EUVIMEDEUVIMED entdecken
KommentarOpen Access · CC BY 4.0

Commentary: Integrated blockchain-deep learning approach for analyzing the electronic health records recommender system

Siwan Noh, Muhammad Firdaus, Kyung-Hyune Rhee

Frontiers in Public Health · 2023

DOI: 10.3389/fpubh.2023.1133142 ↗

Kurz gefasst

Kommentar in der Zeitschrift Frontiers in Public Health, 2023. Verfasst von Siwan Noh und 2 weiteren Autor:innen. Beteiligte Einrichtung: Department of Information Security, Pukyong National University, Busan, Republic of Korea. Hauptthemen laut MEDLINE: Blockchain, Deep Learning. Die Arbeit nennt 7 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Ministry of Science and ICT, South Korea.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Mantey et al. proposed an integrated blockchain-deep learning environment for analyzing electronic health records(EHRs) (1). They mentioned the drawbacks of the traditional EHR management system and argued that the proposed system could address these drawbacks. However, we found some crucial security issues that need to be considered. First, they burdened patients with complicated and wasteful activities by applying access control methods inappropriate for the system environment to control access to EHR data. Second, despite their goal of eliminating a dependency on a centralized system, the algorithm did not guarantee confidentiality for the system administrator. Finally, they used an additional encryption process to ensure the confidentiality of the content identifier, but this approach is impractical because it is unnecessary in both public and private IPFS networks. In this context, their method is inefficient and makes it difficult to determine whether the proposed system adequately meets users' requirements.Mantey et al. used a private IPFS network(4) and a permissioned blockchain in their system (1). They assume that all participants(i.e., patient, doctor, healthcare provider, clinician, etc.) are authenticated by the system administrator to participate in the blockchain network as a peer node, and their information is encapsulated in an X.509 certificate. However, this is inadequate to achieve the enhanced security and privacy that they actively sought in the paper. Therefore, they proposed an algorithm to guarantee the confidentiality of EHRs stored in IPFS and blockchain; however, they did not provide a detailed explanation of the algorithm in (1). Thus, we analyze their previous works(2; 3) with the same system environment and Algorithm 1 Creating and updating health records in Hyperledger blockchain Function 1. Retrieve and Re-encrypt EHR by the Administrator:Input: msk, cid C 0 Output: κ u , C 2 , cid C 1 1.If requester's ID ∈ patient p's ACL, the administrator retrieves p's EHR D i from IPFS 1) C 0 ← retrieveData(cid C 0 ) 2) D i ← decryptS(C 0 , msk) 2.Generate session key sk 3.Re-encrypt D under the new symmetric key sk 1) C 1 ← encryptS(D i , sk) 2) κ u ← encryptA(sk, puk u ) 3) cid C 1 ← uploadData(C 1 ) 4) C 2 ← encryptA(cid C 1 , puk u ) 4.Administrator send κ and C 2 to authorized users(e.g., patient and doctor) Funtion 2. Update EHR by authorized users:Input: κ u , C 2 Output: cid D i+1 1.Decrypt C 2 and retrieve C 1 from IPFS network 1) cid C 1 ← decryptA(C 2 , prk u ) 2) C 1 ← retrieveData(cid C 1 )D i 1) sk ← decryptA(κ u , prk u ) 2) D i ← decryptS(C 1 , sk) 3.Update EHR and upload new EHR 1) D i+1 ← D i + D update 2) C ′ 0 ← encryptS(D i+1 , msk) 3) cid C ′ 0 ← uploadData(C ′ 0 )reconstruct the detailed description of the algorithm based on these papers as shown in Algorithm 1. Table 19 1 shows notations used in this algorithm. First, they used a whitelist-based access control mechanism that was inappropriate for their system 27 environment. Here, whitelisting is a security approach that allows only a few trusted entities to access 28 resources, where the patient is responsible for maintaining a whitelist of entities that can access his EHR.However, applying whitelisting to a large group of dynamic participants is not appropriate because their 30 system takes into account situations in which multiple entities, such as doctors, healthcare providers, and 31 physicians, have access to EHRs. In their system, each patient must create an access control list for every 32 registered participant, which will grow significantly in size with system volume.Second, their algorithms cannot adequately ensure the confidentiality of EHR data. They argued in (1) for the risk of storing patients' sensitive EHR data in the centralized storage environment. They stored encrypted EHR data in the decentralized storage environment(i.e., IPFS network) and designed the system when only authorized users could access it. However, in our opinion, their system still relies on the centralized administrator. As shown in Algorithm 1, the administrator's master secret key msk is used to encrypt the patient's EHR, and when a request for sharing is received, the data is re-encrypted under a 39 new symmetric key sk generated by the administrator and then shared with other users through the IPFS 40 network. After the update process is completed, EHR data is encrypted under msk and stored in the IPFS 41 network. Unfortunately, this algorithm does not guarantee confidentiality to the third-party entity because 42 EHR data is exposed to the administrator during the process of running the algorithm.Thirdly, they encrypt the CID using the requester's public key puk u ; however, simply not disclosing the 44 CID does not guarantee the confidentiality of the data. For example, ipfs-search project 1 and are already protected from these search services. Since the patient's EHR data is being stored on a centralized server or several trusted servers, it is not possible to determine whether their approach has 51 solved the problems of traditional EHR management systems.We believe that further discussion about their proposed research is required, and we hope that this will contribute to their research.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

BlockchainDeep LearningElectronic Health RecordsComputer SecurityConfidentiality

Stichwörter der Autor:innen

access controlblockchain (BC)electronic health record (EHR)information securityinter planetary file system (IPFS)

Autor:innen und Institutionen

  • Department of Information Security, Pukyong National University, Busan, Republic of Korea.
  • Department of Artificial Intelligence Convergence, Pukyong National University, Busan, Republic of Korea.
  • Division of Computer Engineering, Pukyong National University, Busan, Republic of Korea.

Förderung

  • Ministry of Science and ICT, South Korea

Zitieren

APA

Noh, S., Firdaus, M., & Rhee, K. H. (2023). Commentary: Integrated blockchain-deep learning approach for analyzing the electronic health records recommender system. Frontiers in Public Health. https://doi.org/10.3389/fpubh.2023.1133142

Vancouver

Noh S, Firdaus M, Rhee KH. Commentary: Integrated blockchain-deep learning approach for analyzing the electronic health records recommender system. Frontiers in Public Health. 2023. doi:10.3389/fpubh.2023.1133142

BibTeX
@article{Noh2023,
  author = {Noh, Siwan and Firdaus, Muhammad and Rhee, Kyung-Hyune},
  title = {Commentary: Integrated blockchain-deep learning approach for analyzing the electronic health records recommender system},
  journal = {Frontiers in Public Health},
  year = {2023},
  doi = {10.3389/fpubh.2023.1133142}
}
Kostenlos starten →