Background/Aims
The Oxford Cognitive Screen (OCS) is a stroke-specific cognitive screening tool, increasingly being used in clinical practice. Our aim was to determine the sensitivity and specificity of the Australian OCS (OCS-AU) to detect post-stroke cognitive impairment as determined by a gold standard measure of cognition. We also compared the sensitivity and specificity of the OCS-AU with a traditionally used screening measure - the Montreal Cognitive Assessment (MoCA).
Methods
Stroke survivors (n=53) within 12 weeks of stroke completed the OCS-AU, MoCA and a comprehensive cognitive assessment serving as the gold standard. Impairment was defined as failing any OCS-AU domain, scoring below normative MoCA cutoffs, or failing at least two domains on the comprehensive assessment. Analyses were conducted using R.
Results
Domain-specific OCS-AU sensitivity ranged from 0.17 (Memory and Executive) to 0.71 (Spatial Attention) and specificity ranged from 0.81 (Spatial Attention) to 0.96 (Praxis). To general cognitive impairment, the OCS-AU showed sensitivity of 0.64 and specificity of 0.56, compared to the MoCA's higher sensitivity (0.82) but lower specificity (0.44).
Conclusion
Our results suggest that the OCS-AU demonstrates acceptable specificity across domains but variable sensitivity, particularly underperforming in Memory and Executive domains. The MoCA offers higher sensitivity, but lower specificity. Thus, the OCS-AU appears to provide more precise indications of cognitive deficits, with attention to stroke-specific impairments, but may not capture all post-stroke cognitive changes. Research and clinical implications of these findings will be discussed, including potential for assessment protocols to optimise detection and subsequent management of cognitive impairment in stroke survivors.