Translation and psychometric validation of a Danish version of the medication-related quality of life scale

Background Generic and disease specific health-related quality of life scales have been found to be non-responsive to changes in medications in polypharmacy patients. The Taiwanese medication-related quality of life (MRQoL) scale aims to measure the effect of medication use on patients' quality of life. Objective To evaluate the psychometric properties of the Danish translation of MRQoL in a population of patients with polypharmacy. Setting Polypharmacy patients waiting for services at a community pharmacy or hospital in Denmark. Method The original MRQoL included 14 items. It was forward-translated into Danish and backward-translated into Chinese according to a modified translation protocol proposed by Sousa and Rojjanasrirrat et al. The translation was pre-tested, adjusted, and administered to polypharmacy patients. The factor structure was examined using exploratory factor analysis (EFA) and confirmatory factor analysis (CFA). Internal consistency reliability was evaluated, and criterion validity assessed using the beliefs about medicines questionnaire (BMQ) and SF-12v2. Known-group validity was carried out on age, number of medicines and setting. Main outcome measure Validity of the Danish version of the MRQoL-scale. Results 164 patients completed the questionnaire. EFA of all 14 items resulted in a two-factor structure, accounting for 72.8% of the total variance. The two factors were named "Energy/Concentration" (7 items) and "Feelings/Social" (7 items). Items correlating over 0.80 were removed leaving 11 items (Model 1). This model was further reduced to 8 items (Model 2) based on Cronbach's alpha. CFA confirmed the two-factor structure of both models. Model 2 fitted data without having to define covariations between error terms. Both factors showed high internal consistency reliability (Cronbachs' alpha 0.901-0.932). Ceiling effects were detected for both factors. Criterion validity was demonstrated via its significant correlations with SF-12vs2 subscales (Spearman's rho 0.340-0.353) and BMQ Concern (Spearman's rho - 0.451 to - 0.347). There was a statistically significant difference in relation to total scores of the MRQoL for age and number of drugs taken, indicating known-group validity. Conclusion The Danish translation of the MRQoL instrument showed measurement properties indicating a well-defined two-factor structure with high internal reliability, concurrent criterion validity, and known group validity. However, challenges remain with ceiling effects and efforts should be put into further development of the instrument.