Effect of Missed Items on the Reliability of the Kessler Foundation Neglect Assessment Process.

ABSTRACT

OBJECTIVE: To determine the maximum permissible number of missed items on the 10-item Catherine Bergego Scale administered after the Kessler Foundation Neglect Assessment Process (KF-NAP). Secondary objectives were to determine the frequency, characteristics, and most commonly cited reasons reported for missed items. DESIGN: Retrospective diagnostic accuracy study. SETTING: Sixteen inpatient rehabilitation facilities in the United States. PARTICIPANTS: A consecutive clinical sample of 4256 patients (N=4256) with stroke or other neurologic deficits who were assessed for spatial neglect with the KF-NAP. INTERVENTIONS: Not applicable. MAIN OUTCOME MEASURES: Catherine Bergego Scale via KF-NAP. RESULTS: The majority (69.7%) of patients had at least 1 missed item on their KF-NAP. Among those with missed items, it was most common to have 2 missed items (51.4%), and few had more than 3 missed items (11.3%). The most commonly missed items were Collisions (37.2%), Cleaning After Meals (36.1%), Meals (34.0%), and Navigation (19.7%). The most commonly reported reasons for missed items included time constraints, cognitive or communication deficits, and behavior or refusal of the therapy session. These reasons were reported for nearly all item types. Item-specific reasons were also commonly reported, such as a lack of a needed resource for task completion or low functional status of the patient. Prorated scoring of measures with up to 3 missed items maintained an acceptable level of concordance with complete measures (Lin's Concordance Correlation Coefficient=0.96, 95% CI, 0.9478-0.9626) for the combination of 3 missed items with lowest concordance. CONCLUSIONS: Clinicians should make every effort to capture all items on the KF-NAP. However, missed items occur in the majority of cases because of patient factors and barriers inherent to the inpatient hospital setting. When missed items are necessary, clinicians can confidently interpret a prorated score when 7 or more items are scored.