Evolution of Substance Use Disorder Diagnosis Across DSM and ICD Systems

DSM Evolution (DSM-3 through DSM-5-TR)

DSM-3 and DSM-4 maintained a two-category system separating "substance abuse" from "substance dependence" as distinct diagnostic entities 1, 2
The dependence syndrome in these earlier editions was based on substantial treatment research and clinical validation 1
This dichotomous approach presumed that abuse represented milder pathology while dependence indicated more severe disease 1

DSM-5 eliminated the abuse/dependence distinction entirely, collapsing both into a single dimensional diagnosis called "substance use disorder" 3, 1, 2
The new system uses 11 diagnostic criteria, requiring at least 2 criteria met within 12 months for diagnosis 3, 2
Severity grading is purely criterion-based: mild (2-3 criteria), moderate (4-5 criteria), severe (6+ criteria) 3, 2
"Craving" was added as a new 11th criterion based on Item Response Theory (IRT) analyses 3, 1
IRT studies demonstrated that abuse and dependence items could be ordered on a single latent dimension, with some "abuse" items actually indexing greater severity than certain "dependence" items 1
The legal problems criterion from DSM-4 was removed 2
Gambling disorder was moved into the substance-related disorders chapter as the first behavioral addiction 1, 2

DSM-5-TR maintains the same 11-criterion dimensional substance use disorder diagnosis with identical severity grading 3
The two-criterion threshold captures a broader population than traditional dependence diagnoses 3

The collapse of abuse/dependence was justified by IRT analyses showing all criteria form a single unidimensional continuum 3, 1
However, opponents argued this abandoned the validated dependence syndrome and represented a major shift in diagnostic philosophy by considering individual criterion validity rather than incremental validity of the criterion set 1
The craving criterion was added despite being redundant with other criteria because of its conceptual importance and potential as a biomarker 1

ICD-10 maintained separate categories for mental and behavioral disorders related to substance use 4
The system used 11 disorder groupings with traditional categorical classification 4
ICD-10 distinguished between harmful use and dependence as separate entities 5

ICD-11 retains the traditional separation between "harmful substance use" and "substance dependence" as distinct diagnostic categories, fundamentally diverging from DSM-5's unified approach 3, 5
The ICD-11 chapter expanded from 11 to 21 disorder groupings 4
Substance dependence remains the "master diagnosis" in ICD-11 5
ICD-11 introduced dimensional assessments for some disorders while maintaining categorical structure 4
A new "harm to others" feature was added to harmful use criteria, though research shows very low endorsement rates and questionable diagnostic validity 6
Gaming disorder was included alongside gambling disorder in the addictive disorders section 5
The separate childhood/adolescence disorder grouping was eliminated, with disorders redistributed to emphasize developmental continuity across the lifespan 4

The revision prioritized global applicability, scientific validity, and clinical utility through the largest participative process in classification history 4
Empirical support exists for the coherence of substance dependence for alcohol, cannabis, and prescribed opioids, justifying its retention 5
Field studies demonstrated higher reliability and clinical utility of ICD-11 compared to ICD-10 4
Changes were deliberately modest to maintain categorical classification for clinical utility while allowing dimensional assessments in specialized settings 4

Feature	DSM-5-TR	ICD-11
Core Diagnostic Structure	Single dimensional "substance use disorder" [3]	Separate "harmful substance use" and "substance dependence" [3,5]
Number of Criteria	11 criteria [3,2]	Varies by category (dependence vs harmful use) [5]
Diagnostic Threshold	Minimum 2 of 11 criteria in 12 months [3,2]	Category-specific thresholds [5]
Severity Grading	Mild (2-3), Moderate (4-5), Severe (6+) based purely on criterion count [3,2]	Not criterion-count based; dependence is inherently more severe [5]
Craving Criterion	Included as standalone criterion [3,1]	Not specified as separate criterion [5]
Abuse/Dependence Distinction	Eliminated; merged into single disorder [3,1,2]	Retained as separate entities [3,5]
"Harm to Others" Feature	Not included	Added to harmful use (low validity) [6]
Behavioral Addictions	Gambling disorder included; Internet gaming disorder in Section III (further study) [1,2,5]	Both gambling and gaming disorders included [5]
Dimensional Elements	Severity only (mild/moderate/severe) [3]	Optional dimensional assessments for select disorders [4]
Clinical Philosophy	Broader, more inclusive diagnostic net [3]	Narrower, dependence-focused approach [5]
Concordance Between Systems	Excellent concordance with ICD-10 and DSM-4 (κ ≥ 0.9); poor concordance between DSM-5 mild SUD and ICD-11 harmful use [6]	Excellent concordance with ICD-10 and DSM-4 (κ ≥ 0.9); poor concordance with DSM-5 for harmful use category [6]

The most significant divergence is that DSM-5-TR uses a single dimensional disorder while ICD-11 maintains the dependence syndrome as the master diagnosis 3, 5
This creates poor concordance specifically between DSM-5 mild substance use disorder and ICD-11 harmful use 6

The DSM-5-TR two-criterion threshold captures a broader population than ICD-11's dependence-focused approach 3
ICD-11's retention of dependence preserves alignment with decades of treatment research 1, 5
The systems will continue to drift apart, complicating international research and clinical communication 7

Both systems now include gambling disorder in addictive disorders 1, 2, 5
Both show excellent concordance with their respective predecessors (ICD-10, DSM-4) 6
Both maintain primarily categorical approaches despite incorporating some dimensional elements 4, 3