Although there is a high prevalence of delirium and cognitive impairment among hospitalised older adults, short, reliable cognitive measures are rarely used to monitor cognition and potentially alert healthcare professionals to early changes that might signal delirium. We evaluated the reliability, responsiveness, and feasibility of logical memory (LM), immediate verbal recall of a short story, compared to brief tests of attention as a bedside “cognitive vital sign” (CVS). Trained nursing staff performed twice-daily cognitive assessments on 84 clinically stable inpatients in two geriatric units over 3–5 consecutive days using LM and short tests of attention and orientation including months of the year backwards. Scores were compared to those of an expert rater. Inter-rater reliability was excellent with correlation coefficients for LM increasing from r = 0.87 on day 1 to r = 0.97 by day 4 (p < 0.0001). A diurnal fluctuation of two points from a total of 30 was deemed acceptable in clinically stable patients. LM scores were statistically similar (p = 0.98) with repeated testing (suggesting no learning effect). All nurses reported that LM was feasible to score routinely. LM is a reliable measure of cognition showing diurnal variation but minimal learning effects. Further study is required to define the properties of an ideal CVS test, though LM may satisfy these.