Lupine Publishers | Current Trends in Computer Sciences & Applications
Abstract
Medical-hospital records produced by computational management systems must ensure confidentiality, integrity, and availability of information. Literature studies point to the Blockchain technology as a promising candidate to accomplish these needs. Among the various functionalities existing in Blockchain-based systems, one of them is the consensus mechanism. Within this context, this article proposes enhancing the H-BFT consensus algorithm. The operation of this new algorithm is demonstrated by exploiting two case studies which briefly address the Brazilian SUS and the US Medicare-Medicaid, respectively. At last, general considerations and suggestions for future work close this article.
Introduction
The systems responsible for managing medical information records from patients must allow adequate conditions for storage and analysis, providing subsidies for a better diagnosis and treatment [1-3]. These systems are complex [4,5] and must implement requirements to adequately provide integrity, confidentiality, and availability of medical records, besides processing large data volumes [6]. Blockchain technology shows great promise in the implementation of systems capable of dealing with common problems in the field of healthcare management [7- 9]. It is defined as a distributed database, where records are stored in blocks [10], embedding elements such as cryptography, ledger, and consensus algorithms [11], besides owning properties such as immutability, integrity, transparency, availability, decentralization, and disintermediation [12]. Systems developed on Blockchainbased platforms need to establish an agreement between network nodes regarding the validity of transactions then carried out. The consensus occurs from the execution of algorithms that allow communication between the nodes of the Blockchain network, allowing unknown elements, or even competitors, to reach to an agreement regarding the previous and current state of the stored data. Within this context, this article proposes an enhancement of the H-BFT consensus algorithm [13] for use in healthcare management systems. Our proposal is the result of a theoretical study on the most important requirements for healthcare management systems based on Blockchain. The operation of this new algorithm is demonstrated by exploiting two case studies which briefly address the Brazilian SUS [14-16] and the US Medicare-Medicaid [17-19], respectively. The remainder of this article is structured as follows. Section “Literature Proposal” succinctly reviews the H-BFT algorithm. Section “Novel proposal” explains the enhancements we herein propose to the H-BFT algorithm, named as EH-BFT. Section “Performance Analysis” presents an overall discussion to highlight the benefits of the enhanced version of the H-BFT algorithm, by especially delving into two real case studies. This enhancement is named as EH-BFT algorithm. At last, Section “Conclusions and Future Work” present final remarks and gives directions for further research.
Literature Proposal
This Section brings the H-BFT protocol’s motivations, briefly explaining its characteristics and advantages of its use in healthcare management systems.
Healthcare Management System based on Blockchain Requirements
In order to understand the biggest concerns in the development of solutions that could meet the needs of a healthcare management system based on Blockchain technology, a plethora of studies were carried out (e.g., [20-26] in which it was found that there was no definition regarding the type of consensus mechanism that could be used, but some characteristics were present in almost all of them. One of those characteristics was the use of private Blockchain networks, resulting from the need to restrict access to this type of information, valuing the privacy of data produced by medical diagnoses. Another feature was the preference for deterministic consensus algorithms over probabilistic models, for reasons such as the number of participants involved and lower computational cost. With the common characteristics identified in the frameworks aimed at healthcare management, five essential requirements were listed, which guided the specification of the H-BFT: confidentiality, integrity, availability, scalability, and security. Confidentiality is the guarantee of protection against undue access to information [27] and was the main focus of concern, both due to constant cyberattacks and legal issues in countries such as Brazil [28,29] and the USA [30].
Another requirement of great importance was that of integrity, which is linked to the impossibility of changing data by unauthorized individuals, in order to avoid losses due to damage to this asset [31]. The concept of availability refers to the guarantee of access to data by people authorized to do so, whenever necessary, and the system needs to guarantee the continuity of its services and timely responses [8]. Scalability can be defined as the ability of a system to preserve quality in the delivery of its services, even if there is an increase in the number of customers [32], and security pivotally relates to protect the system against any type of attack. For instance, the implementation of a system based on Blockchain technology needs to establish implementations that seek to mitigate already identified vulnerabilities, such as those that allow exploitation of the network by Sybil-type attacks [33,34]. where a node maliciously tries to take control of the network by creating other nodes linked to it [35].
H-B FT algorithm features
The Byzantine Fault Tolerance to Health (H-BFT) consensus mechanism was proposed to meet the identified high-level requirements, incorporating features that aim to meet the already established needs of confidentiality, integrity, availability, scalability and security. It is a voting-based and Byzantine faulttolerant algorithm, inspired by the PAXOS algorithm [36]. Three distinct roles are presented during the execution of the H-BFT: leader, acceptors and verifiers. The leader receives the values and proposes them to obtain consensus; acceptors are responsible for choosing the proposed value, and verifiers are responsible for the list of reliable nodes in the network. A feature of H-BFT is the creation of a list of reliable nodes, asynchronously by verifiers, applying a reputation algorithm [37]. The purpose of this is to mitigate possible Sybil attacks by establishing quality control over the nature of the elements that will be able to vote and receive votes.
To meet the scalability criterion and solve a common problem in voting-based consensus mechanisms, the H-BFT brings the concept of continuous slicing to obtain consensus, which was inspired by the idea of quorum slicing in the FBA algorithm [38,39]. This principle establishes that consensus occurs from a minimum quorum of reliable nodes in the network, expressed by the formula [(T – i)/2] +1, where T is the total number of individuals verified and i the individuals already selected. If consensus is not obtained in a round, a new execution is performed, applying the same criteria for obtaining the minimum quorum. Moreover, H-BFT uses the reputation concept described in the dB FT algorithm [40] and, added to the list of verified nodes, proposes a Peer-to-Peer (P2P) classification mechanism dynamically selecting respectable nodes [37]. With this, it is possible to replace the leader after a certain time has elapsed.
Algorithm Workflow
H-BFT is an algorithm with a three-stage flow, which starts when the leader receives a proposal to insert a new record, then messages are exchanged with the verifiers to provide the list of trusted nodes. After the procedures for obtaining the list, the leader sends a preprepare message to the constant acceptors in the quorum slice. The remaining steps may be then noted in the algorithm execution:
a) When an acceptor receives the message, it checks its local database to confirm its integrity, after which it responds to the leader with a pre-prepared response message, informing that it is ready to start the consensus round.
b) The leader evaluates the number of responses received and, if the quorum is still not enough for consensus and there are still honest nodes in the network, it executes the flow from the beginning; otherwise, it sends a prepare message to each acceptor that responded. The acceptors receive the prepared message and acknowledge it to the leader.
c) The leader receives responses from acceptors and sends a commit message to the acceptors that responded. The acceptors confirm the transaction after receiving the commit message sent by the leader, send a response message and update their bases.
Finally, Figure 1 shows message exchanges between network nodes during the execution of H-BFT, representing the routine to obtain consensus among network nodes.
Figure 1: H-BFT workflow (originally presented in [13].
EH-BFT Algorithm
The Byzantine Fault Tolerance to Health Enhanced (EH-BFT) algorithm has two modifications comparing to the H-BFT algorithm. The proposed modifications are associated with the optimization of the database versioning and adjustments in the execution flow to avoid its anomalous functioning during its execution. The two modifications are shown in the following. Additionally, all its steps are detailed in Table 1. One may note that the proposed modifications increase routines, seeking to improve the original algorithm by granting security and consistency of the stored information, without forgetting the original goal of adequately supporting healthcare management systems.
Table 1: H-BFT Algorithm’s Specification.
a) The first modification is in the role played by verifiers, which were originally only responsible for maintaining the list of trusted nodes. With the EH-BFT, they gain one more attribution, which is to maintain the most current version of the database, and with that, together with the message requesting the list of suitable nodes, the most recent version of the database will be sent. In the end of the consensus round, the verifiers will keep the current state of the ledger.
b) The second modification is in the routine executed by the leader to verify the current status of the records, the proposal received with the existing records the database received from the verifier, seeking to mitigate problems like double spending.
Performance Analysis
This Section presents a general discussion, highlighting the benefits presented by the EH-BFT. To this end, we consider two real case studies, namely the Brazilian Public Healthcare System [35] [36][37] and US Medicare-Medicaid [18,19].
Case: SUS
The Brazilian Public Healthcare System (SUS) is a healthcare system designed to guarantee medical-hospital care to approximately 215,491,518 people [41] free of charge, from simple medical appointments, exams or transplant of organs [42]. Its performance is based on three pillars: promotion, protection and recovery of health, with activities ranging from promoting quality of life, reducing/eliminating health risks to early diagnoses for timely treatment [37]. The structure of the SUS is organized into Basic Healthcare Units, for outpatient care, Emergency Care Units, for less complex urgent and emergency care, and Public Hospitals or clinics, which provide any type of care and have resources to perform complex procedures [43]. The SUS still has public laboratories and maintains an agreement with private healthcare institutions, to complement the services [36]. An Electronic Health Record-EHR is created for every citizen in their first consultation [38], which will accompany them throughout their lives and will be accessed during consultations by the medical professional responsible for the care, except in cases where that the consultation is carried out at a private institution with an agreement. The current model adopted by SUS for EHR management has some shortcomings, which generate bottlenecks in access to the platform that manages patient information, cause insecurity in the maintenance of diagnoses and procedures performed, in addition to not guaranteeing consistency of data stored in its bases. The data produced during medical appointments are accessible only at the place of care, that is, if the patient moves to another state, the healthcare professional who will assist him will not have access to his medical history and access to patient information is limited, carried out without the use of more robust security mechanisms, restricting itself to the use of username and password.
In the context of SUS, the use of the EH-BFT algorithm (Figure 2), would bring the following benefits.
Figure 2: Healthcare Management System Architecture with EH-BFT.
a. The first benefit is linked to the concept of availability of patient information, so that the patient or the healthcare professional responsible for the care can access the medical history quickly, allowing for accurate diagnoses and saving resources, as the treatment would already be more effective, no waste. The decentralized structure, characteristic of Blockchainbased systems, would allow access to the information managed by it anywhere and at any time.
b. Another extremely important factor concerns the integrity of the stored data. The use of EH- BFT would guarantee the consistency and veracity of the records, since the records would be immutable and non-redundant, inserted from the agreement established between the nodes. The consensus mechanism would not only guarantee the veracity of the data, but it would also bring confidence to the professional responsible for the diagnosis.
c. The issue of scalability, something of great impact in a Blockchain network with the size necessary to serve the SUS,is handled by the EH-BFT when it performs the continuous slicing of the quorum necessary for approval of the insertion of the proposed record, which would solve a difficult problem to measure in a healthcare management system of this size.
d. The confidentiality required of the information, through the LGPD [28], would be fully guaranteed, since the data circulating on the network is encrypted, and accessible only to those who need to do so, including the patient himself.
Case: Medicare and Medicaid
In the United States of America, the healthcare service is private [40], however, there are two assistance programs, maintained by the Federal Government. The first of these is Medicare [41], created in 1966 and aimed at people over 65, people with disabilities or those unable to work for some reason. It provides four types of service: Hospital insurance, medical insurance, extension coverage (maintained by companies to serve their employees) and medication coverage. Medicaid [42], on the other hand, was created to serve people below the poverty line, and is maintained entirely by the Federal Government and by the States. Each hospital or independent healthcare professional that provides care to the insured person will receive the reimbursement due, following a specific cost table, which never mirrors the reality of the market.
This model is extremely bureaucratic, since the government establishes different levels of demand, which means that many people do not get the necessary medical care. As there are no public hospitals, all care is provided by professionals and accredited establishments, which exponentially increases decentralization and redundancy in diagnoses. In the two American forms of public healthcare service, there is a large volume of appointments where professionals need to establish new diagnoses at each consultation, without a reliable base of medical histories, without integration between the two healthcare systems, which does not allow the availability of the data, it is not possible to assess the integrity of the records or the confidentiality of the diagnoses.
a) In the context of Medicare and Medicaid, the use of the EH-BFT algorithm (Figure 2), would bring the following benefits.
a. The first benefit of using a healthcare management system based on Blockchain technology would be the integration between the two public models of medical care, allowing the exchange of information produced by diagnoses made by healthcare professionals.
b. The dispersed nature of the services would make each establishment or healthcare professional a node in the Blockchain network, impacting scalability. This problem has already been solved by the EH-BFT during the routines carried out during its execution, in addition to that, the very decentralized model would guarantee availability for access to authorized persons at any time.
c. Keeping a single EHR available to everyone who needs access to it makes the more economical, faster, and more accurate diagnoses, bringing relief to the patient, in addition to savings and efficiency for the Government. This EHR would also be protected by the cryptography used by the Blockchain, which would guarantee the required confidentiality. In both case studies, the implementation of a healthcare management system based on Blockchain, which implements the EH-BFT, would bring performance benefits, cost reduction, accuracy in diagnoses, availability, and integrity of information, bringing benefits to citizens and for governments.
Conclusions and Future Work
This article presents an improvement proposal for enhancing the Byzantine Fault Tolerance to Health (H- BFT) algorithm. This enhancement, called Byzantine Fault Tolerance to Health (H-BFT), is characterized by optimizations in the flow of the algorithm and on its versioning control of the database. The EH-BFT maintains the specifications regarding continuous quorum slicing, role rotation, and generation of the list of suitable nodes, besides, it also assigns to the verifier the responsibility of controlling the version of the database in use by the system, which will be consulted at each necessary round for establishing consensus among network nodes. Additionally, to exemplify the deployment of EH-BFT, two case studies were herein presented, namely the Brazilian Unified Healthcare System and the US public healthcare systems Medicare and Medicaid, respectively. By means of the exploitation of these two scenarios, the effectiveness of the EH-BFT was demonstrated. As future works and being aware of this research’s limitations, we suggest the execution of tests in simulated environments besides the implementation of the EH-BFT algorithm with the goal of validating the processes and results derived herein.
Read More about Lupine Publishers journal of Current Trends in Computer Sciences & Applications Click on the below link: https://computer-sciences-lupine-publishers.blogspot.com/
No comments:
Post a Comment
Note: only a member of this blog may post a comment.