Psychometric properties and validation of the revised Chinese Medication Literacy Scale for Hypertensive Patients (C-MLSHP-R)

Abstract

Background: Medication literacy is one of the key indicators that can affect the self-management of medications and medication safety. This study aimed to revise the Chinese Medication Literacy Scale for hypertensive patients (C-MLSHP) and test the reliability and validity of the revised scale. Methods: We revised the C-MLSHP by several methods, i.e., focus group discussion, expert consultation, patient interview, and pilot study, based on the established evaluation index system of medication literacy for hypertensive patients. Then, a formal survey using the revised Chinese Medication Literacy Scale for hypertensive patients (C-MLSHP-R) was carried out on hypertensive patients from hospitals and community healthcare centers in Changsha city of China to test its reliability and validity. The reliability was evaluated with Cronbach's α coefficient, split-half reliability, and test–retest reliability. The validity was evaluated with content validity, construct validity, convergent validity, discriminant validity, and criterion-related validity. Results: The C-MLSHP-R contained 18 items within four domains, i.e., the knowledge domain included four items, the attitude domain had three items, the skill domain involved seven items, and the practice domain included four items. A total of 339 hypertensive patients participated in the formal survey. The results showed that the Cronbach's α coefficient of C-MLSHP-R was 0.802, and for each domain ranged from 0.639 to 0.815. The split-half reliability coefficient of C-MLSHP-R was 0.709, and for each domain ranged from 0.648 to 0.792. The test–retest reliability coefficient of C-MLSHP-R was 0.851, and for each domain ranged from 0.655 to 0.857. The I-CVI of each item ranged from 0.833 to 1.000, the S-CVI/Ave of C-MLSHP-R was 0.981, the S-CVI/UA was 0.889, and for each domain ranged from 0.958 to 1.000. Confirmatory factor analysis results showed that the model fitted well. The convergent validity of C-MLSHP-R was acceptable, and the discriminant validity was good. The criterion coefficient between C-MLSHP-R and C-MLSHP was 0.797, and for each domain ranged from 0.609 to 0.755. Conclusion: Compared with C-MLSHP, the C-MLSHP-R with 18 items was much shorter for measuring, and had decreased reliability within the acceptable range and better validity, which was more appropriate and time-saving to assess the medication literacy level for hypertensive patients scientifically and conveniently.