Methods: Back-to-back translation was used to produce a bilingual version of the questionnaire. Hand drawings were used to replace photographs from the original questionnaire. Face validity and content validity were assessed, and construct validity was determined by comparing responses from informal caregivers, medical students, and primary care doctors. Finally, the internal consistencies of the subscales were determined.
Results: Pretesting showed that the translated version was sufficiently easy to understand. Internal consistency for the positioning subscale (28 items, Cronbach's α = 0.70) and feeding subscale (15 items, Cronbach's α = 0.70) was good. Mean scores for the positioning subscale for caregivers (mean: 17.1 ± 3.9), medical students (mean: 18.9 ± 3.1), and doctors (mean 21.5 ± 2.2) were significantly different (F = 5.28, P ' = 0.011). Mean scores for the feeding subscale for caregivers (mean 13.1 ± 2.5), medical students (mean 16.1 ± 1.9), and doctors (mean 16.1 ± 2.4) also differed significantly (F = 6.217, P = 0.006).
Conclusions: CKQ-My has good internal consistency and construct validity for the subscales measuring stroke caregivers' knowledge about positioning and feeding of stroke patients. It has potential as an assessment of effectiveness of caregiver training and for future studies on long-term stroke outcomes in Malaysia.