Beijing Law Review, 2012, 3, 145-151
http://dx.doi.org/10.4236/blr.2012.34020 Published Online December 2012 (http://www.SciRP.org/journal/blr) 145
Regulatory Privacy Protection for Biomedical Cloud
Computing
Y. Tony Yang1, Kari Borg2
1Department of Health Administration and Policy, George Mason University, Fairfax, USA; 2Department of Ophthalmology, Poly-
clinic, Seattle, USA.
Received August 7th, 2012; revised September 4th, 2012; accepted September 15th, 2012
ABSTRACT
This article provides background information on biomedical cloud computing. It examines the privacy concerns that
arise from the use of biomedical cloud computing services and then surveys the current state of regulatory privacy safe-
guards for patients and consumers of these services both in the US and abroad. Finally, it identifies opportunities for
legal and technological mechanisms to be implemented or reinforced so that patients and consumers are not forced to
lose control of their information when they use biomedical cloud computing services.
Keywords: Privacy; Law; Regulation; Biomedicine; Cloud Computing
1. Introduction
Advances in technology and progressions in web-based
software have led to the ability to share information with
ease and efficiency. In relation to health care, this has
created the ability to access patient’s electronic medical
records from virtually anywhere. Health care professionals
now have the capability to save more lives with faster
access to pertinent medical information. Medical resear-
chers have also been utilizin g this technology to compile
data as well as allow other users the right to use the data
for further exploration. These researchers are now gather-
ing data from a virtual storage facility where they are
able to store information and permit other researchers to
access it as well. This form of storing data is known as
“cloud computing”: operates through a third party orga-
nization that monitors the use, security, and accessibility
of the stored data. While this has created major innova-
tions in health care research, it brings up a serious con-
cern regarding patient privacy and confidentiality. Re-
cords of patients’ personal medical histories, as well as
other identifying data, have a high risk of being abused
when stored in the cloud. Patient data is now in a com-
plex, virtual world that is constantly threatened by hac-
kers and internal b reaches in security. This ongo ing issue
has created debate on how to properly protect patient pri-
vacy while still allowing professionals to have the access
they need to important data.
2. Biomedical Cloud Computing
Biomedical data is increasingly being stored on the “cloud”.
The cloud allows for multiple users to access a shared
data for research much faster than ever before. Data that
used to take researchers months to gather can now be
extracted from the cloud in minutes or even seconds
(Crawford [1]). One example of a biomedical cloud is
that of Ohio State University. They have created a cloud
system known as the Translational Research Informatics
and Data management grid or “TRIAD” (Crawford [1]).
This software was developed to allow most forms of bio-
medical data to be uploaded and stored in this cloud, with
software then interpreting the data into a standard uni-
versal language (Crawford [1]). “When it comes to bio-
medical research, you have the digital equivalent of the
Tower of Babel. One piece is written in French. And
another is written in Russian . And maybe a third compo-
nent is in Chinese,” says Philip R. O. Payne, a researcher
and chair on the department of biomedical informatics at
the Ohio State University (Crawford [1]). Researchers
can now extract, upload and share data without the need
to spend time and money on tran slatin g th e data into their
own “ language”.
United Health Inc. is also creating a new cloud aimed
at merging the financial aspect of healthcare. For exam-
ple, patients will have their billing combined from vari-
ous locations regarding the same diagnosis or procedure
that was done (Mathews [2]). UnitedHealth is also aim-
ing at creating a cloud with data pooled from multiple
electronic health records to allow doctors to see if their
patient was seen elsewhere, as well as for emergency
rooms seeking the entirety of their patients’ charts while
treating them (Mathews [2]). These private health record

Regulatory Privacy Protection for Biomedical Cloud Computing
146
systems are believed to improve doctor-patient commu-
nication, make understanding and use easier, and reduce
the risk of medical errors (Carrión, Alemán & Toval [3]).
Electronic Health Records (EHRs) that are accessed by
multiple organizations or even different locations, rely on
some form of off-site, or cloud, computing (McCarthy
[4]). Although the benefits of these systems to the medi-
cal research world are significant, the importance of a
solid framework to manage the security of the data can-
not be understated. While all of these interfaces have pri-
vacy policies in place, the virtual nature of the data
means that it is not primarily protected by physical de-
vices such as locks or buildings. The cloud’s virtual sto-
rage lockers are at risk for being hacked into regardless
of the privacy policies put in place to regulate who can
access personal identifying information.
3. Privacy Risks in the Clouds
Organizations that manage data clouds are required to
have privacy policies in place to create a secure network
for their users and for the patients whose data has been
extracted. One of the top concerns with storing sensitive
medical data is risking the patients’ identifiab le informa-
tion (PII). Medical researchers keep data concerning pa-
tients’ medical diagnosis, family history, HIV status, etc.
It also includes patients’ social security numbers, addre-
sses, date of births, among many other records. One form
of preventing identifiers from being stolen and abused is
to have the system erase all unnecessary personal data
when information is being pulled and correlated with
other figures. For example, TRIAD has adopted an inter-
face that allows researchers to connect tissue samples
with medical records that will de-identify the record ma-
king the correlation completely anonymous (Crawford
[1]). By making it difficult for users to attain identifying
information, a significant amount of potential identity theft
is deterred. Biomedical cloud providers must follow strict
guidelines in regards to patients'privacy by the Health
Information Portability and Accountability Act (HIPAA).
For instance, there are eighteen types of identifying in-
formation that must be removed before biomedical data
can be shared with other parties without the consent of
the patient (Klein [5]). Some additional restrictions in-
clude being sure the server for the network is within the
United States and must have physical means of protec-
tion such as cages, back-up power, security guards, etc.
They must also follow guidelin es set forth by the Privacy
Act of 1974 which protects health information and al-
lows patients the right to see and change information on
their records (Osterhaus [6]). In 2009 the Health Infor-
mation Technology for Economic and Clinical Health
(HITECH) [7]. Act was signed to encourage meaningful
use of health care technology and “strengthen the civil
and criminal enforcement of the HIPAA rules”. The
HITECH act made the regulations presen t in HIPAA ap-
plicable to not just healthcare organizations, but also the
cloud service providers that they utilize, closing an im-
mense loophole in security (Delgado [8]).
A group called the Cloud Security Alliance is a non-
profit organization “with a mission to promote the use of
best practices for providing security assurance within
Cloud Computing, as and to provide education on the
uses of Cloud Computing to help secure all other forms
of computing”. The Cloud Security Alliance has deter-
mined the top seven threats to cloud computing as (Klein
[5]): 1) Abuse and nefarious use of cloud computing; 2)
Insecure application programming interfaces; 3) Malici-
ous insiders; 4) Shared technology vulnerabilities; 5) Data
loss or leakages; 6) Account, service and traffic hijacking;
7) Unknown risk profile.
Users require that cloud providers be trustworthy for
storing their data. To create trust there must be security
in place for all of the sensitive information. There are a
number of ways in which to address this complex con-
cern. One is by examining three aspects involved in the
security of these virtual storage facilities: preventative,
detective, and corrective controls (Ko, et al. [9]).
Preventative controls are the measures put in place to
prevent breaches from occurring in the first place (Klein
[5]). These can include encryption technology, firewalls,
lists of approved users, etc. Many of the controls spe-
cified by HIPAA fall under this category, and all can be
readily implemented by cloud service providers to suit
the specialized needs of the health industry (Martin [10 ]).
Each layer of protection that can be added must be evalu-
ated for both its costs and benefits. For instance, one pos-
sible method of preventing data loss is having dedicated
and distinct servers for each client. While this increase in
isolation bolsters security by minimizing “accidental”
thefts when the target was on the same server, this ap-
proach increases the costs to the server provider. Addi-
tionally, this diminishes the benefit of having the infor-
mation stored in multiple locations, which minimizes loss
because of hardware malfunctio ns ( Palan zi [11]).
One more sophisticated, albeit partial, means of pre-
venting unnecessary transfer of information is through a
cloud-based technique called Virtual Machines (VMs).
Virtual machines are minimal terminals that load their
operating system from a server, either through local in-
frastructure or the internet (Grossman & White [12]). Most,
if not all, processing is performed by the server. Rather
than transferring the entire contents of the file system
over the internet, only the pixels displayed on the screen
are transmitted. In the case of viewing of medical images
(which range from 2 - 500 mb) this approach drastically
reduces the amount of bandwidth necessary to view and
manipulate files remotely. The nature of VM systems pre-
vents much of the unnecessary transferring of sensitive

Regulatory Privacy Protection for Biomedical Cloud Computing 147
patient information from the data center to a terminal
when accessing EHRs. Additionally, this approach is
generally more cost effective, reliable and easier to con-
figure than traditional approaches that rely on the instal-
lation and maintenance of one operating system install
per computing system (Philbin, Prior & Nagy [13]). Fu r-
thermore, they provide a uniform computing environment
whether accessed from work, home or on the road.
Detective controls consist of procedures taken to iden-
tify risks in the system (Klein [5]). They include guide-
lines for security administration (such as routinely con-
ducting risk analyses and implementing policies and pro-
cedures to address vulnerabilities), screening and edu-
cating the workforce, and activity audits (Schweitzer [14]).
Corrective controls are actions to resolve breaches that
have already occurred and need to be fixed (Klein [5]). In
having the cloud provider monitor these controls, trust
can flourish by the users knowing which and how secu-
rity measures are being applied to secure their informa-
tion. Creating a strong structure for managing these con-
trols is the key for adherence to the privacy policies go-
verning health data clo uds.
Predictive, detective, and corrective controls are ap-
plicable to both internal and external threats of informa-
tion privacy. Many, if not most, breaches of sensitive data
are perpetrated by internal threats, i.e. employees wor-
king at either the healthcare organization or cloud service
provider. There are two important themes underlying in-
ternal security threats: accountability and auditability.
Of the seven threats outlined by Ko, et al. [9], all but
the 4th and 6th risk can be addressed through increasing
accountability and auditability of the cloud providers
(Klein [5]). They state that to create trust in the cloud
provider, the provider must address these issues to show
its users the measures they are taking to secure data and
private information (Klein [5]). Consumers who are sto-
ring data in the cloud must have full trust in the cloud
provider due to the fact that they are keeping substantial
medical and personal identifying information in this sys-
tem (Carrión, Alemán & Toval [3]). This creates incen-
tive for the provider to show its users how they are ad-
dressing these privacy issues. Cloud providers must be
able to provide accountability for any problems that arise
within their system and display this to their users.
They must also have their system logs and design clearly
presented for their customers, so that users can feel se-
cure in knowing the activity and protection of their ac-
counts—Something that many cloud services currently
fail to do. There are no means to audit a cloud structure if
there is no monitoring system for users so that they can
survey how and when their data is stored and accessed.
Accountability and auditability are crucial for health or-
ganizations to feel comfortable in fully utilizing the po-
tential powers of the cloud; thorough logging is the cor-
ner stone necessary for both (Pearson [15]).
Accountability in a cloud system is necessary for de-
termining which party is responsible wh en a secur ity breach
occurs. If information is taken by an employee of the
health institution, liability typically falls similarly. How-
ever, if the breach is perpetrated by an employee of the
cloud service provider or an external agent, responsibi-
lity should lie with the service provider. A retrospective
analysis of system logs can typically reveal the origin of
a breach, but how to efficiently (in terms of size) and ef-
fectively (in terms of preventing future breaches) log
health information is still an open question. Many clients
that use VM are not fully aware of the linkages between
virtual and physical servers, relationships between virtual
and physical server locations, and how files are written
into both virtual and physical memory addresses (Ko, et
al. [9]). The complexity inherent in virtual machine set-
ups can be a difficult challenge in terms of determining
the origin of a breach. Without proper measures imple-
mented ahead of time, assessing liability may be impos-
sible—Put simply, a legal nightmare.
Auditability refers to the relative ease of auditing a
system or an environment. A system lacking in audit-
ability has poorly maintained records (Soma, et al. [16 ]).
The complex nature of cloud data exchange necessitates
a multilayer approach for modeling information exchan-
ges. At the system level, there is the operating system,
the file system, and the network, as well as the logs cor-
responding to the read/write operations between them. At
the data layer, there is the logging of each detail regar-
ding a single file: its origin, the location(s) where it is
stored, which accounts what permissions to access to it,
and if/when it will expire. The most important bit of in-
formation in logs is the time-stamped record of account
accessing and interacting with files. So long as these de-
tails are both comprehensive and pr operly protected with
encryption, establishing liability is usually possible. Fi-
nally, assessment of the workflow level reveals the ro-
bustness or weakness of the control regarding user/ter-
minal interaction: are employees only accessing the mi-
nimal amount of information they need to accomplish a
task? Through workflow analysis, problem areas and se-
curity loopholes in the clou d are removed or rectified and
control and governance of the cloud processes are im-
proved.
Kirchberg et al. have suggested that a file-centric per-
spective should take prevalence in health-related cloud
computing security (Ko, et al. [9]). Network logs that trace
the life cycle of files (i.e. creation, modification, duplica-
tion and destruction) within clouds best enable account-
ability and auditability. Additionally, event data regar-
ding network activities and actor data (the person or com-
puter component that triggered an event) are necessary
for proper representation of the file life cycle. Estab-

Regulatory Privacy Protection for Biomedical Cloud Computing
148
lishing different network zones is perhaps the best way to
organize network logs. A cloud service provider, for in-
stance, can designate its own network as internal and the
network addresses of the health organization as a safe
zone. Data that is relayed to addresses besides these zones
can be flagged for review by an automated or human su-
pervisor. In the future, file types could be designed to
include some space for the local storage of pertinent
meta-information (such as read/write histories), making
the logging needs of the distributed cloud system less
complex (Ko, Lee & Pearson [17]).
The lack of transparency within the cloud service pro-
vider’s system is the greatest concern for prospective
cloud users. One survey found 88% of potential cloud
customers were concerned about who would have access
to their data (Fujitsu Research Institu te [18]). Some EHR
cloud providers, such as Microsoft Health Vault, fail to
provide thorough accounts of data storage and access for
their users. All contractual relationships between cloud
service providers and medical organizations should be
considered with care; especially in cases when all of the
relevant information for assessing liability may not be
accessible by both parties. For this, and other security
reasons, the organization needs to negotiate a strong con-
tract with the cloud service provider featuring compli-
ance with HIPAA, security controls, and auditab ility (Witt
[19]). If these requirements can be met, the benefit from
shared data sets between health researchers will be im-
mense. In the near future, biomedical data can be mined
for statistical patterns between an individual’s genome,
th erapy, and conditions (Grossman [12]). In creasingly open
and shared data means greater statistical power in de-
tecting predispositions to diseases early.
4. Regulatory Protection for Privacy
4.1. Privacy Law in the US and Abroad
Biomedical enterprises operating in the United States
need to consider HIPAA (45 C.F.R. §§ 160 - 164) and the
Electronic Communications Privacy Act (ECPA) (18 U.S.C.
§§ 2701 - 2712), a federal statutory law that supports
specific protections for electronic communications (in
transit or in storage) to supplement any protections offered
by the Fourth Amendment of the United States Constitu-
tion. Such laws require privacy and data security obliga-
tions.
European-based biomedical enterprises, as well as en-
titles working with providers in or with infrastructure in
Europe, however, need to take into account the expan-
sive requirements under local omnibus data protection
laws that safeguard all personal information, even basic
details like business contact information. These require-
ments can involve informing employees, customers, or
other individuals about the outsourcing and processing of
their data; obligations to co nsult with works councils be-
fore outsourcing employee data; and registering with local
data protection authorities (Lyo n [20]).
4.2. Requirements for Data Security
Even if an enterprise is not subject to these types of pri-
vacy laws, it will want to ensure safeguards for personal
information covered by data security and breach notifica-
tion laws. In the United States, these laws focus on per-
sonal information such as social security numbers, dri-
ver’s license numbers, and credit or debit card or financial
account numbers. One of the key safeguards is encryp-
tion because many (although not all) of the US state
breach notification laws provide an exception for encryp-
ted data. In contrast, many other countries require protec-
tion of all personal information, and do not necessarily
provide an exception for encrypted data. Consequently,
companies operating outside of the United States may
have wider-reaching obligations to protect all personal
information. While data protection obligations vary con-
siderably from law to law, both US and intern ational pri-
vacy laws commonly require the following types of safe-
guards: 1) Conducting appropriate due diligence on pro-
viders; 2) Restricting access, use, and disclosure of per-
sonal information; 3) Establishing technical, organizational,
and administrative safeguards; 4) Executing legally suffi-
cient contracts with providers; and 5) Notifying affec-
ted individuals (and potentially regulators) of a security
breach compromising personal information (Delgado [8]).
The topic of data security in the cloud has received
significant attention lately. The National Institute of Stan-
dards and Technology (NIST) has finalized its first set of
guidelines for managing security and privacy issues in
cloud computing in early 2012. Guidelines on Security
and Privacy in Public Cloud Computing (NIST Special
Publication 800 - 144) provides an overview of the secu-
rity and privacy challenges facing public cloud compu-
ting and presents recommendations that organizations
should consider when outsourcing data, applications and
infrastructure to a public cloud environment (Lyon [20]).
The document provides insights on threats, technology
risks and safeguards related to public cloud env ironments
to help organizations make informed decisions about this
use of this technology. SP 800-144 is geared toward sys-
tem managers, executives and information officers ma-
king decisions about cloud computing initiatives; secu-
rity professional responsible for IT security; IT program
managers concerned with security and privacy measures
for cloud computing; system and net-work administrators;
and users of public cloud computing services (Delgado
[8]).

Regulatory Privacy Protection for Biomedical Cloud Computing 149
4.3. Restrictions on Data Transfers
Internationally
A growing number of countries, especially in Europe,
restrict the transfer or sharing of personal information be-
yond their borders. These restrictions can present sig-
nificant challenges for multinational biomedical enter-
prises seeking to move their data to the cloud. Recogni-
zing these challenges, some providers are starting to offer
geographic-specific clouds, in which the data are main-
tained within a given country or jurisdiction. Some US
providers have also certified to the US-European Union
(EU) Safe Harbor program, in order to accommodate
EU-based customers and patients. However, as the Safe
Harbor only permits transfers from the EU to the United
States, it is not a global solution. Accordingly, a bio-
medical enterprise should assess carefully whether the
options offered by a provider are sufficient to meet the
enterprise’s own legal obligations in the countries where
it operates (Delgado [8]).
5. Critical Recommendations for
Reinforcement of Privacy Protection
5.1. Legal Reform
The law has long recognized the importance of privacy.
Existing statutory privacy law, however, needs a techno-
logical upgrade. For example, ECPA should, but does not
clearly define the statutory protections applicable to cloud
computing services. ECPA does not distinctly express
whether documents stored with many cloud computing
services are protected at all. ECPA, as currently written,
provides protection where con tent is stored with a service
“solely for the purpose of providing storage or computer
processing services to such subscriber or customer, if the
provider is not authorized to access the contents of any
such communications for purposes of providing any ser-
vices other than storage or computer processing.” It is
not clear whether sites that provide collaboration and
sharing functions or employ a targeted advertising busi-
ness model based on information contained in documents
are covered by this clause. Even if ECPA does cover
cloud computing records in a specific situation, the pro-
tections that it provides are insufficient to properly safe-
guard the privacy of sensitive documents being stored
with cloud computing services. Beyond ECPA, there are
questions about whether other specific privacy laws or
regulations fully protect consumers of cloud computing
services. For instance, the HIPAA is designed to protect
the privacy of health records. HIPAA applies to health
care providers, health care clearing-houses, and health
plans (insurers). But, it is not clear whether HIPAA pro-
tections apply to cloud computing services that store
consumer and patient health records. Similarly, does the
Video Privacy Protection Act (18 U.S.C. § 2710), which
provides statutory protection for video rental records and
“other similar material,” protect records of audiovisual
material shared or retrieved through a cloud computing
service?
Therefore, as biomedical cloud computing continues to
evolve, it is imperative th at privacy laws and policies are
updated so that consumers and patients have the clarity
needed to make informed choices and feel confident that
their personal information is being protected (Weissber-
ger [21]). To this end, Congress might consider imple-
menting some of the enfo rcement provisions of the False
Claims Act to strengthen HIPAA. This would allow a
private right of action measure similar to the FCA’s qui-
tam actions (Palanzi [11]). Furthermore, by increasing and
modifying the fine structure for services that store large
numbers of electronic medical records on cloud networks,
the compliance of cloud companies to meet new regula-
tory standards may be financially incentivized (Palanzi
[11]).
5.2. Contractual Protections for
Cloud-Bases Services
Cloud providers must be trus ted to maintain the integrity
and security of the data they a re tasked with sto ring. Esta-
blishing this trust can be achieved by means of a con-
tract which allows the customer the op tio n of strong leg al
recourse in the event of a data breach. The contract will
govern the conditions upon which data is given to the
cloud computing provider and should be drafted in con-
sultation with a local attorney or other third party spe-
cializing in data security laws (Harshbarger [22]).
Contractual provisions for cloud based services should
include an insurance policy for damages resulting from
compromised data, as well as a stipulation of service le-
vels. The former demonstrates that the cloud provider has
the finical ability to pay a breach of data while the latter
clearly delineates the responsibilities on the part of the
cloud service provider. Examples of service level stipula-
tions include having the cloud provider state that it will
have a ninety-nine percent up-time and have its system
free from bugs or other defects (Harshbarger [22]). In
addition, the cloud provider could negotiate additional
operational specifications which might be unique to the
nature of the data the customer needs to be stored. This
would give the customer a defined set of expectations in
relation to the cloud provider’s handli ng of sensitive data.
A contract would also allow for increased transparency
on the part of the cloud provider. A provision for trans-
parency would obligate the cloud provider to make its
data security regime available to the customer so that
specific precautions and safeguards are known (Harsh-
barger [22]). This transparency will increase trust and

Regulatory Privacy Protection for Biomedical Cloud Computing
150
confidence among cloud service purchasers as well as
give tangible meaning to the notions of cloud-based data
security.
5.3. Vigorous Privacy Practices from Biomedical
Enterprises
Biomedical enterprises have the opportunity to ardently
address much of this patient and consumer concern by
following the c ore principles of the Fair Information Pra c-
tices (Federal Trade Commission [23]). This means pro-
viding meaningful notice about how information is used
and to whom it is disclosed, collecting and retaining only
the information that is need ed to provide services, giving
patients and consumers real choice about how any per-
sonal information collected abou t them will be used, pro-
perly safeguarding patient and consumer information
from disclosure and misuse, and enabling them to control,
modify, and delete their own records and accounts (Tu-
row, et al. [24]). Providing patients and consumers with
meaningful control and protection for their personal in-
formation will help give them the confidence to utilize
cloud computing and may also help biomedical enter-
prises avoid negative press, government investigations,
and costly lawsuits (Ozer [25]).
Patients and consumers expect that data stored with a
cloud service provider will stay private; pro viders have a
commercial incentive to make sure that it does. By de-
signing a service with technical measures to protect con-
sumers and patients—Tools that allow them to manage
and protect their own information, encryption and anony-
mity protocols to protect information by default, and ac-
cess controls and data security measures to prevent brea-
ches and inappropriate disclosures—Biomedical cloud com-
puting providers can establish a platform where patients
and consumers are in a position to control their own in-
formation and can feel more confident banking private
content. The key step in giving patients and consumers
control is to build a vigorous and operative interface to
allow them to be in charge of their own content and re-
cords. Consumers and patients should be able to view
and control their entire record. Constructing such an in-
terface is much easier if it is part of the design process of
the service and not added on as an after thought or in
response to consumer and patient demands for greater
control and transparency (Pearson & Charlesworth [26]).
Anonymization and encryption can also protect con-
sumers by reducing the risk of disclosure of information
that is captured and stored by the service. Anonymization
procedur es, however, shou ld ensure that data is irreversi-
bly de-identified (Narayanan & Shmatikov [27]). More-
over, creating a solid data security plan protects not only
patients and customers but also providers. Data breaches
can be disastrous, leading to lawsuits, fines, and lost trust
(Soma, et al. [16]). In order to avoid these outcomes, pro-
viders should use access controls to prevent unauthorized
access to content by both employees and third parties and
take additional steps such as promptly deleting data that
is no longer necessary in order to reduce the risk of
breach. Such practices will help safeguard both patient
and customer privacy and the provider’s bottom line.
Providing technical measures that protect and secure pa-
tient and consumer information may carry both practical
and legal significance. Practically, the measures sug-
gested above, and others that may emerge, reduce the
likelihood of breach or unnecessary disclosure. In addi-
tion, these mechanisms may strengthen the legal positions
of both consumers and providers by making it clear that
the patient or consumer, and not the provider, is the party
with access to and control over any stored content. The
more “locks” a provider puts in the patient or consumer’s
control, the less likely it is that third parties will be as-
king providers for the keys (Wayner [28]).
6. Conclusion
As biomedical cloud computing continues to develop and
expand, it is critical to reinforce regulatory mechanisms
to protect the privacy of consumers and patients. Courts
and policymakers need to recognize the realities of mo-
dern information storage and satisfy the continued ex-
pectations of privacy, regardless of whether the informa-
tion is stored online or offline. Biomedical enterprises
should invest in privacy-friendly technologies and prac-
tices that put consumers and patients in control of their
own private information. They should also support regu-
latory reform to update any outdated statutory under-
standings of online privacy.
REFERENCES
[1] D. Crawford, “Biomedical Research Gets Head Into
Cloud Computing,” 2011.
http://medicalcenter.osu.edu/mediaroom/releases/Pages/B
iomed-
cal-Research-Gets-Head-Into-Cloud-Computing.aspx
[2] A. Mathews, “United Health to Launch Cloud-Based
Data Platform,” Wall Street Journal, 2012.
http://online.wsj.com/article/SB100014240529702040627
04577221551500296744.html
[3] I. Carrión, J. Alemán and A. Toval, “Personal Health
Records: New Means to Safely Handle Our Health Data,”
IEEE Computer Society Digital Library, IEEE Computer
Society, 2012.
http://doi.ieeecomputersociety.org/10.1109/MC.2012.74
[4] C. McCarthy, “Paging Dr. Google: Personal Health Re-
cords and Patient Privacy,” William & Mary Law Review,
Vol. 51, No. 6, 2010, pp. 2243-2268.
http://scholarship.law.wm.edu/wmlr/vol51/iss6/6
[5] C. Klein, “Cloudy Confidentiality: Clinical and Legal Im-

Regulatory Privacy Protection for Biomedical Cloud Computing
151
plications of Cloud Computing in Health Care,” The
Journal of the American Academy of Psychiatry and the
Law, Vol. 39, No. 4, 2011, pp. 571-578.
[6] L. Osterhaus, “Cloud Computing and Health Information,”
The University of Iowa School of Library and Informa-
tion Scienc e Journal, Vol. 19, 2010, pp. 1-9.
[7] HITECH Act Enforcement Interim Final Rule. US De-
partment of Health and Human Services, 2009.
http://www.hhs.gov/ocr/privacy/hipaa/administrative/enfo
rcementrule/hitechenforcementifr.html
[8] M. Delgado, “The Evolution of Health Care IT: Are Cur-
rent US Privacy Policies Ready for the Clouds,” IEEE
World Congress on Services, Washington DC, 4-9 July
2011, pp. 371-378.
http://ieeexplore.ieee.org/xpl/login.jsp?tp=&arnumber=60
12698&url=http%3A%2F%2Fieeexplore.ieee.org%2Fxpl
s%2Fabs_all.jsp%3Farnumber%3D6012698
[9] R. Ko, P. Jagadpramana, M. Mowbray, S. Pearson, M. Kir-
chberg, Q. Liang and B. Lee, “Trust Cloud: A Frame-
work for Accountability and Trust in Cloud Computing,”
2011.
http://www.hpl.hp.com/techreports/2011/HPL-2011-38.p
df
[10] T. Martin, “Hey! You! Get off of My Cloud: Defining
and Protecting the Metes and Bounds o f Privacy, Security,
and Property in Cloud Computing,” 2011.
http://works.bepress.com/timothy_martin/3
[11] A. Palanzi, “Patient Privacy in the Cloud: Why Congress
Should Model HIPAA Enforcement Mechanisms after the
FCA to Meet a New Wave of Privacy Threats from the
Implementation of Cloud-Computing Technologies,” 2012.
http://works.bepress.com/andrew_palanzi/1/
[12] R. Grossman and K. White, “A Vision for Biomedical
Cloud,” Journal of Internal Medicine, Vol. 271, No. 2,
2012, pp. 122-130.
doi:10.1111/j.1365-2796.2011.02491.x
[13] J. Philbin, F. Prior and P. Nagy, “Will the Next Genera-
tion of PACS Be Sitting on a Cloud,” Journal of Digital
Imaging, Vol. 24, No. 2, 2011, pp. 179-183.
doi:10.1007/s10278-010-9331-4
[14] E. Schweitzer, “Reconciliation of the Cloud Computing
Model with US,” Journal of American Medical Informa-
tics Association, Vol. 19, No. 2, 2012, pp. 161-165.
doi:10.1136/amiajnl-2011-000162
[15] S. Pearson, “Toward Accountability in the Cloud,” IEEE
Internet Computing, 2011, pp. 64-69.
http://www.hpl.hp.com/techreports/2011/HPL-2011-138.
html
doi:10.1109/MIC.2011.98
[16] J. Soma, M. Nichols, M. Gates and A. Gutierrez, “Cha-
sing the Clouds without Getting Drenched: A Call for
Fair Practices in Cloud Computing Services,” 2011.
http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2039
439
[17] R. Ko, B. Lee and S. Pearson, “Towards Achieving Ac-
countability, Auditability and Trust in Cloud Compu-
ting,” Advances in Computing and Communicatio ns, Com-
munications in Computer and Information Science, Vol.
193, 2011, pp. 432-444.
doi:10.1007/978-3-642-22726-4_45
[18] Fujitsu Research Institute, “Personal Data in the Cloud: A
Global Survey of Consumer Attitudes,” 2010.
http://www.fujitsu.com/downloads/SOL/fai/reports/fujits
u_personal-data-in-the-cloud.pdf
[19] C. Witt, “HIPAA versus the Cloud,” 2011.
http://healthcare-executive-insight.advanceweb.com/Feat
ures/Articles/HIPAA-Versus-the-Cloud.aspx
[20] C. Lyon and K. Retzer, “Privacy in the Cloud: A Legal
Framework for Moving Personal Data to the Cloud,” 2011.
http://www.nist.gov/manuscript-publication-search.cfm?p
ub_id=909494
[21] A. Weissberger, “ACLU of Northern California, Cloud
Computing: Storm Warning for Privacy,” 2009.
http://viodi.com/2009/02/13/aclu-northern-ca-cloud-comp
uting-storm-warning-for-privacy/
[22] J. Harshbarger, “Cloud Computing Providers and Data
Security Law: Building Trust with United States Compa-
nies,” Journal of Technology Law and Policy, Vol. 16,
No. 2, 2011, pp. 229-254.
[23] Federal Trade Commission, Fair Information Prac tice Prin-
ciples, 2012.
http://www.ftc.gov/reports/privacy3/fairinfo.shtm
[24] J. Turow, J. King, C. Hoofnagle, A. Bleakley and M.
Hennessy, “Americans Reject Tailored Advertising and
Three Activities That Enable It,” 2009.
http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1478
214
[25] N. Ozer, “Privacy and Free Speech: It’s Good for Busi-
ness,” 2009.
http://dotrights.org/business/primer
[26] S. Pearson and A. Charle sworth, “Accountability as a Way
Forward in Privacy Protection in the Cloud,” Cloud Com-
puting, Lecture Notes in Computer Science, Vol. 5931,
2009, pp. 131-144.
[27] A. Narayanan and V. Shmatikov, “Robust De-Anonymi-
zation of Large Sparse Datasets,” 2008.
http://dl.acm.org/citation.cfm?id=1398064
[28] P. Wayner, “You Know about Backups. Now, Do It On-
line,” New York Times, 2008.
http://www.nytimes.com/2008/10/23/technology/personal
tech/23basics1.html