# LLM-Extracted Covariates for Clinical Causal Inference: Rethinking Integration Strategies

arXiv: https://arxiv.org/abs/2604.16763 · HTML: https://mlsanity.com/paper/2604.16763/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 16 of its 57 references that are in the map.
- First posted on arXiv 18 Apr 2026.

## Builds on (most cited first, 16 of 16)

- [Qwen3 Technical Report](https://mlsanity.com/paper/2505.09388/) (2025)
- [Publicly Available Clinical BERT Embeddings](https://mlsanity.com/paper/1904.03323/) (2019)
- [Capabilities of GPT-4 on Medical Challenge Problems](https://mlsanity.com/paper/2303.13375/) (2023)
- [Estimation and Inference of Heterogeneous Treatment Effects using Random Forests](https://mlsanity.com/paper/1510.04342/) (2015)
- [GatorTron: A Large Clinical Language Model to Unlock Patient Information from Unstructured Electronic Health Records](https://mlsanity.com/paper/2203.03540/) (2022)
- [Large Language Models are Few-Shot Clinical Information Extractors](https://mlsanity.com/paper/2205.12689/) (2022)
- [Causal Inference in Natural Language Processing: Estimation, Prediction, Interpretation and Beyond](https://mlsanity.com/paper/2109.00725/) (2021)
- [Text and Causal Inference: A Review of Using Text to Remove Confounding from Causal Estimates](https://mlsanity.com/paper/2005.00649/) (2020)
- [Causal Effects of Linguistic Properties](https://mlsanity.com/paper/2010.12919/) (2020)
- [Adjusting for Confounders with Text: Challenges and an Empirical Evaluation Framework for Causal Inference](https://mlsanity.com/paper/2009.09961/) (2020)
- [Leveraging text data for causal inference using electronic health records](https://mlsanity.com/paper/2307.03687/) (2023)
- [Cause-Effect Deep Information Bottleneck For Systematically Missing Covariates](https://mlsanity.com/paper/1807.02326/) (2018)
- [Estimating Causal Effects With Partial Covariates For Clinical Interpretability](https://mlsanity.com/paper/1811.10347/) (2018)
- [The Medical Deconfounder: Assessing Treatment Effects with Electronic Health Records](https://mlsanity.com/paper/1904.02098/) (2019)
- [Controlling for Unobserved Confounding with Large Language Model Classification of Patient Smoking Status](https://mlsanity.com/paper/2411.03004/) (2024)
- [REACT-LLM: A Benchmark for Evaluating LLM Integration with Causal Features in Clinical Prognostic Tasks](https://mlsanity.com/paper/2511.07127/) (2025)

## Built on by (most cited first, 1 of 1)

- [Information Set Emulation: Causal Certificates for AI Derived EHR Features](https://mlsanity.com/paper/2609.17777/) (2026)

## Abstract (arXiv)

Causal inference from electronic health records (EHR) is fundamentally limited by unmeasured confounding: critical clinical states such as frailty, goals of care, and mental status are documented in free-text notes but absent from structured data. Large language models can extract these latent confounders as interpretable, structured covariates, yet how to effectively integrate them into causal estimation pipelines has not been systematically studied. Using the MIMIC-IV database with 21,859 sepsis patients, we compare seven covariate-integration strategies for estimating the effect of early vasopressor initiation on 28-day mortality, spanning tabular-only baselines, traditional NLP representations, and three LLM-augmented approaches. A central finding is that not all integration strategies are equally effective: directly augmenting the propensity score model with LLM covariates achieves the best performance, while dual-caliper matching on text-derived categorical distances restricts the donor pool and degrades estimation. In semi-synthetic experiments with known ground-truth effects, LLM-augmented propensity scores reduce estimation bias from 0.0143 to 0.0003 relative to tabular-only methods, and this advantage persists under substantial simulated extraction error. On real data, incorporating LLM-extracted covariates reduces the estimated treatment effect from 0.055 to 0.027, directionally consistent with the CLOVERS randomized trial, and a doubly robust estimator yielding 0.031 confirms the robustness of this finding. Our results offer practical guidance on when and how text-derived covariates improve causal estimation in critical care. Our code is available at https://github.com/fpxlei/LLM-Covariates-Causal.