356
G. Chuiko et al.
discrepancy in amplitude value is not necessarily means cautious state. Therefore,
it allows informing patient in advance about possible outputs according to sensor
measurements and to recommend preventive activities to him/her.
In opposite to the device level, generalization of the patient’s actions that affect
medical data a priori cannot include all possible options. Moreover, automatic
retrieval of this information is not possible unless IoT-device supports this feature.
Therefore, this part should also facilitate interface when the patient adds evidence
about improper device usage. The user enters feedback about the incorrect measurements and also marks time interval when data are probably recorded with these
circumstances. The result of this activity is decreased trustworthiness of the data
recorded during the marked period. That also entails that the data cannot be used
during the inference process or, at least, their impact is not so strong.
The case under consideration is tightly connected with capabilities of semantic
technologies, IoT-devices specification, and their limitations. The challenge that
arises when patient is permitted to enter his own explanations is that semantic
reasoning is based on the direct match between tokens in the patient’s response
and data from ontologies. If no match is found, no semantic data can be linked to this
message automatically and it requires manual adjustment. Viable solution for this
challenge is a list with options to select to explain the reason of failed measurement.
However, it should be designed for each sensor/device specifically and even list of
these options have to contain option that allows entering reasons that have not been
predicted during list preparation. Information fetched from the individual sensor has
its own peculiarities and its interpretation should strictly follow documentation.
Nevertheless, there is still probability that some data are left not annotated and
excluded from queries and inference. To overcome this issue we propose a procedure
when doctor annotates data manually and links are created as a result of initial manual
augmentation.
On the other side, we can face situation when single sensor measured value is
actually correct and it indicates cautious physical state of the patient. This scenario
is critical for consideration and automatic denial of the sample could lead to critical consequences. Thus, we propose to exploit the notion “scenario” that describes
possible consequences from the analysis of the retrieved data. In this case, multiple
concurrent scenarios exist in the system. At least one of them marked as “critical”
scenario depicts situation where action should be applied immediately. Typically,
two parallel scenarios will be created. The second case assumes that data might be
received from erroneous source and further analysis is required.
Critical scenario also means that doctor or medical surveillance service are notified
about the possible patient’s state. At the same time, patient is also notified about
his/her critical state and has an option to approve or deny this fact. Hence, two-side
communication mechanism is required by the system. Possibility of such scenario is
the main reason why system design allows non-autonomous communication between
patient and doctor. In our opinion, this indispensable option needs to be present in the
medical healthcare system. Even though semantic technologies significantly enrich
capabilities of the system, they cannot substitute experts experience completely or
to avoid incorrect judgements according to limited information space.
G. Chuiko et al.
discrepancy in amplitude value is not necessarily means cautious state. Therefore,
it allows informing patient in advance about possible outputs according to sensor
measurements and to recommend preventive activities to him/her.
In opposite to the device level, generalization of the patient’s actions that affect
medical data a priori cannot include all possible options. Moreover, automatic
retrieval of this information is not possible unless IoT-device supports this feature.
Therefore, this part should also facilitate interface when the patient adds evidence
about improper device usage. The user enters feedback about the incorrect measurements and also marks time interval when data are probably recorded with these
circumstances. The result of this activity is decreased trustworthiness of the data
recorded during the marked period. That also entails that the data cannot be used
during the inference process or, at least, their impact is not so strong.
The case under consideration is tightly connected with capabilities of semantic
technologies, IoT-devices specification, and their limitations. The challenge that
arises when patient is permitted to enter his own explanations is that semantic
reasoning is based on the direct match between tokens in the patient’s response
and data from ontologies. If no match is found, no semantic data can be linked to this
message automatically and it requires manual adjustment. Viable solution for this
challenge is a list with options to select to explain the reason of failed measurement.
However, it should be designed for each sensor/device specifically and even list of
these options have to contain option that allows entering reasons that have not been
predicted during list preparation. Information fetched from the individual sensor has
its own peculiarities and its interpretation should strictly follow documentation.
Nevertheless, there is still probability that some data are left not annotated and
excluded from queries and inference. To overcome this issue we propose a procedure
when doctor annotates data manually and links are created as a result of initial manual
augmentation.
On the other side, we can face situation when single sensor measured value is
actually correct and it indicates cautious physical state of the patient. This scenario
is critical for consideration and automatic denial of the sample could lead to critical consequences. Thus, we propose to exploit the notion “scenario” that describes
possible consequences from the analysis of the retrieved data. In this case, multiple
concurrent scenarios exist in the system. At least one of them marked as “critical”
scenario depicts situation where action should be applied immediately. Typically,
two parallel scenarios will be created. The second case assumes that data might be
received from erroneous source and further analysis is required.
Critical scenario also means that doctor or medical surveillance service are notified
about the possible patient’s state. At the same time, patient is also notified about
his/her critical state and has an option to approve or deny this fact. Hence, two-side
communication mechanism is required by the system. Possibility of such scenario is
the main reason why system design allows non-autonomous communication between
patient and doctor. In our opinion, this indispensable option needs to be present in the
medical healthcare system. Even though semantic technologies significantly enrich
capabilities of the system, they cannot substitute experts experience completely or
to avoid incorrect judgements according to limited information space.