Health Assistance for Immigrants - Smart Information Systems: Computational Intelligence for Real-Life Applications - page 89

browse through the ontology concepts by expanding concept nodes and traversing
semantic edges between concepts, and also finding documents attached to a concept's
information nodes.
3.4 Evaluation
In order to assess immigrants' expectations and needs for a personalized health
information assistant, we conducted a demand analysis among Turkish migrants
in a mid-sized town in Germany. In addition, we evaluated the development and
implementation of the health assistant. Applying a qualitative research approach,
we recruited a small study cohort from the local Turkish community as well as a
German control group. In the second phase of the evaluation, they were asked to
interact with the system in a supervised scenario and were further asked to assess
the assistant with respect to its technical implementation, relevance of the presented
content, usability in daily life, and its potential to present health information in a
structured manner. In particular, we focused on the following research questions:
(a) How do the participants assess the adaptation of search results based on personal
context? (b) How do they evaluate the system's handling of cultural differences?
(c) How do they evaluate the usability of the graphical user interface (especially the
ability to change language settings)? In the remainder of this section, we outline the
setup of the user evaluation.
3.4.1 Participants and Recruitment
Building on experiences from former migrant studies [ 47 ], we applied a cultural-
sensitive recruitment concept using key persons from existing social networks (fam-
ily and youth centers) in order to establish the access to our study population. We
recruited nine families with a Turkish migrant background and four German families
as a control group. As suggested by Lamnek [ 21 ] and Pelz et al. [ 28 ], we invited
a homogeneous focus group of six participants with similar socioeconomic back-
grounds from two different districts to participate in the user evaluation.
3.4.2 System Interaction
During the user-centric evaluation of the system, the study group was separated into
a German-speaking group and a Turkish-speaking group due to limited knowledge
of the German language of some participants. Both groups interactively tested the
health assistant prototype in a supervised user-scenario. During the test, they were
asked several questions concerning the design, usability, and cultural sensitivity.