Development and testing of a Faces Scale for the assessment of anxiety in critically ill patients

Publication Type:
Journal Article
Journal of Advanced Nursing, 2003, 41 (1), pp. 73 - 79
Issue Date:
Filename Description Size
Thumbnail2003000357.pdf637.17 kB
Adobe PDF
Full metadata record
Background. Many patients experience anxiety during treatment in an intensive care unit, but intensive care patients are often not able to respond to existing validated measures of anxiety such as the Brief Symptom Inventory. We have developed a new single item Faces Anxiety Scale made up of drawings of five faces. Aims. The aims of this study were to: (i) assess the ability of intensive care patients to respond to the Faces Anxiety Scale; and (ii) investigate whether the scale yields ordinal and interval data. Ethics. The project was approved by two Human Research Ethics Committees. Patients were included in the study if they gave informed consent. Instruments and methods. (i) Intensive care patients (n = 40) were asked to respond to the Faces Anxiety Scale, the anxiety subscale of the Brief Symptom Inventory, and a numerical analogue anxiety scale; and (ii) Hospital and University staff and students (n = 75) were asked to place the five faces in rank order. A further 100 staff members and students were asked to place each face at a point on a 60-cm triangular wedge according to the level of anxiety they thought the face showed. Results. The Faces Anxiety Scale elicited more responses from intensive care patients than the numerical analogue anxiety scale the or anxiety subscale of the Brief Symptom Inventory (36 vs. 25 vs. 17, respectively, P < 0.0001). In the testing of the order of the scale items, 93% of respondents independently placed the scale items in the order of least to most anxiety as intended in the design of the scale. In the testing for equality of intervals respondents placed the faces relatively equidistant, with no overlap of 95% confidence intervals around the mean distances. Conclusion. The Faces Anxiety Scale has minimal subject burden, elicits self report from intensive care patients more often than other simple scales, and has evidence of the interval scale properties of rank order and equality between the points on the scale.
Please use this identifier to cite or link to this item: