Kaleidoscope.

Among the criticisms of diagnostic systems is ‘diagnostic inflation’: with time criteria relax, the number of conditions expands and everything is pathologised. It is a common refrain, with the release of DSM-5 attracting particular notoriety. But, is it actually true? Fabiano & Haslam reviewed 123 studies in which a research sample was concurrently diagnosed using two consecutive DSM editions. Meta-analysis determined the average of 476 risk ratios was 1.0: namely, neither diagnostic inflation nor deflation from DSM-III to DSM-5. The overall stringency has been stable over time. This infers that, in general, any reported changes in population prevalences of illnesses are not because diagnostic systems altered ‘capture’ but are either the result of genuine change in numbers or their better detection. However, that is not to say all conditions were the same, with the altered criteria resulting in increases and decreases in prevalence of certain disorders across editions. It might not surprise you that attention-deficit hyperactivity disorder, autism spectrum disorder, eating disorders and substance use disorders showed inflation with time. One imagines that for these, there will be considerable debate as to whether this is just better recognition and detection. Nevertheless, the authors conclude that the notion of ‘concept creep’ across diagnostic systems has been overstated, and is not supported by the evidence.