Sampson et al ^{Reference Sampson, Blanchard, Jones, Tookman and King1} addressed the importance of additional resources attached to the medical care of the growing population of elderly people with dementia or cognitive impairment. Although restricted to the elderly undergoing medical acute admissions, the study highlights the underdiagnosis of dementia, its poor short-term outcome and high mortality rates during admission.

If we translate these findings to a clinical setting, the problem of undiagnosed dementia in medical milieu appears to be much wider. In Newcastle alone, out of nearly 17 000 annual non-elective admissions of over-65-year-olds, only 4.3% are referred to the liaison team for older people's mental health, suggesting that a large proportion of elderly (up to 38%, using Sampson et al ^{Reference Sampson, Blanchard, Jones, Tookman and King1} data) that are not referred to specialist mental health liaison teams, may well have undiagnosed and untreated mental health problems, including dementia and dementia-related health problems.

Hospital-based liaison teams for older people are seeing a number of elderly people with memory problems in various medical settings, and patients with dementia in an acute medical setting may represent only a small portion of all elderly admitted on other medical and surgical wards. Thus, our liaison team (providing hospital mental healthcare for a region including an estimated 41 000 elderly, n = 730–1200 referrals/annum) on average gets 26% of referrals from acute medical wards, with a similar proportion (25%) from care of the elderly wards, and/or rehabilitation wards (16%); an additional 33% comes from various surgical and other specialised medical wards (e.g. dermatology, infectious disease). Of these, 40% are already known to old age psychiatry services. The majority of performed assessments are related to dementia (59%), level of care (25%) and behavioural problems as a result of known memory problems (15%). Importantly, 19% of medically ill patients are obtaining their first diagnosis of dementia via our service, a finding similar to that described by Sampson et al. ^{Reference Sampson, Blanchard, Jones, Tookman and King1} An additional 17% of assessments identify various social issues closely related to the presence of cognitive impairment.

The high rate of elderly people with dementia on medical wards should not come as a surprise, since on average people with dementia (irrespective of the type of dementia) have three or more physical illnesses. ^{Reference Schubert, Boustani, Callahan, Perkins, Carney and Fox2} Furthermore, severity of dementia independently predicts hospitalisation. ^{Reference Albert, Costa, Merchant, Small, Jenders and Stern3} However, the impact of comorbidity on survival appears to be dependent not on severity of dementia, ^{Reference Doraiswamy, Leon, Cummings, Marin and Neumann4} but on the number of medical diseases, which in turn contribute to more rapid dementia decline. ^{Reference Boksay, Boksay, Reisberg, Torossian and Krishnamurthy5}

The high mortality rates described for people with dementia ^{Reference Sampson, Blanchard, Jones, Tookman and King1} also confirm previous findings of the presence of concomitant psychiatric and somatic disorders resulting in poor outcome. ^{Reference Van Dijk, Dippel, Van der Meulen and Habbema6} Furthermore, although the burden of chronic medical conditions was similar in patients with and without dementia, the severity of acute illness (assessed with APACHE II) was higher in individuals with dementia/cognitive impairment. ^{Reference Sampson, Blanchard, Jones, Tookman and King1} This finding is consistent with the reported underdiagnosis of medical problems in patients with dementia which can preclude their early detection and treatment. ^{Reference Zekry, Herrmann, Grandjean, Meynet, Michel and Gold7} Interestingly, Sampson et al included inviduals with delirium episodes in the analysis if these had resolved within 4 days. This may explain the reported high death rates, which are very similar to those reported for delirium in the elderly. ^{Reference Mukaetova-Ladinska and McKeith8} In support of the presence of underlying delirium goes the reported finding of higher burden of acute physiological disturbances in individuals with dementia/cognitive impairment. ^{Reference Sampson, Blanchard, Jones, Tookman and King1}

Attention was drawn in a previous study to difficulties assessing delirium with the Mini-Mental State Examination (MMSE) in acutely medically ill elderly. ^{Reference Yates, Stanley, Cerejeira, Jay and Mukaetova-Ladinska9} In these patients, the most frequent symptoms reported by the Confusion Assessment Method (CAM) are those of memory impairment (55%) and disorientation (37%), whereas the characteristic delirium symptom of altered level of consciousness is reported in only 21%. Similarly, 24% of elderly with an acute medical illness cannot be assessed by the MMSE. Thus, although definition of delirium based on the CAM (DSM–III) or DSM–IV criteria may be adequately suited for delirium assessment in medically ill elderly with cognitive impairment, there still seems to be a lack of standardised instruments specifically developed to be used in this population.

Interestingly, 30% of the participants came from sheltered, residential and/or nursing homes, and this group in particular had a higher mortality rate. ^{Reference Sampson, Blanchard, Jones, Tookman and King1} This raises an additional issue about the healthcare that is provided within these venues and the accessibility to adequate services that in the light of the findings may well need to be provided in situ.

Lastly, we agree with the conclusion that additional mental health liaison services will need to be further developed. Moreover, to cope with the rising numbers of people with dementia, the educational role of such teams is likely to become increasingly important. Although tailored to the learning needs of each group, the focus should be on increasing awareness and understanding of dementia. ¹⁰ The key challenge, which will determine the success of any educational endeavour and ultimately whether outcomes for the older person with dementia are improved, is to ensure that knowledge is successfully transferred into improved practice behaviour. ^{Reference Teodorczuk, Welfare, Corbett and Mukaetova-Ladinska11}