Combining Interventional and Observational Data Using Causal Reductions

Ilse, Maximilian; Forré, Patrick; Welling, Max; Mooij, Joris M.

doi:10.48550/arxiv.2103.04786

Cited by 4 publications

(6 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Recent work has studied combining observational data and data from randomized trials for CATE estimation (Rosenman et al, 2020;Ilse et al, 2022). Under relatively mild assumptions, estimates from combined datasets yield more efficient estimates of CATEs than using RCT data alone.…”

Section: Discussionmentioning

confidence: 99%

“…(3) Despite its simplicity this example is conceptually important for all cases with binary treatment and discrete x as a) when the treatment is binary, any arbitrary confounder can be modeled as a single binary variable while maintaining the same observational and interventional distributions (Ilse et al, 2022); and b) in the limit of infinite data, stratifying the population for each value of x and estimating β 0 in each of the strata is equivalent to non-parametric estimation of β 0 (x) in Equation 2 when u is binary and x is discrete.…”

Section: Example 1: Offset Models Do Not Estimate the Interventional ...mentioning

confidence: 99%

See 1 more Smart Citation

Conditional average treatment effect estimation with treatment offset models

Amsterdam¹,

Ranganath²

2022

Preprint

View full text Add to dashboard Cite

Treatment effect estimates are often available from randomized controlled trials as a single average treatment effect for a certain patient population. Estimates of the conditional average treatment effect (CATE) are more useful for individualized treatment decision making, but randomized trials are often too small to estimate the CATE. There are several examples in medical literature where the assumption of a known constant relative treatment effect (e.g. an odds-ratio) is used to estimate CATE models from large observational datasets. One approach to estimating these CATE models is by using the relative treatment effect as an offset, while estimating the covariate-specific baseline risk. Whether this is a valid approach in the presence of unobserved confounding is unknown.We demonstrate for a simple example that offset models do not recover the true CATE in the presence of unobserved confounding. We then explore the magnitude of this bias in numerical experiments. For virtually all plausible confounding magnitudes, estimating the CATE using offset models is more accurate than assuming a single absolute treatment effect whenever there is sufficient variation in the baseline risk.Next, we observe that the odds-ratios reported in randomized controlled trials are not the odds-ratios that are needed in offset models because trials often report the marginal odds-ratio. We introduce a constraint to better use marginal odds-ratios from randomized controlled trials and find that the newly introduced constrained offset models have lower bias than standard offset models. Finally, we highlight directions for future research for exploiting the assumption of a constant relative treatment effect with offset models.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Example 1: Offset Models Do Not Estimate the Interventional ...mentioning

confidence: 99%

Conditional average treatment effect estimation with treatment offset models

Amsterdam¹,

Ranganath²

2022

Preprint

View full text Add to dashboard Cite

show abstract

“…Several further approaches have been proposed recently, but for different settings than ours. Ilse et al (2021) propose a causal reduction method for combining interventional and observational data. However, they consider a discrete treatment and discrete outcome and rely on linear-Gaussian models.…”

Section: Difference To Our Workmentioning

confidence: 99%

Combining Observational and Randomized Data for Estimating Heterogeneous Treatment Effects

Hatt¹,

Berrevoets²,

Curth³

et al. 2022

Preprint

View full text Add to dashboard Cite

Estimating heterogeneous treatment effects is an important problem across many domains. In order to accurately estimate such treatment effects, one typically relies on data from observational studies or randomized experiments. Currently, most existing works rely exclusively on observational data, which is often confounded and, hence, yields biased estimates. While observational data is confounded, randomized data is unconfounded, but its sample size is usually too small to learn heterogeneous treatment effects. In this paper, we propose to estimate heterogeneous treatment effects by combining large amounts of observational data and small amounts of randomized data via representation learning. In particular, we introduce a two-step framework: first, we use observational data to learn a shared structure (in form of a representation); and then, we use randomized data to learn the data-specific structures. We analyze the finite sample properties of our framework and compare them to several natural baselines. As such, we derive conditions for when combining observational and randomized data is beneficial, and for when it is not. Based on this, we introduce a sample-efficient algorithm, called CorNet. We use extensive simulation studies to verify the theoretical properties of CorNet and multiple real-world datasets to demonstrate our method's superiority compared to existing methods.

show abstract

“…Recent work has studied combining observational data and data from randomized trials for CATE estimation Ilse et al, 2022). Under relatively mild assumptions, estimates from combined datasets yield more efficient estimates of CATEs than using RCT data alone.…”

Section: Discussionmentioning

confidence: 99%

“…Despite its simplicity this example is conceptually important for all cases with binary treatment and discrete x as a) when the treatment is binary, any arbitrary confounder can be modeled as a single binary variable while maintaining the same observational and interventional distributions (Ilse et al, 2022); and b) in the limit of infinite data, stratifying the population for each value of x and estimating β 0 in each of the strata is equivalent to non-parametric estimation of β 0 (x) in Equation 2when u is binary and x is discrete.…”

Section: Example 1: Offset Models Do Not Estimate the Interventional ...mentioning

confidence: 99%

From survival prediction to treatment decision in lung cancer

Amsterdam¹

View full text Add to dashboard Cite

show abstract

Combining Interventional and Observational Data Using Causal Reductions

Cited by 4 publications

References 0 publications

Conditional average treatment effect estimation with treatment offset models

Conditional average treatment effect estimation with treatment offset models

Combining Observational and Randomized Data for Estimating Heterogeneous Treatment Effects

From survival prediction to treatment decision in lung cancer

Contact Info

Product

Resources

About