Validation of the Oxfordshire Community Stroke Project syndrome diagnosis derived from a standard symptom list in acute stroke

Background: The Oxfordshire Community Stroke Project (OCSP) classification allows distinction of stroke subtypes with different prognosis. OCSP classification inferred from clinical signs filled out on patient entry forms has been used to facilitate subgroup analysis in clinical trials. However, such procedure has not been validated against clinical diagnosis. In preparation for an acute stroke trial, we set out to perform such a validation. Methods: An OCSP syndrome diagnosis of 194 acute stroke patients in four hospitals was made within 24 h using a standard list with neurological signs, to be filled out by a stroke physician or neurological resident on duty. This was compared with OCSP diagnosis within 2 days of stroke onset by a (blinded) stroke neurologist ('gold standard'). Results: The proportion of the OCSP syndromes was quite similar between standard list and clinical judgement. Sensitivity, specificity, positive and negative predictive values were respectively: LACS: 0.76, 0.88, 0.72, 0.90; TACS: 0.63, 0.93, 0.62, 0.88; PACS: 0.62, 0.76, 0.63, 0.75; POCS: 0.50, 0.98, 0.60, 0.97. Kappa for agreement was 0.63 (LACS), 0.37 (PACS), 0.50 (TACS). Neuro-imaging falsified stroke subtype diagnosis in 40 cases (20.6%) diagnosed using the standard list, and 42 (21.6%) diagnosed by stroke neurologists. Conclusion: A standard list-derived stroke syndrome diagnosis may be used as a clinical test to make an OCSP syndrome diagnosis in acute stroke. The use of such list in acute stroke trials may facilitate uniformity in early stroke subtype diagnosis. However, to increase such uniformity, ancillary methods such as acute MRI should be evaluated. (C) 2004 Elsevier B.V. All rights reserved.