Bobbie Jacobson's article ('Public Domain', pages 22-23, 9 November) hit several nails on the head. She proposes a model for managing the introduction of clinical governance by concentrating on current national service framework initiatives, and justifies it by considering the prevention paradox of clinical epidemiology.

Such a model is very important, because throwing the cloak of clinical governance too widely is so likely to dilute effort that another plausible initiative will become ineffective and discredited. Concentrating on group or cohort outcomes is surely the way forward.

But she may have more success on epidemiological concepts with GPs than consultants, since that would entail such a shift of treatment paradigm that great resistance may be expected. Specialists' training and experience is designed to support an implicit gold standard of individual consultation and management in clinical practice. We are quintessential discriminators and experts in dealing with clinical uncertainties. With such a strong emphasis on the particular it is difficult to find a middle position, where a 'horses for courses' philosophy allows cohort management when appropriate and individual consultation if not.

The NHS as a whole is actively exploring this with NHS Direct, and whether 'one size will fit all' as far as the refined type of service provided by clinicians is concerned remains to be seen. I expect intractable resistance from clinicians who will feel threatened fundamentally in the skills and roles they have painstakingly acquired for individual management, and which may not survive a more pluralistic healthcare environment. By analogy, banks have generally abandoned the high street manager as ill-suited to addressing current customer concerns. Branch advisers are thought of as GPs, dealing with the 'relationship' component of customer service and referring to a tier of specialists. This offers a balance of individual and categorical service.

In recently assessing our scientific work on algorithms a reviewer asserted that 'the computer will never replace clinical intelligence', as if that was the proposal.

A middle position that uses individual or group approaches depending on the problem may be very slow to develop when clinicians' training and status is based largely on detecting the exceptional, even when that is not the task.

Es Will Consultant nephrologist St James' Hospital Leeds Secretary UK Renal Registry