EGIP: An Evidence-Grounded Interaction Protocol for Clinical Agents

Zecheng Ren; Ruichen Cong; Yi Min Chen; Jing Zhang; Feiyang Chen; Wutong Zhang; Kaixiang Wang; Zihan Wang

cs.CLOct 17, 2026

EGIP: An Evidence-Grounded Interaction Protocol for Clinical Agents

Authors: Zecheng Ren, Ruichen Cong, Yi Min Chen, Jing Zhang, Feiyang Chen, Wutong Zhang, Kaixiang Wang, Zihan Wang,

Organizations: Waseda University, Tokorozawa, Japan · Shanghai Jiao Tong University, Shanghai, China

Abstract

Interactive clinical agents acquire evidence through patient dialogue and diagnostic examinations, yet their performance also depends on how the interaction environment exposes information, available actions, and examination results. When these conditions are implicit or inconsistently controlled, final-answer accuracy can conflate diagnostic reasoning with the effects of the interaction protocol. We propose the Evidence-Grounded Interaction Protocol (EGIP), which enforces role-specific information boundaries, exposes diagnostic actions without revealing their results, retrieves examination results from stored clinical records, and organizes acquired evidence for final decision-making. By separating evidence acquisition from terminal prediction, EGIP enables different decision strategies to operate on the same observations. We evaluate EGIP across two clinical-agent frameworks, five datasets, and six language models. EGIP improves diagnostic accuracy over native baselines across all primary comparisons. A clinician review further shows that the proportion of automatically scored correct responses jointly judged evidence-supported rises from 10.20% to 36.55%. These findings support the effectiveness of EGIP as a unified interaction protocol for improving clinical-agent performance through controlled evidence access and acquisition.

Figures & tables

Figure 1: Evidence-access interfaces in clinical simulation. Top: potential issues in prompt-mediated interaction. Bottom: EGIP’s role-specific views, result-blinded action menus, record lookup, and acquired-evidence interface.

Y

denotes the reference target.

Figure 2: Overview of EGIP. Dark, bold labels and thick borders highlight the four core components: role-specific views (

R

), result-blinded action menus (

B

), record lookup (

O

), and the acquired-evidence interface (

E

). The protocol separates evidence acquisition from terminal prediction. The hidden reference target

Y

is used only for post-prediction evaluation.

Backbone	Method	MedQA	MedQA_Ext	NEJM	NEJM_Ext	MIMIC-IV
Qwen2.5-VL-7B	MedAgentSim	$29.91\pm 7.99$	$31.78\pm 4.15$	$24.44\pm 13.88$	$10.83\pm 0.83$	$57.41\pm 0.40$
	+ EGIP-Direct	$\mathbf{57.94}\pm 2.47$	$\mathbf{55.61}\pm 0.81$	$\mathbf{77.78}\pm 3.85$	$\mathbf{45.00}\pm 3.63$	$\mathbf{87.62}\pm 0.20$
	AgentClinic	$33.33\pm 3.28$	$34.58\pm 2.47$	$26.67\pm 6.67$	$12.78\pm 3.94$	$66.44\pm 2.09$
	+ EGIP-Direct	$\mathbf{57.01}\pm 1.87$	$\mathbf{54.67}\pm 0.00$	$\mathbf{73.33}\pm 0.00$	$\mathbf{51.39}\pm 2.10$	$\mathbf{85.88}\pm 1.22$
Llama3.1-8B	MedAgentSim	$31.15\pm 3.28$	$29.28\pm 1.43$	$13.33\pm 6.67$	$12.50\pm 0.83$	$50.35\pm 1.04$
	+ EGIP-Direct	$\mathbf{59.81}\pm 1.87$	$\mathbf{65.26}\pm 3.57$	$\mathbf{53.33}\pm 6.67$	$\mathbf{47.50}\pm 2.20$	$\mathbf{78.12}\pm 2.28$

Table 1: Accuracy (%) across four backbones, reported as mean

\pm

sample standard deviation over three repetitions. Framework rows denote Native. Shaded entries denote EGIP-Direct. Bold marks the higher mean within each Native–EGIP pair.

Llama3.3-70B-AWQ
Method	MedQA	MIMIC-IV
MedAgentSim	$66.04\pm 1.95$	$\mathbf{79.51}\pm 2.84$
+ EGIP-Direct	$\mathbf{73.83}\pm 2.47$	$72.69\pm 2.56$
AgentClinic	$70.09\pm 5.68$	$\mathbf{80.21}\pm 1.51$
+ EGIP-Direct	$\mathbf{76.01}\pm 1.08$	$\mathbf{80.21}\pm 1.84$
Qwen2.5-72B-AWQ

Table 2: Auxiliary Native–EGIP-Direct comparisons on MedQA and MIMIC-IV. Values are automated accuracy (%), reported as mean

\pm

sample standard deviation over three repetitions. Shading and bold mark the higher mean within each pair, including ties.

Dataset	Variant	Acc. (%)	Drop (pp)	Paired 95% CI
Panel A. Acquisition components
MedQA	EGIP-Direct	$57.94\pm 2.47$	–	–
	w/o action menu	$24.30\pm 0.00$	33.64	[24.30, 42.68]
	Native measurement access	$54.83\pm 1.95$	3.12	[ $-1.56$ , 8.10]
NEJM_Ext	EGIP-Direct	$45.00\pm 3.63$	–	–
	w/o action menu	$40.56\pm 0.96$	4.44	[ $-1.67$ , 10.83]

Table 3: Ablations with Qwen2.5-VL-7B on MedAgentSim. Accuracy (%) is mean

\pm

sample standard deviation over three repetitions. Drop is the shaded control minus the variant within each panel and dataset, in percentage points (pp). Bold drops have paired, pointwise 95% CIs excluding zero. Panel B uses rerun controls. “w/o” denotes removal. MedQA has no candidate labels.

Figure 3: Comparison of Direct, Vote, and Full on the same acquired evidence for Qwen2.5-VL-7B on MedAgentSim. Accuracy and answer coverage are computed over all scheduled outputs.

Support rule / dataset	Native, $n/N$ (%)	Direct, $n/N$ (%)	$\Delta$ (pp)	Paired 95% CI
Panel A. Correct-and-supported yield (539 complete pairs)
Both clinicians $S$ (CSY)	55/539 ( 10.20 )	197/539 ( 36.55 )	+26.35	[20.93, 31.79]
Clinician 1 $S$	112/539 (20.78)	259/539 (48.05)	+27.27	[21.59, 33.09]
Clinician 2 $S$	74/539 (13.73)	233/539 (43.23)	+29.50	[23.89, 35.12]
At least one clinician $S$	131/539 (24.30)	295/539 (54.73)	+30.43	[24.72, 36.19]
Panel B. Evidence support within the shared correct-answer subset (105 pairs)

Table 4: Exploratory review of Native and EGIP-Direct outputs. Panel A reports correct-and-supported yield. Panel B reports support in the post hoc shared correct-answer subset (105 pairs from 60 source cases). Panel C reports joint support irrespective of correctness. CSY requires automated correctness and two supported (

S

) ratings.

n/N

: qualifying outputs/complete pairs.

\Delta

: Direct minus Native in percentage points. CIs cluster by source case. Shading and bold mark joint-support summaries.

Appendix figures & tables4 assets

Supplementary material from the paper’s appendix.

Appendix

Model	Framework	Dataset	$\Delta$	Paired 95% CI	Effect bounds	Conservative envelope
Qwen	MAS	MedQA	28.04	[20.56, 35.83]	[1.25, 27.10]	[-6.54, 34.89]
Qwen	MAS	MedQA_Ext	23.83	[18.22, 29.60]	[0.62, 24.14]	[-5.14, 29.60]
Qwen	MAS	MIMIC-IV	30.21	[26.16, 34.26]	[9.26, 26.97]	[4.98, 30.90]
Qwen	MAS	NEJM	53.33	[20.00, 80.00]	[15.56, 55.56]	[-13.33, 82.22]
Qwen	MAS	NEJM_Ext	34.17	[26.39, 42.22]	[14.72, 40.56]	[6.94, 48.33]
Qwen	AC	MedQA	23.68	[15.26, 32.09]	[15.26, 28.04]	[6.85, 36.14]

Appendix

Table 5: Direct minus Native in percentage points. MAS: MedAgentSim. AC: AgentClinic. CIs are paired, pointwise 95% intervals. Effect bounds are not confidence intervals.

Model	Framework	Dataset	$\Delta$	Paired 95% CI	Effect bounds	Conservative envelope
Qwen	MAS	MedQA	3.12	[-2.81, 9.35]	[-3.12, 8.41]	[-9.97, 14.33]
Qwen	MAS	MedQA_Ext	3.89	[-0.31, 7.79]	[-1.56, 8.10]	[-5.76, 12.15]
Qwen	MAS	MIMIC-IV	11.69	[8.56, 15.05]	[8.56, 15.62]	[4.98, 19.33]
Qwen	MAS	NEJM	22.22	[6.67, 40.00]	[6.67, 22.22]	[-6.67, 40.00]
Qwen	MAS	NEJM_Ext	4.17	[-2.22, 10.83]	[-3.89, 12.78]	[-10.83, 20.00]
Qwen	AC	MedQA	6.23	[-0.31, 12.77]	[0.62, 9.35]	[-6.54, 16.51]

Appendix

Table 6: Direct minus Full in percentage points. MAS: MedAgentSim. AC: AgentClinic. CIs are paired, pointwise 95% intervals. Effect bounds are not confidence intervals.

Stratum	Cases	Model	Framework	Native	Direct	Paired 95% CI	Conservative envelope	Native $U$
Diagnosis	90	Qwen2.5-VL-7B	MAS	13.70	45.93	[23.33, 41.48]	[2.96, 48.89]	—
Diagnosis	90	Qwen2.5-VL-7B	AC	14.81	54.81	[30.00, 49.63]	[8.89, 53.33]	—
Diagnosis	90	Llama3.1-8B	MAS	12.22	49.26	[27.78, 46.30]	[4.81, 51.48]	—
Diagnosis	90	Llama3.1-8B	AC	6.67	45.93	[29.63, 48.52]	[19.26, 58.89]	—
Etiology etc.	13	Qwen2.5-VL-7B	MAS	2.56	28.21	[7.69, 46.15]	[2.56, 56.41]	—
Etiology etc.	13	Qwen2.5-VL-7B	AC	12.82	23.08	[-5.13, 28.21]	[-12.82, 51.28]	—

Appendix

Table 7: NEJM_Ext target strata for the four primary backbones. Values are automated accuracy (%) over three repetitions. Paired 95% CIs and conservative envelopes report Direct minus Native in percentage points for Qwen2.5-VL-7B and Llama3.1-8B.

U

counts Native unscored outputs across repetitions for Qwen3-8B and Qwen3-14B. Dashes denote unreported statistics. All 120 cases are retained, with strata defined by question semantics.

Model	Framework	Dataset	$\Delta$	Paired 95% CI	Holm $p$	$U$
Qwen3-8B	MAS	MedQA	+21.18	[13.71, 28.97]	0.0014	91
Qwen3-8B	MAS	MedQA_Ext	+22.74	[16.98, 28.35]	0.0014	182
Qwen3-8B	MAS	MIMIC-IV	+7.18	[2.08, 12.27]	0.0496	97
Qwen3-8B	MAS	NEJM	+42.22	[15.56, 68.89]	0.0812	8
Qwen3-8B	MAS	NEJM_Ext	+33.06	[24.44, 41.67]	0.0014	86
Qwen3-8B	AC	MedQA	+12.77	[4.98, 20.56]	0.0295	0

Appendix

Table 8: Direct minus Native for Qwen3 and the larger backbones. CIs use case-cluster bootstrap resampling. Exploratory sign-flip

p

values are Holm-adjusted across the 28 comparisons.

Evidence-Grounded Reasoning

Evidence-grounded reasoning studies how AI systems connect conclusions to traceable, relevant evidence and recognize when support is incomplete.

See all 90 papers

MediSkill-Evo: Process-Constrained Self-Evolution for Evidence-Grounded Clinical Interaction
Aug 24, 2026Ruoyu Wu, Shenfu Xie, Yinqian Sun +2Self-Evolving Agents LLM Agent Skill Learning
Brain-inspired Cognitive AI Lab, Institute of Automation, Chinese Academy of Sciences · Beijing Key Laboratory of Safe AI and Superalignment · Beijing Institute of AI Safety and Governance +2
EviDx: Evidence-Aware Active Diagnosis with Scaffolded LLM Agents
Aug 25, 2026Lihang Zeng, Shaoting Zhang, Xiaofan ZhangMedical Diagnosis Clinical Reasoning
Shanghai Jiao Tong University · Shanghai Innovation Institute
MeDxAgent: Multi-Agent Consultation for Interactive Medical Diagnosis
Jun 2, 2026Akshat Sanghvi, Naren Akash, Raza Imam +2Multi-Agent LLM Systems Medical Diagnosis
Microsoft Research India · Institute of AI and Innovation, AIG Hospitals · Mohamed bin Zayed University of Artificial Intelligence

Aug 24, 2026J/K move · Enter open · S save

MediSkill-Evo: Process-Constrained Self-Evolution for Evidence-Grounded Clinical Interaction

Ruoyu Wu, Shenfu Xie, Yinqian Sun, Haibo Tong+1

Brain-inspired Cognitive AI Lab, Institute of Automation, Chinese Academy of Sciences · Beijing Key Laboratory of Safe AI and Superalignment · Beijing Institute of AI Safety and Governance+2

Self-Evolving Agents LLM Agent Skill Learning LLM Agent Safety Clinical Reasoning Agent Memory

Open paper AskPDF arXiv

Interactive clinical agents operate under partial observability, so reliable care depends on reaching the correct diagnosis through evidence-grounded, safe interactions. Yet existing agents struggle to convert experience into reusable process knowledge with explicit provenance and authority. To address this gap, we introduce MediSkill-Evo, which self-evolves governed process knowledge without fine-tuning the backbone. It realizes this self-evolution by updating clinical, process, symbolic, and visual knowledge in four typed banks under type-specific validation and scope rules. The Process-Constrained Preference Harness then turns validated knowledge into action by grounding candidates in evidence and prioritizing safer decisions. We evaluate on 300 MIMIC-IV-derived FullChain encounters, 180 hard-isolation conditions covering six process obligations, and 100 multimodal NEJM image-diagnosis cases. On Qwen FullChain, MediSkill-Evo improves diagnosis accuracy by 7.81% and treatment-intent coverage by 70.67% over the best-performing prior agent, while reducing critical failures by 43.04%. Under stress, it improves the stress-process composite by 7.77% and required-action completion by 12.41% over the best-performing agent for each metric, with stronger patient-fact, temporal-evidence, and triage-red-flag recovery and no controller-scored errors in unavailable-evidence, treatment, and triage safety checks. On multimodal NEJM diagnosis, MediSkill-Evo with optional MedSAM localization improves diagnosis accuracy by 2.56% and core score by 18.96% over the best-performing memory agent. Code is available at https://anonymous.4open.science/r/mediskill-evo_anonymous-68E7.

Abstract

Evidence-Grounded Reasoning

Explore similar work

MediSkill-Evo: Process-Constrained Self-Evolution for Evidence-Grounded Clinical Interaction

EviDx: Evidence-Aware Active Diagnosis with Scaffolded LLM Agents

MeDxAgent: Multi-Agent Consultation for Interactive Medical Diagnosis