Development of a Web Application for Health Product Classification
Article Sidebar
Main Article Content
Abstract
Objective: To develop a web application for the classification of health products regulated by the Food and Drug Administration (FDA). The application serves as a decision support tool, enabling entrepreneurs and researchers to self-classify their products and prepare for the product registration process. Methods: The web application was developed on the Google Apps Script (GAS) platform. It incorporates a rule-based system designed to classify products in accordance with relevant legislation and FDA classification flowcharts. Additionally, it integrates a generative artificial intelligence (AI), specifically a Large Language Model (LLM) utilizing the OpenAI gpt-5 model via API, to provide preliminary guidance. The system's performance was evaluated by comparing the classification outcomes among the rule-based system, the generative AI, and two FDA experts. Furthermore, user satisfaction was assessed among 30 individuals utilizing services at the FDA's One Stop Service Center. Results: The product classifications derived from the rule-based system and the generative AI demonstrated a very high level of agreement with the two experts (Fleiss’ Kappa = 0.862), with no statistically significant differences observed between the assessment methods. The rule-based system exhibited an almost perfect agreement with the experts (K = 0.918), whereas the generative AI showed a substantial level of agreement (K = 0.751). The overall accuracy was 0.933 for the rule-based system and 0.800 for the generative AI. The FDA experts' evaluation of the generative AI's guidance, based on 5 frequently asked questions (repeated for 3 iterations per question), indicated an overall quality at a moderate level (mean score: 3.20 out of 5). User satisfaction with the web application was rated at the highest level (mean score: 4.38 – 4.47 out of 5). Conclusion: The developed web application is an effective decision support tool for health product classification. It serves as a valuable preliminary learning resource and assists entrepreneurs or researchers in preparing for health product registration. Nonetheless, the application should be utilized in conjunction with professional guidance from experts.
Article Details
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
ผลการวิจัยและความคิดเห็นที่ปรากฏในบทความถือเป็นความคิดเห็นและอยู่ในความรับผิดชอบของผู้นิพนธ์ มิใช่ความเห็นหรือความรับผิดชอบของกองบรรณาธิการ หรือคณะเภสัชศาสตร์ มหาวิทยาลัยสงขลานครินทร์ ทั้งนี้ไม่รวมความผิดพลาดอันเกิดจากการพิมพ์ บทความที่ได้รับการเผยแพร่โดยวารสารเภสัชกรรมไทยถือเป็นสิทธิ์ของวารสารฯ
References
Public Health Ministerial Declaration in 2020 on the organization of the office of the Thai Food and Drug Administration. Royal Gazette. No. 137 (Part 8A) (Jan 29, 2020).
Food Act B.E. 2522. Royal Gazette. No. 79 (May 8, 1979).
Narcotics Code. Royal Gazette. No. 138, Part 73A (Nov 8, 2021).
Hazardous Substance Act B.E. 2535. Royal Gazette. No. 57A (Mar 29, 1992).
Drug Act B.E. 2510. Royal Gazette. No. 101 (Oct 20, 1967).
Herbal Product Act B.E. 2562. Royal Gazette. No. 56A (Apr 30, 2019).
Cosmetic Act B.E. 2558. Royal Gazette. No. 132, Part 86A (Sep 8, 2015).
Medical Device Act B.E. 2562. Royal Gazette. No. 2, Part 56A (Apr 30, 2019).
Kasikornbank. The growth of the health business as a new income opportunity in the COVID-19 era [online]. 2021 [cited Apr 4, 2024]. Available from: www.kasikornbank.com/th/personal/the-wisdom/articl es/Pages/BIZ-Empowerment_Onward57_2021.aspx.
Notification of the National Herbal Policy Committee on measures to promote herbal production for use as raw materials in herbal product manufacturing under safe agriculture guidelines. Royal Gazette. No.139, Part 287D special 9Dec 8, 2022).
Department of Thai Traditional and Alternative Medicine. National herbal development action plan, phase 2 (2023–2027). Nonthaburi: Ministry of Public Health; 2023.
Division of Innovative Health Products and Services. Infographic: Guidelines for selecting the appropriate application in the e-consult system [online]. 2023 [cited Apr 20, 2024]. Available from: innovative.fda .moph.go.th/online-consultation/infographic-02.
Moore J, Clark A, McKinley P. Evolutionary robotics on the web with WebGL and Javascript [online]. 2014 [cited Apr 20, 2024]. Available from: doi.org/10.48550 /arXiv.1406.3337
Ginige A, Murugesan S. Web engineering: An introduction. IEEE Multimedia. 2001; 8: 14–8.
Kovari A. AI for decision support: balancing accuracy, transparency, and trust across sectors. Information. 2024; 15: 725.
Ahmed N, Saha AK, Al Noman MA, Jim JR, Mridha MF, Kabir MA. Deep learning-based natural language processing in human–agent interaction: Applications, advancements and challenges. Nat Lang Proc J 2024; 9: 100112.
Cordeschi R. AI turns fifty: revisiting its origins. J Exp Theor Artif Intell 2007; 19: 257-65.
Levin C, Zaboli A, Turcato G, Saban M. Nursing judgment in the age of generative artificial intelligence: A cross-national study on clinical decision-making performance among emergency nurses. Int J Nurs Stud 2025; 172: 105216.
U.S. Food and Drug Administration. Considerations for the use of artificial intelligence to support regulatory decision-making for drug and biological products draft guidance for industry and other interested parties. U.S. Food and Drug Administration; 2025. Report No.: FDA-2024-D-4689.
European Medicines Agency. Reflection paper on the use of Artificial Intelligence (AI) in the medicinal product lifecycle. Amsterdam, The Netherlands: European Medicines Agency; 2024 Sep 9. Report No.: EMA/CHMP/CVMP/83833/2023.
World Health Organization. Ethics and governance of artificial intelligence for health: WHO guidance. Geneva, Switzerland: World Health Organization; 2021.
Gurung G, Shah R, Jaiswal DP. Software develop- ment life cycle models–A comparative study. Int J Sci Res Comput Sci Eng Inf Technol 2020; 6: 30–7.
Google LLC. Apps Script [online]. 2024 [cited Jan 22, 2024]. Available from:developers.google.com/ apps-script?hl=th.
Yang L. Responsive website design and development based on Bootstrap -- taking digital media technology professional learning resource website as an example. In: Proceedings of the 2024 International Conference on Computer and Multimedia Technology (ICCMT '24). New York: Association for Computing Machinery; 2024. p. 190-194. Available from: doi.org/10.1145/3675249.36752 84
Boonrit N, Hopkins AM, Ruanglertboon W. Assess- ment of the appropriateness of responses in Thai from ChatGPT on the questions for recommendations of drug uses in common illnesses. Thai Journal of Pharmacy Practice. 2023; 15: 1134–48.
Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics. 1977; 33: 159–74.
Likert R. A technique for the measurement of attitudes. New York: Columbia University Press; 1932.
Nunnally JC, Bernstein IH. Psychometric theory. New York: McGraw-Hill; 1994.
Halvoník D, Kapusta J. Large language models and rule-based approaches in domain-specific communi- cation. IEEE Access. 2024;12:107046-58
Rashid MRA, Rizzo G, Torchiano M, Mihindukula- sooriya N, Corcho O, García-Castro R. Completeness and consistency analysis for evolving knowledge bases. Web Semant 2019; 54: 48-71.
White J, Fu Q, Hays S, Sandborn M, Olea C, Gilbert H, et al. A prompt pattern catalog to enhance prompt engineering with ChatGPT. arXiv 2023: 2302.11382. doi.org/10.48550/arXiv.2302.11382
Huang JY, Choshen L, Astudillo R, Broderick T, Andreas J. Do LLMs benefit from their own words? arXiv[Preprint]. 2026: 2602.24287.
Anam RK. Prompt engineering and the effectiveness of large language models in enhancing human productivity. arXiv [Preprint]. 2025: 2507.18638.
Büker M, Mercan G. Readability, accuracy and appropriateness and quality of AI chatbot responses as a patient information source on root canal retreatment: A comparative assessment. Int J Med Inform 2025; 201: 105948.
