Developing functional outcome measures for unilateral neglect: a pilot study.

Stroke patients may develop personal neglect, peripersonal neglect or both. Four new measures were tested in a sample of 42 right-handed inpatients (25 male; 17 female, median age 72 years). Participants removed keys from a rack, identified grocery items, washed their face, and cleaned a tray. Prior to this, they were classified as: no neglect (15), personal neglect (8), peripersonal neglect (7), and both personal and peripersonal neglect (12). The sensitivity and specificity of each new measure was determined by agreement with the classification. Test-retest reliability was determined using weighted kappa statistics or limits of agreement. Four occupational therapists (OTs) rated videos of the face and tray measures, and software was developed to measure objectively time spent and area covered on Face and Tray. Keys and Grocery had high specificity, good reliability but poor sensitivity. For the OTs' video ratings, there was good and moderate inter-rater reliability on Tray and Face respectively for area covered, but not time spent. Intra-rater reliability was also better for area than time on Tray. However, the validity of Face and Tray themselves is currently inadequate. A longitudinal study is proposed to modify the measures, increase their sensitivity and evaluate their ability to monitor change over time.