Agreement between DSM-5 and DSM-IV measures of substance use disorders in a sample of adult substance users.

Abstract Background In DSM-5, definitions of substance use disorders (SUD) were changed considerably from DSM-IV, yet little is known about how well DSM-IV and DSM-5 SUD diagnoses agree among substance users. Because data from many studies are based on DSM-IV diagnostic criteria, understanding the agreement between DSM-5 and DSM-IV SUD diagnoses and reasons for discordance between these diagnoses is crucial for comparing results across studies. Measurements Prevalences and chance-corrected agreement of DSM-5 SUD and DSM-IV substance dependence were evaluated in 588 substance users in a suburban inpatient addiction program and an urban medical center, using a semi-structured interview (PRISM-5). Alcohol, tobacco, cannabis, cocaine, heroin, opioid, sedative, and stimulant use disorders were examined. Cohen’s kappa was used to assess agreement between DSM-5 and DSM-IV SUD (abuse or dependence), DSM-5 SUD and DSM-IV dependence, and DSM-5 moderate/severe SUD and DSM-IV dependence. Results Agreement between DSM-5 and DSM-IV SUD was excellent for all substances (κ = 0.84−0.99), except for cannabis and tobacco (κ = 0.75; 0.80, respectively). The most common reason for diagnostic discrepancies was a positive DSM-5 SUD diagnosis but no DSM-IV diagnosis, due to the lowered DSM-5 SUD threshold. Agreement between DSM-5 SUD and DSM-IV dependence was excellent for all substances (κ = 0.88−0.94), except for alcohol, tobacco, and cannabis (κ = 0.63−0.75). Agreement between moderate/severe DSM-5 SUD and DSM-IV dependence was excellent across all substances. Conclusion While care should be used in interpreting results of studies using different methods, studies relying on DSM-IV or DSM-5 SUD diagnostic criteria offer similar information and thus can be compared when accumulating a body of evidence.