TY - JOUR
T1 - Measurement model choice influenced randomized controlled trial results
T2 - Journal of Clinical Epidemiology
AU - Gorter, R.
AU - Fox, J.P.
AU - Apeldoorn, A.
AU - Twisk, J.
N1 - ISI Document Delivery No.: EE9SY Times Cited: 0 Cited Reference Count: 33 Gorter, Rosalie Fox, Jean-Paul Apeldoorn, Adri Twisk, Jos Gorter, Rosalie/0000-0002-2451-9953 EMGO+ Institute for Health and Care Research [WC2009-010] This research was funded by the EMGO+ Institute for Health and Care Research, grant number: WC2009-010. 0 1 ELSEVIER SCIENCE INC NEW YORK J CLIN EPIDEMIOL
PY - 2016
Y1 - 2016
N2 - Objective: In randomized controlled trials (RCTs), outcome variables are of-ten patient-reported outcomes measured with questionnaires. Ideally, all available item information is used for score construction, which requires an item response theory (JRT) measurement model. However, in practice, the classical test theory measurement model (sum scores) is mostly used, and differences between response patterns leading to the same sum score are ignored. The enhanced differentiation between scores with IRT enables more precise estimation of individual trajectories over time and group effects. The objective of this study was to show the advantages of using IRT scores instead of sum scores when analyzing RCTs. Study Design and Setting: Two studies are presented, a real-life RCT, and a simulation study. Both IRT and sum scores are used to measure the construct and are subsequently used as outcomes for effect calculation. Results: The bias in RCT results is conditional on the measurement model that was used to construct the scores. A bias in estimated trend of around one standard deviation was found when sum scores were used, where IRT showed negligible bias. Conclusion: Accurate statistical inferences are made from an RCT study when using IRT to estimate construct measurements. The use of sum scores leads to incorrect RCT results. (C) 2016 Elsevier Inc. All rights reserved.
AB - Objective: In randomized controlled trials (RCTs), outcome variables are of-ten patient-reported outcomes measured with questionnaires. Ideally, all available item information is used for score construction, which requires an item response theory (JRT) measurement model. However, in practice, the classical test theory measurement model (sum scores) is mostly used, and differences between response patterns leading to the same sum score are ignored. The enhanced differentiation between scores with IRT enables more precise estimation of individual trajectories over time and group effects. The objective of this study was to show the advantages of using IRT scores instead of sum scores when analyzing RCTs. Study Design and Setting: Two studies are presented, a real-life RCT, and a simulation study. Both IRT and sum scores are used to measure the construct and are subsequently used as outcomes for effect calculation. Results: The bias in RCT results is conditional on the measurement model that was used to construct the scores. A bias in estimated trend of around one standard deviation was found when sum scores were used, where IRT showed negligible bias. Conclusion: Accurate statistical inferences are made from an RCT study when using IRT to estimate construct measurements. The use of sum scores leads to incorrect RCT results. (C) 2016 Elsevier Inc. All rights reserved.
U2 - https://doi.org/10.1016/j.jclinepi.2016.06.011
DO - https://doi.org/10.1016/j.jclinepi.2016.06.011
M3 - Article
C2 - 27394673
SN - 1878-5921
VL - 79
SP - 140
EP - 149
JO - J Clin Epidemiol
JF - J Clin Epidemiol
ER -