Lawrie and colleagues urge us not to reject the current categorical classification system prematurely. ^{Reference Lawrie, Hall, McIntosh, Owens and Johnstone1} I wish to add to the argument that a categorical system is more likely to be internationally useful. More than 80% of mental illness occurs in middle- and lowincome countries. ² Much of the world's mental illness is seen in overstretched clinics, by practitioners who treat up to 100 patients a day and often have had no training in psychiatry since medical or nursing school. Administering the rating scales necessary for a dimensional system may be possible in high-income countries, but is difficult or impossible elsewhere. The categorical classification system can be used quickly by someone with relatively little training. There is also the problem of translating and validating the rating scales into hundreds of languages. Most published research currently uses the same categorical system, which means that it is useful to doctors all over the world. If the research were to refer only to a dimensional system, then it would not be useful in settings where it is impossible to administer the rating scales. The categorical system gives more people access to evidence-based treatment than any dimensional system would. A classification system that is going to be used all over the world needs to be simple and robust across healthcare systems, languages and cultures, and this is just as important as how closely it resembles the truth.