I believe the quality assurance of the clustering process (using the Mental Health Clustering Tool (MHCT) incorporating the Health of the Nation Outcome Scales (HoNOS)) is a complex field. Bekas & Michev ^{Reference Bekas and Michev1} have approached it from the MHCT ‘red rule’ perspective and the ICD-10 coding perspective. What the results show is that to comply with one you might potentially be in breach of the other. We face this in clinical practice; for example, bipolar affective disorder is considered to be a ‘psychotic’ condition, although as clinicians we all know there are times when patients with bipolar affective disorder are not psychotic. On such occasions, if you rate them on the MHCT they might score ‘0’ and then if you cluster them in a psychotic cluster you breach the ‘red rule’ and if you do not, you breach the ICD-10 coding expectations.

There would be another layer of complexity added when all the ‘care packages’ are agreed between the commissioners and providers. I am sure there would be interest to ensure that the care provided or offered reflects the package agreed. The elusive ‘gold standard’ that the authors allude to, if developed, cannot be one-dimensional. It needs to clarify, when there are conflicting standards, that the clinician has to adhere to the one which takes priority and therefore in my opinion should be hierarchical. In fact, the authors of the MHCT might consider dropping the ‘red rules’ which might have outlived their usefulness when there are agreed care packages in place. Until then the MHCT and the clustering process remain imperfect tools that clinicians have to navigate to communicate with the commissioners.