622
P.-Y. S. Hsueh et al.
Such a long development process is cost-intensive and can hold a high risk of poor
transferability in real world settings. As a result, its development often does not pass
the prototype phase.
Second, the evaluation process of ITS is also challenging as ITS is often complex
and composed of multiple system components. The combination of a variety of AI
methods complicates the task as well. Although different evaluation techniques have
been developed [95], there is still no one-fits-all approach; hence, no guarantee to
the generalizability of the evaluation results. Generally, only the behavior of the
complete system can be observed, and even if a fault is detected during evaluation,
the localization of the fault inside the system can be difficult.
Third, the elusiveness of expert knowledge also poses challenges. When being
inquired on the same question multiple times, the surveyed experts would often
provide different answers to the same question. One approach to tackle this
elusiveness is soft computing, in particular, fuzzy logic.
However, additional challenges arise when ITS is to be implemented in the
healthcare domain to counter the health literacy issues of the population. For one, the
target learners of ITS in this domain start with little domain knowledge. Therefore,
a proper ITS must address a wide range of topics. But with the increasing range
of topics, the domain and tutoring models must also be more comprehensive; yet,
the features and information needed for doing so in the healthcare domain are highly
heterogenous. A comprehensive assessment indicates a significant time commitment
and burden for the patients, which results in the decrease of system usability.
Last but not least, healthcare experts often find it difficult to formalize the
knowledge bases needed for building up ITS knowledge delivery strategies inside
the domain and tutoring models. Furthermore, these are the experts who have very
strict time constraints, and consequently the very short evaluation makes it difficult
to properly evaluate the systems validity.
13.3.4 Case Study: Implicit Knowledge Learning for Nurses
Personal exchange has a strong influence on the learning process in nursing [96],
but heavy workloads and high employee turnover show the necessity of on-thejob training. In this case study, we introduce the Witra Care (a German acronym
for knowledge transfer) project, which aims to employ mobile technologies to
support on-the-job training processes for nurses by recording, categorizing, and
re-providing knowledge snippets in the form of micro-learning modules [97]. At
Hannover Medical School (MHH), experienced nurses support new employees for
several weeks by sharing their knowledge. In Witra Care, nine new employees
and seven experienced nurses created micro-learning modules in a mixed form
of text, images or recorded video and audio via a mobile application. As such,
they collectively created 303 learning modules, which were structured around the
following three categories: daily healthcare activities, health quality management,
and general information for new employees. These categories are not disjunctive,
P.-Y. S. Hsueh et al.
Such a long development process is cost-intensive and can hold a high risk of poor
transferability in real world settings. As a result, its development often does not pass
the prototype phase.
Second, the evaluation process of ITS is also challenging as ITS is often complex
and composed of multiple system components. The combination of a variety of AI
methods complicates the task as well. Although different evaluation techniques have
been developed [95], there is still no one-fits-all approach; hence, no guarantee to
the generalizability of the evaluation results. Generally, only the behavior of the
complete system can be observed, and even if a fault is detected during evaluation,
the localization of the fault inside the system can be difficult.
Third, the elusiveness of expert knowledge also poses challenges. When being
inquired on the same question multiple times, the surveyed experts would often
provide different answers to the same question. One approach to tackle this
elusiveness is soft computing, in particular, fuzzy logic.
However, additional challenges arise when ITS is to be implemented in the
healthcare domain to counter the health literacy issues of the population. For one, the
target learners of ITS in this domain start with little domain knowledge. Therefore,
a proper ITS must address a wide range of topics. But with the increasing range
of topics, the domain and tutoring models must also be more comprehensive; yet,
the features and information needed for doing so in the healthcare domain are highly
heterogenous. A comprehensive assessment indicates a significant time commitment
and burden for the patients, which results in the decrease of system usability.
Last but not least, healthcare experts often find it difficult to formalize the
knowledge bases needed for building up ITS knowledge delivery strategies inside
the domain and tutoring models. Furthermore, these are the experts who have very
strict time constraints, and consequently the very short evaluation makes it difficult
to properly evaluate the systems validity.
13.3.4 Case Study: Implicit Knowledge Learning for Nurses
Personal exchange has a strong influence on the learning process in nursing [96],
but heavy workloads and high employee turnover show the necessity of on-thejob training. In this case study, we introduce the Witra Care (a German acronym
for knowledge transfer) project, which aims to employ mobile technologies to
support on-the-job training processes for nurses by recording, categorizing, and
re-providing knowledge snippets in the form of micro-learning modules [97]. At
Hannover Medical School (MHH), experienced nurses support new employees for
several weeks by sharing their knowledge. In Witra Care, nine new employees
and seven experienced nurses created micro-learning modules in a mixed form
of text, images or recorded video and audio via a mobile application. As such,
they collectively created 303 learning modules, which were structured around the
following three categories: daily healthcare activities, health quality management,
and general information for new employees. These categories are not disjunctive,