Converging blockchain and next-generation artificial intelligence technologies to decentralize and accelerate biomedical research and healthcare

May combine two claims.

Signaling pathway-based dimensionality reduction allowed for more robust performance on the validation set, while classifiers trained on gene expression data demonstrated a significant decrease in predictive accuracy on the validation set compared to training set performance.

“signaling pathway-based dimensionality reduction allowed for the more robust performance on the validation set, while classifiers trained on gene expression data demonstrated a significant decrease in predictive accuracy on the validation set compared to the training set performance.”

• cites [33] Zhavoronkov, 2016, Deep Learning Applications for Predicting Pharmacological Properties of Drugs and Drug Repurposing Using Transcriptomic Data, Mol Pharm sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S2 (checkable), S4 (based). May combine two claims.

Introduction of new technologies, such as artificial intelligence and blockchain, may enhance and scale up progress in health care sciences and lead to effective and cost-efficient healthcare ecosystems.

“Introduction of new technologies, such as an artificial intelligence and blockchain, may enhance and scale up the progress in health care sciences and lead to effective and cost-efficient healthcare ecosystems.”

Probably not a substantial claim: may not pass S3 (load-bearing).

The success of initiatives that provide patients with access to their electronic healthcare records, such as OpenNotes, suggests their potential to improve the quality and efficiency of medical care.

“The success of initiatives that provides patients with the access to their electronic healthcare records, such as OpenNotes, suggests their potential to improve the quality and efficiency of medical care [1, 2].”

• cites [1] Lin, 2004, Use of a patient-accessible electronic medical record in a practice for congestive heart failure: patient and physician experiences, J Am Med Inform Assoc sourceNot checked yet

  Not checked in the free run.

• cites [2] Walker, 2016, Do Patients Who Access Clinical Information on Patient Internet Portals Have More Primary Care Visits?, Med Care sourceNot checked yet

  Not checked in the free run.

May combine two claims.

A DNN trained on pathway scores outperformed a DNN trained on the set of landmark genes and achieved an F1 score of 0.701 for the three drug pharmacological classes.

“DNN trained on the pathway scores outperformed DNN trained on the set of landmark genes and achieved the F1 score, the weighted average of precision and recall, of 0.701 for the three drug pharmacological classes.”

• cites [33] Zhavoronkov, 2016, Deep Learning Applications for Predicting Pharmacological Properties of Drugs and Drug Repurposing Using Transcriptomic Data, Mol Pharm sourceNot checked yet

  Not checked in the free run.

May combine two claims.

An enhanced GAN architecture that included additional molecular parameters such as solubility enabled generation of more chemically diverse molecules and showed improvement in training and generation processes.

“authors proposed an enhanced architecture that also included additional molecular parameters such as solubility and enabled the generation of more chemically diverse molecules [42]. New model clearly showed the improvement in the training and generation processes”

• cites [42] Zhavoronkov, 2017, druGAN: An Advanced Generative Adversarial Autoencoder Model for de Novo Generation of New Molecules with Desired Molecular Properties in Silico, Mol Pharm sourceNot checked yet

  Not checked in the free run.

May combine two claims.

Exonum's service-oriented architecture allows applications to reuse services developed for other Exonum applications, add and reconfigure services, and streamline integration of third-party applications via common transports such as REST+JSON.

“Due to use of service-oriented architecture, the application could easily reuse services developed for other Exonum applications, add and reconfigure services utilized for the application, etc. The service orientation and direct use of common transports (such as REST + JSON) could allow to streamline integration of third-party applications into the ecosystem provided by the Marketplace.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Client-side data validation could allow reducing risks associated with man-in-the-middle attacks, including those performed on the server side.

“Client-side data validation could allow reducing the risks associated with man-in-the-middle attacks, including those when MitM is performed on the server side (e.g., by compromising the user-facing backend of the system).”

May combine two claims.

Networks trained on Tox21 data were tested on SIDER, but none of the one-shot networks achieved any predictive power, highlighting a potential limitation in translation from toxic in vitro assays into the human clinic.

“networks trained on Tox21 data were tested on SIDER, however none of the one-shot networks achieved any predictive power, highlighting the potential limitation in translation from toxic in vitro assays into the human clinic.”

• cites [51] Pande, 2017, Low Data Drug Discovery with One-Shot Learning, ACS Cent Sci sourceNot checked yet

  Not checked in the free run.

May combine two claims.

Exonum's data storage design for auditability could make it easier for clients and auditors, including those with incomplete read access, to audit the system in real time and retrospectively, and the list of auditors could be unknown beforehand and scaled over blockchain operation.

“Because of design of data storage structures for auditability, Exonum could make it easier for clients and auditors (incl. ones with incomplete read access to data) to audit the system both in real time (incl. intermittently) and retrospectively. Further, the list of auditors could be unknown beforehand, and could be scaled over the course of blockchain operation.”

May combine two claims.

RNNs were used to predict heart failure of patients based on clinical events in their records; models trained on 12-month clinical history and tested on 6 months demonstrated an AUC of 0.883 and outperformed shallow models.

“Recently, RNNs were used to predict heart failure of patients based on clinical events in their records [44]. Models trained on 12 month period of clinical history and tested on 6 months demonstrated an Area Under the Curve (AUC) of 0.883 and outperformed shallow models.”

• cites [44] Sun, 2017, Using recurrent neural network models for early detection of heart failure onset, J Am Med Inform Assoc sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S2 (checkable).

Blockchain is very suitable for application in healthcare and other areas where data is very sensitive and strict regulations on how data can be used need to be imposed.

“This makes blockchain very suitable for application in healthcare and other areas where data is very sensitive and strict regulations on how data can be used need to be imposed.”

• AssumptionRests on an assumption: By design, blockchain ensures immutability of data, so once recorded, data in any given block cannot be modified afterwards without alteration of all subsequent blocks and agreement of the members of the network.

May combine two claims.

HIPAA requires an institutional review board to approve the use of data, and this could introduce some degree of complexity to data use.

“HIPAA requires an institutional review board to approve the use of data, and this could simply introduce some degree of complexity to data use [57].”

• cites [57] Just, Whitaker. Addressing the Challenges of Translational and Clinical Research Data Management. Health Catalyst. 2015. Available from: https://www.healthcatalyst.com/addressing-challenges-clinical-research-data-management.Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

The availability of real-time and retrospective authorization tools with guarantees of data authenticity could reduce costs of auditing and monitoring processes.

“The availability of real-time and retrospective authorization tools with guarantees of data authenticity could reduce costs of auditing and monitoring processes.”

Classified as probably not a substantial claim. May combine two claims.

Because of its integrity and immutability, blockchain could be used as an open, distributed ledger and can record transactions between different parties or networked database systems in an efficient, verifiable and permanent manner.

“Because of its integrity and immutability, blockchain could be used as an open, distributed ledger and can record transactions between different parties or networked database systems in an efficient, verifiable and permanent manner.”

• AssumptionRests on an assumption: By design, blockchain ensures immutability of data, so once recorded, data in any given block cannot be modified afterwards without alteration of all subsequent blocks and agreement of the members of the network.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Exonum uses a BFT Bitcoin anchoring algorithm that periodically outputs the hash digest of a recent block to Bitcoin, authenticated by a supermajority, with chaining similar to paper-based anchoring [75].

“Exonum uses a BFT Bitcoin anchoring algorithm, which is packaged as a separate service. The algorithm periodically outputs the hash digest of a recent block on an Exonum blockchain, which commits to the entire blockchain state and transaction history, in a transaction on the Bitcoin blockchain. The anchoring transaction has a well-defined form and must be authenticated by a supermajority of valid”

• cites [75] Buldas, Lipmaa. Schoenmakers - Public Key Cryptography B, 2000. Optimally efficient accountable time-stamping. Springer. 2000. Available from: http://link.springer.com/content/pdf/10.1007/b75033.pdf#page=304. sourceNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S2 (checkable), S4 (based).

One-time combination of data provides a nutritious feed for artificial intelligence, allowing creation of powerful algorithms for effective and precise detection of different human health states.

“One-time combination of date provides a very nutritious feed for artificial intelligence, allowing to create powerful algorithms of effectively and precisely detection different human health states.”

• AssumptionRests on an assumption: Each type of medical data and their combinations have relevant value depending on quality of the medical records and biological significance for certain disease condition.
• AssumptionRests on an assumption: Different types of medical data have their own predictive value, representative sensitivity, prediction rate and weight.

May combine two claims.

Genome analysis provides important heredity information but has low value for prediction of dynamic changes in the organism compared to epigenome [80] or transcriptome [81].

“For example, genome analysis provides an important information on heredity, but due to it relative stability has a low value for prediction of dynamic changes in the organism compared to epigenome [80] or transcriptome [81].”

• cites [80] Tyson, 2013Not checked yet

  Not checked in the free run.

• cites [81] Passos, 2015Not checked yet

  Not checked in the free run.

• AssumptionRests on an assumption: Each type of medical data and their combinations have relevant value depending on quality of the medical records and biological significance for certain disease condition.
• AssumptionRests on an assumption: Different types of medical data have their own predictive value, representative sensitivity, prediction rate and weight.

Classified as probably not a substantial claim.

Following 360° health, the more different parameters analyzed at the same time, the more detailed and voluminous the overall picture [82].

“Following the principles of 360° health introduced by the NHS [82], the more different parameters are analyzed at the same time, the more detailed and voluminous the overall picture is.”

• cites [82] Whitehead. 360° of health data: Harnessing big data for better health. 2014. Available from: http://www.abpi.org.uk/our-work/library/medical-disease/Documents/360%20Degrees%20of%20Health%20Data.pdf.Not checked yet

  Not checked in the free run.

• AssumptionRests on an assumption: Each type of medical data and their combinations have relevant value depending on quality of the medical records and biological significance for certain disease condition.
• AssumptionRests on an assumption: Different types of medical data have their own predictive value, representative sensitivity, prediction rate and weight.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The validator set is reconfigurable; validators can be added or removed by supermajority agreement, and key rotation can be done.

“The validator set is reconfigurable; validators could be added or removed by the agreement of the supermajority of existing validators. The same procedure could be used for key rotation for validators.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Merkelized collections allow creation of compact logarithmic cryptographic proofs of presence and absence of items in the collection.

“This construction allows to create compact (logarithmic wrt the number of elements in the collection) cryptographic proofs of presence (and absence in the case of sets/maps) of items in the collection.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

All Merkelized collections of a service can be committed to a single hash digest, and all services to a blockchain-level hash digest committing to all data in all Merkelized collections, allowing proofs tied to a single root of trust.

“Similarly, commitments of all services within the blockchain could be collected into a single blockchain-level hash digest, which would commit to all data in all Merkelized collections within the blockchain state. For all intents and purposes, the resulting blockchain-level hash digest is the hash (commitment) of the entire blockchain state. This would allow to create proofs of existence or”

May combine two claims.

Combination of data types, especially the sum of low diagnostics data, provides a multi-level overview, better understanding of complex multifactorial conditions, and leads to faster diagnostics [78,79].

“The combination of data types, especially the sum of low diagnostics data, provides a multi-level overview and better understanding of complex multifactorial conditions, and also leads to a faster diagnostics [78, 79].”

• cites [78] Liang, 2016, Nucleic Acid Amplification Testing and Sequencing Combined with Acid-Fast Staining in Needle Biopsy Lung Tissues for the Diagnosis of Smear-Negative Pulmonary Tuberculosis, PLoS One sourceNot checked yet

  Not checked in the free run.

• cites [79] Huang, 2015, Identification of Breast Cancer Using Integrated Information from MRI and Mammography, PLoS One sourceNot checked yet

  Not checked in the free run.

• AssumptionRests on an assumption: Each type of medical data and their combinations have relevant value depending on quality of the medical records and biological significance for certain disease condition.
• AssumptionRests on an assumption: Different types of medical data have their own predictive value, representative sensitivity, prediction rate and weight.

Probably not a substantial claim: may not pass S3 (load-bearing).

Exonum's Requests algorithm allows a validator to restore consensus-related information from other validators, improving system liveness.

“Requests: Requests algorithm allows a validator to restore consensus-related information from other validators by utilizing the fact that all messages are digitally signed. This has a positive effect on system liveness.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Transactions are necessarily authenticated by their originators with public-key digital signatures to ensure integrity and real-time and retrospective universal verifiability.

“Transactions are necessarily authenticated by their originators with the help of public-key digital signatures to ensure their integrity, as well as real-time and retrospective universal verifiability.”

May combine two claims.

The Exonum network continues operating even if up to 1/3 validators are hacked, compromised or switched off, so there is no single point of failure and transaction processing is fully decentralized.

“The Exonum network would continue operating even if up to 1/3 validators are hacked, compromised or switched off. Hence, there is no single point of failure in the network; the whole process of transaction processing is fully decentralized.”

• cites [73] Pease, Shostak, Lamport. Reaching Agreement in the Presence of Faults. J ACM. New York, NY, USA: ACM; 1980; 27: 228-34.Not checked yet

  Not checked in the free run.

May combine two claims.

Due to universal verifiability of transactions and proofs, clients may connect to a single node for all requests; a malicious node cannot forge proofs for read requests but could delay transaction processing.

“Due to universal verifiability of transactions and proofs, clients may connect to a single node for all requests. Note that maliciously acting node cannot forge proofs for read requests; but it could delay transaction processing by not broadcasting transactions received from the client.”

May combine two claims.

Under the assumption of unforgeable public-key digital signatures and a partially synchronous network, Exonum's consensus algorithm provides safety and liveness as defined in [74], with safety not depending on partial synchronicity.

“The consensus algorithm works under the assumption of unforgeable public-key digital signatures and a partially synchronous network. Under these conditions, the algorithm provides safety and liveness as defined in [74], with safety not depending on partial synchronicity.”

• cites [74] Kwon. Tendermint: consensus without mining. 2015. Available from: http://tendermint.com/docs/tendermint.pdf.Not checked yet

  Not checked in the free run.

• AssumptionThe authors state this as an assumption: The consensus algorithm works under the assumption of unforgeable public-key digital signatures and a partially synchronous network.

Probably not a substantial claim: may not pass S3 (load-bearing).

Exonum's unbounded voting rounds help decrease delays when network connection among validators is unstable.

“Unbounded rounds: Voting rounds have a fixed start time, but do not have a definite end time. A round ends only when the next block is received or committed locally. This helps decrease delays when the network connection among validators is unstable.”

May combine two claims.

Interactions in the marketplace are registered on a blockchain as transactions; the blockchain contains hashes used to timestamp and provide non-repudiation, not opened personal information.

“Interactions in the marketplace are registered on a blockchain in the form of transactions. Blockchain by itself does not contain any opened personal information. It contains hashes which could be used to timestamp and provide a reasonable level of non-repudiation for all actions at the marketplace.”

• AssumptionRests on an assumption: One major problem for healthcare is data exchange and ability to use data in research and commercial projects.

May combine two claims.

Combining various biomarkers and linking them to age will provide the basis for a platform able to provide integrative analysis of health status, assess data quality, and identify fake data.

“Hence, combining various biomarkers and linking them to age will provide the basis for platform able to provide integrative analysis of health status, assess data quality and even identify fake data.”

• AssumptionRests on an assumption: Chronological age is one of the most important factors affecting morbidity and mortality in humans.
• AssumptionRests on an assumption: Age is the closest estimate of a health status of a person.

Probably not a substantial claim: may not pass S4 (based). May combine two claims.

The promise of such a solution is the opportunity for users to take ownership of their data and access privileges and sell their data directly to consumers for fair value.

“The promise of such solution is the opportunity for users to take ownership of their data and access priviledges and even allow them to sell their data directly to the consumers of data for the fair value of data.”

• AssumptionRests on an assumption: One major problem for healthcare is data exchange and ability to use data in research and commercial projects.

Probably not a substantial claim: may not pass S4 (based).

DNNs trained on multi-modal data ranging from photographs, videos, blood tests, omics, activity, smell and sweat during aging capture many biologically-relevant features about the group, individual, organ, tissue or molecules.

“The DNNs trained on the multi-modal data ranging from photographs, videos, blood tests, “omics”, activity and even smell and sweat during aging capture the many biologically-relevant features about the group, individual, organ, tissue or even a set of molecules.”

• AssumptionRests on an assumption: Chronological age is one of the most important factors affecting morbidity and mortality in humans.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

For distant relatives, r→0 and almost will not contribute to the cost function; for twins r=1 and at the beginning it will double the cost of data; cost grows with number of close relatives using the platform.

“For distant relatives, r → 0 and almost will not contribute to the cost function of data, however, for a very close relative such as twins, r is equal to 1 and at the beginning, it will double the cost of data. The cost of data will grow with a number of close relatives that are using the platform and contributing their data.”

• AssumptionRests on an assumption: Biomedical data obtained for relatives is commonly more valuable than the same data obtained from unrelated individuals.

Probably not a substantial claim: may not pass S4 (based). May combine two claims.

The proposed high-level architecture with individual DNN predictors per data type and a multi-modal one-shot learning DNN allows accurate age prediction and feature importance analysis, informing relative cost of biomarkers.

“This architecture allows not only for accurate age prediction, but also for feature importance analysis. Results of such analysis across all predictors will tell about the importance of each individual biomarker and may inform its relative ‘cost’.”

• AssumptionRests on an assumption: Chronological age is one of the most important factors affecting morbidity and mortality in humans.
• AssumptionRests on an assumption: Age is the closest estimate of a health status of a person.

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

Application of the data value model makes it possible to convert Big Data into Apprised Data.

“Thus, the application of the data value model makes it possible to convert Big Data into Apprised Data.”

• AssumptionRests on an assumption: Each type of medical data and their combinations have relevant value depending on quality of the medical records and biological significance for certain disease condition.
• AssumptionRests on an assumption: Different types of medical data have their own predictive value, representative sensitivity, prediction rate and weight.

May combine two claims.

Timestamping is achieved with blockchain anchoring [52] and accountable timestamping [59]; non-repudiation with digital signing and a blockchain-based PKI.

“The former is achieved with the help of blockchain anchoring [52] and other accountable timestamping [59] techniques; the latter - with the help of digital signing and a blockchain-based PKI.”

• cites [52] Cory. Cross-Border Data Flows: Where Are the Barriers, and What Do They Cost? Available from: https://itif.org/publications/2017/05/01/cross-border-data-flows-where-are-barriers-and-what-do-they-cost.Not checked yet

  Not checked in the free run.

• cites [59] Office For Civil. Your Rights Under HIPAA. HHS.gov. US Department of Health and Human Services. 2017. Available from: https://www.hhs.gov/hipaa/for-individuals/guidance-materials-for-consumers/index.html.Not checked yet

  Not checked in the free run.

• AssumptionRests on an assumption: One major problem for healthcare is data exchange and ability to use data in research and commercial projects.

Probably not a substantial claim: may not pass S4 (based). May combine two claims.

A blockchain-based system can dramatically simplify data acquisition, allowing users to upload data directly and give permission if bought through the system using a transparent price formula determined by the data value model, and guarantee fair tracking of data usage.

“A blockchain-based system can dramatically simplify data acquisition process. They allow user to upload his data directly to the system and give his permission to use his data if it was bought through the system using transparent price formula determined by data value model. Also it would guarantee fair tracking of all data usage activity.”

• AssumptionRests on an assumption: One major problem for healthcare is data exchange and ability to use data in research and commercial projects.

May combine two claims.

A multitude of biomarkers linked to disease are strongly correlated with age; e.g., triglycerides, HbA1c, waist circumference, IL-6 increase with age, while albumin, IGF and creatinine clearance decrease [84,85].

“For instance, triglycerides, glycated hemoglobin (HbA1c), waist circumference, IL-6 increase with age, but other parameters like albumin, IGF and creatinine clearance go in an opposite direction [84, 85].”

• cites [84] Perls, 2017, Biomarker signatures of aging, Aging Cell sourceNot checked yet

  Not checked in the free run.

• cites [85] Weinstein, 2011, Age-Related Changes in Biomarkers: Longitudinal Data from a Population-Based Sample, Res Aging sourceNot checked yet

  Not checked in the free run.

• AssumptionRests on an assumption: Chronological age is one of the most important factors affecting morbidity and mortality in humans.

Probably not a substantial claim: may not pass S4 (based).

Exchanging data for currency may be problematic due to the need for massive numbers of micro-transactions in multiple countries and among many types of data market participants.

“However, exchanging the data for currency may be problematic for many reasons including the need to perform a massive number of micro-transactions in multiple countries and among a large number of different types of the data market market participants.”

• AssumptionRests on an assumption: One major problem for healthcare is data exchange and ability to use data in research and commercial projects.

May combine two claims.

Treating aging as a disease to train deep neural networks facilitates transfer learning on individual diseases using a much smaller number of samples; first proposed by Zhavoronkov et al in 2015 [93], this technique can reconstruct datasets with missing or incorrect features.

“In addition, treating aging as a disease to train the deep neural networks to capture the most important biological properties of the age-related changes that transpire during aging using the deep neural networks facilitates for transfer learning on individual diseases using a much smaller number of samples. First proposed by Zhavoronkov et al in 2015 [93], this technique can be used to”

• cites [93] Fedintsev, 2017, A review of the biomedical innovations for healthy longevity, Aging (Albany NY) sourceNot checked yet

  Not checked in the free run.

• AssumptionRests on an assumption: Chronological age is one of the most important factors affecting morbidity and mortality in humans.

Probably not a substantial claim: may not pass S4 (based). May combine two claims.

To ensure quality of user data, third-party data validators are needed; they check data quality and provide customers with a guarantee of user data validity.

“To ensure the quality of the data provided by users, third party is needed - data validators, experts who are first buyers of the data. Data validators check the data quality and provide customers with a guarantee of user data validity.”

• AssumptionRests on an assumption: One major problem for healthcare is data exchange and ability to use data in research and commercial projects.

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Every Dataset has its own critical representative level, depending on the type of data, their quality, and patient profile.

“Every Dataset has it’s own critical representative level, critical level depends on the type of data, their quality, and patient profile.”

• AssumptionRests on an assumption: Each type of medical data and their combinations have relevant value depending on quality of the medical records and biological significance for certain disease condition.
• AssumptionRests on an assumption: Different types of medical data have their own predictive value, representative sensitivity, prediction rate and weight.

Probably not a substantial claim: may not pass S4 (based).

Data obtained in the same or short period of time have greater representation and predictive value.

“Data obtained it the same or short period of time have greater representation and predictive value, that’s why we introduce term time value of data.”

• AssumptionRests on an assumption: Each type of medical data and their combinations have relevant value depending on quality of the medical records and biological significance for certain disease condition.
• AssumptionRests on an assumption: Different types of medical data have their own predictive value, representative sensitivity, prediction rate and weight.
• AssumptionRests on an assumption: Sampling time is an important component of any medical analysis, allowing accurate description of the state of human health at the moment.

Probably not a substantial claim: may not pass S3 (load-bearing).

To ensure security and privacy, data uploaded by users to cloud storage would be encrypted on the user side using a threshold encryption scheme [97-100].

“To ensure security and privacy, the data uploaded by users to the cloud storage would be encrypted on the user side using a threshold encryption scheme [97-100].”

• cites [97] Shamir. Communications of the ACM A. 1979. How to share a secret. 1979 Available from: http://dl.acm.org/citation.cfm?id=359176.Not checked yet

  Not checked in the free run.

• cites [98] Blakley. Proceedings of the national computer 1979. Safeguarding cryptographic keys. 1979. Available from: https://pdfs.semanticscholar.org/32d2/1ccc21a807627fcb21ea829d1acdab23be12.pdf.Not checked yet

  Not checked in the free run.

• cites [99] Robling Denning, 1982Not checked yet

  Not checked in the free run.

• cites [100] Desmedt, 1992Not checked yet

  Not checked in the free run.

• AssumptionRests on an assumption: One major problem for healthcare is data exchange and ability to use data in research and commercial projects.

Classified as probably not a substantial claim. May combine two claims.

The formal data value model allows establishing a proper cost for using a given data combination, enabling creation of a formal medical data transaction model and a medical data marketplace.

“In this paper we introduce a formal model that allows to evaluate data value that takes into account combination and time parameters of data. Based on the value model one may establish a proper cost for using the given data combination, and this in turn allows for creation of formal medical data transaction model to create medical data marketplace.”

• AssumptionRests on an assumption: Each type of medical data and their combinations have relevant value depending on quality of the medical records and biological significance for certain disease condition.
• AssumptionRests on an assumption: Different types of medical data have their own predictive value, representative sensitivity, prediction rate and weight.

May combine two claims.

Cholesterol check is valid for only five years or less if a patient is at higher risk of heart disease [83], while genome profile is valid for the whole life, so genome analysis has a longer half-life period than basic cholesterol blood test.

“For example, according to one of largest medical practice and research center, Mayo Clinic, cholesterol check is valid for only five years or less if a patient at the higher risk of heart disease [83]. While, the genome profile is valid for the whole life of the patient, so genome analysis has longer half-life period than basic cholesterol blood test.”

• cites [83] Mayo Clinic. Cholesterol test. Why it’s done. Mayo Clinic. 2016 [cited 2017 Oct 15]. Available from: http://www.mayoclinic.org/tests-procedures/cholesterol-test/details/why-its-done/icc-20169529.Not checked yet

  Not checked in the free run.

• AssumptionRests on an assumption: Each type of medical data and their combinations have relevant value depending on quality of the medical records and biological significance for certain disease condition.
• AssumptionRests on an assumption: Different types of medical data have their own predictive value, representative sensitivity, prediction rate and weight.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Blockchain technology has been proposed as a medium for health care data storage solutions [94,95] and as a tool for improving transparency in clinical trials [96].

“Gained a substantial attention in recent years T, the blockchain technology gained substantial popularity in recent years primarily due to the popularity of the Bitcoin crypto currency, was previously has been proposed as a medium for health care data storage solutions [94, 95] and as a tool for to improving thee transparency in clinical trials [96].”

• cites [94] Ohno-Machado, 2017, Blockchain distributed ledger technologies for biomedical and health care applications, J Am Med Inform Assoc sourceNot checked yet

  Not checked in the free run.

• cites [95] Schulz, 2017, Blockchain Technology: Applications in Health Care, Circ Cardiovasc Qual Outcomes sourceNot checked yet

  Not checked in the free run.

• cites [96] Cimpoesu, 2016, Improving data transparency in clinical trials using blockchain smart contracts, F1000 Res sourceNot checked yet

  Not checked in the free run.

• AssumptionRests on an assumption: One major problem for healthcare is data exchange and ability to use data in research and commercial projects.

Probably not a substantial claim: may not pass S4 (based). May combine two claims.

Blockchain technology enables the creation of a distributed and secure ledger of personal data, where patients are in control, own their data, and monitoring of access privileges and understanding of who looked at the data.

“Blockchain technology enables the creation of a distributed and secure ledger of personal data, where patients are in control, own their data, and monitoring of access privileges and understanding of who looked at the data.”

• AssumptionRests on an assumption: Presently, only a few patients worldwide have comprehensive datasets containing clinical history combined with genetic, blood biochemistry, cell count, lifestyle, drug and supplement use data, because they do not see the value in this data and do not get tested regularly.

Probably not a substantial claim: may not pass S2 (checkable), S4 (based).

Blockchain and AI open new paradigms for health data ecosystems.

“Blockchain and AI open new paradigms for health data ecosystems (Figure 12).”

May combine two claims.

Anomalies in data could be linked to health conditions, so both unsupervised and time-series approaches could be used for pathology detection in health recordings.

“Anomalies in the data could also be linked to health conditions, so both those approaches could be used for pathology detection in health recordings [103, 104].”

• cites [103] Vig, 2015, Anomaly detection in ECG time signals via deep long short-term memory networks, 2015 IEEE International Conference on Data Science and Advanced Analytics (DSAA)Not checked yet

  Not checked in the free run.

• cites [104] Liu, 2016, Research on Healthy Anomaly Detection Model Based on Deep Learning from Multiple Time-Series Physiological Signals, Sci ProgramNot checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S2 (checkable), S4 (based). May combine two claims.

The introduction of the blockchain-based data ecosystem may help ensure that individuals take control over their data and companies and research institutions may acquire data more freely, reducing the need for regulators to interfere.

“The introduction of the blockchain-based data ecosystem may help ensure that the individuals take control over their data and companies and research institutions may acquire data more freely reducing the need for the regulators to interfere.”

• AssumptionRests on an assumption: Data security and data privacy are major challenges in data-driven healthcare.
• AssumptionRests on an assumption: Consumer, healthcare, and research companies require data from many individuals to train their deep neural networks.
• AssumptionRests on an assumption: Companies with the largest datasets acquire data in ways that may not be transparent to individuals, and often companies and individuals do not understand the value of this data.
• AssumptionRests on an assumption: Regulators often set up barriers for consumer data collection and storage, substantially inhibiting the propagation of recent AI advances into clinical practice.

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing). May combine two claims.

Blockchain-based PKI for logically well-known users could base on well-established measures for secure key management and be augmented with blockchain-based smart contracting.

“Blockchain-based PKI for logically “well-known” users (e.g., blockchain validators, data validators, key keepers, etc.) could base on well-established measures for secure key management (e.g., key sharing, use of specialized hardware for key storage, etc.).”

Probably not a substantial claim: may not pass S2 (checkable).

The value of the various data types may depend on the application.

“The value of the various data types may depend on the application.”

• AssumptionRests on an assumption: The value of various types of data, combinations of data types, and time value of data is poorly understood and often debated.

Probably not a substantial claim: may not pass S4 (based).

Blockchain technology allows for the creation of a data-driven marketplace, where patients can earn tangible rewards for making their data available to application developers, pharmaceutical and consumer companies, and research institutions and generating new data through regular and comprehensive tests and checkups.

“Most importantly, blockchain technology allows for the creation of a data-driven marketplace, where patients can earn tangible rewards for making their data available to the application development community, pharmaceutical and consumer companies, and research institutions and generating new data through regular and comprehensive tests and checkups.”

• AssumptionRests on an assumption: Presently, only a few patients worldwide have comprehensive datasets containing clinical history combined with genetic, blood biochemistry, cell count, lifestyle, drug and supplement use data, because they do not see the value in this data and do not get tested regularly.
• AssumptionRests on an assumption: Pharmaceutical and consumer companies are willing to pay substantial amounts for large numbers of personal data records required to train their AI.

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

The use of authenticated communication channels to transmit key shares could achieve forward secrecy, i.e., non-compromise of encrypted user data even if key keepers' long-term asymmetric keys become compromised.

“The use of authenticated communication channels to transmit key shares could allow to achieve forward secrecy, i.e., non-compromise of the encrypted user data even if the key keepers’ long-term asymmetric keys would become compromised (incl. the compromise of the underlying asymmetric cryptosystem, e.g., with the advent of quantum computing).”

Classified as probably not a substantial claim.

The security of the described marketplace setup relies on the security of the hash function, public-key signature schemes, symmetric ciphers, and secret sharing schemes used.

“The security of the described setup relies on the security of utilized crypto-primitives: the hash function and public-key signature scheme(s) utilized in the marketplace blockchain construction, as well as the symmetric cipher(s) and the secret sharing scheme(s) used for encrypting user data.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

The problem of unbalanced labelled datasets for anomaly detection could be solved with zero and one shot learning.

“Still, the problem could be solved with help of zero and one shot learning.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

For insurance companies, the value of a recent picture of the patient may significantly exceed the value of the genome, since it may be more predictive of the patient's age, health status and mortality.

“For example, for the insurance companies, while the cost of data generation may be significantly higher for the genome compared to a selfie, the value of the recent picture of the patient may significantly exceed the value of the genome, since it may be more predictive of the patient’s age, health status and mortality.”

• AssumptionRests on an assumption: Recent advances in AI have enabled highly accurate predictors of biologically relevant features such as age, race, and sex from simple data types such as selfies and blood tests.

Probably not a substantial claim: may not pass S2 (checkable), S4 (based).

The combination of these data types will be considerably more valuable than the value of these data types individually.

“However, the combination of these data types will be considerably more valuable than the value of these data types individually.”

• AssumptionRests on an assumption: Recent advances in AI have enabled highly accurate predictors of biologically relevant features such as age, race, and sex from simple data types such as selfies and blood tests.

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

User-side key management for data encryption and authentication of blockchain transactions may be prone to various risks, including faulty random number generation, inadequate long-term key storage, and compromised user interfaces.

“The user-side key management for data encryption and authentication of blockchain transactions may be prone to various risks (e.g., a faulty random number generator resulting in generation of keys with insufficient entropy; inadequate security for long-term key storage; compromises of the user interface).”

Probably not a substantial claim: may not pass S3 (load-bearing).

One major reason for integrating cloud storage is to provide off-chain storage for large biomedical data files such as CT scans or MRIs, where one file can reach 50 Mb.

“One of the major reasons for integrating cloud storage into the ecosystem is to provide an off-chain storage solution especially for large biomedical data files, such as CT scans or MRIs, where the size of one data file could reach 50 Mb.”

• AssumptionRests on an assumption: One major problem for healthcare is data exchange and ability to use data in research and commercial projects.

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

Cloud storage may require authentication for read and write access, preferably based on PKI established on the marketplace blockchain.

“The cloud storage may require authentication for read and write access to data, which in a preferable setup would be based on the PKI established on the marketplace blockchain.”

• AssumptionRests on an assumption: One major problem for healthcare is data exchange and ability to use data in research and commercial projects.