Back

SurvTEDVAE: A Disentangled Variational Autoencoder for Heterogeneous Treatment Effect Estimation with Time-to-Event Outcomes

Powell, W. J. B.; Zhang, L.

2026-04-28 health informatics
10.64898/2026.04.26.26351790 medRxiv
Show abstract

Estimating heterogeneous treatment effects (HTE) from observational health data is essential for precision medicine, yet existing methods often struggle with high-dimensional covariates and time-to-event outcomes common in electronic health records (EHRs). We propose SurvTEDVAE, a disentangled variational autoencoder designed for causal survival analysis. The model learns latent representations corresponding to instrumental factors, confounders, and outcome-dependent risk factors, and integrates a survival likelihood to model time-to-event outcomes with censoring. The learned representations are used to estimate conditional average treatment effects using downstream causal estimators. We evaluated SurvTEDVAE using a semi-synthetic ACTG dataset and a high-dimensional EHR-based hypertension cohort with over 20,000 covariates. Across both datasets, SurvTEDVAE achieved lower estimation error for heterogeneous treatment effects compared with meta-learning and causal survival forest approaches. These results demonstrate that disentangled representation learning can improve causal effect estimation for survival outcomes in high-dimensional real-world health data.

Matching journals

The top 6 journals account for 50% of the predicted probability mass.

1
Nature Communications
4913 papers in training set
Top 15%
12.2%
2
Nature Biomedical Engineering
42 papers in training set
Top 0.1%
12.0%
3
npj Digital Medicine
97 papers in training set
Top 0.6%
9.8%
4
Nature Machine Intelligence
61 papers in training set
Top 0.5%
6.2%
5
Journal of Biomedical Informatics
45 papers in training set
Top 0.3%
6.2%
6
Bioinformatics
1061 papers in training set
Top 5%
3.9%
50% of probability mass above
7
Journal of the American Medical Informatics Association
61 papers in training set
Top 0.8%
3.6%
8
Science Translational Medicine
111 papers in training set
Top 1%
3.5%
9
Scientific Reports
3102 papers in training set
Top 39%
3.5%
10
Nature Computational Science
50 papers in training set
Top 0.3%
2.7%
11
Advanced Science
249 papers in training set
Top 8%
2.5%
12
Medical Image Analysis
33 papers in training set
Top 0.5%
2.3%
13
Science Advances
1098 papers in training set
Top 14%
2.0%
14
PLOS ONE
4510 papers in training set
Top 55%
1.7%
15
Patterns
70 papers in training set
Top 1%
1.7%
16
Communications Biology
886 papers in training set
Top 11%
1.4%
17
eBioMedicine
130 papers in training set
Top 2%
1.4%
18
Med
38 papers in training set
Top 0.4%
1.4%
19
IEEE Journal of Biomedical and Health Informatics
34 papers in training set
Top 1%
1.3%
20
The American Journal of Human Genetics
206 papers in training set
Top 3%
1.3%
21
Briefings in Bioinformatics
326 papers in training set
Top 5%
1.3%
22
Nature Medicine
117 papers in training set
Top 4%
0.9%
23
Cell Reports Medicine
140 papers in training set
Top 7%
0.9%
24
NeuroImage
813 papers in training set
Top 6%
0.7%
25
Journal of Medical Internet Research
85 papers in training set
Top 5%
0.7%
26
PNAS Nexus
147 papers in training set
Top 2%
0.7%
27
JAMIA Open
37 papers in training set
Top 2%
0.7%
28
Human Brain Mapping
295 papers in training set
Top 5%
0.7%
29
Biological Psychiatry
119 papers in training set
Top 3%
0.6%
30
JMIR Medical Informatics
17 papers in training set
Top 2%
0.6%