scholarly journals COVID-19-induced hyperinflammation, immunosuppression, recovery and survival: how causal inference may help draw robust conclusions

RMD Open ◽  
2021 ◽  
Vol 7 (1) ◽  
pp. e001638
Author(s):  
Robert B M Landewé ◽  
Sofia Ramiro ◽  
Rémy L M Mostard
Keyword(s):  
Supportive Care ◽  
Causal Inference ◽  
Mediation Analysis ◽  
Immunosuppressive Treatment ◽  
Experimental Treatment ◽  
Number Needed To Treat ◽  
Mediated Effects ◽  
Period 2 ◽  
Non Linear ◽  
Quasi Experimental

BackgroundThe CHIC study (COVID-19 High-intensity Immunosuppression in Cytokine storm syndrome) is a quasi-experimental treatment study exploring immunosuppressive treatment versus supportive treatment only in patients with COVID-19 with life-threatening hyperinflammation. Causal inference provides a means of investigating causality in non-randomised experiments. Here we report 14-day improvement as well as 30-day and 90-day mortality.Patients and methodsThe first 86 patients (period 1) received optimal supportive care only; the second 86 patients (period 2) received methylprednisolone and (if necessary) tocilizumab, in addition to optimal supportive care. The main outcomes were 14-day clinical improvement and 30-day and 90-day survival. An 80% decline in C reactive protein (CRP) was recorded on or before day 13 (CRP >100 mg/L was an inclusion criterion). Non-linear mediation analysis was performed to decompose CRP-mediated effects of immunosuppression (defined as natural indirect effects) and non-CRP-mediated effects attributable to natural prognostic differences between periods (defined as natural direct effects).ResultsThe natural direct (non-CRP-mediated) effects for period 2 versus period 1 showed an OR of 1.38 (38% better) for 14-day improvement and an OR of 1.16 (16% better) for 30-day and 90-day survival. The natural indirect (CRP-mediated) effects for period 2 showed an OR of 2.27 (127% better) for 14-day improvement, an OR of 1.60 (60% better) for 30-day survival and an OR of 1.49 (49% better) for 90-day survival. The number needed to treat was 5 for 14-day improvement, 9 for survival on day 30, and 10 for survival on day 90.ConclusionCausal inference with non-linear mediation analysis further substantiates the claim that a brief but intensive treatment with immunosuppressants in patients with COVID-19 and systemic hyperinflammation adds to rapid recovery and saves lives. Causal inference is an alternative to conventional trial analysis, when randomised controlled trials are considered unethical, unfeasible or impracticable.

scholarly journals Improving causal inference of mediation analysis with multiple mediators in randomized experiments

2021 ◽  
Author(s):  
Wen Wei Loh ◽  
Dongning Ren
Keyword(s):  
Causal Inference ◽  
Mediation Analysis ◽  
Causal Structure ◽  
Psychology Research ◽  
Direct Effects ◽  
Randomized Experiments ◽  
Causal Inferences ◽  
Mediated Effects ◽  
Causal Assumptions ◽  
Intermediate Variables

Mediation analysis is an essential tool for investigating how a treatment causally affects an outcome via intermediate variables. However, violations of the (often implicit) causal assumptions can severely threaten the validity of causal inferences of mediation analysis. Psychologists have recently started to raise such concerns, but the discussions have been limited to mediation analysis with a single mediator. In this article, we examine the causal assumptions when there are multiple possible mediators. We pay particular attention to the practice of exploring mediated effects along various paths linking several mediators. Substantive conclusions using such methods are predicated on stringent assumptions about the underlying causal structure that can be indefensible in practice. Therefore, we recommend that researchers shift focus to mediator-specific indirect effects using a recently proposed framework of interventional (in)direct effects. A vital benefit of this approach is that valid causal inference of mediation analysis with multiple mediators does not necessitate correctly assuming the underlying causal structure among the mediators. Finally, we provide a practical guide with suggestions to improve the research practice of mediation analysis at each study stage. We hope this article will encourage explication, justification, and reflection of the causal assumptions underpinning mediation analysis to improve the validity of causal inferences in psychology research.

scholarly journals Recommendations for improving causal inference of mediation analysis

2021 ◽  
Author(s):  
Wen Wei Loh ◽  
Dongning Ren
Keyword(s):  
Causal Inference ◽  
Mediation Analysis ◽  
Indirect Effects ◽  
Causal Structure ◽  
Sensitivity Analyses ◽  
Causal Inferences ◽  
Mediated Effects ◽  
Causal Assumptions ◽  
Pre Treatment ◽  
Indispensable Tool

Mediation analysis is an indispensable tool for investigating how a treatment causally affects an outcome via intermediate variables. Recently, there have been increased concerns about the validity of causal inferences in conclusions drawn using mediation analysis. However, the discussions are limited to a single mediator, and importantly, there is a lack of guidelines on substantiating causal inferences. In this article, we first provide a thorough examination of the causal assumptions underpinning mediation analysis. We pay particular attention to the practice of exploring mediated effects along various paths linking several mediators and the stringent -- yet often overlooked -- assumptions that predicate valid inference. To mitigate the risk of invalid inference, we introduce an alternative approach focusing on mediator-specific indirect effects. An appealing feature of this approach is that valid causal inference of mediation analysis with multiple mediators does not necessitate assuming a (correct) causal structure among the mediators. Finally, we provide a practical guide to improve the research practice of mediation analysis. We clarify when mediation analysis is (in)appropriate; when appropriate, we recommend that researchers preregister (i) justifications of the asserted causal relations in their mediation analysis, (ii) pre-treatment confounders, and (iii) the either path- or mediator-specific indirect effects to be investigated. Confounding adjustment when estimating the preregistered indirect effects and sensitivity analyses for unmeasured confounding can fortify causal inferences in the conclusions. We hope this article will encourage explication, justification, and reflection of the causal assumptions underpinning mediation analysis to improve the validity of causal inferences in psychology research.

scholarly journals Mendelian randomisation for mediation analysis: current methods and challenges for implementation

2021 ◽  
Author(s):  
Alice R. Carter ◽  
Eleanor Sanderson ◽  
Gemma Hammerton ◽  
Rebecca C. Richmond ◽  
George Davey Smith ◽  
...  
Keyword(s):  
Measurement Error ◽  
Causal Inference ◽  
Mediation Analysis ◽  
Instrumental Variable ◽  
Real Data ◽  
Mendelian Randomisation ◽  
Differential Measurement ◽  
Individual Level ◽  
Differential Measurement Error ◽  
Summary Data

AbstractMediation analysis seeks to explain the pathway(s) through which an exposure affects an outcome. Traditional, non-instrumental variable methods for mediation analysis experience a number of methodological difficulties, including bias due to confounding between an exposure, mediator and outcome and measurement error. Mendelian randomisation (MR) can be used to improve causal inference for mediation analysis. We describe two approaches that can be used for estimating mediation analysis with MR: multivariable MR (MVMR) and two-step MR. We outline the approaches and provide code to demonstrate how they can be used in mediation analysis. We review issues that can affect analyses, including confounding, measurement error, weak instrument bias, interactions between exposures and mediators and analysis of multiple mediators. Description of the methods is supplemented by simulated and real data examples. Although MR relies on large sample sizes and strong assumptions, such as having strong instruments and no horizontally pleiotropic pathways, our simulations demonstrate that these methods are unaffected by confounders of the exposure or mediator and the outcome and non-differential measurement error of the exposure or mediator. Both MVMR and two-step MR can be implemented in both individual-level MR and summary data MR. MR mediation methods require different assumptions to be made, compared with non-instrumental variable mediation methods. Where these assumptions are more plausible, MR can be used to improve causal inference in mediation analysis.

scholarly journals Non-linear Causal Inference Using Gaussianity Measures

2019 ◽  
pp. 257-299
Author(s):  
Daniel Hernández-Lobato ◽  
Pablo Morales-Mombiela ◽  
David Lopez-Paz ◽  
Alberto Suárez
Keyword(s):  
Causal Inference ◽  
Non Linear

scholarly journals Mendelian randomisation for mediation analysis: current methods and challenges for implementation

2019 ◽  
Author(s):  
Alice R Carter ◽  
Eleanor Sanderson ◽  
Gemma Hammerton ◽  
Rebecca C Richmond ◽  
George Davey Smith ◽  
...  
Keyword(s):  
Measurement Error ◽  
Causal Inference ◽  
Mediation Analysis ◽  
Real Data ◽  
Mendelian Randomisation ◽  
Reverse Causality ◽  
Differential Measurement ◽  
Individual Level ◽  
Differential Measurement Error ◽  
Summary Data

AbstractMediation analysis seeks to explain the pathway(s) through which an exposure affects an outcome. Mediation analysis experiences a number of methodological difficulties, including bias due to confounding and measurement error. Mendelian randomisation (MR) can be used to improve causal inference for mediation analysis. We describe two approaches that can be used for estimating mediation analysis with MR: multivariable Mendelian randomisation (MVMR) and two-step Mendelian randomisation. We outline the approaches and provide code to demonstrate how they can be used in mediation analysis. We review issues that can affect analyses, including confounding, measurement error, weak instrument bias, and analysis of multiple mediators. Description of the methods is supplemented by simulated and real data examples. Although Mendelian randomisation relies on large sample sizes and strong assumptions, such as having strong instruments and no horizontally pleiotropic pathways, our examples demonstrate that it is unlikely to be affected by confounders of the exposure or mediator and the outcome, reverse causality and non-differential measurement error of the exposure or mediator. Both MVMR and two-step MR can be implemented in both individual-level MR and summary data MR, and can improve causal inference in mediation analysis.

scholarly journals On adaptive propensity score truncation in causal inference

2018 ◽  
Vol 28 (6) ◽  
pp. 1741-1760 ◽  
Author(s):  
Cheng Ju ◽  
Joshua Schwab ◽  
Mark J van der Laan
Keyword(s):  
Propensity Score ◽  
Causal Inference ◽  
Extreme Values ◽  
Likelihood Estimation ◽  
Experimental Treatment ◽  
Point Estimation ◽  
Finite Sample ◽  
Novel Approach ◽  
Confidence Interval Coverage ◽  
Interval Coverage

The positivity assumption, or the experimental treatment assignment (ETA) assumption, is important for identifiability in causal inference. Even if the positivity assumption holds, practical violations of this assumption may jeopardize the finite sample performance of the causal estimator. One of the consequences of practical violations of the positivity assumption is extreme values in the estimated propensity score (PS). A common practice to address this issue is truncating the PS estimate when constructing PS-based estimators. In this study, we propose a novel adaptive truncation method, Positivity-C-TMLE, based on the collaborative targeted maximum likelihood estimation (C-TMLE) methodology. We demonstrate the outstanding performance of our novel approach in a variety of simulations by comparing it with other commonly studied estimators. Results show that by adaptively truncating the estimated PS with a more targeted objective function, the Positivity-C-TMLE estimator achieves the best performance for both point estimation and confidence interval coverage among all estimators considered.

Sign in / Sign up

Export Citation Format

Share Document