51 J Global Clinical Engineering Vol.6 Special Issue 6: 2024 Conference Paper Leveraging Web Scraping and API Integration for Efficient Medical Device Data Management Agapi Konstantina Liontou1,*, Spilios Zisimopoulos2 and Aris Dermitzakis1 1 Biomedical Technology Unit, Department of Medical Physics, School of Medicine, University of Patras, Patras, Greece. 2 Institute of Biomedical Technology, Patras, Greece. * Corresponding Author Email: liontou.kon@gmail.com ABSTRACT Accurate identification and management of medical devices is of particular importance to ensure patient safety and regulatory compliance within healthcare systems. This paper presents a comprehensive exploration of medical device data retrieval, focusing on the integration of web scraping and Application Programming Interface (API) technologies. The utilization of Unique Device Identifiers (UDIs) and the Global Medical Device Nomenclature (GMDN) system is emphasized to enhance device authentication, attribute verification, and accurate categorization. This paper introduces a state-of-the-art code implementation that combines web scraping techniques and API integration to address the challenges of retrieving and verifying device information. The code facilitates both access to data and healthcare professionals and stakeholders to make informed decisions based on reliable and up-to-date information. This is a significant and defining advance in the field, offering a powerful solution that is innovative as well as vital. The paper concludes by discussing the potential impact of these developments on patient safety, regulatory compliance, and the overall advancement of healthcare technology. In addition, the importance of accurate device identification, the role of UDIs and GMDN, and the significance of the provided cutting-edge code are highlighted, providing valuable insights into the field of medical device data retrieval. Keywords—Web scraping, Medical device data management, API integration. Copyright © 2024. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY): Creative Commons - Attribution 4.0 International - CC BY 4.0. The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduc- tion is permitted which does not comply with these terms. http://www.globalce.org http://globalce.org http://globalce.org mailto:liontou.kon@gmail.com https://creativecommons.org/licenses/by/4.0/ https://creativecommons.org/licenses/by/4.0/ J Global Clinical Engineering Vol.6 Special Issue 6: 2024 52 INTRODUCTION In today’s data-driven landscape, the demand for effec- tive data extraction techniques is critical. Patient safety, regulatory compliance, and informed decision-making hinge on the acquisition of accurate and up-to-date in- formation about medical devices. However, obtaining this information can be a daunting task, given the disparate origins and formats of available data. This paper examines the challenges associated with medical device data management and introduces a com- prehensive methodology, proposing a combination of web scraping methods with the integration of Application Programming Interfaces (APIs), with a particular focus on two key elements: Unique Device Identifiers (UDIs) and Global Medical Device Nomenclature (GMDN) terms, pivotal for the identification and categorization of medi- cal devices. Our objective is to present an integrated ap- proach for gathering, validating, and employing medical device data from the AccessGUDID (Global Unique Device Identification Database), which acts as the authoritative source of device information, laying the foundation for our methodology. In addition, our methodology relies on the web-Praxis Medical Equipment Management Soft- ware (MEMS), developed by the Institute of Biomedical Technology (INBIT)1, whose data can be accessed by appropriate healthcare unit users. A dataset of medical device information was provided for the purpose of this study. AccessGUDID database on the other hand is freely accessible. The significance of swift medical equipment (ME) identification in today’s fast-paced healthcare environment cannot be overstated. Particularly in the context of recalls and field safety notices (FSNs) issued by manufacturers, which contain vital information about affected device types, rapid identification plays a crucial role. Clinical engineers are entrusted with the immediate and appro- priate response to these notices, ensuring the correct course of corrective actions or equipment withdrawals to prevent adverse events. To address this pressing need, our developed software serves a dual purpose. It aids in the validation of ME within a hospital’s inventory, ensuring precise matching with the corresponding UDI codes. UDI serves as a standardized system for identifying and tracing medical devices Addi- tionally, it facilitates the accurate classification of ME into GMDN groups, a critical facet of the UDI system. GMDN serves as a standardized system for classifying medical devices worldwide. It is noteworthy that recalls and FSNs frequently include UDIs for affected devices, emphasiz- ing the growing importance of UDI-based traceability. Moreover, according to MDR 2017/7452, the traceability of devices by means of a UDI system should significantly enhance the effectiveness of the post-market safety-related activities for devices. To address the complexities of gathering and verifying device-related data, this paper discloses an innovative code implementation. We combine web scraping techniques with AccessGUDID (Global Unique Device Identification Database) API integration, presenting a powerful solution. Our code adeptly extracts imperative device details from the AccessGUDID website and validates this informa- tion with data provided by web-Praxis. Additionally, our code integrates seamlessly with the AccessGUDID API, simplifying the retrieval of device-specific information dependent on UDIs, ensuring compliance with data usage regulations, and increasing the reliability and accuracy of the retrieved data. In summary, this paper delves into the intricate realm of medical device data management, leveraging advanced techniques for data extraction and integration. Our inte- grated approach not only promises to enhance patient safety and regulatory compliance but also equips healthcare professionals with the tools needed for informed decision- making in an increasingly dynamic healthcare landscape. BACKGROUND In the modern era, accurate and timely management of medical device data is essential for various stakehold- ers in the healthcare industry. In this section, we talk about the background of the key elements of this paper’s methodology: Unique Device Identifiers (UDIs) and Global Medical Device Nomenclature (GMDN), in addition to an overview of web scraping and Application Programming Interface (API) integration. http://www.globalce.org http://globalce.org http://globalce.org 53 J Global Clinical Engineering Vol.6 Special Issue 6: 2024 APIs, on the other hand, serve as intermediaries between software applications, enabling seamless com- munication and data exchange. Web Scraping vs. API Web scraping, with its capability to access data from multiple web pages and repositories, excels in collect- ing large volumes of heterogeneous data efficiently. It offers flexibility in data collection and analysis, making it a valuable tool in domains such as computer vision and natural language processing. APIs, on the other hand, provide a structured and controlled means of accessing specific data from applica- tions or software. While they offer advantages such as standardized interfaces and faster data extraction, they also come with limitations in terms of functionality and access to a single website or predefined functions.5, 7, 8 METHODS AND MATERIALS Python for Web Scraping Python is a highly favored choice for web scraping due to several key reasons. Firstly, Python is a very popular programming language because of its simplicity and learnability, facts that make it accessible even to those with little programming experience. Its extensive com- munity support ensures readily available assistance for overcoming coding challenges. Furthermore, Python excels in web scraping because of the readability of the code. Python code is designed for easy understanding, promoting clear and concise pro- gramming practices. This readability not only enhances productivity during development but also facilitates code understanding, even when revisiting it after some time has passed. This attribute promotes more efficient code maintenance and facilitates code reuse.8 For these reasons, Python is used in all the software we developed. Our Code We present an integrated code solution that combines web scraping via Beautiful Soup (Version 4, Python Li- brary for HTML and XML Parsing, Crummy, Cambridge, MA, USA) with AccessGUDID API integration. Tailored Unique Device Identifiers (UDIs) Unique Device Identifiers or UDIs, are alphanumeric codes for medical devices, offering a standardized global means of identification. They enable precise tracking from manufacturing to post-market surveillance. UDIs encode key device information like manufacturer, model, and production date. Their implementation has significantly advanced healthcare, enhancing patient safety, regulatory compliance, and supply chain management. UDIs empower healthcare professionals to quickly access device data, identify recalls, and respond to adverse events effectively. 3 Global Medical Device Nomenclature (GMDN) The Global Medical Device Nomenclature (GMDN) is an internationally recognized system for naming and categorizing medical devices. It offers a structured clas- sification system that facilitates global communication in healthcare. GMDN codes categorize devices based on their purpose, structure, and operation, enabling precise comparisons. GMDN adoption has streamlined regula- tions, research, and product development in the medical device industry. It promotes consistency in terminology and categorization, as well as seamless compatibility and cooperation among healthcare stakeholders for sharing accurate device information.4 Web Scraping and API Integration In the contemporary era, extracting data from various web sources has become essential for various domains, including healthcare. In this section, a foundational un- derstanding of the main data extraction methods: web scraping and Application Programming Interfaces (APIs), is provided. Web scraping, also known as web extraction or harvesting, involves the automated extraction of data from websites and their subsequent storage for analysis or retrieval. This method, widely recognized for its efficiency and accuracy, has evolved significantly over the years. Modern web scraping tools have become versatile, capable of parsing markup languages, integrating with computer vision and natural language processing techniques, and simulating human browsing behavior.5,6 http://www.globalce.org http://globalce.org http://globalce.org J Global Clinical Engineering Vol.6 Special Issue 6: 2024 54 for healthcare professionals, this tool simplifies access, validation, and analysis of device data based on UDIs and GMDN names while ensuring data compliance. A key aspect of our methodology involves cross-refer- encing data derived from web scraping with information from Praxis, our primary data source. This information is structured in an Excel file and has been collected by humans, checked, and entered into the Praxis database. This quality control step identifies disparities in device attributes, improving data accuracy. For example, dis- crepancies in GMDN names trigger further investigation to address updates or errors. Our comprehensive validation process involves veri- fying GMDN names, selecting the most prevalent name when multiple names exist, and aligning device attributes across sources, ultimately enhancing the reliability of medical device data. Data Import To begin data retrieval, the code begins by importing data from an Excel file provided by Web Praxis. This file includes critical information about medical devices and the specific columns of interest include General Group, Specific Group, Manufacturer, Model, and Comments, where Unique Device Identifiers (UDIs) are often found. The initial Excel file includes 279 records, representing distinct medical devices. UDI Extraction Next, the code extracts the 14-digit UDIs from the appropriate field (if any because not all records have a registered UDI). This step is crucial for subsequent op- erations as UDIs serve as the primary key for accessing device information. Web Scraping and API Integration The heart of the methodology lies in its ability to com- bine web scraping and API integration for comprehensive data collection. Web scraping extracts data based on UDIs, while API calls are made to the AccessGUDID database to retrieve detailed device information. This hybrid approach ensures that even devices without readily available UDIs can be identified and analyzed. Data Validation and Presentation As data is retrieved, the code simultaneously validates and cross-references it to ensure accuracy and reliability. The data are then organized and presented in a structured format for further analysis and reporting. Data Cross-Referencing For enhancing data reliability, the code employs cross- referencing. It verifies device attributes like GMDN terms by comparing data from Praxis with data obtained through web scraping or API integration. This process ensures consistency across multiple sources and is a vital quality control step. The aforementioned accomplishments were realized through the utilization of the following libraries: requests for making HTTP requests to interact with web services and retrieve data from web servers, pandas for data manipulation, tkinter for creating a GUI, re for regular expressions, time for time-related operations, and Beauti- ful Soup for web scraping purposes to pull the data out of HTML and XML files, as it helps in navigating, searching, and modifying the parse tree. Beautiful Soup creates a parse tree from page source code that can be used to extract data easily. RESULTS The results were quite satisfactory. Out of the 279 records initially provided, we found UDI matches for 193 records (69% success rate). By implementing a complex search based on the company name and brand name (this was done because there were cases that we did not have a UDI—device type match a priori), we were able to identify 14 more records which means 207 records in total (74% success rate). The output generated by the code is presented in a structured Excel format (Figure 1), providing a compre- hensive overview of matched UDIs, associated company names, brand names, GMDN names, cross references, and definitions. http://www.globalce.org http://globalce.org http://globalce.org 55 J Global Clinical Engineering Vol.6 Special Issue 6: 2024 demonstrate a fairly accurate recording of medical equip- ment, with correct UDIs, GMDN names, and definitions. Matching all 279 medical records cannot be possible due to inaccuracies inherent in the databases. This creates issues for our model, often leading to many closely related outcomes. Even for a human observer, distinguishing between these outcomes is extremely challenging. So, reaching a perfect 100% match rate is impossible and would demand a lot of human effort from an experienced eye, potentially even reviewing unrelated results. Hence, a 74% match rate is considered satisfactory under these circumstances. Overall, the output provides a comprehensive overview of the matched UDIs, associated company names, brand names, GMDN names, cross-references, and definitions, allowing for further analysis and verification of the re- corded medical equipment data. In general, the implemented code combining web scraping techniques and integration with the AccessGU- DID API has proven to be an essential tool for retrieving, validating, and analyzing medical device data based on UDIs and Global Medical Device Nomenclature (GMDN) names. The need for web scraping arises from the vast amount of device information available on websites like Access- GUDID. Web scraping enables the efficient extraction of specific data elements, such as company names, brand names, GMDN names, cross-references, and definitions, from complex HTML structures. By automating the data retrieval process, this method provides instant access to the most up-to-date device information, ensuring ac- curacy and timeliness. Integrating an API (in this case, the AccessGUDID API) further enhances the code’s functionality and reliability. By leveraging the API, the code establishes a secure con- nection to the comprehensive device database provided by AccessGUDID, ensuring compliance with data usage policies, facilitating seamless data retrieval, and enhanc- ing the reliability and accuracy of the obtained device information. Moreover, the use of UDIs plays a crucial role in device identification and traceability. UDIs provide a standardized system for the unique identification of medical devices, ensuring many benefits for both patient and equipment safety. The code’s ability to cross-check device data with Where: 1st column: UDI, UDIs are listed. It includes both the UDIs that were initially matched and the “N/A” symbol for equipment that had no UDI in the initial data and required an advanced search. 2nd column: Company Name as registered on the website. 3rd column: Brand Name as registered on the website. 4th column: GMDN Name as registered on the website. 5th column: GMDN Cross Reference. The term “same” is displayed for records that have the same GMDN Name on both the website and the initial Excel data. For records that are not the same, a different description is provided. This description was obtained from the initial data. 6th column: GMDN Definition as registered on the website for the devices that were “identified” by the UDI. For the devices that we did not know the UDI beforehand, and we made an advanced search, we had many results. So, the GMDN definition was obtained by examining the first page of search results on the website generated by the advanced search. We identified all the GMDN names and applied a sorting algorithm to select the names that appeared most frequently. FIGURE 1. Format of the output Excel File. DISCUSSION The high success rates in both types of searches in- dicated that the data provided by web-Praxis, although challenging to manage due to its unstructured format, http://www.globalce.org http://globalce.org http://globalce.org J Global Clinical Engineering Vol.6 Special Issue 6: 2024 56 UDIs and GMDN names ensures the authenticity of devices, verifies their attributes, and allows for accurate catego- rization. This cross-checking process allows healthcare professionals to identify and address potential discrep- ancies, ultimately ensuring patient safety and improving informed decision-making. CONCLUSION In conclusion, the combination of web scraping, API integration, and the use of UDIs addresses the challenges of accessing, validating, and analyzing medical device data. The implemented code streamlines these processes, enabling users to efficiently retrieve reliable device in- formation, verify device characteristics, and eventually make informed decisions. The code can be periodically used to ensure the fidelity of UDI codes and GMDN group classification of ME in a medical equipment inventory, to address the dynamic nature of the above systems, and to update the information for new ME types added to the inventory. As technology evolves, more and more advancements in web scraping and API integration will contribute to even more efficient and accurate device data management and analysis in the healthcare industry. Μany suggestions can be taken into account for even greater effectiveness of the solution we proposed. Firstly, validating and cleaning the input data is significant for ensuring data quality in general. Additionally, automating code execution as well as batch processing capabilities can improve efficiency, especially when dealing with large volumes of data. Moreover, integration with healthcare or inventory management systems is significant, in order to synchronize data and improve decision-making capabilities. Collaborations and partnerships with regulatory bod- ies, healthcare institutions, or manufacturers can facilitate data sharing and drive industry-wide improvements in device identification and data management practices. Moreover, continuous data monitoring is recommended both to periodically retrieve and update device informa- tion from reliable sources so as to maintain data accuracy and relevance. Finally, comprehensive documentation and user support materials is necessary to be provided, as they would help users to use the code effectively and maximize its potential. ACKNOWLEDGMENTS We would like to acknowledge the invaluable contribu- tion of the web-Praxis software, as a primary data source for this research. REFERENCES 1. INBIT. Available online: https://www.inbit.gr/en/. 2. Official Journal of the European Union. The European Parliament and The Council of the European Union; 2024. Available online: https://eur-lex.europa.eu/ legal-content/EN/TXT/?uri=CELEX%3A32017R0745. 3. Unique Device Identification System (UDI System), FDA. Available online: https://www.fda.gov/medical-devices/ device-advice-comprehensive-regulatory-assistance/ unique-device-identification-system-udi-system. 4. GMDN. Available online: https://www.gmdnagency.org/. 5. Sathya, A.C., Rao, P. Mastan, Babu, S. Development of a Covid-19 Information Dashboard to Access the Health Care Resources and Requirements Online. In 2022 3rd International Conference on Computing, Analytics and Networks (ICAN), Rajpura, Punjab, India; IEEE; 2022; pp. 1–8. https://doi.org/10.1109/ ICAN56228.2022.10007167. 6. Zhao, B. Web Scraping. In Encyclopedia of Big Data. Laurie A. S., Connie L. M. Springer: Cham, Switzerland; 2017;pp. 1–3. 7. Web Scraping vs API: What’s the Difference and Similar- ity. 2023 Available online: https://www.scrapingdog. com/blog/web-scraping-vs-api/. 8. Khder, M.A. Web scraping or web crawling: State of art, techniques, approaches and application. Int J Adv Soft Computing Appl. 2021;13(3):144–168. https:// doi.org/10.15849/IJASCA.211128.11. http://www.globalce.org http://globalce.org http://globalce.org https://www.inbit.gr/en/ https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A32017R0745 https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A32017R0745 https://www.fda.gov/medical-devices/device-advice-comprehensive-regulatory-assistance/unique-device-identification-system-udi-system https://www.fda.gov/medical-devices/device-advice-comprehensive-regulatory-assistance/unique-device-identification-system-udi-system https://www.fda.gov/medical-devices/device-advice-comprehensive-regulatory-assistance/unique-device-identification-system-udi-system https://www.gmdnagency.org/ https://doi.org/10.1109/ICAN56228.2022.10007167. https://doi.org/10.1109/ICAN56228.2022.10007167. https://www.scrapingdog.com/blog/web-scraping-vs-api/ https://www.scrapingdog.com/blog/web-scraping-vs-api/ https://doi.org/10.15849/IJASCA.211128.11 https://doi.org/10.15849/IJASCA.211128.11 Editor’s Corner Biomedical Technology and Clinical Engineering in Greece after the Pandemic: Highlighted Works from the Panhellenic Conference of Biomedical Technology Aris Dermitzakis1,2,*, Vasiliki Zilidou1,3, Eleftheria Vellidou1,4, Alkinoos Athanasiou1,3 Digital Transformation Management in Health Services: Health Professionals Perceptions as an Implementation Factor Theodoros S. Tanis*, Chryssoula Chatzigeorgiou, Ioanna Simeli, and Evangelia Stalika Validating the ID-GAMING e-Training Toolkit for People with Intellectual Disabilities in Greece Niki Pandria*, Anastasia Barboudi, Vasileia Petronikolou, Panagiotis Antoniou and Panagiotis D. Bamidis Novel Functional Electrical Stimulation Parameter Optimization for Neurorehabilitation Using Both Conventional and AI Techniques Arsenios Arsenidis1, Alexandros Moraitopoulos2, Alkinoos Athanasiou2, Alexandros Vildiridis3, Panagiotis Bamidis2, Petros Stefaneas4 and Alexandros Astaras5 Leveraging Web Scraping and API Integration for Efficient Medical Device Data Management Agapi Konstantina Liontou1,*, Spilios Zisimopoulos2 and Aris Dermitzakis1 Human Muscle State Machine Using Electromyography Classification with Machine Learning George Lyssas1,*, Konstantinos Mitsopoulos1, Dimitris Zantzas2, Anestis Kalfas2, Panagiotis D. Bamidis1 Kinematic and Dynamic Analysis of Lower Limb Movement: Towards the Design of a Wearable Rehabilitation Assistant Device Filippos Margaritis1,*, Konstantinos Mitsopoulos1, Kostas Nizamis2, Alkinoos Athanasiou1 and Panagiotis D. Bamidis1 A Novel Dermatological Diagnosis Support Device Based on Electrical Impedance Spectroscopy Alexandros Moraitopoulos1,*, Konstantinos Mitsopoulos1, Christina Kemanetzi2, Panagiotis Bamidis1 and Alexandros Astaras3 Software Skills Identification: A Multi-Class Classification on Source Code Using Machine Learning Dimitris Bamidis, Ilias Kalouptsoglou, Apostolos Ampatzoglou, Alexandros Chatzigeorgiou* Improvement of Aortic Valve Stenosis Classification in Patients Through Computational Fluid Dynamics Model Ioannis Makropoulos, Dimitris Zantzas, Vasilis Gkoutzamanis, Anestis Kalfas* Kinematic and Dynamic Analysis of the Human Hand’s Articulation for Wearable Soft-Robotic Device Applications Paschalina-Danai Sarra, Vasiliki Fiska, Konstantinos Mitsopoulos, Diamanto Mylopoulou, and Panagiotis D. Bamidis* Deep Learning Classification of Epileptic Magnetoencephalogram Andreas Stylianou1, Lefteris Koumakis2, Maria Hadjinicolaou3, Adam Adamopoulos1,* and Alkinoos Athanasiou4