LongAgent guides variable search to improve medical outcome predictions
LongAgent: History-Guided Agentic Search for Longitudinal Outcome Prediction
Artificial Intelligence
Summary
Predicting future health outcomes using medical data is very hard because the data comes from many sources and changes over time. The authors created LongAgent, which acts like a smart assistant that tries different groups of data and time ranges to find the best way to predict results. It remembers what worked before to make better guesses next time. Tests show LongAgent improves predictions on fake data and works well on real medical information.
What this means in practice
- •For hospital data teams: Automate identification of which medical measurements and timeframes best predict patient outcomes to improve clinical decision support systems.
- •For healthcare analytics teams: Enhance predictive models for patient prognosis by systematically exploring data variables and temporal patterns with guidance from past search results.
Authors
Siyao Wang, Florian Guitton, Shuojie Fu, Guanyu Tao, Kai Sun, Wenjia Bai
Abstract
Extracting informative representations from longitudinal data that can predict future outcomes remains a critical challenge in medicine. Medical datasets are inherently heterogeneous, consisting of a large number of variables collected from different sources, sampled with different temporal spacings, and representing different aspects of human health status. This requires identifying those variables with predictive value, processing longitudinal information, and integrating multiple variables for outcome prediction. Here, we propose a novel agent-based approach, LongAgent, that can autonomously search over combinations of variable sets, temporal windows and longitudinal aggregation functions, and identify candidates with promising predictive performance. LongAgent utilises a history memory of previous searches and numerical evidence to guide subsequent exploration. On synthetic data, LongAgent achieves a mean prediction RMSE of 1.7376 and improves over the strongest non-agent baseline by 0.0151 (95% CI: [0.0045,0.0260]; p=0.0273). On a real clinical dataset, it performs comparably to the best baseline.