Screening for ability to complete a questionnaire: a preliminary evaluation of the AbilityQ and ShoulderQ for assessing shoulder pain in stroke patients

OBJECTIVES: To assess the repeatability of the AbilityQ--a screening tool to assess technical ability to complete a questionnaire--in patients with complex disabilities following stroke. To evaluate the validity of the AbilityQ in predicting repeatability and consistency of response to verbal and visual analogue scale questions in the ShoulderQ in those with shoulder pain. DESIGN: In a cross-sectional study, the AbilityQ and 'ShoulderQ' were administered on two consecutive days by a single investigator. SETTING: Six regional rehabilitation centres in the UK managing patients with severe complex neurological disability. SUBJECTS: Forty-nine patients with anterior or posterior circulation strokes: 13 (63%) had cognitive deficits and 28 (57%) had communicative deficits. RESULTS: Thirty-one (63.3%) patients required some degree of help in completing the AbilityQ. Percentage agreement for individual questions between day 1 and day 2 ranged from 55 to 88% with kappa values from 0.07 to 0.79. Repeatability of ShoulderQ responses was 36-72% for the verbal questions (kappa values 0.16-0.56). For visual analogue scale (VAS) questions, agreement +/- 1 on a 10-point scale was 36-59% with intraclass correlation coefficients 0.50-0.60 (p or = 3/4) and VAS scores (3/3) on the AbilityQ had positive predictive values for good repeatability in the ShoulderQ of 80% and 58% respectively. Mismatched AbilityQ responses to verbal and VAS questions showed a positive predictive value of 92% for similar inconsistency in the ShoulderQ and a negative predictive value of 90%. CONCLUSIONS: In this preliminary evaluation the AbilityQ and the ShoulderQ each demonstrated a moderate level of repeatability in a group of patients expected to have difficulty in completing questionnaires. The AbilityQ appears to have some clinical usefulness in identifying those able to respond to the ShoulderQ and, where responses to verbal and VAS questions conflict, may offer some guidance as to which are most likely to be accurate in that individual. Further development is underway.