Updating DSM-5 to improve diagnostic validity and reliability and clinical utility

The DSM is used by clinicians and researchers to diagnose and classify mental disorders. It was first published by the American Psychiatric Association in 1952 and has been updated four times since the original edition. The most recent edition, DSM-5, was published in 2013 following a 14-year revision process by members of the DSM-5 Task Force and its 13 Work Groups.

The new process enables ongoing revision to DSM-5 to reflect new knowledge

Revising the manual in its entirety at variable intervals has the advantage of promoting stability and consistency in clinical practice and research, said Paul Appelbaum, Professor of Psychiatry, Columbia University, NY, past president of the American Psychiatric Association, and chair of the DSM-5 Steering Committee.

However, such revisions are unable to respond to advances in knowledge and understanding, which occur at a different pace for different disorders. A process has therefore been put in place to enable experts in the field to propose changes so they can be incorporated incrementally in a timely fashion.

Five types of proposed change are anticipated

It is anticipated that the revisions proposed will aim to change an existing diagnostic criteria set, add a new diagnostic category or specifier, delete an existing diagnostic category or specifier, or correct and clarify the current content.

The revisions proposed will aim to change an existing diagnostic criteria set, add a new diagnostic category or specifier, delete an existing diagnostic category or specifier, or correct and clarify the current content

Professor Appelbaum explained that the five types of expected revision are classified as follows:

• Type 1: Changes to an existing diagnostic criteria set

• Type 1a to address the validity of an existing diagnostic criteria set
• Type 1b to address the reliability of a diagnostic criteria set, without reducing the validity
• Type 1c to address the clinical utility of a diagnostic criteria set, without reducing the validity or reliability
• Type 1d to address deleterious consequences associated with a diagnostic criteria set, without reducing validity

• Type 2: Addition of a new diagnostic category or specifier/subtype
• Type 3: Deletion of an existing diagnostic category or specifier/subtype
• Type 4: Corrections and clarifications to improve understanding and application of an ambiguous diagnostic criterion, specifier, or text
• Type 5: Changes to the text to improve clarity

Evidence to support proposed changes

Vice chairs of the DSM-5 Steering Committee, Kenneth Kendler, Professor of Psychiatry, Stanford, CA, and Ellen Leibenluft, Senior Investigator at the National Institute of Mental Health Intramural Research Program, explained that the evidence required to support proposed changes depends on the type of change.

Substantial evidence is required for Type 1, 2 and 3 revisions. Evidence supporting Types 4 and 5 revisions can be less substantial than for Types 1 to 3,

Substantial evidence is required for Type 1 changes to prove that the revision would markedly:

• improve the validity of an existing diagnostic criteria set, the reliability of a diagnostic criteria set without reducing validity, or the clinical utility of a diagnostic criteria set without reducing validity or reliability
• reduce deleterious consequences associated with a diagnostic criteria set, without reducing validity

Substantial evidence is also required for:

• Type 2 changes to ensure the new diagnostic category or specifier meets the criteria for a mental disorder, is valid and reliable, is clinically valuable, does not overlap with existing diagnoses and specifiers/subtypes
• Type 3 changes to delete an existing diagnostic category or specifier/subtype to prove it has weak validity and is of minimal clinical utility

Evidence supporting Types 4 and 5 revisions can be less substantial than for Types 1 to 3, said Dr Leibenluft. She highlighted that these revisions need clear common sense evidence explaining the advantages and disadvantages of the proposed change.

How to submit proposed changes

Anyone can initiate the process for a proposed change to DSM-5 via https://www.psychiatry.org/psychiatrists/practice/dsm/submit-proposals.

The ten broad classes of validator may be further classified as antecedent, concurrent or predictive

Professor Kendler explained that the evidence for Type 1a and 1b revisions is best submitted in a highly detailed, tabulated format providing an overview of the evidence including patients investigated, methods used, key results, year of study and an assessment graded from 1 to 5 on the quality the validators.

He described the following ten broad classes of validating data:

• Antecedent validators: familial, socio-cultural-demographic, environmental, previous psychiatric history
• Concurrent validators: cognitive, biological markers, comorbidity, functional impact
• Predictive validators: diagnostic stability, course of illness, response to treatment

Review process for proposed changes

The Steering Committee will review all completed proposals to decide whether the proposal is substantially likely to meet the criteria for approval. If so it is referred to the relevant Review Committee.

Each of the five Review Committees has expertise in one broad area of psychopathology – neurodevelopmental, serious mental illness (e.g., psychosis, neurodegenerative disease), internalizing mental illness (e.g., depression), externalizing mental illness, and body system disorders.

The proposal is then reviewed again by the Steering Committee before being referred to the APA board for approval.

Future of DSM-5

DSM-5 is a work in progress and is likely to look very different in 20 years’ time

During the discussion at the workshop, the Steering Committee was praised for making the revision process more transparent and open.

One audience member wondered whether the review process should be opened-up to all interested parties.

Professor Appelbaum replied that for now the Steering Committee believe the review process should be restricted in the first instance to Review Committees as currently outlined to avoid excessive noise and to maintain clarity and accuracy.

However, he emphasized that DSM-5 is a work in progress and that the revision process is likely to look very different in 20 years’ time.