Estimating treatment effects from single-arm trials via latent-variable modeling

Manuel Haussmann, Tran Minh Son Le, Viivi Halla-aho, Samu Kurki, Jussi Leinonen, Miika Koskinen, Samuel Kaski, Harri Lähdesmäki
Proceedings of The 27th International Conference on Artificial Intelligence and Statistics, PMLR 238:2926-2934, 2024.

Abstract

Randomized controlled trials (RCTs) are the accepted standard for treatment effect estimation but they can be infeasible due to ethical reasons and prohibitive costs. Single-arm trials, where all patients belong to the treatment group, can be a viable alternative but require access to an external control group. We propose an identifiable deep latent-variable model for this scenario that can also account for missing covariate observations by modeling their structured missingness patterns. Our method uses amortized variational inference to learn both group-specific and identifiable shared latent representations, which can subsequently be used for {\em (i)} patient matching if treatment outcomes are not available for the treatment group, or for {\em (ii)} direct treatment effect estimation assuming outcomes are available for both groups. We evaluate the model on a public benchmark as well as on a data set consisting of a published RCT study and real-world electronic health records. Compared to previous methods, our results show improved performance both for direct treatment effect estimation as well as for effect estimation via patient matching.

Cite this Paper

BibTeX
@InProceedings{pmlr-v238-haussmann24a, title = {Estimating treatment effects from single-arm trials via latent-variable modeling}, author = {Haussmann, Manuel and Minh Son Le, Tran and Halla-aho, Viivi and Kurki, Samu and Leinonen, Jussi and Koskinen, Miika and Kaski, Samuel and L\"{a}hdesm\"{a}ki, Harri}, booktitle = {Proceedings of The 27th International Conference on Artificial Intelligence and Statistics}, pages = {2926--2934}, year = {2024}, editor = {Dasgupta, Sanjoy and Mandt, Stephan and Li, Yingzhen}, volume = {238}, series = {Proceedings of Machine Learning Research}, month = {02--04 May}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v238/haussmann24a/haussmann24a.pdf}, url = {https://proceedings.mlr.press/v238/haussmann24a.html}, abstract = {Randomized controlled trials (RCTs) are the accepted standard for treatment effect estimation but they can be infeasible due to ethical reasons and prohibitive costs. Single-arm trials, where all patients belong to the treatment group, can be a viable alternative but require access to an external control group. We propose an identifiable deep latent-variable model for this scenario that can also account for missing covariate observations by modeling their structured missingness patterns. Our method uses amortized variational inference to learn both group-specific and identifiable shared latent representations, which can subsequently be used for {\em (i)} patient matching if treatment outcomes are not available for the treatment group, or for {\em (ii)} direct treatment effect estimation assuming outcomes are available for both groups. We evaluate the model on a public benchmark as well as on a data set consisting of a published RCT study and real-world electronic health records. Compared to previous methods, our results show improved performance both for direct treatment effect estimation as well as for effect estimation via patient matching.} }
Endnote
%0 Conference Paper %T Estimating treatment effects from single-arm trials via latent-variable modeling %A Manuel Haussmann %A Tran Minh Son Le %A Viivi Halla-aho %A Samu Kurki %A Jussi Leinonen %A Miika Koskinen %A Samuel Kaski %A Harri Lähdesmäki %B Proceedings of The 27th International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2024 %E Sanjoy Dasgupta %E Stephan Mandt %E Yingzhen Li %F pmlr-v238-haussmann24a %I PMLR %P 2926--2934 %U https://proceedings.mlr.press/v238/haussmann24a.html %V 238 %X Randomized controlled trials (RCTs) are the accepted standard for treatment effect estimation but they can be infeasible due to ethical reasons and prohibitive costs. Single-arm trials, where all patients belong to the treatment group, can be a viable alternative but require access to an external control group. We propose an identifiable deep latent-variable model for this scenario that can also account for missing covariate observations by modeling their structured missingness patterns. Our method uses amortized variational inference to learn both group-specific and identifiable shared latent representations, which can subsequently be used for {\em (i)} patient matching if treatment outcomes are not available for the treatment group, or for {\em (ii)} direct treatment effect estimation assuming outcomes are available for both groups. We evaluate the model on a public benchmark as well as on a data set consisting of a published RCT study and real-world electronic health records. Compared to previous methods, our results show improved performance both for direct treatment effect estimation as well as for effect estimation via patient matching.
APA
Haussmann, M., Minh Son Le, T., Halla-aho, V., Kurki, S., Leinonen, J., Koskinen, M., Kaski, S. & Lähdesmäki, H.. (2024). Estimating treatment effects from single-arm trials via latent-variable modeling. Proceedings of The 27th International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 238:2926-2934 Available from https://proceedings.mlr.press/v238/haussmann24a.html.

Related Material