We examined the Treatment Satisfaction Questionnaire for Medication (TSQM), a generic 14-item, 4-domain (Effectiveness, Side Effects, Convenience, Global Satisfaction) scale using ‘real-world’ data from people with relapsing forms of MS (RMS).Methods
Teri-PRO (NCT01895335) included teriflunomide-treated RMS patients (n=1000). A traditional psychometric evaluation examined: data completeness (item-level missing data); scaling assumptions (item means, variances, item-total correlations corrected for overlap, exploratory factor analysis [EFA]); scale-to-sample targeting (score distributions, floor/ceiling effects); internal consistency (Cronbach’s α, homogeneity coefficients); and aspects of construct validity.Results
Item-level missing data were low (<1%). Scaling assumptions were satisfied (similar item means/variances; item-domain correlations 0.63–0.90; EFA supported item groupings). Scale-to-sample targeting was adequate for scale evaluation; however, domain mean scores (66.3–90.3) notably exceeded scale midpoints (50), and ceiling effects for Convenience and Side Effects were high (57%, 58%). Internal consistency reliability was high (Cronbach’s α ≥0.86, homogeneity coefficients≥0.67). Standard errors of measurement were 5.49–7.02. Between-domain correlations (0.17–0.69) supported construct validity.Conclusions
TSQM performance in Teri-PRO mirrors that in TENERE (NCT00883337) and supports the TSQM as a useful measure for RMS. However, high ceiling effects underestimate differences in treatment satisfaction. Better instruments, developed using more sophisticated methods, are needed.Conclusions
Study supported by Sanofi Genzyme.