Harnessing Artificial Intelligence for Hypothesis Generation in Childhood Asthma: Insights from NHANES

Abstract

Although large language models (LLMs) have undergone substantial development, their applicability to epidemiological research has not been sufficiently examined. This study aims to develop and evaluate an LLM-based framework for hypothesis generation and testing, demonstrating its application in childhood asthma in the National Health and Nutrition Examination Survey (NHANES). Pilot study was conducted to explore factors associated with childhood asthma in the 2001–2020 NHANES cycles. A modular agent system was developed, including Database Query, Statistic, Paper Search, and Paper Download tools, along with two LLM models (Key Generator and Hypothesis Tester). Multivariable logistic regression was used to test for the association between each variable and current asthma, generating a tentative affirmative claim. The Key Generator module produced keywords for literature search, the Paper Search and Paper Download tools queried PubMed and retrieved relevant studies, and the Hypothesis Tester module synthesized evidence and determined the support for claims for each variable. Keywords and conclusions were reviewed by researchers and validated using multiple LLMs (ChatGPT, DeepSeek, and Gemini) to ensure consistency and robustness. 25,839 children with ( n = 2928) and without ( n = 22,911) current asthma, and 10,359 variables were included in the multivariable analysis, which yielded 100 variables associated with asthma. Of these, 21 were directly related to asthma (supporting published studies), 43 were indirectly related to asthma (based on background knowledge, though not explicitly discussed in the available publications), and 34 were unrelated to asthma. Two variables were excluded due to a lack of discriminative keywords. This study demonstrates the effectiveness of LLM-based models for generating and testing hypotheses about childhood asthma.

Keywords

Artificial intelligence / Asthma / Children / Risk factors

Acknowledgments

The authors acknowledge support from the University of Pittsburgh Center for Research Computing (RRID:SCR_022735) through the computational resources provided.

Statement of Significance

(1) Problem or Issue: Asthma is the most common chronic respiratory disease of childhood. The causes of childhood asthma remain insufficiently understood.

Statement of the Use of Generative AI and AI-Assisted Technologies in the Writing Process

During the preparation of this manuscript, the author(s) used ChatGPT in order to check for grammatical errors and improve sentence clarity and flow. After using this tool/service, the author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the published article.

Author Contributions

J.L., W.C. and J.C.C. conceived this study. Y.-Y.H. participated in data collection and quality control. X.Y., K.M.G., F.J.R., J.L., W.C., C.Z. and J.C.C. participated in statistical analyses. J.L., W.C. and J.C.C. drafted the manuscript. All authors reviewed the manuscript, contributed substantial intellectual content, and approved the final version of the manuscript.

Ethics Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

The data used in this study are publicly available from the National Health and Nutrition Examination Survey (NHANES). These datasets can be accessed free of charge through the official website of the Centers for Disease Control and Prevention (CDC) at https://www.cdc.gov/nchs/nhanes/ (accessed on 14 April 2026). No additional restrictions apply to the use of these data.

Funding

Dr. Liu was supported by a training grant T32 HL129949 from the U.S. National Institutes of Health (NIH). This work was supported by grants HL152475 and HL168539 from the NIH.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

References

[1] Achiam J, Adler S, Agarwal S, Ahmad L, Akkaya I, Aleman FL, et al. Gpt-4 technical report. arXiv 2023, arXiv:2303.08774. DOI: 10.48550/arXiv.2303.08774

[2] Team G, Anil R, Borgeaud S, Alayrac JB, Yu J, Soricut R, et al. Gemini: A family of highly capable multimodal models. arXiv 2023, arXiv:2312.11805. DOI: 10.48550/arXiv.2312.11805

[3] Anthropic’s Claude. Available online: https://claude.com/product/overview (accessed on 14 April 2026).

[4] Touvron H, Lavril T, Izacard G, Martinet X, Lachaux MA, Lacroix T, et al. Llama: Open and efficient foundation language models. arXiv 2023, arXiv:2302.13971. Available online: https://formacion.actuarios.org/wp-content/uploads/2024/05/2302.13971-LLama-Open-and-Efficient-Foundation-Language-Models.pdf (accessed on 14 April 2026).

[5] Jiang AQ, Sablayrolles A, Roux A, Mensch A, Savary B, Bamford C, et al. Mixtral of experts. arXiv 2024, arXiv:2401.04088. DOI: 10.48550/arXiv.2401.04088

[6] Ali SR, Dobbs TD, Hutchings HA, Whitaker IS. Using ChatGPT to write patient clinic letters. Lancet Digit. Health 2023, 5, e179-e181. DOI: 10.1016/S2589-7500(23)00048-1

[7] Ye W, Li X, Huang Y. Relationship Between Physical Activity and Adult Asthma Control Using NHANES 2011-2020 Data. Med. Sci. Monit. 2023, 29, e939350. DOI: 10.12659/MSM.939350

[8] Han YY, Forno E, Celedón JC. Urinary caffeine and caffeine metabolites, asthma, and lung function in a nationwide study of U.S. adults. J. Asthma 2021, 59, 2127-2134. DOI: 10.1080/02770903.2021.1993250

[9] Han YY, Forno E, Celedón JC. Sex Steroid Hormones and Asthma in a Nationwide Study of U.S. Adults. Am. J. Respir. Crit. Care Med. 2020, 201, 158-166. DOI: 10.1164/rccm.201905-0996OC

[10] Han Y, Rosser F, Forno E, Celedón JC. Exposure to polycyclic aromatic hydrocarbons, vitamin D, and lung function in children with asthma. Pediatr. Pulmonol. 2018, 53, 1362-1368. DOI: 10.1002/ppul.24084

[11] Yang G, Sun T, Han YY, Rosser F, Forno E, Chen W, et al. Serum Cadmium and Lead, Current Wheeze, and Lung Function in a Nationwide Study of Adults in the United States. J. Allergy Clin. Immunol. Pract. 2019, 7, 2653-2660.e3. DOI: 10.1016/j.jaip.2019.05.029

[12] Han X, He X, Hao G, Cao L, Qi Y, Han K. The association of asthma duration with body mass index and Weight-Adjusted-Waist index in a nationwide study of the U.S. adults. Eur. J. Med. Res. 2023, 28, 122. DOI: 10.1186/s40001-023-01089-4

[13] Wang K, Zhao L, Luo H, Deng C, Gong L, Chen Z. Association of serum vitamin C levels with asthma in adults: Results of NHANES 2003-2006 and mendelian randomization study. BMC Pulm. Med. 2024, 24, 4. DOI: 10.1186/s12890-023-02821-w

[14] Hou W, Ji Z. GeneTuring tests GPT models in genomics. BioRxiv 2023. Available online: https://storage.prod.researchhub.com/uploads/papers/2023/10/26/2023.03.11.532238.full.pdf (accessed on 14 April 2026).

[15] Karpas E, Abend O, Belinkov Y, Lenz B, Lieber O, Ratner N, et al. MRKL Systems: A modular, neuro-symbolic architecture that combines large language models, external knowledge sources and discrete reasoning. arXiv 2022, arXiv:2205.00445. DOI: 10.48550/arXiv.2205.00445

[16] Ferruz N, Schmidt S, Höcker B. ProtGPT2 is a deep unsupervised language model for protein design. Nat. Commun. 2022, 13, 4348. DOI: 10.1038/s41467-022-32007-7

[17] Luo R, Sun L, Xia Y, Qin T, Zhang S, Poon H, et al. BioGPT: Generative pre-trained transformer for biomedical text generation and mining. Brief. Bioinform. 2022, 23, bbac409. DOI: 10.1093/bib/bbac409

[18] Jin Q, Yang Y, Chen Q, Lu Z, Wren J. GeneGPT: Augmenting Large Language Models with Domain Tools for Improved Access to Biomedical Information. Bioinformatics 2024, 40, btae075. DOI: 10.1093/bioinformatics/btae075

[19] Lewis P, Perez E, Piktus A, Petroni F, Karpukhin V, Goyal N, et al. Retrieval-augmented generation for knowledge-intensive NLP tasks. In Proceedings of the 34th International Conference on Neural Information Processing Systems, Vancouver, BC, Canada, 6-12 December 2020; p. 793.

[20] Peng B, Galley M, He P, Cheng H, Xie Y, Hu Y, et al. Check your facts and try again: Improving large language models with external knowledge and automated feedback. arXiv 2023, arXiv:2302.12813. DOI: 10.48550/arXiv.2302.12813

[21] Wei J, Wang X, Schuurmans D, Bosma M, Xia F, Chi E, et al. Chain-of-thought prompting elicits reasoning in large language models. In Proceedings of the 36th International Conference on Neural Information Processing Systems, New Orleans, LA, USA, 28 November-9 December 2022; p. 1800.

[22] Lála J, O’Donoghue O, Shtedritski A, Cox S, Rodriques SG, White AD. Paperqa: Retrieval-augmented generative agent for scientific research. arXiv 2023, arXiv:2312.07559. DOI: 10.48550/arXiv.2312.07559

[23] Chase H. LangChain. 17 October 2022. Available online: https://github.com/langchain-ai/langchain (accessed on 14 April 2026).

[24] Simonti CN, Vernot B, Bastarache L, Bottinger E, Carrell DS, Chisholm RL, et al. The phenotypic legacy of admixture between modern humans and Neandertals. Science 2016, 351, 737-741. DOI: 10.1126/science.aad2149

[25] Huang L, Yu W, Ma W, Zhong W, Feng Z, Wang H, et al. A Survey on Hallucination in Large Language Models: Principles, Taxonomy, Challenges, and Open Questions. ACM Trans. Inf. Syst. 2025, 43, 42. DOI: 10.1145/3703155