Intermediate care had the characteristics of much new national policy when it was launched; it lacked definition, had no evidence for its effectiveness, and was aimed at solving a problem that it was unlikely to impact on—in this case overcrowding in hospitals. It is therefore good to see new evidence on the effectiveness of a service that fits under the umbrella term “intermediate care” compared with care in a district general hospital department for the care of elderly people, as reported by Green et al.1

The relation between older people and hospitals is a complex one. A suspicion among several eminent commentators is that ageism is still endemic in the NHS, keeping older people out of the bright sparkling general hospitals because they overstay their welcome.2 A further suspicion is that the National Service Framework for older people is complicit in this, with its emphasis on an extra 5000 intermediate care beds, with no real definition or evidence for such an approach or for the number chosen (it would seem to be a heavy dose of digit preference). This suspicion is added to by the suggestion in the National Service Framework that acute admission rates for people aged over 75 should rise less than 2% per annum, again with no evidence to suggest that this percentage will match clinical need or that likely geographical variations should be taken into account. A fixed percentage seems especially odd in a country that has a successfully ageing population. Most geriatricians would regard 75 as quite young.

Work on methods of diverting older people into intermediate care straight from home has not proved successful so far using a non-randomised cohort approach.3 A systematic review of a nurse led intermediate care facility for post-acute rehabilitation, however, suggested that this is effective, although the authors urge caution as this may be due to an increased length of stay in the nurse unit. They also make the point that the safety of the approach has not been fully proved.4 An economic evaluation using participants from a randomised trial has shown that such a nurse led unit was more expensive than a traditional approach.5

Green et al show that another approach—a single community hospital, in Bradford—gave better results in one variable, of many, compared with a hospital based unit for elderly people. The organisation of the study had lots of difficulties, but such health services research is always beset by such issues. In an area where resources are scarce, people are loath to allow their favourite patient groups to be entered into the lottery of a randomised trial. Many of us have had similar experiences. So the study group was unusual by the time randomisation took place. This, and the small proportion of positive outcomes, compared with the number measured, makes the work difficult to generalise elsewhere.

Overall, the paper does a little more to clothe the Emperor of intermediate care, but he is still not really fit to be seen out in public.