Sir: We read with interest of the protocol-based approach for prescribing donezepil described by Jani and Prettyman (Psychiatric Bulletin, May 2000, 25, 174-177).

However, the criteria suggested as guidelines for making the diagnosis of Alzheimer's disease and tests used to determine the therapeutic outcome are not appropriate for assessment of the population with learning disability, in whom, particularly in those with Down's syndrome, there is a high prevalence of dementia. There are, however, scales such as the Dementia Questionnaire for Mentally Retarded Persons and the Dementia Scale for Down's syndrome, which can give useful measurements. It is therefore unfortunate that the guidelines on prescribing donezepil and similar treatments recently produced by the National Institute for Clinical Excellence lean so heavily on the Mini Mental State Examination, which is not a validated instrument for this purpose in those with learning disability, who will score poorly whether they have dementia or not. They also seemingly limit the initiation of such treatments to old age psychiatrists, neurologists and care of the elderly physicians, many of whom do not deal with those with learning disabilities. This policy would seem to clash with the recent White Paper Valuing People, which states that all health services should be available to those with learning disabilities with a significant role for learning disability psychiatrists such as ourselves, who know this patient group best, and should surely also be authorised to initiate these treatments.