OBJECTIVE: To assess whether the measurement error, and recall bias can reach magnitudes comparable to minimally important difference (MID) in symptoms scores used in chronic venous disease-specific quality of life QOL instruments, such as Specific Quality of Life & Outcomes Response-Venous (SQOR-V) questionnaire. METHODS: Prospective non-randomized study of 150 patients with primary chronic venous disease. SQOR-V questionnaire was administered prior to clinical visit (in 32 patients twice), and 1 month post-treatment. Patients were asked to recall their symptoms 12 months later. Measurement error (SEM) was calculated from repeated baseline measurements. MID was derived from change in the symptom score (SS) part of the SQOR-V questionnaire at 1 month. Recall bias was calculated from 12-month recall data. RESULTS: SEM was 1.91, the recall bias was -3.16 (95% CI from -4.08 to -2.24), and the MID was 3.7. In the treatment group the recall bias was negative (recall more severe symptoms than they actually were at the baseline), patients in the observation group had positive recall bias. SS change moderately correlated with transition rating index (Spearman rank-order correlation 0.526, P < 0.0001), and strongly correlated with the baseline value (Pearson r = 0.84). CONCLUSION: The measurement error of patient-reported symptoms scores included in disease-specific QOL instruments is small, but recall bias is close in magnitude to MID.
OBJECTIVE: To assess whether the measurement error, and recall bias can reach magnitudes comparable to minimally important difference (MID) in symptoms scores used in chronic venous disease-specific quality of life QOL instruments, such as Specific Quality of Life & Outcomes Response-Venous (SQOR-V) questionnaire. METHODS: Prospective non-randomized study of 150 patients with primary chronic venous disease. SQOR-V questionnaire was administered prior to clinical visit (in 32 patients twice), and 1 month post-treatment. Patients were asked to recall their symptoms 12 months later. Measurement error (SEM) was calculated from repeated baseline measurements. MID was derived from change in the symptom score (SS) part of the SQOR-V questionnaire at 1 month. Recall bias was calculated from 12-month recall data. RESULTS: SEM was 1.91, the recall bias was -3.16 (95% CI from -4.08 to -2.24), and the MID was 3.7. In the treatment group the recall bias was negative (recall more severe symptoms than they actually were at the baseline), patients in the observation group had positive recall bias. SS change moderately correlated with transition rating index (Spearman rank-order correlation 0.526, P < 0.0001), and strongly correlated with the baseline value (Pearson r = 0.84). CONCLUSION: The measurement error of patient-reported symptoms scores included in disease-specific QOL instruments is small, but recall bias is close in magnitude to MID.
Authors: Bo Eklöf; Robert B Rutherford; John J Bergan; Patrick H Carpentier; Peter Gloviczki; Robert L Kistner; Mark H Meissner; Gregory L Moneta; Kenneth Myers; Frank T Padberg; Michel Perrin; C Vaughan Ruckley; Philip Coleridge Smith; Thomas W Wakefield Journal: J Vasc Surg Date: 2004-12 Impact factor: 4.268
Authors: F Lurie; D Creton; B Eklof; L S Kabnick; R L Kistner; O Pichot; C Sessa; S Schuller-Petrovic Journal: Eur J Vasc Endovasc Surg Date: 2005-01 Impact factor: 7.069
Authors: Caroline B Terwee; Leo D Roorda; Joost Dekker; Sita M Bierma-Zeinstra; George Peat; Kelvin P Jordan; Peter Croft; Henrica C W de Vet Journal: J Clin Epidemiol Date: 2009-11-18 Impact factor: 6.437