Generalizing Clinical Trials with Convex Hulls

Eric Strobl, Thomas A Lasko
Proceedings of the Second Conference on Causal Learning and Reasoning, PMLR 213:197-221, 2023.

Abstract

Randomized clinical trials eliminate confounding but impose strict exclusion criteria that limit recruitment to a subset of the population. Observational datasets are more inclusive but suffer from confounding – often providing overly optimistic estimates of treatment response over time due to partially optimized physician prescribing patterns. We therefore assume that the unconfounded treatment response lies somewhere in between the observational estimate before and the observational estimate after treatment assignment. This assumption allows us to extrapolate results from exclusive trials to the broader population by analyzing observational and trial data simultaneously using an algorithm called Optimum in Convex Hulls (OCH). OCH represents the treatment effect either in terms of convex hulls of conditional expectations or convex hulls (also known as mixtures) of conditional densities. The algorithm first learns the component expectations or densities using the observational data and then learns the linear mixing coefficients using trial data in order to approximate the true treatment effect; theory importantly explains why this linear combination should hold. OCH estimates the treatment effect in terms both expectations and densities with state of the art accuracy.

Cite this Paper

BibTeX
@InProceedings{pmlr-v213-strobl23a, title = {Generalizing Clinical Trials with Convex Hulls}, author = {Strobl, Eric and Lasko, Thomas A}, booktitle = {Proceedings of the Second Conference on Causal Learning and Reasoning}, pages = {197--221}, year = {2023}, editor = {van der Schaar, Mihaela and Zhang, Cheng and Janzing, Dominik}, volume = {213}, series = {Proceedings of Machine Learning Research}, month = {11--14 Apr}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v213/strobl23a/strobl23a.pdf}, url = {https://proceedings.mlr.press/v213/strobl23a.html}, abstract = {Randomized clinical trials eliminate confounding but impose strict exclusion criteria that limit recruitment to a subset of the population. Observational datasets are more inclusive but suffer from confounding – often providing overly optimistic estimates of treatment response over time due to partially optimized physician prescribing patterns. We therefore assume that the unconfounded treatment response lies somewhere in between the observational estimate before and the observational estimate after treatment assignment. This assumption allows us to extrapolate results from exclusive trials to the broader population by analyzing observational and trial data simultaneously using an algorithm called Optimum in Convex Hulls (OCH). OCH represents the treatment effect either in terms of convex hulls of conditional expectations or convex hulls (also known as mixtures) of conditional densities. The algorithm first learns the component expectations or densities using the observational data and then learns the linear mixing coefficients using trial data in order to approximate the true treatment effect; theory importantly explains why this linear combination should hold. OCH estimates the treatment effect in terms both expectations and densities with state of the art accuracy.} }
Endnote
%0 Conference Paper %T Generalizing Clinical Trials with Convex Hulls %A Eric Strobl %A Thomas A Lasko %B Proceedings of the Second Conference on Causal Learning and Reasoning %C Proceedings of Machine Learning Research %D 2023 %E Mihaela van der Schaar %E Cheng Zhang %E Dominik Janzing %F pmlr-v213-strobl23a %I PMLR %P 197--221 %U https://proceedings.mlr.press/v213/strobl23a.html %V 213 %X Randomized clinical trials eliminate confounding but impose strict exclusion criteria that limit recruitment to a subset of the population. Observational datasets are more inclusive but suffer from confounding – often providing overly optimistic estimates of treatment response over time due to partially optimized physician prescribing patterns. We therefore assume that the unconfounded treatment response lies somewhere in between the observational estimate before and the observational estimate after treatment assignment. This assumption allows us to extrapolate results from exclusive trials to the broader population by analyzing observational and trial data simultaneously using an algorithm called Optimum in Convex Hulls (OCH). OCH represents the treatment effect either in terms of convex hulls of conditional expectations or convex hulls (also known as mixtures) of conditional densities. The algorithm first learns the component expectations or densities using the observational data and then learns the linear mixing coefficients using trial data in order to approximate the true treatment effect; theory importantly explains why this linear combination should hold. OCH estimates the treatment effect in terms both expectations and densities with state of the art accuracy.
APA
Strobl, E. & Lasko, T.A.. (2023). Generalizing Clinical Trials with Convex Hulls. Proceedings of the Second Conference on Causal Learning and Reasoning, in Proceedings of Machine Learning Research 213:197-221 Available from https://proceedings.mlr.press/v213/strobl23a.html.

Related Material