Estimating the Individual Treatment Effect on Survival Time Based on Prior Knowledge and Counterfactual Prediction
The estimation of the Individual Treatment Effect (ITE) on survival time is an important research topic in clinics-based causal inference. Various representation learning methods have been proposed to deal with its three key problems, i.e., reducing selection bias, handling censored survival data, a...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-07-01
|
| Series: | Entropy |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1099-4300/24/7/975 |
| _version_ | 1797433607114981376 |
|---|---|
| author | Yijie Zhao Hao Zhou Jin Gu Hao Ye |
| author_facet | Yijie Zhao Hao Zhou Jin Gu Hao Ye |
| author_sort | Yijie Zhao |
| collection | DOAJ |
| description | The estimation of the Individual Treatment Effect (ITE) on survival time is an important research topic in clinics-based causal inference. Various representation learning methods have been proposed to deal with its three key problems, i.e., reducing selection bias, handling censored survival data, and avoiding balancing non-confounders. However, none of them consider all three problems in a single method. In this study, by combining the Counterfactual Survival Analysis (CSA) model and Dragonnet from the literature, we first propose a CSA–Dragonnet to deal with the three problems simultaneously. Moreover, we found that conclusions from traditional Randomized Controlled Trials (RCTs) or Retrospective Cohort Studies (RCSs) can offer valuable bound information to the counterfactual learning of ITE, which has never been used by existing ITE estimation methods. Hence, we further propose a CSA–Dragonnet with Embedded Prior Knowledge (CDNEPK) by formulating a unified expression of the prior knowledge given by RCTs or RCSs, inserting counterfactual prediction nets into CSA–Dragonnet and defining loss items based on the bounds for the ITE extracted from prior knowledge. Semi-synthetic data experiments showed that CDNEPK has superior performance. Real-world experiments indicated that CDNEPK can offer meaningful treatment advice. |
| first_indexed | 2024-03-09T10:19:24Z |
| format | Article |
| id | doaj.art-906cdbb2ec96439884ba7212597179a8 |
| institution | Directory Open Access Journal |
| issn | 1099-4300 |
| language | English |
| last_indexed | 2024-03-09T10:19:24Z |
| publishDate | 2022-07-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Entropy |
| spelling | doaj.art-906cdbb2ec96439884ba7212597179a82023-12-01T22:06:58ZengMDPI AGEntropy1099-43002022-07-0124797510.3390/e24070975Estimating the Individual Treatment Effect on Survival Time Based on Prior Knowledge and Counterfactual PredictionYijie Zhao0Hao Zhou1Jin Gu2Hao Ye3Department of Automation, Beijing National Research Center for Information Science and Technology (BNRist), Tsinghua University, Beijing 100084, ChinaDepartment of Automation, Beijing National Research Center for Information Science and Technology (BNRist), Tsinghua University, Beijing 100084, ChinaMOE Key Laboratory for Bioinformatics, BNRIST Bioinformatics Division, Department of Automation, Tsinghua University, Beijing 100084, ChinaDepartment of Automation, Beijing National Research Center for Information Science and Technology (BNRist), Tsinghua University, Beijing 100084, ChinaThe estimation of the Individual Treatment Effect (ITE) on survival time is an important research topic in clinics-based causal inference. Various representation learning methods have been proposed to deal with its three key problems, i.e., reducing selection bias, handling censored survival data, and avoiding balancing non-confounders. However, none of them consider all three problems in a single method. In this study, by combining the Counterfactual Survival Analysis (CSA) model and Dragonnet from the literature, we first propose a CSA–Dragonnet to deal with the three problems simultaneously. Moreover, we found that conclusions from traditional Randomized Controlled Trials (RCTs) or Retrospective Cohort Studies (RCSs) can offer valuable bound information to the counterfactual learning of ITE, which has never been used by existing ITE estimation methods. Hence, we further propose a CSA–Dragonnet with Embedded Prior Knowledge (CDNEPK) by formulating a unified expression of the prior knowledge given by RCTs or RCSs, inserting counterfactual prediction nets into CSA–Dragonnet and defining loss items based on the bounds for the ITE extracted from prior knowledge. Semi-synthetic data experiments showed that CDNEPK has superior performance. Real-world experiments indicated that CDNEPK can offer meaningful treatment advice.https://www.mdpi.com/1099-4300/24/7/975individual treatment effectsurvival datacounterfactual predictionprior knowledge |
| spellingShingle | Yijie Zhao Hao Zhou Jin Gu Hao Ye Estimating the Individual Treatment Effect on Survival Time Based on Prior Knowledge and Counterfactual Prediction Entropy individual treatment effect survival data counterfactual prediction prior knowledge |
| title | Estimating the Individual Treatment Effect on Survival Time Based on Prior Knowledge and Counterfactual Prediction |
| title_full | Estimating the Individual Treatment Effect on Survival Time Based on Prior Knowledge and Counterfactual Prediction |
| title_fullStr | Estimating the Individual Treatment Effect on Survival Time Based on Prior Knowledge and Counterfactual Prediction |
| title_full_unstemmed | Estimating the Individual Treatment Effect on Survival Time Based on Prior Knowledge and Counterfactual Prediction |
| title_short | Estimating the Individual Treatment Effect on Survival Time Based on Prior Knowledge and Counterfactual Prediction |
| title_sort | estimating the individual treatment effect on survival time based on prior knowledge and counterfactual prediction |
| topic | individual treatment effect survival data counterfactual prediction prior knowledge |
| url | https://www.mdpi.com/1099-4300/24/7/975 |
| work_keys_str_mv | AT yijiezhao estimatingtheindividualtreatmenteffectonsurvivaltimebasedonpriorknowledgeandcounterfactualprediction AT haozhou estimatingtheindividualtreatmenteffectonsurvivaltimebasedonpriorknowledgeandcounterfactualprediction AT jingu estimatingtheindividualtreatmenteffectonsurvivaltimebasedonpriorknowledgeandcounterfactualprediction AT haoye estimatingtheindividualtreatmenteffectonsurvivaltimebasedonpriorknowledgeandcounterfactualprediction |