single person is still in control of this particular key.
Unauthorized Transfer of Medical Data. In (SGB,
2007, SGBV, 294a) it is stated that by law,
depending on the medical issue, full medical data
has to be given to the insurance company without
the patient's consent. This exception conflicts with a
basic requirement found in (gematik, 2006a, p. 63)
stating that no one is allowed to access medical data
without the permission of the insured person.
Furthermore (SGB, 2007, 2007, SGBV, 291a, para.
3) states that insurers must inform the affected
patient about the transfer of data. This possible
vulnerability remains as long as (SGB, 2007, SGBV,
294a) is not changed, for example by involving the
HTI in data transfer. Using the eHC, it would be
possible to securely transfer the data to the insurance
provider, while involving the insured in the process
using the eHC to grant permission.
Missing Backup Method for Electronic
Prescriptions. According to (IBM et al., 2004, p.
19), an alternate backup process must be created for
every electronic HTI process. (gematik, 2006b, p.
28) however states that there will be no backup
process for the filling of prescriptions by
pharmacies. Although this statement was purged
from (gematik, 2008e), there is still no trace in
(gematik, 2008b) and (gematik, 2008c) of a backup
process. This means that a patient with a prescription
stored on an eHC would have to either revisit their
doctor to get a paper prescription, go to another
pharmacy with a working HTI connection or wait
until the HTI is working again. This is obviously an
inconvenient and potentially dangerous situation,
depending on the person's health.
Possibility to Issue the Same Prescription Twice.
If the above-mentioned backup process is
implemented using paper-based prescriptions, a
patient would possess a prescription in both digital
and paper form, allowing them to fill both in
different pharmacies (by filling the paper-
prescription at a pharmacy with HTI problems and
the electronic one in a pharmacy where the HTI is
working). The possibility of prescriptions being
filled twice would violate requirements concerning
accountability and non-repudiation. This is,
however, currently not an issue as no backup process
exists.
No Security Verification for the ”Zone-Concept”.
In (gematik, 2007b), the HTI is divided into several
zones to allow a separate view of each specific
security zone. These zones are treated as closed
areas as mentioned in (gematik, 2007b, p. 32),
meaning that security vulnerabilities in one zone
shouldn't affect adjacent zones. But the zones of the
HTI are physically connected for data transfer, and
thus there is still a possibility of unauthorized traffic
between zones. The statement itself is still present in
(gematik, 2008k, p. 40) but (gematik, 2007a, p. 9)
explains that it is possible to break into the HTI
network and compromise adjacent systems, which
contradicts statements about closed zones.
Therefore, this statement should be replaced by a
more accurate one explaining the connection
between zones.
Adjustment of Security Standards. According to
(gematik, 2007b, p. 28), the minimum security
standards for the HTI have to be checked and
adjusted once a year. This time span remains
unaltered in the current version of the document, but
considering the CERT (http://www.cert.org/stats)
statistics of about 8,000 vulnerabilities per year, a
time span of one year seems unnecessarily long. A
shorter period between minimum security standard
adjustments would improve the security of the HTI.
Inadequate Security Assumptions about the HTI.
gematik considers all systems inside the VPN of the
HTI to be secure (gematik, 2006b, p. 60). Therefore,
the time servers within the HTI are used by the
connector without any authentication. In fact, there
are no completely secure IS systems (Sharman et al.,
2004), so the argument is not valid.
Nevertheless, it is still used in (gematik, 2008e,
p. 114). There is also contradicting information
regarding the time servers. While (gematik, 2008a,
p. 49) explains that the time sync is mainly for
chronological logging purposes and therefore no
authentication is needed, (gematik, 2008c, p.15)
points out that electronic prescriptions on the eHC
use the primary system's time, which is synced with
connector time, which is synced with the time
server's time. This leads to the conclusion that the
time server's time is used for time stamps of secure
medical data on the eHC. If a manipulated time
server could be used to manipulate timestamps on
the eHC, there must be authentication to prevent
such a scenario.
Security by Obscurity. The concept of security by
obscurity isn't a proper way of securing medical IS
systems, as pointed out by gematik in (gematik,
2007b, p. 246). However, parts of the software used
within the HTI are classified as highly confidential
by the same document, due to copyright issues, as
HEALTHINF 2010 - International Conference on Health Informatics
190