Ilja Cornelisz1, Pim Cuijpers2,3, Tara Donker2,3, Chris van Klaveren1. 1. Amsterdam Center for Learning Analytics, Vrije Universiteit Amsterdam, Amsterdam, The Netherlands. 2. Department of Clinical, Neuro- and Developmental Psychology, Section Clinical Psychology, Vrije Universiteit Amsterdam, Amsterdam, The Netherlands. 3. Amsterdam Public Health Research Institute, Amsterdam, The Netherlands.
Abstract
BACKGROUND: The importance of randomization in clinical trials has long been acknowledged for avoiding selection bias. Yet, bias concerns re-emerge with selective attrition. This study takes a causal inference perspective in addressing distinct scenarios of missing outcome data (MCAR, MAR and MNAR). METHODS: This study adopts a causal inference perspective in providing an overview of empirical strategies to estimate the average treatment effect, improve precision of the estimator, and to test whether the underlying identifying assumptions hold. We propose to use Random Forest Lee Bounds (RFLB) to address selective attrition and to obtain more precise average treatment effect intervals. RESULTS: When assuming MCAR or MAR, the often untenable identifying assumptions with respect to causal inference can hardly be verified empirically. Instead, missing outcome data in clinical trials should be considered as potentially non-random unobserved events (i.e. MNAR). Using simulated attrition data, we show how average treatment effect intervals can be tightened considerably using RFLB, by exploiting both continuous and discrete attrition predictor variables. CONCLUSIONS: Bounding approaches should be used to acknowledge selective attrition in randomized clinical trials in acknowledging the resulting uncertainty with respect to causal inference. As such, Random Forest Lee Bounds estimates are more informative than point estimates obtained assuming MCAR or MAR.
BACKGROUND: The importance of randomization in clinical trials has long been acknowledged for avoiding selection bias. Yet, bias concerns re-emerge with selective attrition. This study takes a causal inference perspective in addressing distinct scenarios of missing outcome data (MCAR, MAR and MNAR). METHODS: This study adopts a causal inference perspective in providing an overview of empirical strategies to estimate the average treatment effect, improve precision of the estimator, and to test whether the underlying identifying assumptions hold. We propose to use Random Forest Lee Bounds (RFLB) to address selective attrition and to obtain more precise average treatment effect intervals. RESULTS: When assuming MCAR or MAR, the often untenable identifying assumptions with respect to causal inference can hardly be verified empirically. Instead, missing outcome data in clinical trials should be considered as potentially non-random unobserved events (i.e. MNAR). Using simulated attrition data, we show how average treatment effect intervals can be tightened considerably using RFLB, by exploiting both continuous and discrete attrition predictor variables. CONCLUSIONS: Bounding approaches should be used to acknowledge selective attrition in randomized clinical trials in acknowledging the resulting uncertainty with respect to causal inference. As such, Random Forest Lee Bounds estimates are more informative than point estimates obtained assuming MCAR or MAR.
Authors: Pim Cuijpers; Eva Heim; Jinane Abi Ramia; Sebastian Burchert; Kenneth Carswell; Ilja Cornelisz; Christine Knaevelsrud; Philip Noun; Chris van Klaveren; Edith Van't Hof; Edwina Zoghbi; Mark van Ommeren; Rabih El Chammay Journal: PLoS Med Date: 2022-06-23 Impact factor: 11.613
Authors: Marieke B J Toffolo; Jamie R Fehribach; Chris P B J van Klaveren; Ilja Cornelisz; Annemieke van Straten; Jean-Louis van Gelder; Tara Donker Journal: PLoS One Date: 2022-07-13 Impact factor: 3.752