Background There is currently no widely accepted instrument for measuring preoperative anxiety. The objective of this study was to develop a simple visual facial anxiety scale (VFAS) for assessing acute preoperative anxiety. Methods The initial VFAS was comprised of 11 similarly styled stick-figure reflecting different types of facial expressions (Fig 1). After obtaining IRB approval, a total of 265 participant-healthcare providers (e.g., anesthesiologists, anesthesiology residents, and perioperative nurses) were recruited to participate in this study. The participants were asked to: (1) rank the 11 faces from 0-10 (0 = no anxiety, while 10 = highest anxiety) and then to (2) match one of the 11 facial expression with a numeric verbal rating scale (NVRS) (0 = no anxiety and 10 = highest level of anxiety) and a specific categorical level of anxiety, namely no anxiety, mild, mild-moderate, moderate, moderate-high or highest anxiety. Based on these data, the Spearman correlation and frequencies of the 11 faces in relation to the 11-point numerical anxiety scale and 6 categorical anxiety levels were calculated. The highest frequency of a face assigned to a level of the numerical anxiety scale resulted in a finalized order of faces corresponding to the 11-point numeric rating scale. Results The highest frequency for each of the NVRS anxiety scores were as follow: A0, A1, A2, A3, A4, A5, A7, A6, A8, A9 and A10 (Fig 2). For the six categorical anxiety levels, a total of 260 (98.1%) participants chose the face A0 as representing 'no' anxiety, 250 (94.3%) participants chose the face A0 as representing 'highest' anxiety and 147 (55.5%) participants chose the face A8 as representing 'moderate-high' anxiety. Spearman analysis showed a significant correlation between the faces A3 and A5 assigned to the mild-moderate anxiety category (r = 0.58), but A5 was ultimately chosen due to its higher frequency compared tothe frequency of A3 (30.6% vs 24.9%)(Fig 3). Similarly, the correlation of the faces A7 and A6 was significantly correlated with moderate anxiety (r = 0.87), but A7 remained because of its higher frequency (35.9% vs 22.6%). Using frequency and Spearman correlations, the final order of the faces assigned to the categories none, mild, mild-moderate, moderate, moderate -high and highest anxiety levels was A0, A1, A5, A7, A8 and A10, respectively (Fig 4). Conclusion The proposed VFAS was a valid tool for assessing the severity of acute [state] anxiety, and could be easy to administer in routine clinical practice.
