Estimation of Causal Effects with Small Data in the Presence of Trapdoor Variables

Helske, Jouni; Tikka, Santtu; Karvanen, Juha

doi:10.1111/rssa.12699

Cited by 7 publications

(9 citation statements)

References 31 publications

(36 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…While such a scenario could, sometimes, justify control for indication even if it was a true IV (i.e. indication has no effect on the outcome) – to alleviate part of the imposed confounding by U – its implications must be considered in the light of recent work on trapdoor variables [28]. Nonetheless, we have favoured simplicity to enhance transportability outside our considered mechanisms.…”

Section: Discussionmentioning

confidence: 99%

Bias amplification of unobserved confounding in pharmacoepidemiological studies using indication-based sampling: there is no free lunch in restricting the sample to those with a particular drug-indication

Ahlqvist

Madley‐Dowd

et al. 2022

Preprint

View full text Add to dashboard Cite

Estimating causal effects in observational pharmacoepidemiology is a challenging task, as it is often plagued by confounding by indication. Restricting the sample to those with an indication for drug use is a commonly performed procedure; indication-based sampling ensures that the exposed and unexposed are exchangeable on the indication - limiting the potential for confounding by indication. However, indication-based sampling has received little scrutiny, despite the hazards of exposure-related covariate control. Using causal diagrams, simulations, and empirical examples, we demonstrate that indication-based sampling in the presence of unobserved confounding can give rise to bias amplification, a self-inflicted phenomenon where one inflates pre-existing bias through inappropriate covariate control. Additionally, we show that indication-based sampling generally leads to a greater net bias than alternative approaches, such as regression adjustment. Finally, we expand on how bias amplification should be reasoned about when distinct clinically relevant effects on the outcome among those with an indication (effect-heterogeneity) exist. We conclude that studies using indication-based sampling should have robust justification - and that it should by no means be considered a 'free lunch' to adopt such approaches. As such, we suggest that future observational studies stay wary of bias amplification when considering drug indications.

show abstract

Section: Discussionmentioning

confidence: 99%

Bias amplification of unobserved confounding in pharmacoepidemiological studies using indication-based sampling: there is no free lunch in restricting the sample to those with a particular drug-indication

Ahlqvist

Madley‐Dowd

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…Accurate estimation of connectivity measures requires sufficient sample sizes [ 193 ]. Guidelines for sufficient sample sizes have been presented for different scenarios [ 194 , 195 , 196 ], whereas solutions for different applications with small samples have been proposed [ 197 , 198 , 199 ].…”

Section: Limitations and Pitfalls Of Connectivity Measuresmentioning

confidence: 99%

Connectivity Analysis for Multivariate Time Series: Correlation vs. Causality

Papana

2021

Entropy

View full text Add to dashboard Cite

The study of the interdependence relationships of the variables of an examined system is of great importance and remains a challenging task. There are two distinct cases of interdependence. In the first case, the variables evolve in synchrony, connections are undirected and the connectivity is examined based on symmetric measures, such as correlation. In the second case, a variable drives another one and they are connected with a causal relationship. Therefore, directed connections entail the determination of the interrelationships based on causality measures. The main open question that arises is the following: can symmetric correlation measures or directional causality measures be applied to infer the connectivity network of an examined system? Using simulations, we demonstrate the performance of different connectivity measures in case of contemporaneous or/and temporal dependencies. Results suggest the sensitivity of correlation measures when temporal dependencies exist in the data. On the other hand, causality measures do not spuriously indicate causal effects when data present only contemporaneous dependencies. Finally, the necessity of introducing effective instantaneous causality measures is highlighted since they are able to handle both contemporaneous and causal effects at the same time. Results based on instantaneous causality measures are promising; however, further investigation is required in order to achieve an overall satisfactory performance.

show abstract

“…, z 13 }, with variables r = {z 1 , z 2 , z 4 , z 10 } as receivers and e = {z 1 , z 3 , z 4 , z 6 , z 7 , z 8 , z 10 , z 12 , z 13 } as emitters (with variables z 5 , z 9 , and z 11 being neither). This leads to a simplified graph in Figure 4 As an example of peripheral extension and the robustness of the estimation strategies, consider a causal graph shown in Figure 5, studied earlier in (Helske et al, 2021), where the interest is in the causal effect of the education level X e on income X i . Now assume that there are a number of other background variables X A = (X a 1 , .…”

Section: Illustrationmentioning

confidence: 99%

“…the identifiability of the causal effect p(x i | do(x e )). While the causal effect estimates can depend on the availability of data on x T , the obtained estimator is robust to these changes in a sense that the same methodology as in (Helske et al, 2021) can be used to estimate the effect in all of these cases.…”

Section: Illustrationmentioning

confidence: 99%

“…Directed acyclic graphs (DAGs) and their extensions are commonly used to describe causal relations between variables in epidemiology and other fields (Pearl, 1995;Greenland et al, 1999;Tennant et al, 2021). The power of graphs lies in their ability to visualize the assumed structure, and at the same time, to serve as well-defined inputs for algorithms such as those that solve the nonparametric identifiability of causal effects (Shpitser and Pearl, 2006;Lee et al, 2019;Lee and Shpitser, 2020;Tikka et al, 2021). The graphical approach has been criticized by proponents of potential outcome framework (Rubin, 1974) for its impracticality when a large number of variables is considered (Imbens, 2020).…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Clustering and Structural Robustness in Causal Diagrams

Tikka¹,

Helske²,

Karvanen³

2021

Preprint

Self Cite

View full text Add to dashboard Cite

Graphs are commonly used to represent and visualize causal relations. For a small number of variables, this approach provides a succinct and clear view of the scenario at hand. As the number of variables under study increases, the graphical approach may become impractical, and the clarity of the representation is lost. Clustering of variables is a natural way to reduce the size of the causal diagram but it may erroneously change the essential properties of the causal relations if implemented arbitrarily. We define a specific type of cluster, called transit cluster, that is guaranteed to preserve the identifiability properties of causal effects under certain conditions. We provide a sound and complete algorithm for finding all transit clusters in a given graph and demonstrate how clustering can simplify the identification of causal effects. We also study the inverse problem, where one starts with a clustered graph and looks for extended graphs where the identifiability properties of causal effects remain unchanged. We show that this kind of structural robustness is closely related to transit clusters.

show abstract

Estimation of Causal Effects with Small Data in the Presence of Trapdoor Variables

Abstract: This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.

Cited by 7 publications

References 31 publications

Bias amplification of unobserved confounding in pharmacoepidemiological studies using indication-based sampling: there is no free lunch in restricting the sample to those with a particular drug-indication

Bias amplification of unobserved confounding in pharmacoepidemiological studies using indication-based sampling: there is no free lunch in restricting the sample to those with a particular drug-indication

Connectivity Analysis for Multivariate Time Series: Correlation vs. Causality

Clustering and Structural Robustness in Causal Diagrams

Contact Info

Product

Resources

About