Large Language Models for Structured Clinical Data Analysis: Dual-Agent Grounding and Validation

Erfan D. Dehkalani; Seetha Shankaran; Abbot R. Laptook; C. Michael Cotten; P. Ellen Grant; Yangming Ou

cs.AISep 28, 2026

Large Language Models for Structured Clinical Data Analysis: Dual-Agent Grounding and Validation

Authors: Erfan D. Dehkalani, Seetha Shankaran, Abbot R. Laptook, C. Michael Cotten, P. Ellen Grant, Yangming Ou

Organizations: Fetal-Neonatal Neuroimaging and Developmental Science Center, Boston Children’s Hospital, and Harvard Medical School, Boston, MA, USA · Department of Pediatrics, Wayne State University School of Medicine, Detroit, MI, USA · Department of Pediatrics, Women & Infants Hospital of Rhode Island, and Warren Alpert Medical School of Brown University, Providence, RI, USA · Division of Neonatology, Department of Pediatrics, Duke University School of Medicine, Durham, NC, USA

Abstract

Objective: To develop and characterize CLEAR-Med, a dual-agent framework for natural-language analysis of structured clinical data that separates SQL-based invocation from independent validation. Methods: CLEAR-Med uses one agent to translate a question into executable Structured Query Language (SQL), retain the executed query and database result, and produce a draft. Deterministic checks and a separately invoked cross-provider Validation Agent then accept the draft, request one bounded repair, or abstain. We formalized the system as a bounded selective pipeline and evaluated CLEAR-Med's configuration and scalability, and the Invocation Agent's accuracy and consistency on a 25-query development benchmark, using a harmonized 21-site neonatal hypoxic-ischemic encephalopathy table containing 532 de-identified infant records and approximately 1,300 variables. Results: CLEAR-Med completed all six nominal scalability configurations, including 500x1300. Across 25 development-benchmark queries repeated five times, the Invocation Agent answered 83 of 125 responses correctly (66.4%; query-cluster bootstrap 95% CI, 48.0-83.2%), compared with 15 of 125 (12.0%; 95% CI, 3.2-22.4%) for the ungrounded ChatGPT baseline, a paired improvement of 54.4 percentage points (95% CI, 36.8-72.0%). Conclusion: CLEAR-Med provides a general architecture for traceable analysis of structured clinical data: numerical claims remain linked to executed SQL, and unresolved cases can fail closed. The reported experiments characterize CLEAR-Med's configuration and scalability and the Invocation Agent's accuracy, while the formal analysis establishes the encoded-property guarantee of the complete control flow; a prospective full-pipeline evaluation of the validation and abstention stages is the next stage of this work.

Figures & tables

Problem or Issue	Natural-language analysis of governed clinical tables can produce fluent numerical answers without exposing the cohort, calculation, or supporting rows.
What is Already Known	Language models can generate SQL and use external tools, but query generation alone does not establish that a reported number preserves the executed result and intended clinical constraints.
What this Paper Adds	CLEAR-Med separates schema-guided SQL invocation from deterministic checks and an independently invoked Validation Agent. The pipeline retains SQL provenance, permits one bounded repair, abstains on unresolved failures, and provides a property-specific encoded-validity guarantee.
Who would benefit from the new knowledge in this paper	Biomedical-informatics researchers and clinical-data teams building auditable natural-language interfaces to governed relational data.

Table 1 : Statement of Significance

System	50 $\times$ 50	50 $\times$ 100	100 $\times$ 50	100 $\times$ 100	50 $\times$ 1000	500 $\times$ 1300	Max Time (s)	Notes
OpenBioLLM-8B [ 6 ]	$\checkmark$	$\checkmark$	$\checkmark$	$\checkmark$	$\times$	$\times$	2.19	CUDA OOM
open-bio-med-8B [ 8 ]	$\checkmark$	$\checkmark$	$\checkmark$	$\checkmark$	$\times$	$\times$	11.78	CUDA OOM
Med-ChimeraLlama-3-8B [ 7 ]	$\checkmark$	$\checkmark$	$\checkmark$	$\checkmark$	$\times$	$\times$	26.05	CUDA OOM
Daredevil-8B [ 22 ]	$\checkmark$	$\checkmark$	$\checkmark$	$\checkmark$	$\times$	$\times$	26.00	CUDA OOM
DISC-MedLLM [ 11 ]	$\checkmark$	$\times$	$\times$	$\times$	$\times$	$\times$	32.86	Timeout
CLEAR-Med	$\checkmark$	$\checkmark$	$\checkmark$	$\checkmark$	$\checkmark$	$\checkmark$	28.12	–

Table 2 : End-to-end scalability across six fixed table configurations under the deployment interfaces described in Section 3.9. The table reports deterministic completion status and elapsed time for each tested configuration.

Condition	Err (%) ↓	5%-GT (%) ↑	SD (%) ↓	Range (%) ↓
ChatGPT baseline (OpenAI API)	55.74	6.67	58.80	146.58
Invocation Agent	33.67	65.00	35.07	78.55

Table 3 : Accuracy and consistency on numerical queries across repeated trials (

K=5

repetitions, 12 percentage-valued queries). Err = mean absolute percentage error vs. the deterministic data oracle; 5%-GT = share of responses within 5% of ground truth; SD and Range are computed across repetitions and averaged over queries. Arrows (

\downarrow

\uparrow

) indicate the preferred direction.

Task type	$\boldsymbol{n}$ /condition	ChatGPT baseline	Invocation Agent
comparison	30	0.0	30.0
count	25	36.0	100.0
multi-condition	20	5.0	50.0
odds ratio	10	20.0	100.0
percentage	40	7.5	72.5
Overall (95% CI)	125	12.0 [3.2, 22.4]	66.4 [48.0, 83.2]

Table 4 : Per-task accuracy across 25 queries

\times

5 repetitions.

n

is the number of responses per condition in each subgroup. Overall accuracy (

n=125

per condition) is reported with a query-cluster bootstrap 95% confidence interval.

Appendix figures & tables4 assets

Supplementary material from the paper’s appendix.

Appendix

Metric	Operational definition	Goal
Time-based diagnostics
Setup environment time	Time to initialize the computational environment	$\downarrow$
Load files to database	Time to load data files into the database	$\downarrow$
Setup database time	Time to configure the database for query processing	$\downarrow$
Setup language model time	Time to initialize the language model	$\downarrow$
Create agent time	Time to instantiate the SQL agent	$\downarrow$

Appendix

Table A.1 : Supporting definitions and results. Panel A: operational definitions for the efficiency and execution diagnostics used to compare ChatGPT configurations. The final column gives the preferred direction; an em dash denotes a descriptive metric. All time measures are in seconds.

Metric	Operational definition	Goal
Performance diagnostics
Execution time	Duration from invocation start to final response	$\downarrow$
Total time	Execution time plus all setup operations	$\downarrow$
Number of database invocations	Count of successful database connections and queries	—
Number of errors	Count of SQL syntax errors and execution failures	$\downarrow$
Query diagnostics

Appendix

Table A.1 : Supporting definitions and results (continued). Panel C: diagnostics used in the one-factor-at-a-time configuration screen. The final column gives the preferred direction where applicable; an em dash denotes a descriptive diagnostic.

Factor	Setting	Exec. (s)	Total (s)	Errors	DB inv.	SQL len.	SQL compl.	Selected
Max iterations	10	317.12	317.30	–	3	503	5
	50	134.19	134.71	–	1	555	6	$\checkmark$
	100	139.08	139.28	–	1	208	2
Max execution time	10 s	38.52	38.71	3	0	–	–
	30 s	49.90	50.13	3	0	–	–
	60 s	85.40	85.60	3	1	–	–	$\checkmark$

Appendix

Table A.1 : Supporting definitions and results (continued). Panel D: one-factor-at-a-time configuration screen for CLEAR-Med. The check mark identifies the setting selected for the reported configuration. Selection first required successful SQL generation and execution, then used errors and execution cost; SQL length and complexity are descriptive. An em dash indicates that the diagnostic was not recorded for that factor.

Difficulty	$n$ /condition	ChatGPT baseline	Invocation Agent
easy	45	24.4	86.7
medium	25	8.0	80.0
hard	55	3.6	43.6

Appendix

Table A.1 : Supporting definitions and results (continued). Panel E: accuracy by query difficulty.

n

is the number of responses per condition; the clustered interval for the overall result appears in Table 4.

Explore similar work

ProtoMedAgent: Multimodal Clinical Interpretability via Privacy-Aware Agentic Workflows
May 13, 2026Alvaro Lopez Pellicer, Plamen Angelov, Marwan Bukhari +3Multimodal Clinical Data Interpretability
School of Computing and Communications, Lancaster University · Lancaster Medical School, Lancaster University · PUC-Rio, Puc-Behring Institute for AI
ClinLens: Towards Long-Horizon Coding Agents for Longitudinal Multimodal Clinical Data Science
Jul 28, 2026Yuan Zhu, Ethan B. Liu, Frank Nie +1Multimodal Clinical Data Longitudinal Patient Records
Shandong University, China
Medi-Gemma: A Hybrid Clinical Decision Support System Integrating Deterministic EMR Analytics and Retrieval-Augmented Generation
Jul 6, 2026Mohammed Saim Ahmed Quadri, Yunzhe Xue, Justin W. Ady +1Clinical Decision Support Electronic Health Records
Department of Computer Science, New Jersey Institute of Technology Newark, NJ, USA · Department of Data Science, New Jersey Institute of Technology Newark, NJ, USA · Vascular and Endovascular Surgery, Robert Wood Johnson Hospital New Brunswick, NJ, USA

May 13, 2026, cs.CVJ/K move · Enter open · S save

ProtoMedAgent: Multimodal Clinical Interpretability via Privacy-Aware Agentic Workflows

Alvaro Lopez Pellicer, Plamen Angelov, Marwan Bukhari, Yi Li+2

School of Computing and Communications, Lancaster University · Lancaster Medical School, Lancaster University · PUC-Rio, Puc-Behring Institute for AI

Multimodal Clinical Data Interpretability Agentic Workflows

Open paper AskPDF arXiv

While interpretable prototype networks offer compelling case-based reasoning for clinical diagnostics, their raw continuous outputs lack the semantic structure required for medical documentation. Bridging this gap via standard Retrieval-Augmented Generation (RAG) routinely triggers ``retrieval sycophancy,'' where Large Language Models (LLMs) hallucinate post-hoc rationalizations to align with visual predictions. We introduce ProtoMedAgent, a framework that formalizes multimodal clinical reporting as an iterative, zero-gradient test-time optimization problem over a strict neuro-symbolic bottleneck. Operating on a frozen prototype backbone, we distill latent visual and tabular features into a discrete semantic memory. Online generation is strictly constrained by exact set-theoretic differentials and a reflective Scribe-Critic loop, mathematically precluding unsupported narrative claims. To safely bound data disclosure, we introduce a semantic privacy gate governed by

k

-anonymity and

\ell

-diversity. Evaluated on a 4,160-patient clinical cohort, ProtoMedAgent achieves 91.2% Comparison Set Faithfulness where it fundamentally outperforms standard RAG (46.2%). ProtoMedAgent additionally leverages a binding

\ell

-diversity phase transition to systematically reduce artifact-level membership inference risks by an absolute 9.8%.

Abstract

Explore similar work

ProtoMedAgent: Multimodal Clinical Interpretability via Privacy-Aware Agentic Workflows

ClinLens: Towards Long-Horizon Coding Agents for Longitudinal Multimodal Clinical Data Science

Medi-Gemma: A Hybrid Clinical Decision Support System Integrating Deterministic EMR Analytics and Retrieval-Augmented Generation