Statistics from Altmetric.com
Paul et al 1 have assessed the clinical outcomes, early hospital admission rates and hospital care costs in clinic referred and hospital referred minor stroke patients in a prospective population based study. They did not find significant differences in the 30 day admission rate in clinic patients compared with the 30 day readmission rate after discharge in hospital treated patients (16/237 vs 9/150).
The 30 day recurrent stroke risk in patients with minor ischaemic stroke was also similar in those discharged from clinic compared with hospital patients (3.8% vs 5.3%; p=0.61). The recurrent stroke risk remained similar in patients treated at the clinic compared with those managed in hospital at 1 year (11.5% vs 12.4%; p=0.86) and at the 5 year follow-up (21.2% vs 20.5%; p=0.84). The cost of outpatient management of minor stroke was substantially lower than that of inpatient care, saving £4800 per patient.
The authors conclude that “rapid assessment and treatment of minor stroke can be achieved in the outpatient clinic and the complication rate of outpatient management is relatively low”, and the rates of secondary prevention interventions were at least as good in clinic managed as in hospital managed patients.
However, I feel it necessary to warn readers against the automatic application of the Oxfordshire experience because the study had several limitations: first, it was a non-randomised comparison of different management strategies; and second, the noticeable results were achieved through an extraordinarily well organised system of care with an optimal physician/individual ratio (63 general practitioners to 91 105 registered individuals). The study was carried out with rigorous ascertainment and follow-up by experienced physicians with all of the necessary instrumentation at their disposal; this scenario does not exist in many European countries.
With regard to the non-random nature of subject allocation, there were a number of factors associated with hospital versus clinic referral, including case severity, patient's decision where to self-present, living alone, etc. Studies where this is the case typically suffer from the phenomenon referred to in epidemiology as confounding by indication—that is, comparison of two alternatives based on subjects’ outcomes is complicated by the fact that probably most, even though arguably not all, subjects were directed towards the alternative that was reasonably expected to produce better outcomes for them. Hence one cannot safely assume that had a subject ended up in the opposite alternative, their expected outcome would have been the average outcome observed there. For this reason, it would be difficult to substantiate proposals that preference should be given to outpatient management. A proposal for a randomised comparison of the two settings seems easier to argue for, and the authors do so. This entails the assumption that we have no reason to consider the two alternatives to be different, to the best of our knowledge, and exposing volunteers to chance in their choice of how/where they will be treated does not violate the principle of first do no harm.
The risk of worsening and stroke related complications is high during the first days after a minor stroke.2 The risk can multiply if more than one coexisting diseases are present (unstable hypertension, high blood glucose, paroxysmal or permanent arrhythmias, early post-stroke seizures).3 Early haemorrhagic transformation occurs in about 9% of patients, especially in cardioembolic stroke or in patients with high blood glucose.4
Minor stroke (a focal neurological deficit lasting more than 24 h and with a National Institutes of Health Stroke Scale score of ≤3 at baseline), as a diagnostic entity, can cover a variety of pathophysiological backgrounds that represent a wide range of potential complications. A case of a lacunar infarct with mild hemiparesis, no arrhythmia, and normal or near normal metabolic parameters would be at one extreme of this spectrum and a case of an extended cardiogenic posterior cerebral artery infarct with hemianopia, unstable blood pressure and high blood glucose, at the opposite extreme. The second patient has a significantly higher risk of complications (eg, haemorrhagic transformation, early post-stroke seizure) and might significantly benefit from hospital management. Therefore, only a thorough combined evaluation of clinical symptoms, imaging signs and concomitant diseases can help us select the optimal method of management once a subject has been allocated to either alternative in a randomised study. The requisite highly trained community of general practitioners and other health care professionals may exist in Oxfordshire but not necessarily in many other settings.5 So, while I agree with the authors that “A randomised comparison of either treatment setting would be justified”, they do envisage that future study to be of a particularly challenging nature.
Competing interests None.
Contributors LC is the sole author of this paper.
Provenance and peer review Commissioned; not externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.