinicians) or organizations (e.g., psychiatric subspecialty groups, advocacy organizations, APA components) that are external to the APA committee overseeing the DSM revision process. This is in marked contrast to the prior DSM revision efforts, in which the proposals were drafted by workgroup members, who were also responsible for providing supportive literature reviews and conducting data re‐analyses.

The revision process will be overseen by a Steering Committee (analogous to the DSM Task Force) whose members have expertise in psychiatric nosology, psychiatric research, clinical psychiatry, and the DSM. Five standing Review Committees (analogous to the DSM Workgroups), which cover broad domains of psychiatric diagnosis, will work with the Steering Committee to review the proposals and draft the revisions. For example, a single review committee will cover so‐called “serious mental illnesses”, which include schizophrenic spectrum and other psychotic disorders, bipolar disorders, and neurodegenerative disorders. Final review of the proposals (which will be posted for public comment) in terms of whether criteria for approval have been met will be conducted by the Steering Committee and, if so, will be referred to the APA Board of Trustees for official approval. Once approved, each change will be publicized by the APA and digital versions of the manual will be updated to reflect the change.

Three types of proposals that require substantial empirical support have been identified, each with explicit criteria regarding the type of evidence that is expected to be submitted. Type 1 proposals involve changes to an existing diagnostic criteria set. Submitted evidence should document that the change would markedly improve the validity, reliability, or clinical utility of a criteria set, or that it would substantially reduce identified deleterious consequences associated with a criteria set. Type 2 proposals involve the addition of a new diagnostic category, subtype or specifier, and supporting evidence should document that the new category: a) meets the criteria for a mental disorder provided in the DSM‐51; b) has strong evidence of validity; c) can be applied reliably; d) has substantial clinical value (e.g., it identifies a group of patients now not receiving appropriate clinical attention); e) avoids substantial overlap with existing diagnoses; and f) has a positive benefit/harm ratio (e.g., low risk of harm due to social or forensic consequences). Type 3 proposals entail deleting an existing category or subtype/specifier, and require evidence that the proposed item to be deleted has only weak validity, minimal utility (e.g., it is rarely used in clinical practice or research) or is better conceptualized as a subtype of an existing diagnosis.

Proposals involving corrections and clarifications of existing criteria that do not require empirical support will be considered on an expedited basis by a subcommittee of the Steering Committee. Included are instances of lack of clarity or ambiguity in the meaning of the wording of criteria or text; inconsistencies or contradictions within text or criteria (e.g., diagnostic criteria in conflict with descriptive text), and errors of omission or inadvertent inclusion (e.g., omission of a disorder in the “not better explained by” list in an exclusion criterion).

In conclusion, it is hoped that the implementation of an empirically rigorous continuous improvement process for the DSM will facilitate the inclusion of scientific advances in a timelier manner than was possible using the current revision process, which should ultimately result in a more valid and clinically useful diagnostic classification.

Michael B. First
Columbia University Department of Psychiatry, New York State Psychiatric Institute,
New York, NY, USA