Use of Information and Communication Technology in Healthcare Context by Older Adults in Germany: Initial Results of the Tech4Age Long-Term Study

2.1 Questionnaire Construction

In order to map user- and context-dependent parameters together with technology-specific usage aspects the Tech4Age questionnaire consists of topically different parts representing each aspect. Existing validated tools were combined with standard demographic- and health-related items as well as with newly included items. The topical order of the questionnaire remained the same for all participants, starting with ten demographic items and 25 items on personal ICT usage. Subsequently 19 theory-based items on information need and behavior and 62 items on (health) application usage (mobile applications, user tasks, mobile interaction) were followed by standardized questionnaires on computer and health literacy as well as on technological affinity. The social desirability scale aimed at examining the latter. Most of the questions had closed answers, where participants had to mark one or multiple predefined answers or rate statements on a Likert scale. Open questions were applied in some cases to gain insight on motivational background and reasons for a certain behavior or to collect answers beyond already known ones. Open-ended answers were coded by two independent investigators for analysis.

As mentioned, the survey started with demographic questions: year of birth, gender, living conditions like who lives within the same household, urban or rural residential area, educational achievement, years of being in retirement, (previous) profession and a list of chronic diseases. These were followed by a section asking about ICT and internet use: (mobile) internet usage at home and average usage per day, followed by a list to indicate usage of devices like desktop computer, smartphone (defined as having a touchscreen), mobile phone (defined as working with hardware buttons), tablet, smartwatch and how long participants have owned these devices. Furthermore, usage of health-related devices like fitness tracker, blood glucose meter, home emergency call system, navigation, e-book or smart home were queried, together with corresponding purchasing reasons. Finally, the last question of this block queried who supports technical service of the mentioned devices (like installing apps) and device usage frequency (daily, every two to three days, weekly, monthly, never).

The next section had the topic of health-related information need and behavior. It started with an adapted version of the EU-HLS-47 consisting of 15 instead of 47 items to assess participants’ health competence selected in an expert workshop. Following this, general health information need and trust in information sources (TV, magazines, internet, doctor and pharmacist, family and friends or other) were investigated. Subsequently, questions were asked on topic-related information need (diagnosis, therapy, interdependency of medicine) resulting from pre-study interviews. This section closed with a question on information sources participants use to access (personal) health information and with a question on characteristics of health information search behavior considered as relevant in theoretical models [13].

The subsequent section focused on (health) application usage. Here, it was asked how many applications participants have installed on their smartphone and on their tablet and how often they use them. Furthermore, it was asked if they sort smartphone applications following special rules and where they get information about apps. Then usage of different types of health application were queried, like applications on measuring pulse or calories, or first aid apps, diagnosis, fitness and medicine reminders, applications enabling communication with the doctor or doctor rating applications. Furthermore, participants were asked how many health-related apps they have installed on their smartphone or tablet and how often they use them. After that it was asked what participants consider as reasons against the use of health applications. Thereby, a more detailed picture should be drawn of how older adults already use or are willing to use health related applications. This attempt is a more in depth as was found in related surveys. Demographics and personal information asked in the first block were used as reference for more detailed analysis and insights.

The health application related block was followed by a block about general application and information and communication technology usage. First a list of user goals and activities (navigation, playing, shopping, banking, writing letters and messages, taking photos) was sketched against possible media (desktop computer, smartphone, tablet, other) to conduct each mentioned activity. The idea behind this question was to determine which parts of everyday life are already digitally assisted and which ones are preferred to be performed in person. Then came questions about the experience of controlling devices with different input technologies (mouse, touch), about how satisfied they were with those input technologies and about the hand position during touchscreen interaction for reading and writing shorter or longer text.

Finally a section of standard questionnaires followed: the computer literacy scale (CLS, [7]), which indicates how familiar participants are with computer symbols, a section on technological affinity [3], a section on social desirability [15] and the adapted version of the health literacy scale [8]. At the end of the questionnaire participants had space for further remarks.

2.2 Procedure

Together with a one-page cover letter describing the reason for the contact and the request for participation in this study the 15-page paper-based questionnaire was sent to N = 5,000 adults with an age of 60 years and older in June 2016. Potential participants’ addresses were selected randomly restricted by age criteria (equal or above 60 years) and a balanced gender distribution within Germany by a service provider. Returned surveys were digitalized through Remark Office software. In cases of uncertainty of the Remark output a questionnaire was visually inspected and values were transcribed by hand. The data matrix output of Remark was subsequently analyzed with SPSS 22.

2.3 Restrictions

A total of 586 participants responded to our questionnaire. Thirty-five participants had to be excluded because they were younger than 60 years, resulting in a final sample size of N=551 and a response rate of 11%.

Beyond this paper-based questionnaire a corresponding online version was created. Participants were given the option to answer the online version instead of the paper-based one. Furthermore, the online questionnaire was advertised in social media. 141 participants answered that way. However, this sample is not part of the present paper as the age structure of the participants differed widely and ranged from 20–80 years and as it is known that participants of online questionnaires do in general differ from participants in paper-based questionnaires [10].

2.4 Participants

The mean age of the participants was 69.17 (SD = 5.787) and ranged from 60–90 years (see figure 1). The gender ratio of the sample was balanced including 51.3% males and 48.7% females. Most of them lived in their own house or flat (n = 391; 71%), second most in a rented apartment (n = 151; 27.4%). Only two participants lived in assisted accommodation. 333 participants (60.4%) lived in an urban area and 216 (39.2%) in a rural area. Table 1 shows the highest educational achievement of our participants. 441 participants (80%) were already retired, while 109 (19.8%) were still working. Table 2 shows participants’ (previous) professional domains.

Figure 1

Participants’ age distribution.

The global technological affinity had an average value of 3.23 (SD = .701) on a five-point Likert scale. While the average on the subscale acceptance was M = 2.71 (SD = .969), participants’ competence scored on average M = 3.36 (SD = 1.025) and the locus of control dimension had a mean value of M = 3.59 (SD = .874). The social desirability of participants’ answers was on average 5.18 (SD = .672) on a seven-point Likert scale. The subscales self-delusion (M = 5.08, SD = .753) and external delusion (M = 5.28, SD = .969) had similar average values. Technological affinity [3] and social desirability [15] will be discussed later (chapter 3.5) while in this paragraph mean values are presented to characterize the sample.

3.1 Information and Communication Technology Used by Older Adults

One aspect of the Tech4Age long-term study is to investigate the integration of information and communication technology (ICT) within older adults’ everyday life. Pre-studies revealed aspects regarding the actual use as well as the acquisition reasons of ICT [5], [6]. Based on these pre-studies descriptive results of the long-term study will provide a more profound and detailed insight by providing answers to the following research questions: (1) How is ICT used among older adults and what are the reasons for technology acquisition? (2) If used, who administrates ICT products? (3) Do older adults use the internet and if so in which situations? (4) Which everyday activities do older adults perform with the help of ICT?

3.2 Use of Health-Related Information and Levels of Trust

Seeking information about one’s health is increasingly documented as a key coping strategy in health-promotive activities and psychosocial adjustment to illness. The way people seek information influences the use of digital health technology and the corresponding levels of acceptance. Health literacy (HL) is considered the capability to obtain, comprehend and use health information in terms of decision-making [8]. Unlike health information seeking behavior (HISB) it includes individual skills [13], [12], [14]. Both concepts were considered as a framework for investigating the elderly’s requirements and needs in terms of digital health applications. Relations between both concepts and interactions with technology usage by the elderly (see section 3.1.1) will also be investigated. Current results focus on descriptives and will answer the following research questions: (1) What health information needs do older adults have? (2) How do older adults seek information about health? (3) What level of health competence do older adults in Germany have?

3.2.5 Health Literacy: What Do Older Adults Know About Health?

Descriptive results on participants’ health literacy suggest that older adults in Germany have no problems understanding information on health topics. The questions of the health literacy scale queried how the participants would rank the easiness with which information on different topics could be interpreted, see table 9. On a scale from 1 = ‘very good’ to 5 = ‘very bad’ they only ranked ‘information on advantages and disadvantages of treatment methods’ (M = 3.06, SD = 0.92, n = 527) and ‘credibility of the media’ (M = 3.25, SD = 0.96, n = 527) with an average score above three. Topics they were able to understand best included ‘following medication intake’ (M = 1.93, SD = 0.79, n = 527) and the ‘handling of health-harming behavior’ (M = 1.84, SD = 0.71, n = 526).

Figure 2

Number of apps installed on devices sorted according to apps in general and mHealth apps in particular.

3.3 General and mHealth Applications Used by Older Adults

The third aim of this study is to investigate whether the full potential of modern ICT products is used by older adults. Therefore the use of apps in general and the use of mHealth apps in particular were investigated. Descriptive results will provide a more profound and detailed insight by means of the following research questions: (1) Do older adults use applications and do they use mHealth applications? (2) What are the reasons for not using mHealth applications? (3) If mHealth applications are used by older adults, what type of applications do they use? (4) How do older adults find out about applications?

3.3.1 Use of Applications and mHealth Applications

Among older adults the use of applications is quite limited in general. The majority only use up to ten general applications on a smartphone or tablet. In the case of mHealth applications the majority do not use any (see figure 2).

Although the majority only uses a small number of apps the frequency of use is quite high. About every third participant uses an application on a daily basis. Due to question design it is unclear whether it is always the same app that they are using, but nevertheless this frequency suggests quite a high level of integration of applications into everyday life. In the case of mHealth applications the frequency of use is different. The majority of the small group of mHealth application users uses them on a weekly or monthly basis (see figure 3).

Figure 3

Frequency of use for apps in general and mHealth apps in particular.

3.4 Analysis of Handheld Positions During Touchscreen Interaction

As a result of technical progress, the use of touchscreen devices has become more common in recent years. This may have far-reaching consequences as recent studies have shown that visual information close to the hands is perceived in a different way than information further away from the hands [1]. In order to gain insights on how users hold touchscreen devices during interaction, the present research focuses on a task-dependent analysis of handheld positions, distinguishing between smartphones and tablets.

Touchscreen devices can be operated either vertically or horizontally, with one or both thumbs, or one or more fingers of the same hand or the other one. It is essential to differentiate between tablets and smartphones since they have different sizes, weights and shapes. Research conducted so far investigated approaches of finger placement and hand grasp during touchscreen interaction. With regard to the interaction with tablets, Oulasvirta et al. [4] recommended a symmetric bimanual grip while holding a tablet in landscape orientation as being the most appropriate for ergonomic text input. Regarding smartphone interaction, Wobbrock, Myers and Aung [16] studied a total of eight different postures, four different two-handed postures and four different one-handed postures. They found that the posture of the hand has a significant effect on users’ touch performance and that interaction with the device should not only take place by placing fingers on the front of the device but also that the back of the device should be investigated as a feasible means for interaction.

While previous work provides some approaches to hand position analysis, none of them takes older aged subjects, a thorough differentiation of hand positions, and a comparison between smartphone and tablet interaction into account. Therefore, our survey contained questions regarding hand postures while interacting with smartphones and tablets for three different tasks: typing short text, typing long text, and reading. The hand positions that were studied were:

1. Holding the device vertically and interacting with the thumb of the same hand.

2. Holding the device vertically with one hand and interacting with a finger of the other hand.

3. Holding the device vertically with both hands and interacting with both hands.

4. Holding the device horizontally and interacting with both thumbs.

5. Holding the device horizontally with one hand and interacting with a finger of the other hand.

3.4.1 Handheld Positions

Chi-square tests were used to study the effects of handheld positions for the three different tasks separately, as the measurement was at the nominal level. The level of significance was set to α=0.05.

Results of the analysis of handheld positions with regard to smartphone interaction of the generation older than 60 years of age showed significant effects for typing short text (χ2(4,n=52)=93.58,p<.001), typing long text (χ2(4,n=52)=90.12,p<.001) and reading text (χ2(4,n=52)=31.46,p<.001). The visual analysis of the bar graphs (figure 4) shows that, once again, the most prominent position for all three tasks is holding the smartphone vertically and typing with a finger of the other hand.

Results of the analysis of handheld positions with regard to tablet interaction of the generation older than 60 years of age showed significant effects for typing short text (χ2(4,n=52)=38.81,p<.001), typing long text (χ2(4,n=52)=27.81,p<.001) and reading text (χ2(4,n=52)=17.04,p<.001). The visual analysis of the bar graphs (figure 4) shows that the most prominent position for typing long text and reading is holding the tablet horizontally and typing with a finger of the other hand, whereas for typing short text the most prominent position is holding the tablet vertically and typing with a finger of the other hand.

Figure 4

Bar graphs for the three different tasks in relation to handheld positions for smartphone and tablet interaction for the generation older than 60 years of age.

3.5 Correlation of Factors

Beneath these detailed descriptive results on different aspects given in the chapters above, this subchapter will focus on correlations of the scales like social desirability, technical affinity and computer literacy with items like use of internet and apps to characterize the sample more detailed. For all those correlations Spearman’s-Rho is used as most variables like ‘number of apps installed’ use ordinal data sorted in categories like ‘1–5’, ‘6–10’, ‘11–15’ and so on rather than concrete numbers as values.

As was expected technical affinity shows many correlations with other variables and has therefore also in the target group of the elderly a large influence in the use of information and communication technology. Not surprisingly, the scales of technical affinity and computer literacy do highly correlate in a positive manner (rs = .403, p = .000), meaning that those participants with high technical affinity do also have high computer literacy. High positive correlations of technical affinity can be found with the duration of using the internet at home (rs = .474, p = .000) or mobile (rs = .335, p = .000), the number of apps on the smartphone (rs = .349, p = .000) and tablet (rs = .291, p = .000) as well as the frequency of app using on the smartphone (rs = .341, p = .000) or tablet (rs = .345, p = .000). Furthermore technical affinity do correlate negative with age (rs=−.172, p = .000), indicating that older persons within the given 60+ sample has less technical affinity, and with gender (rs=−.233, p = .000), indicating that woman in general have less technical affinity.

Beneath the correlation with technical affinity the variables of ICT-use correlate highly among each other: Those participants who use the internet often at home use it also often in mobile situations (rs = .475, p = .000) and those with many apps installed on the smartphone have also many apps installed on tablet (rs = .711, p = .000) and it is the same with the frequency of app use on smartphone and tablet (rs = .520, p = .000). These facts do indicate that the use of ICT is platform independent (smartphone – tablet) and independent from the situation (mobile – at home) but mostly moderated by personal traits like technical affinity, which is slightly moderated by age (as mentioned above).

Finally correlations between social desirability and other results of the questionnaire are reported as this tool was integrated as a control indicator whether participants answered truly or if they try to answer how they think the researchers want them to answer. Analysis revealed a correlation between technical affinity and social desirability (rs = .107, p = .016), meaning that those participants with high technical affinity also show high social desirability scores. This can be seen as an indicator that some amount of the technical affinity score is based on the wish of the participants to answer how they think that the researcher wants them to answer. So the mean technical affinity score (3.23 on a five-point Likert scale as mentioned under 2.4) might be biased and a little higher as in reality. However, the only other variable social desirability correlates with – and only in a very small manner – is the number of chronical diseases (rs = .095, p = .030). Here again, as those participants who count more diseases do also have higher social desirability scores, participants might have mentioned chronic diseases although they just suffer in mild or even no manor from these. The fact that no other variables correlate with social desirability shows that at least within that sample participants have answered most honestly and not as they think they should have to answer. The social desirability can help to get insight into this phenomenon and should be integrated in more surveys, especially as it is based on 6 questions only.

Regarding mHealth apps it is worth noticing that neither the number of installed mHealth apps nor the frequency of use on smartphone or tablet-PC correlates with the number of chronical diseases or the amount of being satisfied with the given health information. Instead the number of installed mHealth apps correlates again with technical affinity (smartphone: rs = .294, p = .000; tablet: rs = .167, p = .022). Those participants with many mHealth apps on the smartphone do also have more mHealth apps on the tablet (rs = .556, p = .000) and those participants who use mHealth apps frequently on the smartphone do use them also more often on the tablet (rs = .657, p = .000). Taken together this indicates that the use of mHealth apps is platform independent and linked to the individual technical affinity rather than to the number of chronical diseases or satisfaction with health information.

4.1 Comparison with Other Surveys

In this last paragraph we now compare exemplarily some of our results to those well-known other surveys focusing on elderly, ICT-use and health mentioned in the introduction. One problem in such a comparison is, that those surveys often have another main topic, other age groups (focusing on age 60+ is rare) and last but not least other answer categories even if using the same questions.

The often cited ‘ARD-ZDF-online Studie’^[1] (online survey of the German public service broadcast companies) focused on the use of internet for all ages and had 1508 participants in 2016. They found a mean duration of being online for 85 minutes a day for persons aged between 50 and 69 years and for 28 minutes a day for persons being 70 or older. As our survey did not ask for concrete minutes of being online but had categories (‘0 = no use’, ‘1 = up to 15 minutes’, ‘2 = 15–30 minutes’, ‘3 = 30–60 minutes’, ‘4 = 1–2 hours’, ‘5 = more than 2 hours’), the comparison includes a bit estimation. For the age group of 60–69 we find a mean value of the categories of 2.41 representing something between 15 to 60 minutes of being online, which is less than in the ARD-ZDF-online study but also in an older age category (60–69 instead of 50–69). For the age group from 70 upwards the mean of the categorical data is 1.88 representing something between 1 and 30 minutes of being online. So compared to this study our participants are less online, meaning that at least the time of being online for elderly is not overestimated in our study. Asked for what they are doing online the age group 70+ states in the ARD-ZDF-online study that for 17% they communicated, for 15% they were using media, for 22% they are searching for information, for 30% they play games and for 15% they purchase goods in the internet. Our categories are more detailed, but show similar results, as shown in table 8.

The ‘Altersstudie Generali’^[2] (age survey of the insurance company Generali) interviewed about 2000 participants face-to-face, focused on living condition of participants aged between 65 and 85. ICT use and being online were no addressed topics. Results of the ‘Altersstudie Generali’ show that more than 50% live in their own house, which is a bit less than our finding with about 70% living in their own house or flat.

The destatis report ‘Ältere Menschen in Deutschland und der EU, 2016’^[3] (report of the german federal office for statistics) state that about 52–55% of the age group 60–79 years are female, whereas this number rises by age. Our sample (aged 60–90 years) included 48.7% woman, which is less than the federal average. Due to destatis 72% of households with persons over 65 own a desktop computer, which is pretty in line with our findings of 77% owning a desktop computer in our sample. According to destatis about 49% of people older than 65 years were online in 2015. In our sample, collected 2016, about 80% of sample were online in 2016. Beneath the process of getting more and more people online each year, a gap remains showing that in our sample more participants are online than compared to the official statistics. However, comparison between this study and official statistics of destatis regarding activities and goals older adults use ICT for shows no differences: 90% write emails, 85% searching for information, 67% do consume online media, 44% do online banking, 39–49% do online shopping (in different categories). Taking into account that those percentages by destatis only refer to online users (not the whole sample) the activities match with those presented here in table 8.

To sum it up: Although surveys on a voluntary basis can never be representative (as they have selective sample, based on the voluntary return of surveys) the Tech4Age long-term study is mostly in line with other well-known statistics and surveys. The sample analyzed in this paper included a bit more (5%) male person answered, some more owners of a house or flat (70% instead of 50% in common) and more persons being online (80% instead of 50%) as the average statistics and representative surveys showed. However, activities and goals using ICT for were same as compared to the ARD-ZDF-online study and the destatis data.

Nevertheless, by utilizing the information gathered in the Tech4Age long-term study one of the biggest obstacles of integrating digital technology into medical care processes with older adults can already be overcome or at least reduced at the product development stage, as product designers and developers get a deeper insight into the lifestyle habits and mental models of prospective elderly users.