Education And Debate

Controversies in Management: Costs outweigh benefits

BMJ 1994; 309 doi: (Published 03 December 1994) Cite this as: BMJ 1994;309:1499
  1. S P Allison, consultant physiciana
  1. a University Hospital, Nottingham NG7 2UH


    In Britain 100 000 strokes occur annually, and patients with stroke occupy a fifth of all medical and a quarter of all long term beds as well as consuming 5% of the NHS budget. At present half of all acute strokes are managed by general practitioners at home. Last year 891 patients were admitted to the University Hospital, Nottingham, with a diagnosis of stroke; 210 had computed tomography at a cost of £80 per head. If computed tomography of all patients with stroke became mandatory the first effect would be to double the admission rate to include all those currently managed at home. The additional cost of scanning these and the rest of the 891 patients admitted would be roughly £125000 a year without the costs of admitting 800 more medical patients. The medicolegal consequences of making computed tomography mandatory in all stroke patients are easily imagined. Computed tomography for stroke must be proved to be extremely effective for such a cost to be justified. I believe that the case has yet to be made.

    Reasons for investigation

    Medical investigations can be justified on several counts. In research they can be used to validate a new test against clinical symptoms and signs or against another standard test; to improve understanding of the disease process; or to define patient groups for intervention studies and prospective trials. In clinical practice they are used to define the diagnosis more clearly, allowing more effective and appropriate treatment, and to assist in prognosis.

    Computed tomography is currently used to define patient groups for trials of anticoagulant and fibrinolytic drugs. It is used to select patients with thrombotic strokes for treatment and exclude those with haemorrhagic strokes in whom the treatment may be harmful Ebrahim compared computed tomography and aspirin in the acute phase with aspirin at one month without computed tomography for managing all strokes in Britain in one year (table).1 The cost and outcome issues are plain.

    Comparison of two strategies for management of all United Kingdom strokes in a year (n = 100000)1

    View this table:

    Accuracy of clinical diagnosis

    Apart from cases where intervention is pointless, clinicians may be faced with three questions. The first is, has this patient had a stroke or is there another condition? Reports differ concerning the relative value of clinical diagnosis and computed tomography. The sensitivity of the history and examination can be as high as 95%, but specificity varies from 66% to 97%.2 3 In a recent report from Finland 1191 consecutive patients with a putative diagnosis of stroke were assessed clinically before computed tomography. An expert clinician had a false positive rate of 3.1% and false negative rate of 4.7% in diagnosing stroke. Brown et al assessed 821 consecutive admissions to a stroke ward and found that computed tomography did not affect the rate of misdiagnosis. They emphasised the importance of clinical experience in increasing diagnostic accuracy. These reports seem to argue the case for investing in acute stroke units rather than for routine scanning, particularly as such units have been shown to reduce mortality from stroke by 20-30% without increasing morbidity.6

    The second question concerns the differentiation between thrombosis and haemorrhage when early antithrombotic treatment is contemplated. Thrombosis and haemorrhage are difficult to differentiate clinically; there is a sensitivity of 68% and a specificity of 67%.2 3 But does it matter in terms of treatment? Trials are in progress to determine the role of early treatment, but in the light of current information, delaying aspirin treatment for one month is a reasonable and cost effective strategy. In any case, the computed tomographic appearances of haemorrhage and thrombosis are difficult to distinguish after three weeks. The international stroke trial pilot study of aspirin versus low dose heparin within 48 hours of onset suggested that if patients with clinically obvious haemorrhage are excluded—that is, those with severe headache at onset, vomiting, or coma—early and long term mortality was similar in patients who did and did not have computed tomography before beginning treatment.

    Commentary: computed tomography is needed for accurate diagnosis

    Some time ago I was amused to hear a neurologist talking about the enthusiasm of his department to get closer to the “front line” of medicine and emphasising how keen they all were to be contacted when patients with acute neurological problems were admitted to hospital.

    However, he did add the rider that two conditions were not in his opinion within the realms of acute neurology: one was epilepsy and the other was stroke. This encapsulates the problem since, with a few notable exceptions, no one has regarded stroke as their special interest and the condition is so common that doctors from almost every medical specialty routinely manage patients with stroke. When nothing much could be done other than waiting and seeing it perhaps did not matter about drawing a distinction between cerebral haemorrhage and infarction. But now that aspirin has been shown to be effective in the secondary prevention of cerebral infarction and with many drugs being investigated for treating acute stroke it matters very much. For the foreseeable future most patients with stroke will be managed by doctors who do not specialise in the area. Since even those who are expert cannot reliably distinguish different types of stroke on clinical grounds the case for routine computed tomography seems very strong, and 90% of British physicians apparently feel the same. -PETER C RUBIN, professor of therapeutics, University of Nottingham

    The third question concerns the differentiation of haemorrhagic from embolic stroke in patients with atrial fibrillation. Primary prevention trials in atrial fibrillation have shown a reduction in stroke or death rate of 67% with warfarin and 32% with aspirin.7 The role of early treatment in secondary prevention remains to be established. Pessin et al described 12 patients who were already receiving warfarin but who suffered embolic infarcts with secondary haemorrhage.8 Anticoagulants were continued without ill effects. The case for excluding haemorrhage by computed tomography in such patients is therefore not established.


    1. 1.
    2. 2.
    3. 3.
    4. 4.
    5. 5.
    6. 6.
    7. 7.
    8. 8.
    View Abstract

    Sign in

    Log in through your institution

    Free trial

    Register for a free trial to to receive unlimited access to all content on for 14 days.
    Sign up for a free trial