AUTHENTICITY AND INTEGRITY OF PORTABLE
ELECTRONIC HEALTH RECORDS
Chung-Yueh Lien
Institute of Biomedical Engineering, National Yang-Ming University, No. 155, Sec. 2, Linong st.
Beitou District, Taipei 112, Taiwan
Chia-Hung Hsiao
Department of Medical Informatics, Tzu Chi University, Hualien City, Taiwan
Lu-Chou Huang, Tsair Kao
Institute of Biomedical Engineering, National Yang-Ming University, Taipei, Taiwan
Keywords: Security, Portable electronic health record, Digital signature, Integrity, Authenticity.
Abstract: In this paper, we proposed a method to secure an electronic health record stored on a portable data storage
media (CDs/DVDs, diskettes, flash drives). We applied cryptography to realize the authenticity and
integrity of the portable health record. A manifest signature mechanism was used to reduce the computation
time of the signing and verifying processes. A DICOM DIR consists of 166 DICOM MR images was tested
as an example of a portable medical record. The performance of this method is faster than the regular digital
signature mechanism.
1 INTRODUCTION
Electronic health records (EHRs) may be generated
by hospitals, examination laboratories, insurance
companies or patients themselves (Bates et al. 2003;
Wang et al. 2004; Pharow & Bolbel 2005; Tang et
al. 2006). To realize clinical data exchange between
healthcare providers, a trusted conduit is needed for
the EHR systems and users. The integrity and
authenticity of EHRs can be validated using a digital
signature mechanism (Ruotsalainen & Manning
2007; Schütze et al. 2006). A digest of the digital
document is calculated from an irreversible one-way
hash function. The hash check of digital data is
commonly used on the Internet to prevent
unauthorized modification. The digital signature can
be implemented by a combination of the hash
algorithm and public-key cryptography such as the
RSA algorithm. When the RSA algorithm is used to
calculate a digital signature, the signer encrypts the
digest of the digital document with his/her own
private key. The recipient, with access to the
signer’s public key then verifies the digital
signature.
The implementation of EHRs has to conform to
security regulations, laws and standards, such as
Digital Imaging and Communication in Medicine
(DICOM), Health Level 7 (HL7), World Wide Web
Consortium (W3C), Health Insurance Portability
Accountability Act (HIPAA) and ISO/TS 17090.
According to the healthcare standards, a legal signed
EHR must contain one digest, one digital signature
and one timestamp signature. Under public-key
infrastructure (PKI), the use of the RSA algorithm
makes it possible to work with the certificate and
trusted third party (TTP) to process inter-
institutional applications such as the verification of
an EHR and referral information (Lekkas & Gritzalis
2007). An EHR may contain hundreds of digital
files, and then require the same number of digital
signatures. However, it is impractical to implement
these lengthy signing and verifying processes in the
real world due to the high computation time.
In this paper, a manifest signature mechanism is
proposed to reduce the computation time of the
signing and verifying processes used when dealing
67
Lien C., Hsiao C., Huang L. and Kao T. (2008).
AUTHENTICITY AND INTEGRITY OF PORTABLE ELECTRONIC HEALTH RECORDS.
In Proceedings of the First International Conference on Health Informatics, pages 67-71
Copyright
c
SciTePress
with a portable EHR. A DICOM DIR consists of
166 DICOM MR images was tested as an example
of a portable EHR.
2 METHODS
We used a smart card system that supports the
Microsoft Cryptography Service Provider (CSP) as
the digital signature module. The use of the health
professional card (HPC) with a smart card-based
certificate is a good example and can be found in the
healthcare environments of Taiwan (Yang et al.
2006), Germany (Schütze et al. 2006; Schurig,
Heuser & Wedekind 2001), Belgium (France,
Bangels & De Clercq 2007), etc. A health
organization certificate card (HOC) holds the digital
official seal of every health organization and can be
used for EHRs exchange among organizations. The
Health Certificate Authority Timestamping
Authority (HCA-TSA) provides the timestamp
service as the TTP.
2.1 The Characteristics of a Portable
EHR
The flowchart of the security protection process is
shown in Fig. 1. The clinical documents, integrated
from various systems in a hospital stored in the
portable storage media. The signed list has to be
managed through the metadata that define these data
in a portable EHR. The metadata has to identify the
information, structures and formats to meet the
needs of multimedia data exchange. The signed list
is used to sign the clinical documents from the
signing hospital; the documents described in the
signed list will be signed using the hospital
certificate with a digital time signature. According to
the signed list, the digest values of the clinical
documents are calculated and packaged as the digest
of these documents. The hospital and TTP generate
the digital signature and digital time signature to
verify the authenticity of the clinical documents. The
hospital certificate is used as identification on the
exchanged clinical documents and to authenticate
the source site, and the site receiving the data can
then verify whether or not the exchanged clinical
documents are valid. Table 1 summarizes the use of
the characteristics of a portable EHR.
Document M
N
Digest M
1
Document M
2
Document M
1
Digest M
2
Digest M
N
Manifest
Digital time
signature
Portable EHR
Manifest
Resources
Reference Link
.
.
.
.
.
.
Manifest
Signature
SIG digest
Timestamp
Signed List
Metadata
Metadata
Manifest
Signature
Figure 1: The characteristics of a portable EHR
architecture.
Table 1: The use of the characteristics of a portable EHR.
Characteristics Use
Metadata Definition of the portable EHR
Signed List Information of signed EHR
Digital Time
Signature
Authenticity
Manifest Integrity
Manifest Signature Authenticity
Resource The physical EHR structure
2.2 The Manifest Signature Mechanism
The manifest signature process is shown in Fig. 2. In
the signed list, there are N documents M
1
, M
2
,
, M
N
need to be signed. For each file
Ref(M
n
)=ID(M
n
)^H(M
n
) is calculated. The
symbol ’^’ means cascade. Ref(M
n
) represents the
metadata of M
n
, and some information related to M
n
can be defined in ID(M
n
), such as data type, creation
time, purpose, etc., for exchange among various
EHR systems. H(M
n
) is the digest of M
n
, where H is
the hash function, such as MD5, SHA1, etc. We can
reconstruct Ref(M
n
) as the manifest of N documents’
digest MD
D
defined as:
MD
D
= Ref(M
1
)^Ref(M
2
)^...^Ref(M
N
)
Based on the RSA algorithm, using the signer’s
private key P
r
to encrypt H(MD
D
), the digest value
MD
D
, as the digital sitnature of MD
D
, we then define
the digital signature process as follows:
SIG(MD
D
)= RSA
Enc
(P
r,
H(MD
D
)),
Where RSA
Enc
is the RSA encryption function.
SIG(MD
D
) is the manifest signature. We send the
digest H(SIG(MD
D
)) to a TSA to obtain a qualified
digital time signature TSA
SIG
, defined as:
TSA
SIG
= RSA
Enc
((P
rTSA,
H(SIG(MD
D
)||T
SS
))
The symbol ’||’ represents the concatenation,
P
rTSA,
is the TSA’s private key and T
SS
is the
Greenwich Mean Time (GMT) timestamp.
This is an efficient mechanism to ensure the
integrity and authenticity of a portable EHR. From
the aspect of data integrity, Ref(M
n
) ensures the
HEALTHINF 2008 - International Conference on Health Informatics
68
integrity of a document M
n
. The manifest signature
SIG(MD
D
) and the digital time signature TSA
SIG
provide verification of the authenticity for the
documents in the signed list and the T
SS
confirms the
synchronized time.
Figure 2: Manifest of the documents’ digest process.
2.3 Performance Analysis of the
Manifest Signature Mechanism
Assuming the time required for hashing, signing and
digital time signature retrieval are T
n
, T
r
and S
n
,
respectively, where 1
n N. The timestamp
retrieval time depends on the network status; here S
n
is just for reference. If we create digital signature for
each document one by one, the total calculation time
is:
T
total
=
=
N
n
n
T
1
+N×T
r
+
=
N
n
n
S
1
However, using the method proposed in this stuy,
the total time will be reduced to:
T
total
=
=
N
n
n
T
1
+S
1
+T
r
+T
M
T
M
is the computation time of H(MD
D
). The
number of calculations needed for verification
process is still the same. Table 2 shows the number
of calculations in hashing, signing and digital time
signature retrieval for different methods.
Table 2: Comparison of the number of calculations
between the proposed method and the one-by-one method.
Calculation time
Operations
Proposed
method (sec)
One-by-one
method (sec)
Hashing N+1 N
Signing 1 N
Digital time
signature retrieval
1 N
*N is the number of documents in the signed list.
3 RESULTS
3.1 The Manifest Signature
Architecture Implemented using
XML
We tested an example of the manifest signatures of
166 DICOM MR images (packaged as DICOM DIR)
to be signed using both our method and the one-by-
one method regulated in the DICOM standards.
First, system will search all of files in selected
folder and mark the URI of files to generate the
signed list. According to the URI of the signed list,
system will find the direction pathway of file and
calculate the digest value of each file. The ID(M
n
) is
created from the DICOM head tags such as Transfer
syntax UID (0020, 0010), SOP Instance UID (0008,
0018)…etc, and we use XML encoding to present
Ref(M
n
)=ID(M
n
)^H(M
n
).
Fig. 3 shows the Ref(M
n
) structure presented as
XML. In Fig. 3, a unit of Ref(M
n
) is represented by
the tag name “Reference”. The attribute “URI” of
<Reference> is the related directory pathway in
addition to an identification of the resource file.
H(M
n
) is represented by <DigestValue> and ID(M
n
)
is represented by <DigestMethod Algorithm> and
<Transforms>. Transforms means the namespace of
this referenced document, which identifies the data
format. In this example, the value in <Transform
Algorithm> is urn:oid:1.2.840.10008.1.2.1, which
expresses the explicit little endian coding for
DICOM. This attribute can deal with different cases
of different data formats for EHR exchange between
hospitals; in addition, it can be extended for multiple
data formats, which are defined by the user.
We reconstruct Ref(M
n
) as the manifest digest
MD
D
, and put MD
D
into <Object> tag as the signing
range of manifest signature and the attribute “Id” of
<Manifest> is represent the identifier of MD
D
. The
cascade of the element in XML as the manifest of
these DICOM files is shown in Fig. 4. We calculate
H(MD
D
) and using HOC’s private key to encrypt
H(MD
D
) and generate the manifest signature
SIG(MD
D
) complies with the W3C XML enveloped
signature standard is shown in fig. 5. All value is
encode by Baase64, the value “Object” in attribute
“URI” of <Reference> means sign the <Object>
described above, H(SIG(MD
D
)) is represented by
<DigestValue>; the manifest signature is
represented by <SignatureValue> the singing
certificate is put in <X509Certificate> tag.
AUTHENTICITY AND INTEGRITY OF PORTABLE ELECTRONIC HEALTH RECORDS
69
Figure 3: A unit of Ref(Mn) of one image in DICOM DIR.
Figure 4: The manifest structure of DICOM DIR presented
as XML.
Figure 5: The manifest digital signature presented as W3C
XML enveloped signature standard.
3.2 Performance Test
Table 3 shows a comparison of the operation time
for test files by different methods. In the one-by-one
method, each image has to be hashed, RSA
encrypted and timestamp retrieved once, for which
the total time needed was about 167.89 seconds. In
the proposed method, however, the total time was
reduced to 2.81 seconds. The demonstrating PC is a
Pentium D 2.8 GHz, with 4 GB memory; the size of
each image is about 516KB~1.58 MB; the total data
size of the images is 82.3 MB; the type of the smart
card is e-gate. The test of digital time signature time
is retrieved from the Taiwan HCA-TSA.
Table 3: Comparison of the time needed for signing 166
DICOM MR images between the proposed method and the
one-by-one method.
Calculation time
Operations
Proposed
method (sec)
One-by-one
method (sec)
Hashing 1.34 1.34
Signing 1.27 148
Digital time
signature retrieval
0.2 18.55
Total time 2.81 167.89
4 DISCUSSION
If a portable EHR is to be exchanged with other
organizations, the presence of a removable storage
media is also required. To verify the authenticity of
the EHR, the sender site creates the digital signature
of the EHR as the organizational stamp. The receiver
can then verify whether the EHR is valid by
Certificate Authority (CA) and TTP. The digital
signature created by the hospital and the digital time
signature created by the TSA can record the
authenticity of the EHR for inter-organization
exchange. In terms of cryptography, the digital
timestamp mechanism is not used to provide a
qualified digital signature, but to certify a qualified
time signature. Some infrastructures, such as patient
identification, certificate management, and standards
should be established as well. And some security
issues should be noticed in implementation: e.g.,
data backup, audit trail, register loss, maintenance,
recovery, etc.
A portable EHR contains clinical data and
related setting data, the combination of these data
can reconstruct representation of EHR. The signing
range should contain all of the data related to clinical
data. It is very important to ensure the integrity of
representation of EHR. If only signing clinical data
and related setting data is not singed, it could be
happened in inconsistency of representation of EHR
while setting data had been modified. The security
protection of EHR should include all of data in
portable storage media.
In general, most of the medical information
standards and national regulations regulating the
legal EHR do not use the manifest signature. If
existing medical information digital signature rules
are followed, as the practice is not feasible due to the
high computational time of the signing and verifying
processes. However, the signing time and timestamp
retrieval time need to be reduced because a portable
EHR may contain many clinical documents and
HEALTHINF 2008 - International Conference on Health Informatics
70
images, all of which need to be protected. For
example, a study may contain hundreds of DICOM
images, and following the DICOM standards in
these cases is impracticable in real-world clinical
operations. The manifest signature can be used not
only for the exchange of EHR, but also for EHR
long-term storage in hospitals.
Fast and reliable proof of authenticity and
integrity is needed for security considerations when
an EHR became portable. It is common that patients
collect their own health records from different
hospitals and manage the process by themselves.
The implementation of a centralized health record
containing personal health records is very difficult
when taking into considerations of the physicians’
intellectual property rights and patient privacy.
5 CONCLUSIONS
The computational time of this prototype is much
lower than that of the one by one digital signature
method. Following the existing medical information
digital signature rules, the practice is not feasible.
Using the proposed method, the computational time
is reduced. In addition, this method can be used not
only for the exchange of EHRs, but also for their
long-term storage.
ACKNOWLEDGEMENTS
This work was supported by the National Science
Council of Taiwan under Grant NSC 95-2221-E010-
003.
REFERENCES
Bates, DW, Ebell M, Gotlieb, E, Zapp, J, & Mullins, HC
2003, ‘A Proposal for Electronic Medical Records in
U.S. Primary Care’, Journal of the American Medical
Informatics Association, vol. 10, no. 1, pp. 1-10.
France, FHR, Bangels, M & De Clercq, E 2007, ‘Purposes
of health identification cards and role of a secure
access platform (Be-Health) in Belgium’,
International Journal of Medical Informatics, vol. 76,
no. 2-3, pp. 84-88.
Lekkas, D & Gritzalis, D 2007, ‘Long-term verifiability of
the electronic healthcare records’ authenticity’,
International Journal of Medical Informatics, vol. 76,
no. 5-6, pp. 442-448.
Lekkas, D, Gritzalis, S & Katsikas, S 2002, ‘Quality
assured trusted third parties for deploying secure
internet-based healthcare applications’, International
Journal of Medical Informatics, vol. 65, no. 2, pp. 79-
96.
Makoul, G, Curry, RH & Tang, PC 2001, ‘The Use of
Electronic Medical Records: Communication Patterns
in Outpatient Encounters’, Journal of the American
Medical Informatics Association, vol. 8, no. 6, pp.
610-615.
Pharow, P & Blobel, B 2005, ‘Electronic signatures for
long-lasting storage purposes in electronic archives’,
International Journal of Medical Informatics, vol. 74,
no. 2-4, pp. 279-287.
Schurig, A, Heuser, H & Wedekind, R 2001, ‘Introduction
of the health professional card into the
SAXTELEMED-Project’, International Congress
Series, vol. 1230, pp. 867-871.
Tang, PC, Ash, JS, Bates, DW, Overhage, JMS & Sands,
DZ 2006, ‘Personal Health Records: Definitions,
Benefits, and Strategies for Overcoming Barriers to
Adoption’, Journal of the American Medical
Informatics Association, vol. 13, no. 2, pp. 121-126.
Wang, M, Lau, C, Matsen, FAIII & Kim, Y 2004,
‘Personal health information management system and
its application in referral management’, IEEE
Transactions on Information Technology in
Biomedicine, vol. 8, no. 3, pp. 287-297.
Yang, CM, Lin, HC, Chang P & Jian, WS 2006, ‘Taiwan’s
perspective on electronic medical records’ security and
privacy protection: Lessons learned from HIPAA’,
Computer Methods and Programs in Biomedicine, vol.
82, no. 3, pp. 277-282.
Zhou, XQ, Huang, HK & Lou, SL 2001. ‘Authenticity and
integrity of digital mammography images’, IEEE
Transactions on Medical Imaging, vol. 20, no. 8, pp.
784-791.
AUTHENTICITY AND INTEGRITY OF PORTABLE ELECTRONIC HEALTH RECORDS
71