Introduction Clinical experience suggests that patients admitted to general medical wards will rarely undergo a neurological examination if the presenting complaint is perceived to be primarily non–neurological. With evolution of the presentation, any later formal neurological assessment not benefit from a record of the temporal evolution of the neurological signs. We present a rapid neurological screen which may ultimately be used in patients presenting without neurological symptoms and may improve recording of the neurological examination. In this study we performed initial validation for subjects with neurological symptoms presenting to our neurology unit.

Methods Two protocols were devised to accommodate both ambulant and non–ambulant patients (Figure 1). A third year medical student (FR) performed the screening examination blind to the formal examination. These results were compared to results from the formal neurological examination performed in either our acute neurology ward, day investigation unit, or the outpatient clinic. Results from formal examination by a neurology registrar or consultant were determined from retrospective review of case notes. For inpatients only, we stipulated that the inter–examination interval should be <24h. The results from the screening examination were classified as normal or abnormal for the following domains: overall examination, eyes, face, arms, and legs. Statistical analyses for sensitivity and specificity assumed that the formal neurological examination was the gold standard.

Results Twelve non–ambulant and eighteen ambulant patients were identified. For non–ambulant patients, the screening examination detected any abnormality with sensitivity and specificity of 100%. Agreement was excellent (Scott's Pi=1; CI=0.42 to 1). Results for the subgroups of the exam were as follows: Eyes; sensitivity=100%, specificity=87.5%: face; sensitivity=50%, specificity=93%: arms; sensitivity=67%, specificity=100%: legs; sensitivity=83%, specificity=83%.

For ambulant patients the screening examination detected any abnormality with a sensitivity of 69% and a specificity of 60%. Agreement was fair (Scott's Pi=0.25; CI=–0.17 to 0.63). Results for the subgroups of the exam were as follows; Eyes–sensitivity=75%, specificity=78%; face–sensitivity=100%, specificity=100%; arms–sensitivity=100%, specificity=100%; legs–sensitivity=100%, specificity=77%.

Conclusion A screening examination by a medical student showed excellent agreement with formal examination in those with significant neurological impairment (non–ambulant patients). In ambulant patients it performed less well, but with a sensitivity of 69 still represents and advance on ‘examination not done’. We argue that further studies to improve sensitivity of this proposed screening examination are indicated to promote adoption in a general medical setting for patients without neurological symptoms.