Contributions
Abstract: P351
Type: Poster
Abstract Category: Clinical aspects of MS - Clinical assessment tools
Background: The Treatment Satisfaction Questionnaire for Medication (TSQM) has 14 items and generates 4 domain scores: Effectiveness (3 items); Side Effects (5 items); Convenience (3 items); Global Satisfaction (4 items). Although not developed specifically for use in MS, the TSQM demonstrated good measurement performance in an MS clinical trial setting, the TENERE phase 3 trial of teriflunomide (NCT00883337) in patients with relapsing forms of MS (RMS). TSQM measurement performance has yet to be evaluated in routine RMS clinical practice.
Objective(s): To evaluate the measurement performance of the TSQM (version 1.4), using traditional psychometric methods and Teri-PRO (NCT01895335) study data.
Methods: Teri-PRO included patients with RMS (n=1000) starting treatment with teriflunomide 7 mg or 14 mg as per local labelling. TSQM data were collected at baseline and Week 48/end of treatment (EOT); data for Week 48/EOT were used. Aspects of 5 measurement properties were examined: data completeness (item-level missing data); scaling assumptions (item means, variances and item-total correlations corrected for overlap, exploratory factor analysis [EFA]); scale-to-sample targeting (score distributions: floor=lowest/ceiling=highest satisfaction); internal consistency reliability (Cronbach"s α, homogeneity coefficients); construct validity (within-scale/with other variables).
Results: Item-level missing data were low (< 1%). Domain scores were computable for ≥88% of patients. Scaling assumptions were satisfied (similar item means and variances; item-domain correlations 0.63-0.90; EFA supported item groupings). Scale-to-sample targeting was adequate overall: domain scores exceeded scale midpoints, floor scores noted in ≤3% of patients. However, ceiling effects were 57% for Convenience and 58% for Side Effects. Internal consistency reliability was high (Cronbach"s α ≥0.86, homogeneity coefficients ≥0.67) and standard errors of measurement were small (5.49-7.02). Construct validity was supported by the characteristics of between-domain correlations (0.17-0.69). Scores were not biased by sample characteristics (correlations -0.23-0.013).
Conclusions: This analysis of TSQM performance, using Teri-PRO “real-world” data, complements our prior analysis in TENERE, and gives similar findings: good overall measurement properties with the potential to improve scale-to-sample targeting. Together, our analyses support the TSQM as a useful measure of treatment satisfaction in patients with RMS.
Disclosure: Study supported by Sanofi Genzyme.
JH: Honoraria, consulting fees (Acorda, Biogen Idec, Critical Path Institute, LORA group, MAPI Research Institute, Sanofi Genzyme); license fee payments or royalty payments (Plymouth University receives fees for the use of rating scales developed as part of author"s research); research support (Biogen Idec, Novartis, Merck Serono).
KT: Employee of Sanofi Genzyme.
SH: Employee of Sanofi Genzyme.
SC: Nothing to disclose.
PR: Employee of Sanofi Genzyme.
Abstract: P351
Type: Poster
Abstract Category: Clinical aspects of MS - Clinical assessment tools
Background: The Treatment Satisfaction Questionnaire for Medication (TSQM) has 14 items and generates 4 domain scores: Effectiveness (3 items); Side Effects (5 items); Convenience (3 items); Global Satisfaction (4 items). Although not developed specifically for use in MS, the TSQM demonstrated good measurement performance in an MS clinical trial setting, the TENERE phase 3 trial of teriflunomide (NCT00883337) in patients with relapsing forms of MS (RMS). TSQM measurement performance has yet to be evaluated in routine RMS clinical practice.
Objective(s): To evaluate the measurement performance of the TSQM (version 1.4), using traditional psychometric methods and Teri-PRO (NCT01895335) study data.
Methods: Teri-PRO included patients with RMS (n=1000) starting treatment with teriflunomide 7 mg or 14 mg as per local labelling. TSQM data were collected at baseline and Week 48/end of treatment (EOT); data for Week 48/EOT were used. Aspects of 5 measurement properties were examined: data completeness (item-level missing data); scaling assumptions (item means, variances and item-total correlations corrected for overlap, exploratory factor analysis [EFA]); scale-to-sample targeting (score distributions: floor=lowest/ceiling=highest satisfaction); internal consistency reliability (Cronbach"s α, homogeneity coefficients); construct validity (within-scale/with other variables).
Results: Item-level missing data were low (< 1%). Domain scores were computable for ≥88% of patients. Scaling assumptions were satisfied (similar item means and variances; item-domain correlations 0.63-0.90; EFA supported item groupings). Scale-to-sample targeting was adequate overall: domain scores exceeded scale midpoints, floor scores noted in ≤3% of patients. However, ceiling effects were 57% for Convenience and 58% for Side Effects. Internal consistency reliability was high (Cronbach"s α ≥0.86, homogeneity coefficients ≥0.67) and standard errors of measurement were small (5.49-7.02). Construct validity was supported by the characteristics of between-domain correlations (0.17-0.69). Scores were not biased by sample characteristics (correlations -0.23-0.013).
Conclusions: This analysis of TSQM performance, using Teri-PRO “real-world” data, complements our prior analysis in TENERE, and gives similar findings: good overall measurement properties with the potential to improve scale-to-sample targeting. Together, our analyses support the TSQM as a useful measure of treatment satisfaction in patients with RMS.
Disclosure: Study supported by Sanofi Genzyme.
JH: Honoraria, consulting fees (Acorda, Biogen Idec, Critical Path Institute, LORA group, MAPI Research Institute, Sanofi Genzyme); license fee payments or royalty payments (Plymouth University receives fees for the use of rating scales developed as part of author"s research); research support (Biogen Idec, Novartis, Merck Serono).
KT: Employee of Sanofi Genzyme.
SH: Employee of Sanofi Genzyme.
SC: Nothing to disclose.
PR: Employee of Sanofi Genzyme.