
A Blockchain-Based Approach to Health Information Exchange
Networks
Kevin Peterson, Rammohan Deeduvanu, Pradip Kanjamala, and Kelly Boles
Mayo Clinic
Abstract
Sharing healthcare data between institutions is challenging. Heterogeneous data structures may
preclude compatibility, while disparate use of healthcare terminology limits data comprehension. Even
if structure and semantics could be agreed upon, both security and data consistency concerns abound.
Centralized data stores and authority providers are attractive targets for cyber attack, and establishing a
consistent view of the patient record across a data sharing network is problematic. In this work we present
a Blockchain-based approach to sharing patient data. This approach trades a single centralized source
of trust in favor of network consensus, and predicates consensus on pro of of structural and semantic
interoperability.
1 Problem Statement
Cross-institutional sharing of healthcare data is a complex undertaking with the potential to significantly
increase research and clinical effectiveness[1]. First and foremost, institutions often are reluctant to share
data because of privacy concerns[2], and may fear that sending information will give others a competitive
advantage[3]. Next, even if privacy concerns could be addressed, there is no broad consensus around the
specific technical infrastructure needed to support such a task[4]. Finally, healthcare data itself is complex,
and sending information across institutional boundaries requires a shared understanding of both data struc-
tures and meaning. Even assuming data can be shared efficiently and securely, these interoperability issues
left unchecked will limit the utility of the data. Despite evidence that the value of healthcare data exchange
is large[5], these issues, described below, remain significant barriers.
1.1 Security
Failing to secure the patient record has financial and legal consequences, as well as the potential to impact
patient care. Securing the electronic medical record is a challenging task[6], and the ramifications of a breach
are a strong disincentive to sharing data. For this work, we focus on both privacy and anonymity and how
they apply to data sharing.
Data privacy involves ensuring only authorized parties may access the medical record. This impacts any
healthcare system, as patient privacy is not only an ethical responsibility, but a legal mandate[7]. Patient
data is also an asset to the institution, and unauthorized access could compromise competitive advantages
or reveal proprietary practices.
Data anonymity may also be used to secure the record. In this way, identifiable information is left out,
and only summary/partial data is shared. This can be acceptable, but is challenging, as it requires a large
number of attributes with potential resource or patient care value to be removed from the record in order
for it to be considered de-identified[8].
1.2 Infrastructure
A significant hurdle to sharing data is the agreement of the supporting technical architecture and infrastruc-
ture. Many attempts at data sharing require either (1) a centralized data source, or (2) the transmission
1