Work 62 (2019) 443–457 443

DOI:10.3233/WOR-192878
IOS Press

Predicting technology usage by health

information need of older adults:
Implications for eHealth technology
Sabine Theisa,∗ , Dajana Schäfera , Christina Bröhla , Katharina Schäfera , Peter Raschea ,
Mattias Willea , Christopher Brandla , Nicole Jochemsb , Verena Nitscha and Alexander Mertensa
a Institute of Industrial Engineering and Ergonomics, RWTH Aachen University, Aachen, Germany
b Institute for Multimedia and Interactive Systems (IMIS), University of Lübeck, Lübeck, Germany

Received 30 October 2018

Accepted 17 January 2019

Abstract.
BACKGROUND: Information and communication technology increasingly addresses the information needs patients have
regarding their personal health. While an understanding of older adults’ needs is crucial for developing successful eHealth
technology, user research results hardly apply to different systems.
OBJECTIVE: The present study aims at: (1) describing and analysing the context of digital health systems in a general
manner, (2) investigating if information need of older adults influences their technology usage to show the relevance of the
concept for a general context analysis and (3) testing which demographic variables intervene with their health information
need.
METHODS: Survey data from a longitudinal study with older adults (N = 551) were reported descriptively. After showing a
significant relationship during chi-square tests, we quantified the ones between general health information need and technology
usage, as well as between general health information need and the demographic variables age, education, chronic diseases
and gender by means of (multiple) linear regression models.
RESULTS: We predicted older adults’ technology usage based on their health information need. The results confirmed this
relationship. Higher information need led to a more frequent usage of apps installed on the tablet personal computer (PC), to
a frequent use of smartwatches and to the possession of a computer or laptop. Users’ education has a higher impact on health
information need than amount of chronic diseases, gender and age.
CONCLUSIONS: Information need emerged as a useful object for investigation of context and user requirement analysis
across different systems: it predicted technology usage so that design recommendations derived from the descriptive gained
in importance.

Keywords: Information-seeking behavior, technology usage, requirement analysis, human factors, digital health, user study,
usability

1. Introduction adults has increased. This is because western pop-

ulations are not only getting progressively older, but
In times of demographic change, the need for inno- also because the incidence and prevalence of illnesses
vative and comprehensive care concepts for older and the resulting limitations can greatly reduce the
quality of life for patients [1]. The application of digi-
∗ Address for correspondence: Sabine Theis, Human Factors
tal information and communication technology (ICT)
Engineering and Ergonomics in Healthcare, Chair and Institute
of Industrial Engineering and Ergonomics, RWTH Aachen Uni-
and related services in existing and future health
versity, Bergdriesch 27, 52062 Aachen, Germany. Tel.: +49 241 services represents an extremely promising oppor-
80 99469; E-mail: [email protected]. tunity. Digital health has the potential to address the

1051-9815/19/$35.00 © 2019 – IOS Press and the authors. All rights reserved
444 S. Theis et al. / Predicting technology usage by health information need of older adults

challenges posed by the societal age shift in Germany, Although there are user requirements that are of
but also in other western countries [2]. The decou- a general nature, these are given little attention.
pling of medical care from the local and temporal Research activities instead concentrate on user needs
availability of medical personnel, which is associated for one specific system. We consider user research
with digital health systems, makes it possible to pro- results as “general” if they hold either across mul-
vide patient-specific and cost-efficient services and tiple systems or across multiple users. Such general
therapies. Through digital monitoring and appropri- context and requirement description could identify
ate support systems, elderly people’s wish to remain opportunities for digital health information systems
longer and more independent in their home envi- and initiate their development. It could furthermore
ronment can be accommodated. Unfortunately, the identify the potential for new types of applications
nationwide distribution of so-called eHealth systems and reveal potentials for increasing user acceptance
is the exception in Germany and other European and adherence. Developers and health system com-
countries. Legal and organizational obstacles as well panies could use a general description of the needs of
as a lack of knowledge about the requirements of older adults to make more informed product devel-
older adults regarding the design of eHealth and opment decisions and they could use it as starting
eHealth services could be identified as obstacles in point for detailed context- and requirement analysis,
post-hoc analysis of facilitators and enablers of digi- especially if this involves demographic factors of the
tal health systems [3]. Existing systems and services target group.
often lack acceptance and adherence due to usability
issues. One effective mean to support the acceptance 1.2. Information need as part of user centered
and usability of digital health information systems is development
the user-centered design process [4, 5].
Information technology helps the user to find, use
1.1. User research and generalizability and transfer information. Looking at the information
a user needs to fulfill her or his goal, has implications
User-Centered Design (UCD) processes involve for different kind of technologies, which are able to
users in the design of technology by identifying transfer this required information. User’s information
usability goals, user characteristics, environment, need (IN) becomes a concept describing a proportion
tasks and the user interaction with a product, ser- of user requirements across systems while a user’s
vice or process through different steps of an iterative information seeking behavior (ISB) describes the cur-
process [5]. UCD serves to develop a specific techni- rent context of information acquisition. ISB usually
cal system according to the individual needs of users represents the consequence of a previously existing
[6–9]. User research is one pre-requisite for UCD need for information.
and methods depend on the type of technology to be To give an example, imagine the case of a patient
developed, accessible resources, and the evaluation receiving a medical diagnosis. Consequently, this
and application environment. User research results patient may now develop a need to seek additional
are also made available within the framework of clas- information to his or her diagnosis. Accordingly, to
sical research. Here, with each new technology, a new satisfy this need, the patient will want to employ
research question arises. However, by its nature, con- information sources or information retrieval systems
text and requirement analysis are rarely generalizable such as a search engine for the internet, using a com-
across applications and users. puter, a smartphone or other compatible devices. The
Besides this research-related issue, in applied set- patient either can find required information or not.
tings the problem arises that UCD evaluation of every The source could lead to either a success or a failure
new product requires resources. In addition, context of the information search. In the case of success, the
and requirements analysis outcomes rarely involve patient will use the gathered information so that the
alternatives to technology or the interaction of tech- initial IN will be satisfied or not. Success or failure
nical and non-technical information sources in their of one information source influences the use of other
effects on users. When the product decision is to sources.
develop a smartphone app, according requirement Accordingly, we want to investigate the potential
analysis will ask e.g. what the users goals are and of the concepts health information need (HIN) and
will consider how to support them with the app. It is health information seeking behavior (HISB) for a
unlikely that a book or a human will be the solution. general context and requirement analysis for digital
 S. Theis et al. / Predicting technology usage by health information need of older adults 445

health systems. To the best of our knowledge, the rela- systems or similar ones. On the other hand, these
tion between information need (IN) and technology can be systems that focus on non-informatory tasks
use is lacking so far, even if it is oftentimes implicitly and provide information as a secondary task. One
assumed by studies. example for such a system would be a real estate
organization. Its primary goal is to sell houses. How-
ever, if one searches for suitable areas to live in a
2. Related work city, this organization would be a good information
source to approach. Alternatively, the user may seek
One of the first studies of its kind investigated the information from people, rather than from systems.
IN of urban residents [10] to identify how it could This reciprocal process is considered an informa-
be better addressed by public institutions in a city. tion exchange. Regardless of the source that the user
Nevertheless, it was criticized due to a lack of defini- approaches, as the previous example of a patient illus-
tion of the concept “information need”. Wilson [11] trates, the ISB could lead either to success or to
therefore defined and modeled the concepts of IN and failure. Information usage might include perception,
ISB. He defines information, as “facts, advice or opin- inspection and storage but most importantly an eval-
ions”. For him, IN is a subjective need for information uation with regard to the perceived need. After using
or knowledge perceived by a user. the information, the user can be satisfied or dissatis-
Various theoretical models exist to describe IN and fied with the information at hand. Being subjective in
ISB [12–17]. One developed in the health information nature, IN is not directly obvious and the researcher
context is Wilson’s model of ISB [18–20], depicted has to derive it from a person’s behavior or report.
in Fig. 1. Various intervening factors of IN and ISB exist
Perceiving the need for certain information ini- – as they do for human behavior in general. Those
tiates an ISB. In order to obtain the required intervening factors can be personal characteristics
information, the user approaches one or multiple of physiological, cognitive [21, 22] or emotional
information sources. On the one hand, these can kind [23]. In a study about the information-seeking
be systems with the primary task of providing behavior of cancer patients, the authors found hearing
information, such as libraries, information retrieval loss, a lack of medical knowledge, verbal limitations

Fig. 1. Wilson’s information seeking behavior model [18].

446 S. Theis et al. / Predicting technology usage by health information need of older adults

and nervousness to determine information-seeking in New York City [33]. They found good adoption of
during the consultation with the doctor. In addi- the system by low-income participants, but small per-
tion, demographic variables, educational variables, sistent racial disparities in both the likelihood to be
social/interpersonal variables, economic variables offered an access code and the likelihood to actually
and source characteristics may play a role. Harris use the system. Thus, income itself did not appear to
[24] found out that medical terminology was a barrier be a limiting factor in adoption of patient portals,
to information exchanges with doctors. Furthermore, but may have a modifying influence in combina-
Bettmann and Park [25] detected that more educated tion with other demographic factors. However, these
people feel less need to search for more information, results may not be generalized to rural populations.
while Harris [26] proposed that the more knowledge- In a qualitative study dealing with eHealth users in
able the individual, the easier they would find it to rural Nebraska, Fruehling reports that usability and
encode information, thereby making further infor- clarity were important in the development of such sys-
mation acquisition easier. Finally, Moormann and tems, and noted that security and privacy issues were
Matulich [27] found that high knowledge levels did a major concern with this population [34]. Finally,
indeed facilitate information acquisition. Connell and it has to be considered that IN is not a fundamental
Crawford [28] found that the amount of health infor- need like the ones proposed by [35]; instead Wil-
mation received by urban residents from all sources son considers IN as supporting the needs listed in
declined with age, this was less the case for older rural Maslow’s theory of human motivation. Wilson noted
women while they also made use of a larger variety that research on the topic of IN and ISB, can stand
of sources. An age difference of information obtained on its own as an area of applied research where the
was also apparent between younger and older men: motive for the investigation is pragmatically related
younger men searched less than older men. to system design and development. That is why we
Intervening factors were later included in the consider it as suitable for user studies for eHealth
revised model of IN [29]. Besides personal char- system development.
acteristics of psychological, cognitive or emotional
kind, demographic, educational, social/interpersonal
and economic variables were mentioned. But not only 3. Research purpose
user characteristic are expected to influence IN and
seeking behavior, also source characteristics have to Related work shows that IN and technology usage
be taken into account. Moreover, Wilson’s revised are assumed to relate in some way. However, while
model now also takes into account how information- the investigations of HIN primarily focus on specific
seeking behavior can occur, ranging from passive diseases and contexts [36–40], a description of HIN
attention to more active and ongoing search. Addi- and HISB of older adults is lacking within the evi-
tionally, the construct of self-efficacy that can affect dence literature. In order to close this knowledge gap,
how successfully information sources are used, is also we want to address the following research questions
considered. However, different stress-coping mecha- (RQ):
nisms can also arise when e.g. getting an unpleasant
(RQ1) How large is older adults’ general and topic
medical diagnosis [30]. Some people may show an
specific HIN?
avoidant behavior while others may be more likely to
(RQ2) How do older adults acquire and use infor-
focus on the possible threat.
mation on their personal health and on health
IN and seeking behavior was investigated for
topics in general?
numerous application contexts. Greysen et al. have
investigated how functional impairment affects inter- The second challenge the present study aims to
net use among older adults and found that this had address is a perceived lack of generalizability in user
negative implications for meaningful use of patient centered development research of health information
portals [31]. Yagil et al. also found that using every- and communication technology (ICT). While it is not
day technologies generated stress for older adults, our aim to generalize our results to different popu-
which also suggested a potential barrier to patient lations, different contexts, or to different times, we
portal adoption and use [32]. Ancker et al. have inves- intend to describe the context and user requirements
tigated how disadvantaged populations use patient not only for one eHealth system, but also to as many
portals, based on system access logs of more than eHealth systems as possible. While concepts such as
74,000 patients at a Federally Qualified Health Center user goals or activities would be just as suitable here
 S. Theis et al. / Predicting technology usage by health information need of older adults 447

[19], present work will concentrate on the subject of authorities, which are provided with individual
health information. Since information transfer is an characteristics of people. From these 46 million
integral part of health information and communica- addresses, 5,000 addresses of people older than
tion technology, we consider information needs and 60 years with an equally distributed gender ratio
information search behavior to be suitable for gener- were randomly selected. Of the 5,000 addresses the
ating general context and user requirements. In order paper-based questionnaire was sent to, 586 partici-
to document the potential of these constructs (HIN, pants responded. 35 participants had to be excluded
HISB) to this purpose, empirical evidence for their because they were younger than 60 years, resulting
relation has yet to be made. In order to address this in a final sample size of N = 551 and a response rate
knowledge gap and in order to test Wilson’s Model of 11%. By investigating elderly’s HIN and ISB,
on ISB we ask: we assessed general HIN and trust in information
sources (TV, magazines, internet, doctor and phar-
(RQ3) What demographic factors influence HIN macist, family and friends or other). Subsequently,
and seeking behavior? questions on topic-related IN (diagnosis, therapy,
(RQ4) Is there a relationship between HIN/HISB interdependency of medicine) had to be answered.
and technology usage? This section closed with a question on information
sources participants use to access (personal) health
information and with a question on characteristics
4. Method of HISB considered as relevant in Wilson’s model
on ISB. Returned surveys were digitalized through
The aim of the study was to describe the HIN Remark Office software.
of older adults as the starting point and motiva-
tion for investigating ergonomic aspects of personal 4.3. Statistical analysis
health data access [41–43]. An Institutional Review
Board was obtained prior to the start of the study. A SPSS 24 served to conduct the statistical analy-
detailed description of the study method and descrip- sis of the data resulting from survey digitization with
tive results can be found in Mertens, Rasche, Theis, the help of Remark software. Initially we computed
Bröhl, Wille (2017) [44]. In contrast to the mentioned descriptive results for all HIN and HISB items (RQ1
article, the present paper does not deal with all the- and RQ2). A chi-square test of independence was
matic blocks of the initial survey in its descriptive performed to examine the relation between HIN and
documentation, but only with those of the HIN and HISB and demographic variables (RQ3). The chi-
HISB. The focus is also placed on the relationships square test was followed by a multiple linear regres-
between mentioned concepts and their determination sion involving all demographic variables showing a
of influencing factors. significant relation with HIN during chi-square tests.
To investigate the relation between HIN and tech-
4.1. Sample nology usage (RQ4) we computed the point-biserial
correlation coefficient between HIN and technology
The mean age of the participants (N = 551) was usage items. Each item on technology usage that
69.17 years (SD = 5.79) and ranged from 60–90 years. showed a significant relation with HIN, was included
The gender ratio of the sample was balanced includ- into a simple linear regression with HIN.
ing 51.3% males and 48.7% females. 441 participants
(80%) were already retired, while 109 (19.8%) were
still working. 5. Results

4.2. Procedure 5.1. General and topical HIN of older adults

The 15-page paper-based questionnaire was sent General HIN was measured by one item asking
via conventional mail to N = 5,000 adults aged 60 participants how satisfied they are with health infor-
years and older in June 2016. The addresses to mation at their disposal. Thematic HIN was measured
which the questionnaires were sent were rented by satisfaction regarding information on a certain
from Deutsche Post. This company has more topic. The answer scale ranged from one to five
than 46 million private addresses from registration where one stands for very satisfied and five for
448 S. Theis et al. / Predicting technology usage by health information need of older adults

very dissatisfied. High satisfaction represents a low and 11.8% (n = 65) share it with health insurance
HIN. Older adults in Germany are satisfied with the companies. A negligible number of participants share
information they get about (personal) health. 64% health-related information on social networks, with
(N = 334) claim to be ‘very satisfied’ and ‘satisfied’ self-help groups, or with people they do not know at
while 27.6% (N = 129) consider their satisfaction all. From N = 551 valid answers only twenty people
with the health information they get as ‘neutral’. do not share health-related information at all.
Only 36 participants are not satisfied or very unsat- In Germany, older adults describe their search
isfied. The health topics where older adults have the for information about health in general or their per-
highest HIN are information on health billings from sonal health as active and incidental. On a scale with
doctors and health insurances (n = 516, M = 3.05, 1 = ‘fully agree’ and 5 = ‘fully disagree’, the average
SD = 0.93). Besides billing transparency, older adults rating of the statement that their information search is
require information about medical diagnosis (n = 540, active is M = 2.7 (SD = 1.3, n = 387) while it is M = 2.7
M = 2.27, SD = 0.75), information about the meaning (SD = 1.22, n = 424) for the statement that their infor-
of examination results (n = 535, M = 2.37, SD = 0.72) mation search is causal. The least agreement is with
as well as on information on therapy and treatment the statement ‘I search regularly for health informa-
options (n = 530, M = 2.66, SD = 0.83). tion’ with a rating of M = 3.3 (SD = 1.9, n = 361).

5.2. HISB of older adults 5.3. Relation between HIN / HISB and
demographic variables
Descriptive results suggest that older adults in Ger-
many get information about health primarily from A chi-square test (s. Table 1) found a significant
their doctor and pharmacist or from TV. The internet association between the general HIN and the educa-
is the least used information source when it comes tional level. People with lower educational level are
to searching and finding health-related information. less satisfied with the information they have regarding
In total, 45.6% of N = 551 valid answers find health their personal health.
information on TV, 45.6% (n = 251) find health infor- The participant without education was satisfied the
mation in magazines, while 37.4% (n = 206) look on least with the information about medical billing (see
the internet. Doctors and pharmacists are the primary Table 2). Differences between all other levels of edu-
source for 47.9% (n = 264), while friends and fam- cational attainments were non-significant. Another
ily are consulted by 38.8% (n = 214). Participants significant association could be identified regarding
trust their doctor and pharmacist the most (n = 539, the satisfaction with information about the experience
M = 1.96, SD = 0.67,), while their trust in magazines that others have with their health and their educa-
and newspapers is the lowest (n = 492, M = 3.30, tion (see Table 2). Older adults with a lower level of
SD = 0.80,) on a scale ranging from one (very high) education were less satisfied with the available infor-
to five (very low). mation than older adults that are more educated. As
Older adults primarily share their information with there was also significant association between the sat-
family and friends (89.7%, n = 494) and with their isfaction concerning information about (interaction
doctor or their pharmacist (73.5%, n = 403). Just over of) medications and the number of chronic diseases.
a quarter (25.2%, n = 139) of older adults in Germany Our results suggest that people with a higher amount
share health-related information with acquaintances of chronic diseases are much more satisfied with the

Table 1
HIN of older adults (1 = very satisfied, 5 = very unsatisfied)
Education Satisfaction with health information (HIN, N = 520)
Very satisfied Satisfied Neutral Dissatisfied Very dissatisfied Total
None 0 0 0 0 1 (100%) 1
Lower second. education 4(8%) 21(40%) 18(35%) 9(17%) 0(0%) 52
Second. education 2(2%) 42(49%) 36(42%) 5(6%) 1(1%) 86
Vocational training 12(7%) 109(67%) 34(21%) 8(5%) 1(1%) 164
A-Levels 0(0%) 17(57%) 12(40%) 1(3%) 0(0%) 30
University degree 9(5%) 109(62%) 49(28%) 8(5%) 0(0%) 175
Other 1(8%) 8(66%) 2(17%) 1(8%) 0(0%) 12
χ2 (24) = 209,385, p < 0.000; Phi = 0.635; Cramer-V = 0.317; Contingency coefficient = 0.536.
 S. Theis et al. / Predicting technology usage by health information need of older adults 449

Table 2
Chi-square test statistics of associated variables (crosstabs)
Associated variables: HIN (general and thematic) χ2 DF p
General HIN × education 209.38 24 <0.001
HIN on billing information × education 60.48 24 <0.001
HIN on experiences of others × education 52.46 20 <0.001
HIN on medication interactions × no. chronic diseases 47.93 24 <0.001
HIN on course of disease × education 36.44 24 <0.05
HIN on therapeutic options × gender 12.05 4 <0.05
HIN on meaning of examination results × gender 11.08 4 <0.05
Associated variables HISB: trust in sources
Internet trustworthiness × postal codes west/east 12.08 4 <0.05
Trustworthiness doctor/pharmacist × education 59.25 18 <0.001
Associated variables HISB: information sharing
With family/friends × age group 15.2 6 <0.05
With acquaintance × living situation 9.76 4 <0.05
With health fund × age group 23.97 6 0.001
With doctor/pharmacist × postal code (west/east) 5.52 1 <0.05
Associated variables HISB: search characteristics
Activity of search × gender 11.56 4 <0.05
Persistence of search × number chronic diseases 47.93 24 <0.01

kind of information they get than people with no or health information is in all cases higher than the
less chronic diseases. The same applies to informa- amount who do not share their health information
tion about the course of diseases. Finally, we found with their health fund. The only exception here is
women are more satisfied with the information they the age group of the 85 to 90 year old adults. Here,
get than men. This is the opposite for the understand- more people share information about their health with
ing older adults have about their medical diagnosis. their health insurance. Older adults in the east of Ger-
Women indicated that understanding the meaning many share health information less with their doctor
of a diagnosis is very easy while more men than or pharmacist than in the western part.
women indicated a neutral position to this ease of
understanding. 5.3.3. Health information search characteristics
Significantly, more women than men describe their
5.3.1. Trust in health information sources search for health information as active. Finally, yet
Surprisingly, people in the eastern part of Ger- importantly, results indicate that the major part of
many rate the internet as slightly less trustworthy than older adults without a chronic disease reports that
people in the western part. Similarly, we found a sig- they do not regularly search health information while
nificant association between the trustworthiness in the people with two or more chronic diseases report
doctor/pharmacist and their education. Older adults the opposite.
with university degree rate this information source as
less trustworthy while people with lower educational 5.4. Relation between HIN and technology usage
level trust this source more.
There was a significant negative relationship
5.3.2. Health information sharing between HIN and the possession of a computer or
The older the participants, the less they share laptop rs = –0.11, T 95%, BCa CI [–0.19, –0.01],
health-related information with friends and family. p = 0.02. Hence, based on the point biserial cor-
In nearly all age groups, the number of people who relation, we conclude that the possession of a
share their information was higher than the number laptop or computer accounts to 0.11% to the IN
of people who do not. Furthermore, results reveal that represented by the satisfaction with available health
people who live in a supervised community for older information. We additionally found a significant
adults share health information with acquaintances. negative correlation between HIN and frequency
There was also a significant association between shar- of using a smartwatch, rs = –0.30, T 95%, BCa CI
ing health information with health insurance and age [–0.51, –0.10], p = 0.04. That means, the lower the
group. The amount of older adults who do not share satisfaction with available health information, the
450 S. Theis et al. / Predicting technology usage by health information need of older adults

Table 3
Demographic descriptor variables for the prediction of older adults HIN
Predictor variables Statistical measures (N = 514)
B SE B ß
Step 1 (Constant) 2.333 0.052
University vs. lower secondary education 0.294 0.110 0.123
University vs. secondary education 0.213 0.091 0.111
University vs. vocational training –0.083 0.075 –0.054
University vs. A-levels 0.133 0.137 0.044
University vs. others –0.083 0.207 –0.018
University vs. no education 2.667 0.695 0.164
Step 2 (Constant) 1.281 0.382
University vs. lower secondary education 0.250 0.111 0.105
University vs. secondary education 0.178 0.091 0.093
University vs. vocational training –0.116 0.076 –0.076
University vs. A-levels 0.124 0.137 0.041
University vs. others –0.052 0.206 –0.011
University vs. no education 2.633 0.693 0.162
Gender 0.139 0.061 0.097
Age 0.013 0.005 0.102
Amount of chronic diseases –0.004 0.026 –0.006
R2 = 0.25 for Step 1, R2 = 0.28 for Step 2, ∗ p < 0.001, ∗ p < 0.02∗∗ .

Fig. 2. HIN of older adults depending on education.

more frequently older adults use smartwatches. Fur- on their tablet PC, rs = –0.17, T 95%, BCa CI [–0.34,
thermore, a significant negative relationship could be 0.01], p = 0.04. This indicates that lower satisfaction
identified between HIN and older adults’ frequency with health information available corresponds with
of app usage on tablet PCs, rs = –0.15, T 95%, BCa a higher amount of health apps installed on older
CI [–0.29, 0.01], p = 0.04 as well as between HIN adults’ tablet PC. This corresponds with the relation
and amount of health apps older adults have installed between IN and frequency of app usage.
 S. Theis et al. / Predicting technology usage by health information need of older adults 451

5.5. Predicting HIN by age, gender, chronic vs. secondary education HIN (regression coeffi-
diseases and education cient) = 0.17–1.28 = –1.12), the change of satisfaction
with health information decreases by 1.12 points as
In a second step, in order to investigate the influ- a person changes from having a university degree to
ences on the general need for health information a person with secondary education degree. For the
more closely, variables significantly correlating with dummy variable of university vs. vocational training
general HIN were examined by a multiple linear and university vs. A-level the change in HIN is the
regression with forced choice method. We choose same for a person changing from university degree
this method as we have good theoretical reasons for to vocational training or to A-level. Only the differ-
including the chosen demographic predictors. And ence in the group means for the university group and
like Studenmund and Cassidy, (1987) we consider the participant with no education achievements was
stepwise techniques as influenced by random varia- significant (p < 0.001). Change of HIN (University vs.
tion in the data. no education = 0.01–1.28 = –1.27) increases (because
After transforming each educational level vari- satisfaction with information decreases) when a per-
able into a dichotomous category in relation to the son changes from having a university degree to having
educational level variable “university”. Dummy cod- no educational achievement.
ing was conducted, as multiple linear regression Within the model, participant’s satisfaction with
requires interval or ratio scaled variables (Field, health information increases by 0.01 points per
2018). While qualitative educational variables have additional chronic disease (p = 0.89), 0.01 points
a rank order, their distances are not equal which by each year of age (p = 0.02); male’s satisfaction
makes them an ordinal variable [45]. Whether or with health information was 0.13 points higher than
not participants attended the university was cho- females (p = 0.02). All together age, gender, number
sen as reference category, as sample sizes are very of chronic diseases, and educational level were sig-
unequal in the groups. As it can become problem- nificant predictors for older adult’s satisfaction with
atic to use very small groups as the reference (like health information
no education), we choose just the largest category as Participants’ satisfaction with health information
the reference. is equal to 1.28 (constant) + 0.25 (university vs.
Multiple regression shows that demographic vari- lower secondary education) + 0.18 (university vs. sec-
ables (amount of chronic diseases, age, gender and ondary education) + 0.112 (university vs. vocational
educational level) explain together 25%*, 28% of the training) + 0.12 (university vs. A-levels) –0.05 (uni-
variance in HIN. A regression was found to be statis- versity vs. others) + 2.63 (university vs. no education)
tically significant (R2 = 0.25 for Step 1, R2 = 0.28 + 0.14 (gender) + 0.01 (age) + (–0.01) (amount of
for Step 2, ∗ p < 0.001, ∗ p < 0.02∗∗ ), with an R2 chronic diseases) = 4.68.
of 0.28 (Table 3). Based on the standardized ß value, which is an indi-
The first dummy variable (university vs. lower cator for a predictor’s influence on the output variable,
secondary education) shows that the mean IN dif- we learn that education in general influences older
fers significantly between participants with university adults IN the most.
degree and lower secondary education (p = 0.03). University vs. no education (standardized ß = 0.69)
That means satisfaction with health information indicates that as educational level changes from
decreases as a person changes from having a a university degree to no education, satisfaction
university degree to having a lower secondary edu- with health information increases by 0.69 standard
cation degree 0.10 (regression coefficient B of deviations. The standard deviation for HIN is 0.77
university vs. lower secondary education) –1.28 (standardized ß 0.69 × 0.77 SD HIN). Therefore,
(constant) = –0.17). It is important to consider that changing the educational level from university degree
a decrease of satisfaction with health information to no education (the biggest difference in education
represents an increase of IN. This means that IN you could have) would increase the satisfaction of a
increases significantly more in adults with lower person with the information they have by 0.53 points
secondary education compared to those with uni- on a 5-Point Likert scale. Consider that an increase of
versity degree. Computing the significant (p = 0.05) satisfaction with health information corresponds to a
difference in the group means of satisfaction with decrease in IN.
health information for the group with secondary Changing the educational level from university
education (regression coefficient B of university degree to other educational levels, would increase
452 S. Theis et al. / Predicting technology usage by health information need of older adults

the satisfaction of a person with the information they would account for approximately 8% less variance in
have by 0.16 points on a 5-Point Likert scale (stan- the outcome.
dardized ß = 0.21 × 0.77 SD HIN = 0.16). Changing
a person’s educational level from university degree 5.6. Predicting technology usage with HIN
to A-levels would increase the satisfaction with
health information by 0.11 on a 5-Point Likert scale Simple linear regression was carried out to
(standardized ß 0.14 × 0.71 SD HIN = 0.11). This investigate the relationship between HIN as pre-
value would be 0.09 when changing it to lower sec- dictor (5-Point Likert scale) and possession of a
ondary education (standardized ß 0.11 × 0.77 SD computer/laptop (dichotomous) as output variable,
HIN = 0.09), 0.07 when changing it to secondary edu- between HIN and frequency of using a smartwatch
cation (standardized ß 0.091 × 0.771 SD HIN = 0.07) and between HIN and frequency of using apps
and 0.06 for vocational training (0.078 (standardized on a tablet PC. All according scatterplots showed
ß) × 0.77 = 0.06). that there was a strong positive linear relationship
The standardized ß value of gender (0.01) indi- between the pairs of variables, which was confirmed
cates that as the gender changes from men to women, by preceding correlation analysis documented in
satisfaction with health information would increase chapter 5.5.
by 0.05 points on a 5-Point Likert scale (0.06 Simple linear regression showed a significant rela-
(standardized ß) × 0.77 (SD HIN) = 0.05). With each tionship between older adults’ health HIN and the
additional number of chronic disease, the satisfaction possession of a computer or laptop (p = 0.01). The
with health information need increases about 0.02 slope coefficient for HIN was –0.07 so that the score
points on a 5-Point Likert scale (0.03 (standardized for possessing a computer or laptop changed by –0.07
ß) × 0.77 (SD HIN) = 0.02, while per year of age the from 1 (=possessing a laptop) to 0 (=not possessing
satisfaction with health information increases by 0.01 a laptop) when increasing the score for satisfaction
points on a 5-Point Likert scale (standardized beta with health information on a 5-Point Likert scale
0.005 × 0.771 = 0.01). value by one point (1 = very satisfied/very low HIN
If we look at the predictors of the second model to 5 = very dissatisfied/very high HIN). The R2 value
[18] then the highest correlation exists between edu- was 0.12, indicating that the model explains 12% of
cational level and satisfaction with health information the variation in HIN of older adults.
(r = –0.14, p = 0.001). The amount of chronic dis- Simple linear regression showed a significant rela-
eases (r = 0.03, p = 0.28) is not significantly related tionship between older adults health HIN and the
to IN, while age (r = 0.97, p = 0.01) and gender frequency of using a smartwatch (p = 0.01). The slope
(r = 0.10, p = 0.01) positively correlate with satisfac- coefficient for HIN was –0.32 so that the usage fre-
tion with health information. As the correlation of quency decreased 0.32 points from 4 (=daily) to 1
educational level and IN is negative, this denotes (=never), when increasing the score for satisfaction
that less educated older adults are more satis- with health information on a 5-Point Likert scale
fied with the health information they have at their value by one point. (1 = very satisfied/very low HIN
disposal and thus have a lower need for informa- to 5 = very dissatisfied/very high HIN). The R2 value
tion. Despite the significance of this correlation, was 0.18, indicating that the model explains 18% of
the coefficient is quite small, though it looks like the variation in older adults’ frequency of using a
the predictors are measuring different things and smartwatch.
that there is little collinearity. As out of all pre- Simple linear regression also showed a significant
dictor variables, educational level correlates best relationship between older adults health HIN and the
with the outcome variable, it is likely that educa- frequency apps on a tablet PC (p = 0.01). The slope
tional level will best predict IN. For all predictors coefficient for HIN was –0.45 so that the usage fre-
the difference between the regression beta coeffi- quency decreased 0.45 points from 4 (=daily) to 1
cient and the standardized beta varies between 0.047 (=never), when increasing the score for satisfaction
(age), 0.56 (educational level) and 0.038 (gender). with health information on a 5-Point Likert scale
Although educational level is the most important value by one point (1 = very satisfied/very low HIN
predictor, its results are the least generalizable. A to 5 = very dissatisfied/very high HIN). The R2 value
difference of 0.08 between the R2 and the adjusted was 0.21 so the model can explain 21% of the vari-
R2 of the model indicate that if the model was ation in older adults’ frequency of using an app on a
derived from the population rather than a sample it tablet PC.
 S. Theis et al. / Predicting technology usage by health information need of older adults 453

6. Conclusion treatment process as well as (billing) legal adapta-

tions, for example, could be sufficient in this case. For
In summary, results of a large-scale survey with the design of eHealth systems, the following recom-
adults older than 60 years in Germany suggest that mendations for action result from the research results
older adults are mainly satisfied with the informa- to answer the first research question:
tion they get about health in general as well as about
their personal health. Consequently, in addition to • eHealth technology should provide patients with
usability and legal hurdles, a lack of HIN might insight into treatment costs and processes.
also contribute to a lack of digital health system • eHealth technology should facilitate a compari-
usage [3]. HIN was quantified according to Wilson son of therapy and treatment options.
[20] as satisfaction with available information. To • eHealth technology should document treatment
the best of our knowledge, the need for information results and provide the patient with interpre-
has not been investigated in general and independent tation aids as well as reliable and trustworthy
of technology in other studies, so that a comparison additional information.
is challenging. General HIN revealed no significant As answer to RQ2: how do older adults acquire
correlation to social desirability (r = –0.05, n = 504, and use information on their personal health and on
p = 0.30), but with computer literacy – although to health topics in general? – the descriptive results iden-
a lesser extent (CLS, r = –0.15, n = 445, p = 0.01). tified doctors and pharmacists as the primary source
The less elderly people are familiar with the use of health information while only 37% of the partic-
of computers, the higher the need for health-related ipants access the internet. Since television serves as
information. Here we can already see the connec- a source of information for almost the same amount
tion between the handling and the attitude toward of older people, our findings contradict the statement
technology and the HIN. that older people prefer people as sources of infor-
According to described findings, the answer to mation to things like [50]. Trust is shaped according
RQ1, how large is older adults’ general and topic to the use of the source: older people mostly trust the
specific health information need? is that the general doctor or pharmacist. When exchanging information,
HIN of older adults is low. Thematic HIN include a however, friends and family take priority, followed by
comparably great need for information on doctors’ the doctor and pharmacist. Nearly every participant
and health insurance companies’ bills, diagnoses, the shared information. The results with regard to the
importance of examination results and information on important role of friends and families in the exchange
therapy and treatment options. Reason for the low IN of information suggest that the bi-directionality of
could be attributed to the fact that all expressed needs the information channels is more strongly perceived
are sufficiently addressed by doctors as it counts or practiced here than with the doctor or pharmacist.
among a component of their core tasks (diagnosis). A Participants describe their HISB as causal, active and
further explanation for low HIN could lie in the over- not regular. Wilson [29] describes these characteris-
idealized role the doctor has particularly for older tics as closely related to the interaction with sources.
patients [46] by which decisions are placed in the In addition to a pure description of the HISB through
hands of the doctor and any action is initiated by the characteristics mentioned above and with regard
him. This interpretation aligns with previous results to technology development, it would also be neces-
about older adult treatment behaviors, which indicate sary to consider the preferred characteristics in the
that as people age, they are less inclined toward self- future. As older adults consider professional medical
directed forms of health management [47–49]. Since sources as most trustworthy, involving these actors in
these studies do not include longitudinal results, it the introduction and application of digital health tech-
could also be that the older elders show a different nology might increase the acceptance and adherence
behavior (e.g. questioning authority less) than the of digital health systems. For the design of eHealth
younger people. Based on our results in answer to system it therefore should be considered that:
RQ1 we thus conclude that digital information sys-
tems purely geared towards the communication of • Digital health systems should be applied in coop-
health-related information – such as information on eration with the doctor, supporting his core tasks
the Internet - do not meet the requirement of older and facilitating communication with him.
adults as the usage of personal sources already deliv- • Older adults need health information irregularly
ers satisfactory health information. Changes in the and casually. The long-term use of digital health
454 S. Theis et al. / Predicting technology usage by health information need of older adults

systems should therefore require as little activity, empirically demonstrated a relationship between
effort and attention of the user as possible. HIN and older adults’ technology usage. Dissatisfac-
tion with available health information relates to the
In order to find out what demographic factors influ- possession of a computer or laptop, and to a higher
ence HIN and seeking behavior? (RQ3) chi-square frequency of using a smartwatch (n = 6) as well as to a
tests were conducted. According results support the higher frequency of using health apps on a tablet PC.
theoretically stated relation between demographic However, it has to be considered that the significant
variables, HIN and HISB. It is not surprising that both correlation of mentioned variables does not necessar-
objectives strongly relate to educational attainment. ily imply causation. This fact indicates the theoretical
Furthermore, the understanding of health informa- relationship stated by Wilson [20] that an informa-
tion and level of activity in information seeking was tion source’s failure to deliver desired information
revealed to differ depending on gender. In addition, influences the usage of other sources. Although the
the occurrence of chronic diseases plays a major role results do not allow for the conclusion that the source
for the information search continuity. Particularly of health information most frequently consulted by
unexpected the trust in information sources as well older people - the doctor or pharmacist - does not
as the information sharing with the doctor seem to provide the desired information and that the result-
be influenced by the political history. In the areas of ing need is the cause of the use of technology, it does
the former GDR, the mistrust for the internet seemed suggest this assumption. With a more detailed investi-
to be higher. In addition, older adults coming from gation of this relationship, the question arises whether
those areas share personal health information to a the information which the doctor cannot supply are
smaller extent with the doctor or pharmacist. Further at his/her disposal or not. It is conceivable that the
investigations into this aspect are necessary. Sharing doctor will not have access to this information or that
information about one’s own health relates demon- the resources and contextual conditions will be an
strably to age. The older people get, the less they obstacle to the transfer of this information. In par-
share information with family and friends. However, ticular, the relation between the HIN and the use of
it seems reasonable to assume that this is not only smartwatches suggests that the required information
due to the willingness and motivation, but also to is about personal behavior or vital functions, which,
mortality of family members and friends. at least in an outpatient context, are very difficult to
A more in-depth investigation by means of mul- acquire for the doctor himself.
tiple linear regression analysis confirmed a relation
between age, educational level and gender. In line • Providing information not available through
with previous research [25–27] education of a per- other sources can encourage to use an eHealth
son had the strongest impact on HIN. Lower educated technology
people have a higher HIN. With increasing age, the IN • eHealth should provide the physician and the
of a person declines. The HIN of men is lower than the patient with objective health-relevant data from
one of women. The initial relation between amount patients daily lives in real life settings
of chronic diseases and general HIN was not con-
firmed. These results allow the following guidance 6.1. Limitations
for eHealth system design:

• eHealth systems should have adjustable or even Described results have to be interpreted with regard
adaptive levels of information complexity as the to internal validity, imprecision of measurement and
depth of required and needed information can be other limitations. Internal validity of the present study
highly divergent is particularly influenced by the application of a non-
• eHealth systems for very old people should dras- standardized and self-formulated questionnaire.
tically decrease the amount of information and Results of HINs influence on smartwatch and tablet
required interaction by the user PCs have to be interpreted for a comparably small
• Personal experiences through social, societal number of people who actually used smartwatches
and political history should be taken into account (n = 6). These results suggest that especially tablet
in the user-oriented design of eHealth systems PCs, Smartwatches and Computers serve to provide
older adults with health relevant information. Over-
Results to answer RQ4 (Is there a relation- all, it has to be stated the proportion of variance
ship between HIN/HISB and technology usage?) explained by the model is relatively low at about
 S. Theis et al. / Predicting technology usage by health information need of older adults 455

7.6%. The use of technology thus appears to be use of technology often see in health currently – a
characterized by many influencing variables that are much more collaborative approach among the doc-
not taken into account in our model. Among other tors, pharmacists, family and friends, and the older
things, the understanding of the term information may adult. We suspect the reason for this is that the user’s
have contributed to this. Another reason might have perspective is still too rarely taken into account when
been the fact that the examination of the HIN con- deciding on the use of a particular technology. In
stituted a peripheral topic within a larger study on addition, collaborative approaches are accompanied
older adults technology usage. Considering a subset by a distribution of responsibility and competence. If
of all variables mentioned in Wilson’s model kept the these are in the hands of medical experts, a depen-
questionnaire short and concise. Identified influence dency gap arises that is more conducive to profit
of HIN on the use of technology are a motivation maximization than to patient autonomy and personal
and a first step to further validate and refine Wilson’s responsibility. In doing so, collaborative approaches
model. supported by technology could represent exactly the
When interpreting the results, it should also be con- relief for health systems that are so often demanded.
sidered that we are dealing with a snapshot of people However, implications and recommendations derived
in Germany who were older than 60 years in 2016. from the HIN only partially represent user needs and
This generation did not grow up with digital tech- contextual aspects. Their use alone does not yet lead
nology, and technology usage of people who grew to generalizable context and requirements analysis
up without digital tools differ from those of digital results. Further efforts and approaches in this direc-
natives [51]. For cancer survivors it has already been tion remain necessary. The aim of this work, however,
documented that older cancer survivors were less was to clarify the significance and benefits of general
likely to seek out information compared with their requirements analyses and to demonstrate a way to
younger counterparts and more likely to seek infor- address this challenge.
mation from friends, family, or books than from the
Internet [52]. Recurring observations of the present
type are thus required. In order to address this prob- Acknowledgments
lem, a repetition of the study is planned in the coming
years. This publication is part of the research project
Finally, the self-selected sample must be pointed “TECH4AGE,” financed by the Federal Ministry of
out. Although the randomized selection of question- Education and Research (BMBF, under Grant No:
naire addressees from the Federal German region was 16SV7111) and promoted by VDI/VDE Innovation
taken into account, each participant decided on his or + Technik GmbH.
her own collaboration. Generalizability of the results
is to be assessed against this background.
Conflict of interest
6.2. Summary
None to report.
The shown influence of HIN on technology usage
provides empirical evidence that IN is a feasible con-
struct to describe the context of eHealth systems
References
across applications and to investigate general user
research results.
[1] Akner G. Analysis of multimorbidity in individual elderly
This study describes the health-related informa- nursing home residents. Development of a multimor-
tion needs of older adults in Germany, as well as bidity matrix. Arch Gerontol Geriatr 2009;49(3):413-9
their theoretically assumed influences. The results [https://doi.org/10.1016/j.archger.2008.12.009][PMID: 19
could be used to formulate practical design guidelines 185359]
[2] Jochen P, Hartmut B, Ernst K, Hans GM, Martina M, Heim-
for eHealth systems, so that user characteristics and fried W. Future Report Demographic change: Innovation
needs can be better-accommodated in future digital ability in an ageing society. Bonn: DLR-PT; 1999.
work systems in the health sector. Thereby, this work [3] Ziefle M, Röcker C. Acceptance of pervasive healthcare
systems: A comparison of different implementation con-
contributes to a more efficient, effective and satisfac-
cepts. In: Acceptance of pervasive healthcare systems: A
tory use of digital work systems. It is very interesting comparison of different implementation concepts; 2010; pp.
that the recommendations are very different from the 1-6.
456 S. Theis et al. / Predicting technology usage by health information need of older adults

[4] Donald AN. Psychology of everyday things: The design of [26] MacInnis DJ, Jaworski BJ. Enhancing and measuring con-
veryday things - Revised and expanded edition. sumers’ motivation, opportunity, and ability to process
[5] Nielsen J. Usability engineering. 18. [print.] Amsterdam brand information from advertisements. Journal of Market-
u.a.: Morgan Kaufmann, 2010. ing. 1991;(53):1-23.
[6] McDonald N, Ward M, Morrison R. Achieving impact in [27] Moorman C, Matulich E. A model of consumers’ preven-
ergonomic research. Work. 2012;41:209-18. [https://doi. tive health behaviours: The role of health motivation and
org/10.3233/WOR-2012-0158-209] health ability. Journal of Consumer Research. 1993;(20):
[7] Talsma L, Molenbroek JFM. User-centered ecotourism 208-28.
development. Work. 2012;41:2147-54. [https://doi.org/ [28] Connell CM, Crawford CO. How people obtain their health
10.3233/WOR-2012-1019-2147] information: A survey in two Pennsylvania counties. Public
[8] Broberg O, Edwards K. User-driven innovation of Health Reports. 1988;(103):189-95.
an outpatient department. Work. 2012;41:101-6. [29] Wilson TD. A re-examination of information seeking
[https://doi.org/10.3233/WOR-2012-0142-101] behaviour in the context of activity theory. Information
[9] Almeida e Silva CM, Okimoto MLRL, Tanure RLZ. Similar Research. 2005;11(4):1.
methodological analysis involving the user experience. [30] Holton MK, Barry AE, Chaney JD. Employee stress
Work. 2012;41:968-75. [https://doi.org/10.3233/WOR- management: An examination of adaptive and mal-
2012-0272-968] adaptive coping strategies on employee health. Work.
[10] Warner ES, Murray AD, Palmour VE. Information needs 2016;53(2):299-305.
of urban residents. Final report on Contract No: OEC-0-71- [31] Greysen S, Chin GC, Sudore R, Cenzer I, Covinsky
4555 to US Office of Education; 1973. K. Functional impairment and internet use among
[11] Wilson TD. On User Studies and Information Needs. Jour- older adults: Implications for meaningful use of
nal of Documentation 1981;37(1):3-15. [https://doi.org/ patient portals. JAMA Intern Med. 2014;174(7):1187-8.
10.1108/eb026702] [https://doi.org/10.1001/jamainternmed.2014.1803][PMID:
[12] Kuhlthau CC. Inside the search process: Information seek- 24841330]
ing from the user’s perspective. Journal of the American [32] Yagil D, Cohen M, Beer JD. Older adults’ coping with
Society for Information Science. 1991;42(no. 5):361-71. the stress involved in the use of everyday technolo-
[13] Ellis D, Deborah C, Katherine H. A comparison of the gies. J Appl Gerontol. 2016;35(2):131-49. [https://doi.org/
information seeking patterns of researchers in the physical 10.1177/0733464813515089][PMID: 24652919]
and social sciences. Journal of Documentation. 1993;49(no. [33] Ancker JS, Barrón Y, Rockoff ML, et al. Use of an electronic
4):356-69. patient portal among disadvantaged populations. J Gen
[14] Brenda D. Sense-making theory and practice: An overview Intern Med. 2011;26(10):1117-23. [https://doi.org/10.1007/
of user interests in knowledge seeking and use. Journal of s11606-011-1749-y][PMID: 21647748]
Knowledge Management. 1998;2(2):36-46. [34] Fruhling AL. Perceptions of e-health in rural communities.
[15] Cheuk Wai-Yi B. An information seeking and using process In: Patient-Centered E-Health. IGI Global. 2008;157-67.
model in the workplace: A constructivist approach. Asian [35] Abraham M. A theory of human motivation. Psychological
Libraries. 1998;7(12):375-90. Review. 1943;50(4):370.
[16] Sandstrom PE. An optimal foraging approach to information [36] Azadeh F, Ghasemi S. Investigating information-seeking
seeking and use. The Library Quarterly. 1994;64(4):414-49. behavior of faculty members based on wilson’s model:
[17] Choo CW, Detlor B, Turnbull D. A Behavioral Model of Case study of PNU University, Mazandaran, Iran.
Information Seeking on the Web–Preliminary Results of a Global Journal of Health Science. 2015;8(9):26-36.
Study of How Managers and IT Specialists Use the Web.; [https://doi.org/10.5539/gjhs.v8n9p26]
1998. [37] JA H, PA B. Understanding the information behaviours
[18] Wilson TD. Human information behavior. Informing Sci- of carers of people with dementia: A critical review
ence. 2000;3(2):49-56. of models from information science. Aging & Men-
[19] Wilson TD. Activity theory and information seeking. Ann tal Health. 2008;12(4):467-77. [https://doi.org/10.1080/
Rev Info Sci Tech. 2008;42(1):119-61. 13607860802224300]
[20] Wilson TD, Walsh C. Information behaviour: An inter- [38] A KP, S P, C L, J P, N K. Music, musicians and infor-
disciplinary perspective: A review of the literature. British mation seeking behaviour: A case study on a community
Library Research and Innovation Centre London, 1996. concert band. Journal of Documentation. 2015;71(1):3-24.
[21] Festinger L. A theory of cognitive dissonance. Stanford, [https://doi.org/10.1108/JD-07-2013-0083]
California: Stanford University Press, 1957. [39] Tury S, Robinson L, Bawden D. The information seek-
[22] Rogers EM. Diffusion of Innovation: Third Edition. New ing behaviour of distance learners: A case study of
York: A Division of Macmillan Publishing C., Inc. 1962. the University of London international programmes. The
[23] Kassulke D, Stenner DK, Doory M, Ring I. Information Journal of Academic Librarianship. 2015;41(3):312-21.
seeking behavior and sources of health information: Asso- [https://doi.org/10.1016/j.acalib.2015.03.008]
ciations with risk factors status in an analyiss of three [40] W C, X Z, K X, Y W. Modeling online health
Queensland electorates. Australien Journal of Public Health. information-seeking behavior in China: The roles of
1993;(17):51-7. source characteristics, reward assessment, and internet
[24] Harris J. You can’t ask if you don’t know what to ask: self-efficacy. Health Communication. 2016;31(9):1105-14.
A survey of the information needs and resources of [https://doi.org/10.1080/10410236.2015.1045236]
hospital outpatients. The New Zealand Medical Journal. [41] Theis S, Rasche P, Mertens A, Schlick CM. An Age-
1992;105(934):199-202. Differentiated Perspective on Visualizations of Personal
[25] Bettman JR, Park CW. Effects of prior knowledge and expe- Health Data. In: Advances in Ergonomic Design of Sys-
rience and phase of the choice process on consumer decision tems, Products and Processes. Springer, Berlin, Heidelberg
processes. Journal of Consumer Research. 1980;(7):234-48. 2017; pp. 289-308.
 S. Theis et al. / Predicting technology usage by health information need of older adults 457

[42] Theis S, Bröhl C, Wille M, et al. Ergonomic Considera- [48] Turk-Charles S, Meyerowitz BE, Gatz M. Age dif-
tions for the Design and the Evaluation of Uncertain Data ferences in information-seeking among cancer patients.
Visualizations. In: Human Interface and the Management of The International Journal of Aging and Human Devel-
Information:Information, Design and Interaction. Springer opment. 1997;45(2):85-98. [https://doi.org/10.2190/7CBT-
International Publishing 2016; pp. 191-203. 12K3-GA8H-F68R]
[43] Theis S, Bröhl C, Rasche P, Wille M, Schlick CM, Mertens [49] Jenkins V, et al. Information needs of patients with cancer:
A. Age-dependent health data visualizations: A research Results from a large study in UK cancer centres.
agenda. Mensch und Computer 2016-Workshopband, 2016. [50] Chaudhuri S, Le T, White C, Thompson H, Demiris
[44] Mertens A, Rasche P, Theis S, Bröhl C, Wille M. Use of G. Examining health information-seeking behaviors of
information and communication technology in healthcare older adults. Comput Inform Nurs. 2013;31(11):547-53.
context by older adults in germany: Initial results of the [https://doi.org/10.1097/01.NCN.0000432131.92020.42]
tech4Age long-term study. I-com. 2017;16(2):165-80. [PMID: 23974574]
[45] Research Methods Knowledge Base. [51] Sackmann R, Winkler O. Technology generations revis-
[46] Johnson MF, Lin M, Mangalik S, Murphy DJ, Kramer AM. ited: The internet generation. Gerontechnology. 2013;11(4).
Patients’ perceptions of physicians’ recommendations for [https://doi.org/10.4017/gt.2013.11.4.002.00]
comfort care differ by patient age and gender. J Gen Intern [52] Bennett JA, Cameron LD, Whitehead LC, Porter
Med. 2000;15(4):248-55. [https://doi.org/10.1111/j.1525- D. Differences between older and younger cancer
1497.2000.07004.x] survivors in seeking cancer information and using
[47] Woodward NJ, Wallston BS. Age and health care beliefs: complementary/alternative medicine. J Gen Intern Med.
Self-efficacy as a mediator of low desire for control. Psychol 2009;24(10):1089-94. [https://doi.org/10.1007/s11606-
Aging. 1987;2(1):3. 009-0979-8][PMID: 19685099]