mlsanityData to 7 Oct 2026

Human-AI Co-design for Clinical Prediction Models

arXiv 2601.09072 ↗First posted 14 Jan 2026cs.AI cs.CL stat.MEcited by 9

Not on the map. Why?

Abstract (arXiv)

Developing safe, effective, and practically useful clinical prediction models (CPMs) traditionally requires iterative collaboration between clinical experts, data scientists, and informaticists. This process refines the often small but critical details of the model building process, such as which features/patients to include and how clinical categories should be defined. However, this traditional collaboration process is extremely time- and resource-intensive, resulting in only a small fraction of CPMs reaching clinical practice. This challenge intensifies when teams attempt to incorporate unstructured clinical notes, which can contain an enormous number of concepts. To address this challenge, we introduce HACHI, an iterative human-in-the-loop framework that uses AI agents to accelerate the development of fully interpretable CPMs by enabling the exploration of concepts in clinical notes. HACHI alternates between (i) an AI agent rapidly exploring and evaluating candidate concepts in clinical notes and (ii) clinical and domain experts providing feedback to improve the CPM learning process. HACHI defines concepts as simple yes-no questions that are used in linear models, allowing the clinical AI team to transparently review, refine, and validate the CPM learned in each round. In two real-world prediction tasks (acute kidney injury and traumatic brain injury), HACHI outperforms existing approaches, surfaces new clinically relevant concepts not included in commonly-used CPMs, and improves model generalizability across clinical sites and time periods. Furthermore, HACHI reveals the critical role of the clinical AI team, such as directing the AI agent to explore concepts that it had not previously considered, adjusting the granularity of concepts it considers, changing the objective function to better align with the clinical objectives, and identifying issues of data bias and leakage.

Key facts. Cited by 9 papers in the map (data through 7 Oct 2026). Builds on 24 of its 69 references that are in the map. First posted on arXiv 14 Jan 2026.

Builds on

24 of 69 references.

References · 24

  1. 2020Concept Bottleneck Modelscited by 1,022
  2. 2025AlphaEvolve: A coding agent for scientific and algorithmic discoverycited by 763
  3. 2023Label-Free Concept Bottleneck Modelscited by 282
  4. 2022GatorTron: A Large Clinical Language Model to Unlock Patient Information from Unstructured Electronic Health Recordscited by 238
  5. 2022Large Language Models are Few-Shot Clinical Information Extractorscited by 206
  6. 2025CollabLLM: From Passive Responders to Active Collaboratorscited by 117
  7. 2022Interactive Concept Bottleneck Modelscited by 79
  8. 2015Supersparse Linear Integer Models for Optimized Medical Scoring Systemscited by 76
  9. 2023Bridging the Gulf of Envisioning: Cognitive Design Challenges in LLM Interfacescited by 71
  10. 2016Scalable Bayesian Rule Listscited by 64
  11. 2023Large Language Models to Identify Social Determinants of Health in Electronic Health Recordscited by 55
  12. 2022Overlooked factors in concept-based explanations: Dataset choice, concept learnability, and human capabilitycited by 46
12 more
  1. 2024A Taxonomy for Human-LLM Interaction Modes: An Initial Explorationcited by 38
  2. 2023Learning Interpretable Style Embeddings via Prompting LLMscited by 29
  3. 2020Learning Interpretable Concept-Based Models with Human Feedbackcited by 17
  4. 2023Tree Prompting: Efficient Task Adaptation without Fine-Tuningcited by 17
  5. 2024Integrating Clinical Knowledge into Concept Bottleneck Modelscited by 17
  6. 2024Bayesian Concept Bottleneck Models with LLM Priorscited by 16
  7. 2023CHiLL: Zero-shot Custom Interpretable Feature Extraction from Clinical Notes with Large Language Modelscited by 15
  8. 2024A General Framework for Producing Interpretable Semantic Text Embeddingscited by 10
  9. 2024Tree-Based Leakage Inspection and Control in Concept Bottleneck Modelscited by 8
  10. 2024Constructing Concept-based Models to Mitigate Spurious Correlations with Minimal Human Effortcited by 7
  11. 2025Tempo: Helping Data Scientists and Domain Experts Collaboratively Specify Predictive Modeling Taskscited by 6
  12. 2025ACT: Agentic Classification Treecited by 1

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 9.

9 papers.

  1. 2026Test-time Recursive Thinking: Self-Improvement without External Feedbackcited by 7
  2. 2026Do explanations generalize across large reasoning models?cited by 5
  3. 2026Sanity Checks for Agentic Data Sciencecited by 4
  4. 2026Expert Evaluation of Clinical AI Tools on Real Point-of-Care Clinical Queriescited by 4
  5. 2026Interpreting and Steering State-Space Models via Activation Subspace Bottleneckscited by 3
  6. 2026Selecting Feature Interactions for Generalized Additive Models by Distilling Foundation Modelscited by 1
  7. 2026Agentic-imodels: Evolving agentic interpretability tools via autoresearchcited by 1
  8. 2026From Fuzzy to Formal: Scaling Hospital Quality Improvement with AIcited by 0
  9. 2026Data-Efficient Adaptation of LLMs via Attention Head Reweightingcited by 0