The Effect of Diagnostic Criteria on Dementia Prevalence - A Population-Based Study From Gothenburg, Sweden

To examine how the use of different diagnostic criteria (Diagnostic and Statistical Manual of Mental Disorders third revised, fourth, and fifth editions [DSM-III-R, DSM-IV, and DSM-5], and the 10th and 11th editions of the International Classification of Diseases [ICD-10 and ICD-11] influences the reported prevalence of dementia.

Two cross-sectional population-based studies of systematically selected 85-year-olds in Gothenburg, Sweden, (N = 774), were examined in comprehensive health examinations including comprehensive neurocognitive examinations. Five algorithms based on the diagnostic criteria in the DSM-III-R, DSM-IV, DSM-5, ICD-10, and ICD-11 were created, including 105 different variables that were operationalized in different ways to match the criteria of each classification system.

ICD-11 yielded the highest prevalence of dementia (36.4%), followed by DSM-5 (32.9%), DSM-IV (30.7%), the clinical consensus DSM-III-R diagnosis (26.7%), DSM-III-R (21.4%), and ICD-10 (20.5%). The agreement between the DSM-5 and the ICD-11 was κ = 0.9. All other kappa values ranged between 0.6 and 0.9.

The choice of diagnostic criteria has a large effect on the estimated prevalence of dementia. We found that the recent editions, the DSM-5 and ICD-11, gave a higher prevalence of dementia than older editions. We also show that the attempts to harmonize DSM and ICD have in part been successful, however, there are still differences between the systems.