The Alignment Paradox of Medical Large Language Models in Infertility Care: Decoupling Algorithmic Improvement from Clinical Decision-making Quality
This work addresses the challenge of aligning LLMs with clinical decision-making in medicine, revealing a paradox where algorithmic improvements may not enhance clinical trust, which is incremental but important for medical AI applications.
The study evaluated four alignment strategies for medical LLMs in infertility care, finding that while GRPO achieved the highest algorithmic accuracy, clinicians preferred the SFT model for its clearer reasoning and higher therapeutic feasibility, with SFT outperforming other models and even physicians' original decisions in blinded comparisons.
Large language models (LLMs) are increasingly adopted in clinical decision support, yet aligning them with the multifaceted reasoning pathways of real-world medicine remains a major challenge. Using more than 8,000 infertility treatment records, we systematically evaluate four alignment strategies: Supervised Fine-Tuning (SFT), Direct Preference Optimization (DPO), Group Relative Policy Optimization (GRPO), and In-Context Learning (ICL) through a dual-layer framework combining automatic benchmarks with blinded doctor-in-the-loop assessments. GRPO achieves the highest algorithmic accuracy across multiple decision layers, confirming the value of reinforcement-based optimization for structured prediction tasks. However, clinicians consistently prefer the SFT model, citing clearer reasoning processes (p = 0.035) and higher therapeutic feasibility (p = 0.019). In blinded pairwise comparisons, SFT attains the highest winning rate (51.2%), outperforming both GRPO (26.2%) and even physicians' original decisions (22.7%). These results reveal an alignment paradox: algorithmic improvements do not necessarily translate into higher clinical trust, and may diverge from human-centered preferences. Our findings highlight the need for alignment strategies that prioritize clinically interpretable and practically feasible reasoning, rather than solely optimizing decision-level accuracy.