Preserving Electronic Health Records using Federated Learning

[Audio] Preserving Electronic Health Records using Federated Learning Presented By: Muhammad Fahad Qureshi ECI-IT-20-051 Faizan Ali Afsar ECI-IT-20-052 Project Supervisor: Miss Sakha Project Co-Supervisor: Dr. Shaheer Muhammad Department of Computing Faculty of Engineering Science & Technology Hamdard University Islamabad Campus.

[Audio] Overview of Presentation Introduction Aims & Objectives Problem Statement Literature Review Project Scope Methodology Functional & Non-Functional Requirements Tools & Technology Project Proposed Time Schedule References Preserving Electronic Health Records using Federated Learning.

[Audio] Introduction(1/2) Electronic Health Records (EHRs): Electronic Healthcare Record (EHR) systems were introduced to store patient records electronically in a database so that they could be accessed from anywhere and by anyone, with granted access, using the Internet. Data Breaching in EHRs: A data breach happens when unauthorized individuals or entities access sensitive and private data cloud user , which result in the theft of private data. Data breaches in electronic health records happens when someone unauthorized access to sensitive patient information stored in cloud. Breaches can happen due to weak security measures such as weak passwords, outdated software, or lack of encryption. Employees at healthcare organizations can access and leak information about patients accidentally or for some other purpose, This can lead to breaches. Preserving Electronic Health Records using Federated Learning.

[Audio] Introduction(2/2) Security and privacy Security and privacy are two crucial issues in the protection of health information. The purpose is to keep the medical privacy of confidential information about the patient. Privacy: How you allow your personal information to be accessed and viewed. Security :Protection of your private data and information. The main difference between privacy and security is that privacy involves how your data is used and controlled, while security protects this data. Federated Learning Federated learning is a distributed machine learning that enables machine learning models to be trained on decentralized data sources without compromising privacy or security. This allows the server to aggregate the models and update the global model while preserving the privacy and security of the data. Preserving Electronic Health Records using Federated Learning by allowing multiple parties to train models on their local data while sharing the learned parameters with the central serve.

[Audio] Aims & Objectives Aim: The aim is study to introduce the methods for determining the data leaks and implements state of the art data privacy and security technique. Objectives: Preserving Patient Privacy One of the primary objectives of implementing Federated Learning in electronic health records (EHR) is to enhance and safeguard patient privacy. Securing Sensitive Health Data. The security of electronic health records is crucial to prevent unauthorized access and protect patient confidentiality..

[Audio] Problem Statement The Electronic Health Record (EHR) system faces significant security and privacy challenges, risking sensitive patient data exposure. The existing EHR systems lack in security measures, posing a serious threat to patient data confidentiality. Federating learning is distributed Machine learning model can be implemented on decentralized data sources However implementation of Federated learning can be challenge..

[Audio] Literature Review(1/4) In 2010 (Zhang, et al.) proposed a authentication technique for Healthcare Application Clouds. It is important to any organization since it protects user identities, allows permission to use network resources, and verifies that a user is really who he is pretending to be. We illustrate the development of the EHR security reference model through a use-case scenario and describe the corresponding security countermeasures and state of art security techniques that can be applied as basic security guards [1]. In 2016 (N. M. Shrestha, et al.) proposed a secured e-health framework. In this framework, patient centric personal data and access control scheme with enhanced encryption method has been considered. They propose methodology of identified the personal health information by digital signature and patient pseudo identity. The SPSS tool was used to analyze the data. [2]. In 2019 (Hathaliya, et al.) proposed a model on securing electronics health record, and many security risks and difficulties were found when trying to access the EHR from the database store They propose methodology of biometric-based authentication scheme to ensure secure access of the patients EHR from any location. The secure biometric-based scheme is designed which is validated using the Automated Validation of Internet Security Protocols and Applications (AVISPA) tool [3]..

[Audio] Literature Review (2/4) In 2023 (Almalawi, et al.) proposed a novel Lionized remora optimization-based serpent (LRO-S) encryption method to encrypt sensitive data and reduce privacy breaches and cyberattacks from unauthorized users and hackers. The LRO-S method is the combination of hybrid metaheuristic optimization and improved security algorithm. The LRO-S technique encrypts sensitive patient data before storing it in the cloud. The primary goal of this study is to improve the safety and adaptability of medical professionals’ access to cloud-based patient-sensitive data more securely [5] In 2020 (Hathaliya, et al.) proposed a mobile-based healthcare system for the Healthcare 4.0 era. The proposed approach enables patients to self-authenticate, using their own mobile and wearable devices, establishing a session key between owned devices. After mutual authentication, the cloud server verifies each user. The security of the approach is described through the AVISPA tool. We analyze the embedded security and functionality which improves the healthcare system. The Results show that the proposed approach provides greater security compared to other state-of-the-art schemes [6]. In 2021 (Mothukuri , et al.) have comprehensive study on concerning FL’s security and privacy aspects that can help bridge the gap between the current state of federated AI and a future in which mass adoption is possible. We present an illustrative description of approaches and various implementation styles with an examination of the current challenges in FL and establish a detailed review of security and privacy concerns that need to be considered in a thorough and clear context.. The most specific security threats currently are communication bottlenecks, poisoning, and backdoor attacks while inference-based attacks are the most critical to the privacy of FL [7].

[Audio] Literature Review (3/4) In 2022 (TK Dang,, et al.) data-driven medical research, multi-center studies have long been preferred over single-center ones due to a single institute sometimes not having enough data to obtain sufficient statistical power for certain hypothesis testing’s as well as predictive and subgroup studies. The wide adoption of electronic health records (EHRs) has made Mult institutional collaboration much more feasible. However, concerns over infrastructures, regulations, privacy, and data standardization present a challenge to data sharing across healthcare institutions. Federated Learning (FL), which allows multiple sites to collaboratively train a global model without directly sharing data, has become a promising paradigm to break the data isolation. In this study, we surveyed existing works on FL applications in EHRs and evaluated the performance of current state-of-the-art FL algorithms on two EHR machine learning tasks of significant clinical importance on a real-world multi-center EHR dataset.[8] In 2020 (Li, et al.) Federated learning involves localizing data while training statistical models across dispersed devices or isolated data centers, such cell phones or hospitals. Training in diverse and perhaps large-scale networks presents new difficulties that need for a fundamental shift from conventional methods for distributed optimization, large-scale machine learning, and privacy-preserving data analysis. In this paper, we address the special qualities and difficulties of federated learning, give a general review of existing methods, and suggest a number of future research topics that will be of interest to many research communities [9]..

[Audio] Literature Review (4/4) In 2020 (Xu, et al.), a growing amount of healthcare data is becoming easily accessible from clinical facilities, patients, insurance companies, and the pharmaceutical industry, among others, due to the quick development of computer software and hardware technologies. This Entry gives data science technologies a previously unheard-of chance to gain data-driven insights and raise the standard of care delivery. However, the majority of healthcare data are private and fragmented, which makes it challenging to provide reliable findings across demographics. For instance, the electronic health records (EHR) of various patient populations are owned by different institutions, and because to the sensitive nature of these records, sharing them between hospitals is challenging. This poses a significant obstacle to the development of generalizable, effective analytical techniques that require diverse, "big data." A method of developing a shared global model with federated learning [10]..

[Audio] Project Scope This project's scope includes the design and implementation of security and privacy features in Electronic Health Record (EHR) systems using federated learning. It encompasses the development of a federated learning model, integration with existing EHR infrastructure, and compliance with healthcare data protection regulation as per HIPAA..

[Audio] Methodology Data Acquisition: Collect a dataset of Electronics Health Records like Patients information, Covid datasets, diabetes datasets etc. Data Preprocessing: Handle missing values and remove duplicates in the dataset. Federated Learning Framework Selection: Choose a suitable federated learning framework, such as TensorFlow Federated or PySyft, based on the requirements of the EHR system. Data Encryption: Implement end-to-end encryption for data at rest and during transmission between the central server and individual nodes. Model Evaluation: Evaluate the trained model's performance on the testing dataset..

[Audio] Methodology Diagram Dataset Data Preprocessing Security and Privacy Federated Learning Model Evaluation Testing Preserving Electronic Health Records using Federated Learning.

[Audio] Functional / Non Functional Requirements Functional Requirements User Authentication and Authorization Encrypted Data Transmission Adaptive Security Measures (HIPAA) Non-Functional Requirements Performance Security Maintainability Preserving Electronic Health Records using Federated Learning Functional Requirements: 1. Authorization and Authentication by users function similarly to two locks on a door. Authentication functions similarly to a key that opens a door by confirming that you are who you claim you are. Authorization determines what you can and cannot do while inside, allowing you to enter just the rooms that you are authorized to enter. 2. Encrypting data transmission is similar to having a special code to secure information as it moves between devices when it comes to protecting the security and privacy of Electronic Health Records during federated learning on the cloud. In doing so, the patient records are protected from strangers and remain hidden. The Health Insurance Portability and Accountability Act, or HIPAA, is a US rules that helps maintain the security and privacy of your medical records. It provides you more choice over who can access your personal health data and establishes guidelines for how healthcare providers must secure it. 3. Adaptive security measures under HIPAA, is the use of smart and adaptable methods to safeguard your personal health information, modifying and upgrading as necessary to meet new threats and challenges. Non-Functional Requirements: 1.Performance: To fulfill the real-time goals of healthcare professionals using electronic health records, design the system with maximum speed and responsiveness in mind. 2.Security: It is the care and maintenance of things safety and protection. When it comes to keeping Electronic Health Records secure, this means making sure that data is protected from threats and unwanted access, kept private, and available to authorized personnel only. 3.Maintainability: means making something simple to take care of and keep in great condition. Within the setting of a framework or project, it includes planning it in a way that creates it straightforward to update, fix, and manage over time without causing disturbances or complications..

[Audio] Tools/ Technology Software Requirements Operating system: Windows 10 Coding Language: Python Compiler: Google Colab Hardware Requirements Laptop: Core i5 /Processor /GPUs Hardware: SSD 256GB Ram: 8 GB Size:15”.

[Audio] Project Proposed Time Schedule Activity Jan/Feb 2024 Mar 2024 Apr 2024 May 2024 Jun 2024 Jul 2024 Jul/Aug 2024 Aug 2024 Collection of Literature ✓ Study of Literature ✓ Analysis of Proposed Scheme ✓ Preparation of Schemes/Model ✓ Implementation of Schemes/Model ✓ ✓ Analysis and Simulation ✓ Result Formulation ✓ Final Write-up and Thesis Submission ✓.

[Audio] Budget Description Personnel & Resources: Covering researcher costs, datasets, Health Records and essential resources. Technology Infrastructure: Expenses for computing equipment, software licenses. Publication & Dissemination: Budget for publishing in journals and sharing findings in workshops..

[Audio] References (1/2) [1]. Hathaliya, Jigna J., et al. "Securing electronics healthcare records in healthcare 4.0: A biometric-based approach." Computers & Electrical Engineering 76 (2019): 398-410. [2]. Rahman, Mohamed Abdur, et al. "Secure and provenance enhanced internet of health things framework: A blockchain managed federated learning approach." Ieee Access 8 (2020): 205071- 205087. [3]. M. Puppala, T. He, X. Yu, S. Chen, R. Ogunti and S. T. C. Wong, "Data security and privacy management in healthcare applications and clinical data warehouse environment," 2016 IEEE-EMBS International Conference on Biomedical and Health Informatics (BHI), Las Vegas [4]. Sweeney, Latanya. "Achieving k-anonymity privacy protection using generalization and suppression." International Journal of Uncertainty, Fuzziness and Knowledge-Based Systems 10.05 (2002): 571-588. [5]. Almalawi, Abdulmohsen, et al. "Managing Security of Healthcare Data for a Modern Healthcare System." Sensors 23.7 (2023): [6]. Hathaliya, Jigna J., Sudeep Tanwar, and Richard Evans. "Securing electronic healthcare records: A mobile-based biometric authentication approach." Journal of Information Security and Applications 53 (2020): 102528..

[Audio] References (2/2) [7]. Wang, C-K. "Security and privacy of personal health record, electronic medical record and health information." Problems and perspectives in management 13, [8]. Dang, Trung Kien, et al. "Federated learning for electronic health records." ACM Transactions on Intelligent Systems and Technology (TIST) 13.5 (2022): [9]. Mothukuri, Viraaji, et al. "A survey on security and privacy of federated learning." Future Generation Computer Systems 115 (2021): 619-640. [10]. Li, T., Sahu, A.K., Talwalkar, A. and Smith, V., 2020. Federated learning: Challenges, methods, and future directions. IEEE signal processing magazine, 37(3), pp.50-60. [11]. Xu, Jie, et al. "Federated learning for healthcare informatics." Journal of Healthcare Informatics Research 5 (2021): 1-19..

Video undefined. [Audio] Video. Video. Preserving Electronic Health Records using Federated Learning.