Journal of Computer Information Systems

ISSN: 0887-4417 (Print) 2380-2057 (Online) Journal homepage: http://www.tandfonline.com/loi/ucis20

A Service-Oriented Model for Personal Health

Records

Jingquan Li

To cite this article: Jingquan Li (2018): A Service-Oriented Model for Personal Health Records,
Journal of Computer Information Systems, DOI: 10.1080/08874417.2018.1483213

To link to this article: https://doi.org/10.1080/08874417.2018.1483213

Published online: 18 Jun 2018.

Submit your article to this journal

Article views: 12

View Crossmark data

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=ucis20
JOURNAL OF COMPUTER INFORMATION SYSTEMS
https://doi.org/10.1080/08874417.2018.1483213

A Service-Oriented Model for Personal Health Records

Jingquan Li
Hofstra University, Hempstead, NY, USA

ABSTRACT KEYWORDS
Personal health records (PHRs) have received widespread attention as a tool for people to store and Personal health records;
manage their own or their family’s health records. Today PHRs are provided by a variety of players, service-oriented
including health care systems or hospitals, health plans, employers, and technology vendors. Traditional architecture; cloud
PHR systems are institution specific and have not been concerned with the interoperability and privacy computing; web services;
interoperability; data privacy
of PHRs when interacting with other systems. The paper reported here seeks to address this challenge
by presenting a service-oriented model for PHRs. In this study, we examine the challenges faced in the
integration of existing electronic medical record (EMR) systems and health information exchanges (HIEs).
We then propose a service-oriented model for integrating PHR and EMR systems. Furthermore, we
analyze the services in SOA and explore the potential of service-oriented personal health cloud. We
present a practical case study to verify and validate our the proposed approach. We conclude by
discussing practical and policy implications and future research directions.

Introduction complete PHR and share it with other medical providers and
Personal health records (PHRs) have received increasing recogni- clinicians. Furthermore, sharing of PHRs may also pose
tion in the health care industry today. PHR systems allow indivi- potential risks to the privacy and confidentiality of highly
duals to receive copies of all or part of their electronic medical sensitive personal information. However, people, especially
records (EMRs),1 the test results from labs, the medication records these with multiple diseases, can have many different medical
of pharmacies, the claims data from insurers, and individuals’ own providers, including primary care physicians, specialists,
notes, such as information about their weight changes or blood therapists, and even alternative medicine practitioners to
sugar levels. While a PHR system is likely to be tied to an EMR serve their different medical needs. They can also use mobile
system, web-based standalone PHRs are also provided by technol- health (mHealth) applications, wearable devices, and home
ogy companies, such as Microsoft, WebMD, and Dossia monitoring devices for determining health status or managing
Consortium. Electronic PHRs support patient-centered e-Health disease conditions. Consequently, the information pertaining
(PCEH) by making medical records and patients self-manage- to a person’s past medical history typically is scattered among
ment tools available to patients.2,3 PHRs can help people with all the places where care has been given.
chronic diseases to manage their conditions, maintain an ongoing Under the Health Information Technology for Economic
connection with their doctors, and improve overall health.4,5 Even and Clinical Health (HITECH) Act of 2009, the United States
for healthy people, PHRs will empower individuals to manage founded billions of dollars incentives for clinicians and hos-
their own health care, reduce dangerous medical errors, improve pitals to use health information technology, including EMRs.
communications with doctors, lower health care costs, and pro- The law also clarifies that systems must be in place to aggre-
vide vital and complete information for emergency care.5–8 gate the records for each patient into a comprehensive PHR.
Nowadays, PHRs are provided by various players, including med- To accomplish the latter, many communities across the nation
ical providers, health insurance plans, employers, and IT have built regional health information exchanges (HIEs) to
companies. facilitate the exchange of patient records across different
However, the gathering and sharing of PHRs are still big health care systems or hospitals. However, the regional HIEs
challenges. The United States health care system is costly, are very complex and unreliable.
fragmented, incompatible, and still partially paper based. The United States is making substantial investments to
Since most EMR and PHR systems are institution specific accelerate the adoption and use of interoperable EMR and
(limited to a certain organization), they have generally not PHR systems. However, there seems to be no standard frame-
been concerned with the portability of a user’s health infor- work for the systems. Accordingly, it is always difficult for
mation. Health records from different organizations are consumers to gather their medical records from many differ-
incompatible formats and there are no uniform standards ent health care systems and share the data with their current
for interoperability, in most cases a patient cannot gather a physicians. In addition, medical practitioners have and

CONTACT Jingquan Li [email protected] Department of Information Systems and Business Analytics, Frank G. Zarb School of Business, Hofstra
University, 134 Hofstra University, Hempstead, NY 11549, USA
Color versions of one or more of the figures in the article can be found online at www.tandfonline.com/ucis.
© 2018 International Association for Computer Information Systems
2 J. LI

continue to use various non-interoperable EMR systems to a service-oriented PHR model and explore the potential of
document rendered care. There is a strong need to design an service-oriented personal health cloud. Second, we verify and
interoperable PHR system that achieves secure exchange with validate the proposed approach with a practical case study.
various EMR and PHR systems by using communication and Finally, we develop more robust understanding of the inter-
health vocabulary standards. operability and privacy issues with PHR/EMR systems and
To implement the enacted Directive 2011/24/EU on offer insights into the design and implementation of service-
patients’ right in cross-border healthcare, the European oriented PHR systems. Our manuscript proceeds as follows:
Commission aims to enable all Europeans to have access to First, we identify the challenges faced in the integration of
online medical records anywhere in Europe by 2020. Cross- traditional EMR systems and HIEs. Second, we present a
border and interoperable EMR and PHR systems are taking SOA-based approach for PHR systems and identify the pri-
place in Europe. The systems make confidential health records mary focus areas for improving the interoperability and secur-
more easily and immediately accessible to a wider audience ity of PHR systems. Third, we analyze the functional services
and increase the risk that personal health data could be of a service-oriented PHR system and explore the potential of
accidentally exposed or easily disseminated to unauthorized service-oriented personal health cloud. Fourth, we present a
parties by enabling greater access to a compilation of EMR practical case study that demonstrates how the service-
and PHR systems, from different sources, and throughout a oriented approach is used in a large health care system.
lifetime.9 Finally, we conclude by discussing policy and practical impli-
A service-oriented PHR model based on both service- cations and future research in service-oriented PHR systems.
oriented architecture (SOA) and cloud computing has the
potential to address the challenge of interoperability in health
care. This can be more easily seen by taking a look at a couple Traditional health care information systems
of service-oriented EMR and PHR systems. As an example, While web-based standalone PHRs, such as WebMD and
Kart, Moser, and Melliar-Smith described a distributed Microsoft are emerging, many PHR systems are linked to a
E-Health system that used SOA to enforce basic software specific health care organization’s EMR system. A traditional
architecture principles and provided interoperability between PHR system essentially extends an existing EMR system. Most
diverse computing applications that communicated with each health care organizations have chosen an institution-centric
other.10 The system supported the clinic, pharmacy, and architecture for their hosting EMR and PHR systems. HIE is
patient service modules. In the system, different devices used to electronically move clinical data among disparate
could interact with the service modules, including Personal EMR and PHR systems. However, the data interfaces between
Digital Assistants and smart phones, desktop, and server different EMR and PHR systems were point-to-point, with
computers, and even electronic medical devices, such as each system having its own data format. Standard interface
blood pressure monitors. For another example, the formats, such as Health Level Seven International (HL7), have
MammoGrid was a service-oriented EMR application. It was been adopted only by larger health care providers.17
a SOA-based medical grid application intended for the man- Figure 1 shows a common HIE architecture based on
agement and coordination of medical image data with feder- point-to-point connections to legacy EMR systems. When a
ated mammogram databases across Europe.11 Bahga and clinician’s EMR system requests a patient’s prior medical
Madisetti proposed a cloud health information systems tech- records from the HIE, this clinician’s EMR is added to the
nology architecture (CHISTAR) that achieves semantic inter- index for future queries for the patient (if not already pre-
operable EMR systems.12 CHISTAR application components sent). The HIE then send multiple queries to all other EMR
were designed using the cloud component model approach systems of prior care recorded in the HIE index. After patient
that comprises of loosely coupled components that commu- consent is verified at each “other” system, each system must
nicates asynchronously. return medical records for the patient to the exchange and the
A service-oriented PHR system has the potential to achieve
easy, accurate, and secure exchange with various EMR and
PHR systems by using communication and health vocabulary
standards. Gold and Ball described a conceptual health record
banking model based on regional and national HIE.5 Li
described the limitations of different types of PHR systems
and has presented a framework for addressing privacy and
security issues in PHR systems.7 With the rapid growth of
SOA, the use of cloud computing and web services became
the main trend to achieve SOA. Cloud computing can provide
cost-effective, flexible, and scalable solutions to businesses.
Cloud computing was proposed for integrating EMR
systems.13–15 But none of prior work explicitly addressed the
interoperability and security challenges faced by PHR systems.
This study significantly extends our previous work from
Ref. 16 and makes several contributions. First, we directly
incorporate the principles of SOA and cloud computing into Figure 1. Common health information exchange.
 JOURNAL OF COMPUTER INFORMATION SYSTEMS 3

exchange must wait for all responses. Ideally, the exchange is So far there are no industry standards relating to the exact
able to assemble the returned records, resolve any inconsis- composition of a SOA, although some industries have pub-
tencies or incompatibilities between records, and send the lished their own principles. Some of the key principles of SOA
records to the clinician’s EMR system.18 After the episode of include the following. First, regarding the service reusability
care, the new information is stored in the clinician’s EMR principle, application logic is divided into services with the
system only. Because information exchange between the clin- intention of promoting reuse. Depending on their granularity,
ician’s EMR system and the HIE is fully controlled by the services can be used by multiple processes and other coarse-
health care organization, what information patients will be grained services through a standard interface. Second, the
able to access may vary considerably, relying on the standardized service contract design principle keeps service
institution. contracts independent from their implementation. The service
The HIE has serious limitations. First, each EMR/PHR contracts must be documented to formalize the required
system is limited to a specific institution; it has generally not processing resources by the individual functional capabilities.
been concerned with the exchange and interoperability of Services interact by sending messages according to a commu-
EMRs and PHRs. Because EMRs and PHRs from different nications contract, as defined collectively by one or more
organizations have incompatible data formats, in most cases service-description documents. Third, services are loosely
an individual cannot transfer data from other data sources to coupled. SOA only documents the service contract and any
the patient’s EMR/PHR or share the EMR/PHR with other published service-level agreement (SLA) but it makes no
medical providers. Second, because the connections between assumptions about internal details of each service. This service
different EMR/PHR systems were point-to-point, this archi- abstract principle preserves flexibility in how services are
tecture imposes substantial processing burdens on the HIE. implemented and deployed. SOAs, for example, may use
Third, the privacy and security of electronic health informa- legacy system capabilities as services. Fourth, services are
tion is also an important concern. Patients are worried about autonomous. Services have control over the logic they encap-
what will happen when their medical information leaves their sulate, from a design-time and runtime perspective, but they
doctor’s databases, where it is protected by the Health are not subservient to other code. A service reacts to a mes-
Insurance Portability and Accountability Act of 1996 sage, but how that message was created and what will happen
(HIPAA) and the HITECH Act, and lands in the server of to any response the service creates is immaterial to the action
the HIE, controlled by technology vendors. Moreover, the that the service will take. Fifth, services are platform-indepen-
incentives for a PHR system may be not well aligned with dent. Both the consuming and SOA service systems can oper-
the institution’s goal. For example, non-visit care is not gen- ate on any platform that supports the service transport and
erally reimbursed, so strong incentives exist for the institution interface requirements. Sixth, services are discoverable and
to delay PHR implementation, even if the institution already location independent. Services are located through a service
has an EMR system with PHR functionality. registry and accessed via universal resource locators, and
therefore may move over time without disruption to consum-
ing systems. Finally, services are built, consumed, and
described using standards, such as WSDL, SOAP, and REST.
A service-oriented model
Service-oriented architecture
SOA is an architectural approach to building complex, inter- Motivations to use a service-oriented model
operable software systems from a collection of reusable mod- In health care, there is a strong need to support broader and more
ules (services) that obey service-orientation principles. A consistent integration of information systems. On one hand,
service is a function that is self-contained, well-defined, and consumers can have many different medical providers, including
does not rely on the context or state of other services. Services primary care physicians, specialists, therapists, and even alterna-
can be combined with existing software systems to provide the tive medicine practitioners to serve their diverse medical needs.
complete functionality of a large software system. Services link Consumers also take advantage of a variety of health-related
together using the technology of Web services. Web services online services that are usually provided by private companies
describe a standard way of integrating web-based applications direct to consumer, such as PatientsLikeMe, WebMD, Microsoft
using the Extensive Markup Language (XML), Simple Object HealthVault, and Google genomics for personal health informa-
Access Protocol (SOAP), Representational State Transfer tion management. Consequently, the information pertaining to a
(REST), Web Services Description Language (WSDL), and person’s past medical history typically is scattered among all the
Universal Description, Discovery, and Integration (UDDI). places where care has been given. Remarkably, no health care
The combination of services—internal and external to an institutions are legally responsible for ensuring that a complete
organization—makes up a SOA. A focus of SOA is on the PHR is available for each person when care is needed. On the
definition of service interfaces and predictable service beha- other hand, health care organizations today are challenged to
viors. Through the reuse, interoperability, and governances of build interoperable and secure PHR systems. The rapid rise of
services across internal and external organizational and pro- heterogeneous technologies and their application to heteroge-
gram boundaries, SOA can help businesses to be more neous IT infrastructures has caused health care organizations to
“agile”—in other words, enable faster and more cost-effective collect an accumulation of non-interoperable systems that not
responses to changing conditions.19 only need to work together within the organization, but also
4 J. LI

interoperate securely with other organizations to support care domains.20 Although there are industry standards for data
delivery. representation in health care, such as HL7, a fundamental
A service-oriented model offers system design and man- problem with the standards is their varied interpretation in
agement principles that support reuse and sharing of various software. Therefore, it is important for PHR systems to stan-
patient records across different units within the organization dardize the software interpretation and then implementation
or across organizational boundaries. SOA does not require the of representation and translation of patient records. The most
re-engineering of existing systems, such as EMR, clinical cost-effective way to do this is through a standardized set of
decision support system (CDSS), and computerized physician core business services that represent health records.
order entry (CPOE) systems. With SOA, existing processing Figure 2 shows a service-oriented PHR model that uses
can be combined with new capabilities to build a library of SOA to enforce basic software architecture principles and
services that are used as a part of solutions. Using shared, provides a service bus that allows for robust message trans-
standardized services that are aligned with business processes, formation and communication across different systems. The
a service-oriented model strengthens interoperability and actual services revolve around the EMR, radiology images,
security while reducing the need to synchronize data between PHR, pharmacy, and patient modules. Various devices can
heterogeneous systems. interact with the modules, including computers, smart
phones, and even mHealth devices, such as blood pressure
monitors. As Figure 2 shows, implementing SOA services to
A service-oriented PHR system manage a standardized implementation of data representa-
The inability to access patient records in multiple systems tions reduces the number of systems interface points by an
across separate health care systems has been the frustration order of magnitude.17 Instead of each system having to create
of the health care industry for years. In some communities, and sustain a point-to-point connection to another system, all
HIEs, such as Chesapeake Regional Information System for a system needs to do is transform the system representation to
our Patients (www.crisphealth.org) for Maryland and Indiana the one specified by the service, which defines the canonical
Health Information Exchange (www.ihie.com) for Indiana, form for the specific data being exchanged (such as patient
have been established to support interoperability by synchro- portal, PHRs, EMRs, Medications, and so on). Therefore, SOA
nizing data between various hospitals and health care systems. allows for the integration of diverse systems because IT
In the systems, EMR databases are synchronized using point- resources—whether applications or systems—can be made
to-point interfaces and, for less critical systems, duplicate data available through standard interfaces that do not require the
entry. The development of infrastructures that could success- specific communication protocols that an operating system
fully deliver needed patient records inside and outside the might.
complicated HIEs could equate to a costly, inefficient mess. In the health IT industry, the exchange of health data
Increasingly, many large health care systems, such as the currently focuses on two primary focus areas: the standardi-
Veterans Affairs and the Mayo Clinic, find that SOA is an zation of information/data that gets exchanged and the devel-
ideal architecture for HIE. opment of standards for connecting disparate health IT
SOA attempts to integrate many disparate systems by systems. Regarding information/data standards, medical pro-
developing functionality as a suite of interoperable services viders and technology vendors are focusing most of their
that adhere to the various protocols necessary for various effort on achieving Heath Level Seven International’s (HL7)

Figure 2. A service-oriented model for PHRs.

 JOURNAL OF COMPUTER INFORMATION SYSTEMS 5

clinical document architecture (CDA), a specification devel- Communication services

oped for clinical document exchanging between health care
Communication services include secure patient–physician
expressions that is based on a service-oriented environment.
communication and exchange of health information across
In the matter of connecting disparate systems for the
different EMR and PHR systems. First, each PHR system
exchange of health data, a service-oriented approach uses
must support patient–physician communication. The patient
standard protocols and conventional interfaces that are typi-
would sign up and compose a message, which include the
cally provided as Web services, which allow access to data and
information from the PHR or permission to access the PHR,
information from diverse systems. The benefits come in the
and send the message to the physician. The system would
way of creating efficiencies as they relate to the development
then transmit this message to the physician. Any responses
of standardized interfaces and protocols for data exchange.
from the physician would be accessible to the patient, and the
Integrating the Healthcare Enterprise (IHE) is an initiative by
physician may originate messages, including attachments with
health care professionals and technology vendors to improve
information that would be deposited automatically in the
the interoperability of healthcare information systems. The
patient’s PHR account.18
IHE Cross-Enterprise Document Sharing (XDS) profile is a
In addition, with the patient’s authorization, each PHR
specification developed for clinical document exchange
system must transmit the patient’ PHR to another EMR
between health care enterprises that is based on SOA
system so that the health care provider gets key information
environment.21 SOA is currently the preferred model for
like the patient’s medical history, allergies, persistent medical
deploying CDA content. The VA and military health system
problems, and current medications and active treatments at
both built their gateways through XDS.
the point of care. Different PHR and EMR systems connects
SOA also lends itself to more robust software because IT
each other by using Web services.
vendors can write a program functionality once and reuse it
many times and modify it when needed.
Deposit services
A PHR system is a trusted repository for people to gather,
Services store, and share personal health information. Deposits from a
A service-oriented PHR system emphasize the service con- medical provider are made securely to the PHR system (for
cept. By definition, a service is performance of work by one example, using HL 7 specifications via Web services). Patients
for another. This section presents some examples of business- may enter information through an interactive web interface or
related services in a service-oriented PHR system. Services are upload their medical records or use other methods. Patients
developed that can serve different applications, such as EMRs, can also automatically upload health information from con-
PHRs, and clinical DSS. The services can be configured and nected devices, such as pedometers and smart phones. For
shared among different software applications. data integrity, all deposits are marked with the source of the
data.

Administrative services Access to records with permission

Administrative services include authentication, authorization, Doctors or other authorized people can access patient infor-
etc. First, patients are authenticated by their own PHR system mation via authorization supplied by the patient or from the
to access their own account. Since patient records are highly ALS managed by a trusted party, such as a trusted service
sensitive, the system should authenticate using a standard registry, which is used to find a patient PHR account if the
procedure, which typically would involve two or more patient does not have or is unable to present information
factors.8 about their account (e.g., in an emergency).18 After obtaining
Second, authorization for accessing a PHR account is genuine informed consent from a patient, a physician or other
given by the account holder by logging into the PHR authorized people can access to or download information
account and setting or changing permissions. The system from the patient’s PHR account. All accesses to a PHR
sends authorization credentials electronically to medical accounts are recorded in an audit trial that is accessible in
providers that are approved to access any portion of the an easily understandable form to the account holder.
patient’s account. Each PHR system determines the level of
granularity of information shared subject to patient permis-
Data analytics services
sion, but at a minimum allows differentiations of classes of
information (e.g., doctor notes, labs, medical images, A PHR system shall provide data analytics services that can
patient notes, problems, etc). Medical provider directories help improve individual health and population health, user
are available to assist patients in transmitting authoriza- experience analysis, and software quality analysis. A PHR
tions. In an emergency, a medical provider could find a system can also be used to deliver tailored health education
patient PHR account with an account locator service (ALS), or interventions to the right people and find research subjects
which could be operated by a trustworthy organization, for clinical trials. A trusted organization could distribute
such as a government or state agency, or a trusted service clinical trial subject request queries to multiple PHR systems.
registry.18 Patients with the desired characteristics would receive
6 J. LI

notification that they are qualified for the trial, including computing technology is considered as the cost-effective, flex-
contact information for the researchers. ible, and scalable solution for building a service-oriented
Moreover, a PHR system can also be used to generate system. Its main objective is to leverage internet or intranet
anonymized and aggregated reports for researchers and public for users to share services and resources.25 National Institute
health. A trusted organization could aggregate query results of Standards and Technology (NIST) defines cloud computing
from multiple PHR and EMR systems. The results would as a model for enabling convenient, on-demand network
represent summary of patient characteristics and would be access, to a shared pool of configurable computing resources
subject to statistical disclosure control to minimize the re- (e.g., networks, servers, storage, applications, and services)
identification risk of individual patients.22 that can be rapidly provisioned and released with manage-
ment effort or service provider interaction.26 When a PHR is
operated through cloud computing it is called personal health
Privacy and compliance
cloud. Personal health cloud is a style of computing that
The privacy and security of personal health information is a enables people to capture, organize, and share portable health
stringent requirement for the PHR. A service-oriented PHR records of themselves and their family with health profes-
system must comply with the HIPAA Privacy Rule and other sionals or caregivers over the internet.
federal and state laws. The system shall guarantee that con- Cloud computing and SOA are complementary activities,
sumers would retain sole authority over access to any portion and both can play an important role in a service-oriented
of the PHR and who they trust to share the PHR with. To PHR system. SOA provides interoperability and privacy
many patients, individuals’ authorization of all uses of their between different systems that communicate with each
PHRs is the essence of privacy and thus a comprehensive other. Cloud computing’s platform and storage service offer-
privacy framework would set out circumstances in which ings can provide scalability, ubiquity, and elasticity features
informed consent must be obtained.7,23,24 No medical infor- for SOA efforts. A service-oriented personal health cloud that
mation could be disclosed without genuine informed consent, combines both cloud computing and SOA provides a promis-
especially for nonmedical purposes. There would also be ing and cost-effective way to achieve a scalable PHR system.
regular independent privacy audits, analogous to auditing In a service-oriented personal health cloud system, either
requirements for financial institutions. By checking audit patient PHR or physician EMR is stored on external servers
logs, consumers would be able to see exactly who else has and can be accessed via the web, requiring only a computer or
access to the PHR and when. a smart phone with an Internet connection.

Security services Benefits of service-oriented personal health cloud

All the security and identity management requirements of the Service-oriented personal health cloud presents many bene-
HIPAA Security Rule could be satisfied according to multiple fits. First, cloud computing makes SOA implementation sim-
technical solutions for the PHR, including authentication and ple and easily accessible to all stakeholders. PHR software
authorization, access control, user registry, encryption, digital runs on the web instead of the server, meaning no hardware
signature, immutable logging, network security, malware or software installation. Most of individual or small group
detection, and backup and disaster recovery plan. practices are looking for an implementation process much
Appropriate administrative, physical, and technical safeguards quicker than traditional client-server system. Second, ser-
must be in place to ensure the confidentiality, integrity, and vice-oriented personal health cloud also provides superior
availability of the PHR. accessibility for patients because patients and their doctors
can securely log in to the system from anywhere they have
internet connection. In combination with the mobile device
Moving PHRs to the cloud
personal health cloud could be particularly powerful and
PHR systems are divided into two categories: client server or accessible. Third, small medical practices can realize tremen-
cloud-based. Client-server systems store data in house, requir- dous cost savings from service-oriented personal health cloud.
ing a server, hardware and software be installed in the physi- One of the largest hurdles for small practices is the initial cost
cian’s office. Systems developers are not incentivized to have of PHR installation. Both PHR implementation cost and IT
their system integrate with other systems. Therefore, many resource requirements can be significantly reduced when
client-server systems do not provide a method for deploying practices choose to move medical records to the cloud.27
an SOA infrastructure. While in-house servers have tradition- Lastly, service-oriented personal health cloud makes PHR
ally been the norm, many small to medium size medical systems scalable and interoperable. Cloud computing provides
practices are looking for cloud-based solutions that can be almost unlimited CPU, storage space and bandwidth due to
taken out of the box, plugged into the Internet, and run on resource virtualization and makes it easy to add new users,
their own. doctors, or medical practices.

Service-oriented personal health cloud Comparison of PHR solution models

PHR storage and application services in the cloud promises Each PHR system shall enable seamless and secure access to
anywhere, anytime access to critical health data. Cloud integrated health records whenever and wherever needed.
 JOURNAL OF COMPUTER INFORMATION SYSTEMS 7

Table 1. Comparison of the properties of PHR solution models.

Solution models
Attribute Traditional PHR system Service-oriented PHR model Service-oriented personal health cloud
Integration A traditional PHR system is institution- A service-oriented PHR model provides system-to- Service-oriented personal health cloud
specific and not interoperable with other system integration by creating standards-based, provides system-to-system integration using
healthcare systems. interoperable services. standards-based services.
Interoperability Not interoperable Interoperable Interoperable
Reusability Not reusable Services are reusable Resources are reusable
Scalability Not scalable SOA is scalable since the system-to-system services Personal health cloud is scalable since cloud
are on demand and as needed. services are provided to the users on demand
and as needed.
Compliance Covered under HIPAA The service-oriented PHR system provided by Not covered under HIPAA
with legal consumer’s primary care site are covered under
mandates HIPAA.
Privacy control Managed by consumer’s primary care site Data privacy is controlled by privacy and Data privacy is controlled by the cloud
and the health information exchange (HIE) compliance services. provider.
Security Secure patient portal Data security is controlled by security services. Data security is controlled by the cloud
governance provider.

However, the most obvious obstacle when implementing HealtheVet account. The downloadable PHR via Blue Button can
PHRs is integrating PHR data from different electronic provide patient records from many data sources and VA patients
health systems. Where PHR and EMR systems are con- with a premium account can access data from the VA EMR
cerned, we will probably run legacy client server solutions system and communicate electronically with their healthcare
for a long time, while the latest solution models continue to team using secure messaging.
make use of more cloud services and SOA technologies VA’s EMR system, VistA, is the primary source of EMRs
become the bridge to the cloud. The SOA integration plat- that are transferred to the PHR associated with the veteran’s
form plays a key role in integrating the existing PHR system My HealtheVet account. Additional data from other sources
into cloud services and between different EMR/PHR sys- can be added to the My HealtheVet account, including self-
tems by creating loosely coupled, standards-based, intero- entered information, such as self-entered medications, Over-
perable services. Cloud computing is internet-based the-Counters (OTCs), herbals, and supplements.
computing, whereby shared resources, software, and infor- My HealtheVet portal is a PHR system implemented with a
mation are provided to users on demand and as needed. SOA. Vast majority of the systems that would hold or origi-
However, cloud data privacy could be a problem due to nate the veteran’s medical profile and data would be siloed
outsourcing health data to a cloud computing platform. In and distributed across military hospital record systems (prior
particular for a healthcare system that archives and pro- active duty records), the VA records, and private records from
cesses sensitive data. Table 1 presents a comparison of the pre- and post-military service. My HealtheVet portal must
properties of PHR solution models. Integration, interoper- integrate with back-end systems seamlessly and securely via
ability, reusability, and scalability define the solution model Web services. Moreover, since My HealtheVet portal has
in terms of operational characteristics. Legal bindings, priv- requirements to scale to a broad and diverse population, the
acy, and security help to define the solution model in terms adopted technology solution must be scalable. Therefore, My
of privacy and security characteristics. HealtheVet, VA Blue Button, and other applications leverage
Oracle’s WebLogic Portal and a SOA-based infrastructure for
integration to the VistA EMR system and different data
Case study: My HealtheVet
sources.28
This section presents a case study of My HealtheVet PHR My HealtheVet is a secure and private patient portal that
system of the Veterans Health Administration (VHA) and its provide the veteran with full control over his/her PHR. For
associated Veterans Information Systems and Technology example, the veteran could give permission to a non-VA
Architecture (VistA). physician to access the results of some procedure into the
My HealtheVet system. It makes veteran-defined role-based
and user-based access available both inside and outside the
My HealtheVet
VA system. Doctors or others can only view or edit the
My HealtheVet portal, launched on Veterans Day, 11th veteran’s PHR with an explicit permission. Furthermore, it
November 2003, is an online PHR system that enables veterans, can securely handle all logistical items including prescription
family members, and their clinicians to access and update their ordering, appointment scheduling, and health monitoring and
health records, communicate with their care providers, access providing decision analytics to particular aspects of the veter-
health education resources, refill prescriptions, and schedule an’s health.28
appointments. In August 2010, the Department of Veterans
Affairs (VA) launched the Blue Button in My HealtheVet portal.
VistA electronic medical record system
VA Blue Button allows patients to access a single, downloadable
file with their VA medical records and self-entered information. The Veterans Information Systems and Technology Architecture
All VA patients are eligible to complete a one-time process of (VistA) is an EMR system developed and deployed by VA
identity authentication to obtain an upgraded or premium My throughout the United States to all 1200+ healthcare sites of the
8 J. LI

VHA. VistA has been ranked among the best EMR systems by
physicians. Both My HealtheVet and VA Blue Button are tethered
to data from the VistA EMR system.
VistA is built on a client-server architecture, which ties
together workstations and personal computers with graphical
user interfaces at VHA facilities. VistA includes nearly 180
applications for clinical, financial, administrative, and infra-
structure need in VA integrated into a single, common data-
base, permitting all VA applications to share one single,
authoritative data source for all veteran-related care and ser-
vices. VistA EMR, more formally known as the VistA
Computerized Patient Record System (CPRS), provides a cli-
ent-server interface that allows healthcare providers to review
and update a patient’s EMR.25 VistA, which supports HL7
standards for HIE, is used in both hospital, ambulatory, and
long-term care settings. It includes modules for CPOE, Figure 3. Service-oriented architecture for My HealtheVet and VistA.
laboratory tests, MRIs, patient care diet orders, e-prescribing,
and other clinical documentation, but does not include mod-
ules for billing because VA health care is free for veterans who applications may leverage existing Medical Domain Web
quality for medical services.29 It provides an integrated inpa- Services (MDWS) (to be retired in the to-be VistA SOA environ-
tient and outpatient EMR for VA patients, and administrative ment) or new VistA SOA services developed by the VistA Service
tools to help VA deliver a high-quality medical care for Assembler (VSA) toolkit. MDWS is a fully open source collection
military veterans in the United States. of web services providing access to VistA data and common
VistA’s data store is implemented in the Massachusetts services via industry standard Simple Object Access Protocol
General Hospital Utility Multi-programming System (SOAP) while future VistA SOA services provide asynchronous
(MUMPS, also referred to M), an application environment and stateless services using the Representational State Transfer
that integrates both programming logic and data storage. (REST) protocols. Both My HealtheVet portal and VA Blue
MUMPS excels in multi-user database-driven applications, Button access VistA via MDWS web services. All major program-
such as health information systems and financial systems. ming languages have tools to turn a SOAP web service’s WSDL
The data stored in VistA’s database can be accessed by a document into code. Client applications developers can develop
variety of methods, tightly integrated custom code to loosely an application quickly via MDWS services. MDWS is simply a
coupled standard web services to networking-spanning thin web service layer that exposes the functionality of Medical
exchange mechanisms, such as eHealth Exchange.29 Domain Objects (MDOs). MDO is available as a Dynamic Link
Library (DLL) that interacts with each of the different data
sources, queries and assembles the results.
VistA evolution The service-oriented approach provides secure, controlled
VA sets out to evolve its VistA system from a set of decentralized access for authorized users while providing for fine-grained
legacy systems to an integrated, modern service-oriented environ- protection against unauthorized access or usage. Built on
ment. This service-oriented approach leverages a combination of MDWS, a robust IAM service, an integrated Master Veteran
enterprise application and SOA support infrastructure services Index (MVI) service, and well-designed network security sup-
(including service registry, secure messaging, identity and access port the high degree of security and privacy are embedded
management (IAM), master veterans index, etc.).30 The goal of into the system. Unauthorized system access is prevented, and
VistA evolution is twofold: to create interoperability and health- policy-based authorization selectively enables or prevents
care information sharing among the Department of Defense from access to individual records or data fields, as well as
(DoD), VA, and private healthcare providers, and to meet the disallowed behaviors (such as a doctor attempting to self-
highest security and privacy standards. VistA evolution will rely medication).
upon secure infrastructure, data models, and web services that VA’s IAM architecture separates identity authentication from
support an open, modular, and extensible EMR and PHR role management. User identity is established through multiple
platform. two-factor authentication methods, including methods supported
Constructing an integrated patient profile from an array of by DOD, such as Personal Identity Verification (PIV), Common
data sources, devices, and applications necessitates a high level of Access Card (CAC), and DoD Self Service Logon (DS Logon). Role
intra- and inter-system interoperability. VA has developed open management, including provisioning, attribute query, and enfor-
web services interfaces to VistA that fully empower developers to cement are implemented via open standards, such as LDAP,
develop applications that meet the varied needs of VA clinicians SAML and XACML, and SPML enabling interaction with new
and patients using pervasive, standards-based software constructs. applications including Commercial Off-the-shelf (COTS)
VistA applications generally have a presentation layer at the top, a applications.
services layer in the middle, and a data services layer at the bottom My HealtheVet portal is an important part of VA’s VistA
(Figure 3). VistA supports a variety of applications, including My evolution. The SOA-based architectural approach enables
HealtheVet, VA Blue Button, VistWeb, CPRS, etc. These PHR interoperability, security, and innovation. VistA
 JOURNAL OF COMPUTER INFORMATION SYSTEMS 9

evolution will provide or enhance the following for My and expensive to develop and implement. Strong incentives exist
HealtheVet portal. First, it provides the ability to view, down- for PHR vendors not to implement it without cost efficiency.
load, and/or transmit to third-party a summary of care record, Second, another challenge is privacy, security, and patient
such as human readable consolidated Clinical Document safety when dealing with personal health cloud. Some consumers
Architecture (CCDA). Second, it provides the ability to view may worry about what will occur when they store their sensitive
the user activity history log for views, downloads, and trans- health information in the “blackbox” of a cloud offering. Although
missions of a summary of care record. Furthermore, it enables different organizations tend to have different views on privacy and
a patient to securely send messages to and receive messages security, it is important to consider direct misuse and improper
from health care providers.30 disclosure of personal health information, especially in a complex
mobile cloud environment.
Third, most SOA implementations in the health care industry
Move to the cloud are currently for high-end projects like My HealtheVet and Mayo
VA plans a move to use cloud computing and mobile technologies Clinic. Since the majority of physicians in the United States are
to modernize VistA, Blue Button, and My HealtheVet programs.31 individual or small group practices, most are looking for solutions
Currently, VA uses VA-owned, VA-provided services and devices. that can be taken out of the box, plugged into the internet, and play
This limits flexibility and elasticity of application services, devices, on their own. Therefore, there is a strong need to develop cost-
and staffing. The deployment environment for VistA and My effective, flexible and scalable PHR solutions that works well in
HealtheVet will adhere to a “cloud first” policy, with mechanisms both high- and low-end environments. The proposed service-
and provisions in place to consider the most appropriate cloud oriented personal health cloud provides interesting and cost effec-
deployment option, including private cloud offerings or infra- tive opportunities to both high- and low-end environments.
structure/platform hosted on public clouds. And last not least, there is a lack of guidance and governance
Cloud computing will be tested at three levels.32 At the first around a standardized approach to service-oriented solution
level, VA has created an “internal cloud” for consolidating operat- models. HL7’s CDA for standardization of content works well in
ing of the VistA EMR system in its data centers, rather than storing both high- and low-end environments. However, in the case of
the information at hospitals. The second level of the VA’s cloud connecting systems for data exchange, there are no standards for
computing is that the department has outsourced data centers and interoperability. While the IHE XDS profile has the potential to
hardware for cost saving benefits. VA plans for expansion of improve the interoperability of large healthcare information sys-
Infrastructure as a Service (IaaS), Software as a Service (SaaS), tems, there currently is no standardized cloud computing frame-
and Network as a Service (NaaS). At the third level, VA will design work that works well for connectivity and the retrieval of low-end
personal cloud in which the thousands of VA physicians have data.
access to application services and data on their personal Black
Berries, iPhones, iPads, Android devices, and other mobile Limitations and suggestions for future research
devices.
This paper has limitations that create some interesting opportu-
nities for future research. First, we discuss a service-oriented
Discussion and conclusion model for PHRs without detailed design specification. Further
research in this area should focus on reference architecture and
In this paper, we particularly examine a service-oriented implementation of cloud-based PHR systems that will support a
model for PHR systems and explore the potential of service- large variety of clinical settings, ranging from individual or small
oriented personal health cloud. We present a case study of My clinics to large medical care delivery systems or Web services.
HealtheVet portal. The debate on methods for PHR systems In addition, many PHR systems consider using sophisticated
development and integration is far from over, but this study consumer decision support tools and mobile devices for chronic
provides insights into this emerging class of health informa- disease management, but this paper does not consider the impact
tion systems. of the innovations on the interoperability and privacy of PHRs.
The rapidly developing uses of shared decision making, data
mining, big data, and mobile technologies in business and health
Policy and practical implications
care signal even more demand for cloud-based PHR systems. The
A service-oriented PHR model can be developed to overcome literature on the integration of big data and its impact on data
barriers, such as interoperability, reusability, flexibility, and scal- privacy and security is quite limited. It is important to continue
ability along with security and privacy. As with any systems examining the service-oriented PHR system in the context of big
development and integration project, there are policy and practical data and cloud computing.
implications to keep in mind when designing a service-oriented
PHR model. First, one primary challenge of SOA adoption is that
Conclusion
PHR providers are not incentivized to have their systems integrate
with other systems. It is relatively easy to architect a tethered PHR Despite these limitations, this paper makes important contribu-
system, such as My HealtheVet and Mayo Clinic, because such a tions. This study incorporated the principles of SOA and cloud
system is tethered to data from a single health care organization. computing into a service-oriented PHR model and proposed a
However, a service-oriented environment across organizational new service-oriented personal health cloud model. It verified and
boundaries, such as SOA and HIE, can be extremely complex validated the proposed approach via a practical case study of the
10 J. LI

VHA’s My HealtheVet. It developed more robust understanding 435-438. IEEE Symposium on Service-Oriented System Engineering
of the interoperability and privacy issues with PHR/EMR systems (SOSE); 2017 Jun; p. 38–46. doi:10.1109/HealthCom.2015.7454539.
14. Wu Y, He F, Zhang D. Service-oriented feature-based data
and offered insights into systems development and integration.
exchange for cloud-based design and manufacturing. IEEE Trans
Extending health care services through PHRs, eHealth, and Serv Comput. 2018;11(2):341–53. doi:10.1109/TSC.2015.2501981.
mHealth are becoming increasingly popular in health care, parti- 15. Zaied ANH, Elmogy M, Elkader SA. A proposed cloud-based frame-
cularly under the urgent demand due to rapidly aging populations work for integrating electronic health records. Paper presented at: 10th
throughout the world. The benefits of ubiquitous health care International Conference on Informatics and Systems; 2016 May 09–
11; p. 139–45. doi:10.11465/2908446.2908478.
services are very clear. A service-oriented PHR model offers a
16. Li J. A service-oriented approach to interoperable and secure
flexible and scalable method of integrating health data as informa- personal health record systems. Paper presented at: 11th IEEE
tion is collected from heterogeneous systems. Future research in Symposium on Service-Oriented System Engineering (SOSE);
this area should focus on reference architecture and implementa- 2017 Jun; 38–46. doi:10.1109/SOSE.2017.23.
tion of service-oriented personal health cloud that will support a 17. Juneja G, Dournaee B, Natoli J, Birkel S. Improving performance of
healthcare systems with service oriented architecture. 2008 [accessed
large variety of clinical settings, ranging from individual or small
2018 May 20]. http://www.infog.com/articles/soa-healthcare.
clinics to large medical care delivery systems. 18. Health Record Banking Alliance White Paper. A proposed national
infrastructure for HIE usign personally controlled records. 2013
[accessed 2018 May 20]. http://www.healthbanking.org/docs/HRBA
ORCID %20Architecture%20White%20Paper%20Ja.n%202013.pdf.
Jingquan Li http://orcid.org/0000-0002-9654-1770 19. Zhang D. Web services composition for process management
in E-Business process. J Comput Inf Syst. 2005;45:83–91.
20. Nano O, Zisman A. Realizing service-centric software systems.
References IEEE Software. 2007;24(6):28–30. doi:10.1109/MS.2007.166.
21. Chavis S. A service-oriented approach. For Rec. 2011;23:10.
1. Roberts N, Mellott M. Electronic medical record system avoidance 22. Li J. Data protection in healthcare social networks. IEEE Software.
in a turbulent environment. Inf Manag. 2016;53(5):581–90. 2014;31(1):46–53. doi:10.1109/MS.2013.99.
doi:10.1016/j.im.2016.01.003. 23. Li J. Privacy policies for health social networking sites. J Am Med
2. Wilson V, Strong D. Editors’ introduction to the special section Inform Assoc. 2013;20(4):704–07. doi:10.1136/amiajnl-2012-001500.
on patient-centered e-Health: research opportunities and chal- 24. Li J, Li X. Privacy preserving data analysis in mental health
lenges. Commun Assoc Inf Syst. 2014;34:323–36. research. Paper presented at: First IEEE International
3. Archer N, Fevrier T, McKibbon K, Straus S. Personal health Conference on Big Data; 2007 Jul; p. 95–101. doi:10.1109/
records: a scoping review. J Am Med Inform Assoc. 2011;18 BigDataCongress.2015.23.
(4):515–22. doi:10.1136/amiajnl-2011-000105. 25. Zhang L, Zhou Q. Cloud computing open architecture. Paper
4. Frasier HS, Biondich P, Moodley D, Choi S, Mamlin BW, presented at: IEEE International Conferen on Web Services; Jul
Szolovits P. Implementing electronic medical record systems in 2009, Los Angeles, CA: p. 607–17.
developing countries. Inform Prim Care. 2005;13:83–95. 26. Mell P, Grance T. NIST definition of cloud computing.
5. Li J. Improving chronic disease self-management through social Gaithersburg, MD, USA: National Institute of Standards and
networks. Popul Health Manag. 2013;15(5):285–87. doi:10.1089/ Technology; 2011.
pop.2012.0110. 27. CareCloud. 5 advantages of a cloud-based EHR for small prac-
6. Gold JD, Ball MJ. The health record banking imperative: a conceptual tices. 2011 [accessed 20 Feb 2018]. http://www.carecloud.com/
model. IBM Syst J. 2007;46(1):43–55. doi:10.1147/sj.461.0043. blog/5-advantages-of-a-cloud-based-ehr-for-small-practices/.
7. Li J. Ensuring privacy in a personal health record system. 28. Oracle White Paper. How providers can lower costs and
Computer. 2015;48(2):20–27. doi:10.1109/MC.2015.43. improve patient care using evidence based medicine. 2009
8. Li J. A privacy preservation model for health-related social networking [accessed 2018 May 20]. http://www.oracle.com/us/industries/
sites. J Med Internet Res. 2015;17(7):e168. doi:10.2196/jmir.3973. 018896.pdf.
9. Kierkegaard P. Electronic health record: wiring Europe’s health- 29. The Department of Veterans Affairs. VistA monograph. 2013 Oct
care. Comput Law Security Rev. 2011;27(5):503–15. doi:10.1016/j. [accessed 2018 May 20]. http://www.ehealth.va.gov/VistA_
clsr.2011.07.013. Monograph.asp.
10. Kart F, Moser LE, Melliar-Smith PM. Building a distributed 30. The Department of Veterans Affairs. SOA design patterns for
e-healthcare system using SOA. IT Prof. 2008;10(2):24–30. VistA evolution: web technologies data sharing for VistA evolu-
doi:10.1109/MITP.2008.22. tion. 2014 [accessed 2018 May 20]. http://www.techstrategies.oit.
11. Amendolia SR, Estrella F, Hassen W, Hauer T, Manset D, va.gov/docs/designpatterns/AchC_ENTDP_Inc2Rel1_
McClatchey R, Rogulin D, Solomonides T. Mammogrid: a service WebTechDataVistAEv.pdf.
oriented architecture based medical grid application. Lecture 31. The Department of Veterans Affairs. OneVA enterprise technol-
Notes Comput Sci. 2004;3251:939–42. ogy strategic plan. 2014 [accessed 2018 May 20]. http://www.ea.
12. Bahga A, Madisetti V. A cloud-based approach for interoperable oit.va.gov/EAOIT/docs/OneVA_Enterprise_Technology_
electronic health records (EHRs). IEEE J Biomed Health Inform. Strategic_Plan_Feb_28_2014_final1.pdf.
2013;17(5):894–906. doi:10.1109/JBHI.2013.2257818. 32. Lipowicz A. VA CIO describes cloud, blue button and EHR
13. Bahrami M, Singhal M. A dynamic cloud computing platform for programs. 2011. [accessed 2018 May 20]. https://fcw.com/arti
eHealth systems. Paper presented at: 17th International Conference on cles/2011/05/05/va-cio-roger-baker-describes-cloud-blue-button-
E-health Networking, Application & Services (HealthCom); 2015: and-ehr-initiatives.aspx.