Objective: To investigate the reliability and validity of the Italian version of ECOS-16 (Assessment of health related quality of life in osteoporosis) in comparison to other questionnaires in postmenopausal women with osteoporosis.
Methods: A cross-sectional multicentre study was carried out among postmenopausal women with osteoporosis who were attending primary care centres and hospital outpatient clinics. The patient group included 234 females (mean age 69 years, range 48-89) who presented vertebral fractures due to osteoporosis. The control group consisted of 244 asymptomatic osteoporotic subjects matched for age with the patient group. The psychometric properties of the questionnaires were evaluated in terms of feasibility, validity (construct validity and discriminant validity) and internal consistency. Test-retest reliability was analysed for 196 outpatients who reported that their general health status due to osteoporosis had not changed after one week. In all patients the ECOS-16, the SF-36 (Medical Outcomes Study Short Form-36), EUROQoL (EQ-5D), mini-OQLQ (mini-Osteoporosis Quality of Life Questionnaire), and RMDQ (Roland-Morris Disability Questionnaire) were administered, and all clinical variables and sociodemographic variables were taken into account. Construct and discriminant validity were assessed by Spearman's correlations, the Wilcoxon rank sum test, the Kruskal Wallis test and by receiver operating characteristic (ROC) curves. Internal consistency was evaluated using Cronbach's alpha and the test-retest reliability was evaluated by intra-class correlation coefficients (ICCs).
Results: 96.9% of the patients answered all items of the ECOS-16 questionnaires. The mean administration time was 10 minutes. Factor analysis yielded two factors that accounted for 88.4% of the explained variance in the ECOS-16 questionnaire. The first factor was the ECOS-16 Physical Component Score (PCS) (45.9% of the explained variance) and the second factor was the ECOS-16 Mental Component Score (MCS) (42.4% of the explained variance). The inter-item correlation between the two factors was 0.48. Significant correlations were found between the scores of similar domains or subscales of the ECOS-16 and SF-36, EQ-5D and mini-OQLQ, supporting the concept of convergent construct validity. The total ECOS-16 score progressively increased with the number of prevelant vertebral fractures (p<0.001) and the effect of the first fracture was already statistically significant (p<0.01). On ROC curve analysis the total ECOS-16 score showed the highest performance among the different questionnaires in discriminating between patients with vertebral fractures and controls with no fractures. In the reliability study, internal consistency within the domain of ECOS-16 was generally good, with Chronbach's alpha values ranging from 0.81 to 0.89. Test-retest reliability was 0.87 for the total ECOS-16 score.
Conclusion: The Italian version of the ECOS-16 questionnaire was demonstrated to have good psychometric properties and could offer a useful tool in research and routine clinical practice to evaluate HRQoL in post-menopausal women with osteoporosis. A full validation of the psychometric properties will require data on its sensitivity to change.