This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR Human Factors, is properly cited. The complete bibliographic information, a link to the original publication on https://humanfactors.jmir.org, as well as this copyright and license information must be included.
Maintaining nutrition and exercise strategies after rehabilitation can be difficult for older people with malnutrition or limited mobility. A technical assistance system such as an e-coach could help to positively influence changes in dietary and exercise behavior and contribute to a sustainable improvement in one’s nutrition and mobility status. Most apps do not provide a combination of nutrition and exercise content. In most cases, these apps were evaluated with healthy individuals aged <70 years, making transferability to vulnerable patients, with functional limitations and an assumed lower affinity for technology, in geriatric rehabilitation unlikely.
This study aims to identify the potential for optimization and enhance usability through iterative test phases to develop a nutrition and mobility e-coach suitable for older adults (≥65 years) based on individual health behavior change stages in a rehabilitation setting.
Iterative testing was performed with patients aged ≥65 years in a rehabilitation center. During testing, participants used an e-coach prototype with educational elements and active input options on nutrition and mobility as a 1-time application test. The participants performed navigation and comprehension tasks and subsequently provided feedback on the design aspects. Hints were provided by the study team when required, documented, and used for improvements. After testing, the participants were asked to rate the usability of the prototype using the System Usability Scale (SUS).
In all, 3 iterative test phases (T1-T3) were conducted with 49 participants (24/49, 49% female; mean 77.8, SD 6.2 years). Improvements were made after each test phase, such as adding explanatory notes on overview screens or using consistent chart types. The use of the user-centered design in this specific target group facilitated an increase in the average SUS score from 69.3 (SD 16.3; median 65) at T1 to 78.1 (SD 11.8; median 82.5) at T3. Fewer hints were required for navigation tasks (T1: 14.1%; T2: 26.5%; T3: 17.2%) than for comprehension questions (T1: 30.5%; T2: 21.6%; T3: 20%). However, the proportion of unsolved tasks, calculated across all participants in all tasks, was higher for navigation tasks (T1: 0%, T2: 15.2%, T3: 4.3%) than for comprehension tasks (T1: 1.9%, T2: 0%, T3: 2.5%).
The extensive addition of explanatory sentences and terms, instead of shorter keywords, to make it easier for users to navigate and comprehend the content was a major adjustment. Thus, good usability (SUS: 80th-84th percentile) was achieved using iterative optimizations within the user-centered design. Long-term usability and any possible effects on nutritional and physical activity behavior need to be evaluated in an additional study in which patients should be able to use the e-coach with increasing independence, thereby helping them to gain access to content that could support their long-term behavior change.
Different demographic, clinical, biological, and lifestyle factors contribute to the development of frailty and sarcopenia in older populations. The accumulation of these risk factors leads to a reduction in resistance to health stressors. In addition to a decline in independence, there is an increased risk of falls and mortality [
Technical assistance systems, such as health apps, could help ensure that dietary and exercise behaviors are implemented and changed after rehabilitation. In a survey in Germany with older adults (aged >65 years) in 2020, up to 22% reported that they used health apps to track fitness data, and 16% used apps to obtain information about health, fitness, and nutrition topics. However, the proportion of seniors interviewed who could envision using such apps was more than twice as high for both types of health apps [
Apps are one type of technical tool for improving, assisting, or supporting people. The various realizations, such as tele-visits, exergames, or health websites, can be summarized under the term eHealth. A review of the use of eHealth in the context of geriatric rehabilitation revealed that most studies (68%) involved people with neurological diseases. In addition, only 8% of all identified studies assessed the use of health apps as an eHealth intervention. However, the results on the applicability of eHealth with the target group indicated that interventions are feasible if adequate training takes place, and if the eHealth intervention is simple and has good usability [
As previously described, a nutrition intervention in combination with an exercise intervention could lead to more significant effects than would an exercise intervention alone in older people with frailty or sarcopenia [
To develop an age-adapted device and health app (e-coach) for older adults with deficits in nutritional status and physical activity needs, we followed the German International Organization for Standardization 9241-210:2019
These concepts first require an analysis of the context of use and, as a next step, a specification of the use requirements. The context of use was described in a previous study by performing and analyzing focus groups with older adults as well as experts [
The information gained was used to derive the user and design requirements for the e-coach. For geriatric patients in rehabilitation, educative content from the areas of nutrition and physical activity focused on the changes and demands of aging that should be included in the e-coach. The older adults would also like the e-coach to be able to provide them with exercises and thus support them in their training. The feedback and evaluation of input regarding nutrition and exercise are described as helpful but should not be an admonition. The results indicate that, as many patients in this age group have little experience with technology and usually use other sources of information, it is important to develop a nutrition and mobility e-coach, particularly given the easy handling and provision of clear information to individual users on the advantages of the e-coach. It is also important for older adults to avoid barriers, such as small font, low video volume, or poor contrast.
The e-coach needs to be integrated into users’ daily lives without stressing or restricting them. Moreover, it must be possible to adapt the content to the physical abilities of the users, and because of the heterogeneity of older people in terms of previous knowledge and willingness to change their behavior, appropriate strategies should be used. A recent umbrella review of eHealth interventions suggested that applications involving behavior change techniques may have promising effects on physical activity, sedentary behavior, and healthy eating. However, it is not yet known which theoretical construct is the most effective [
On the basis of the findings from the focus groups, it was possible to further differentiate the settings for teaching the use of the e-coach. To give older adults time to familiarize themselves with the system and generally introduce them to its use, this introduction should already take place in the rehabilitation center. In a real health care situation, it would be easier to explain the technology to the patient; in case of questions or if further explanations are necessary, it would be more uncomplicated to address these points in a personal appointment. In addition, patients would also be able to repeat relevant content that they may not have been able to remember completely from the seminars at their own pace. Therapists should have the possibility to adapt the e-coach to the needs of the patients and their TTM phase. As, in the context of rehabilitation, the therapies take place directly between the physiotherapist or nutritionist and the patient, and the interventions are also strongly influenced by the interactions between the professionals and patients, complete automation of the e-coach would not be efficient. Adaptations of the e-coach to the patient’s previous knowledge and support needs should therefore be made by a physiotherapist or a nutritionist.
This paper aims to describe the design process and an iterative evaluation of the developed content. The aim of this study is to identify optimization potentials and enhance usability through iterative test phases to develop a nutrition and mobility e-coach based on individual health behavior change stages, usable for older adults (≥65 years) in a rehabilitation setting.
The e-coach prototypes were evaluated with older adults in 3 iterative test phases, using a between-subject design. User experience was reflected by the System Usability Scale (SUS) [
To detect and analyze usability problems in more detail, at least 10 patients were included in each iterative test phase [
The study was approved by the Ethical Review Board of the Carl von Ossietzky University Oldenburg (registration number: 2018–132). We conducted the study in accordance with the Declaration of Helsinki and the underlying data protection regulation.
Patients in rehabilitation, from geriatric and cardiology wards, were eligible based on the following inclusion criteria: (1) participants aged ≥65 years and (2) participants were able to speak and understand German. Exclusion criteria were (1) severe visual impairment (eg, inability to read large font on a screen), (2) severe hearing impairment (eg, deafness), or (3) inability to understand study information and provide informed written consent (eg, aphasia or severe cognitive impairment or dementia). Participants were recruited by placing flyers in the patients’ wards.
The e-coach screens were designed in Adobe XD (version 34; Adobe Systems) for a 10-inch tablet in landscape mode. Design guidelines for apps for older adults were used to take into account specific requirements, such as a decline in vision or decreased motor abilities when developing the prototype [
The e-coach contained two main topics:
Structure of the e-coach modules.
Data were collected in the form of in-person testing in patients’ rooms at the rehabilitation center. After the patients had been informed of the content and the procedure of the study by a study team member (LH or MS) and had signed the informed consent, a survey was conducted on sociodemographic data, data on nutrition and physical activity, the phases of behavioral change, and technology commitment. A usability test was subsequently performed. The entire process of data collection took approximately 45 minutes to 60 minutes per patient.
Nutritional status was assessed using the Mini Nutritional Assessment-Short Form (MNA-SF) [
Patients were classified separately into the TTM for physical activity and nutrition based on the data from the dietary protocol and the PASE. Patients who did not achieve the defined target criteria in the areas of physical activity or nutrition were asked whether they had thought about changing their behavior (contemplation) and, if so, whether they had already planned to do so in any specific way (planning). Depending on their answers, patients were then categorized into the phases of precontemplation, contemplation, or planning. Patients who were already performing the target behavior were asked whether they had been doing so for a short time (action) or for a longer period (maintenance). On the basis of their answers, the patients were classified into the phases of action or maintenance.
In addition, the patients’ technology commitment was assessed using a questionnaire developed by Neyer et al [
The test procedure was explained in detail, and the contents and navigation options were shown in advance. Before testing the usability task, the patients were told that the aim of the study was not to test their abilities, but the quality and usability of the e-coach [
Testing also took place in the patient’s room at the rehabilitation center. During the test, the examiner (LH or MS) and the participant sat at a table. The tablet with the app could be placed on the integrated stand of the device, placed on the table, held in the hand, or laid down by the participant. The examiner read the tasks to the participant and then observed the participant.
Usability tasks in three different domains were defined and used for each iterative testing period: navigation, comprehension, and design. Navigation tasks were used to determine whether users were able to find their way through the e-coach and use the buttons correctly. In the case of quizzes, we tested whether the screens were structured such that they could be used successfully by the participants. On quiz screens, the question was highlighted at the top of the screen (eg, What is the recommended minimum number of small portions of dairy products to eat per day?) and below it, two to three answer options were shown (eg, You should eat at least two servings of dairy products per day and You should eat at least four servings of dairy products a day) along with a prompt (press the correct answer). Comprehension tasks were intended to test whether screen content and information were correctly interpreted and understood. The aim of the design questions was to identify visual barriers such as a font that was too small or an acoustic problem, such as an extremely fast rate of speech in videos. An example of questions and tasks is shown in
Task types in the different iterative testing phases as the total number of tasks and percentages per iterative testing phase.
Task type | Navigation task, T1a, n (%) | Navigation task, T2b, n (%) | Navigation task, T3c, n (%) | Comprehension question, T1, n (%) | Comprehension question, T2, n (%) | Comprehension question, T3, n (%) | |||||
|
N/Ad | N/A | N/A | ||||||||
|
Next screen (1 screen) | 16 (57) | 5 (24) | 3 (30) |
|
|
|
||||
|
Further screen (≥2 screens) | 4 (14) | 5 (24) | 4 (40) |
|
|
|
||||
|
Use of back button | 5 (18) | 5 (24) | 2 (20) |
|
|
|
||||
|
Use of different tabs (text elements) | 2 (7) | 2 (10) | 0 (0) |
|
|
|
||||
|
Use of the help button | 0 (0) | 3 (14) | 1 (10) |
|
|
|
||||
|
Use of quizzes | 1 (4) | 0 (0) | 0 (0) |
|
|
|
||||
|
Use of the exercise diary | 0 (0) | 1 (5) | 0 (0) |
|
|
|
||||
|
N/A | N/A | N/A |
|
|||||||
|
Purpose of the screen |
|
|
|
5 (28) | 2 (20) | 1 (25) | ||||
|
Foresight of content |
|
|
|
3 (17) | 0 (0) | 1 (25) | ||||
|
Nutrition diary |
|
|
|
1 (6) | 0 (0) | 0 (0) | ||||
|
Interpretation of content |
|
|
|
7 (39) | 8 (80) | 2 (50) | ||||
|
Understanding of quizzes |
|
|
|
2 (11) | 0 (0) | 0 (0) |
aT1: iterative phase 1.
bT2: iterative phase 2.
cT3: iterative phase 3.
dN/A: not applicable.
The results from the usability tasks were reported in three different categories (success rate, number of hints, and content of hints) to evaluate usability problems in more detail. The performance of the particular task was evaluated in the categories of
For the usability test, patients were instructed to simultaneously speak their thoughts aloud while performing the tasks. The concurrent think-aloud method was intended to immediately identify and specify problems for older adults using the e-coach [
The tasks in the test sequence were always carried out in the same order, but patients had the option of skipping tasks at any time or stopping the usability test. The tests and verbal feedback during the tasks were recorded by taking notes.
After the usability test, patients were finally interviewed using the SUS. The questionnaire contained 10 statements about the usability of a system rated on a 5-point Likert scale. The values of individual items were added together and then multiplied by 2.5, resulting in a score between 0 and 100. The results provided a general overview of product usability [
A total of 3 iterative test phases were used. Before each test phase, we refined the content, design, and potential functionality of the elements. The first test phase mainly tested the basic functionality of the chosen navigation structure with the target group, with
German original version of screens from the 3 iterative test phases. Explanations and translations of changed elements based on the results of the test phase are shown in the text boxes.
Statistical analyses were performed using the software SPSS (version 27.0; SPSS Inc).
Participants’ characteristics, success rates of the tasks, and SUS scores were analyzed using descriptive statistics; they were presented as frequencies, means, SDs, and percentages.
The hints given during the usability test were presented as the total number of hints required for this task type in the iterative phase. The total number of unsolved tasks and percentages of unsolved tasks for all participants that performed this task type in the iterative phase were also reported.
The notes from the concurrent thinking aloud during the usability test were used to derive aspects that the participants noticed during the test. These aspects were discussed by the study team (LH and MS) to identify specific problems in the tasks and to derive possibilities for optimization.
An explorative ANOVA was conducted to compare the SUS scores among the 3 iterative test phases. The robust Welch
The group was divided into 2 equally sized subgroups to test for group differences in successful task completion. Half of the participants who performed the tasks better were compared with the other half. The variables of age, sex, BMI, MNA-SF classification, technology affinity, TTM phase nutrition, TTM phase mobility, and SUS score were tested for group differences. The explorative analysis was performed using the Mann–Whitney
A total of 49 patients who were aged 66-94 years (24/49, 49% female; mean 77.8, SD 6.2 years) participated in the study. Patient characteristics per iterative test phase are presented in
Overview of participants’ characteristics within each iterative test phase (N=49).
Characteristics | Iterative phase 1 | Iterative phase 2 | Iterative phase 3a | ||||||||
|
Nutrition | Mobility | Nutrition | Mobility | Nutrition | ||||||
|
|||||||||||
|
Total, n | 15 | 12 | 13 | 13 | 12 | |||||
|
Female, n (%) | 8 (53) | 6 (50) | 9 (69) | 5 (39) | 5 (42) | |||||
BMI (kg/m2), mean (SD) | 26.3 (4.4) | 26.6 (5.6) | 27.6 (5.8) | 26.6 (4.6) | 27.6 (4.3) | ||||||
Age (years), mean (SD) | 79.1 (6.8) | 78.4 (5.5) | 78.5 (7.3) | 76.3 (6.8) | 76.8 (5.2) | ||||||
|
|||||||||||
|
Precontemplation | 3 (20) | 1 (8) | 4 (31) | 0 (0) | 3 (25) | |||||
|
Contemplation | 2 (13) | 5 (42) | 3 (23) | 5 (39) | 2 (17) | |||||
|
Planning | 4 (27) | 3 (25) | 1 (8) | 4 (31) | 5 (42) | |||||
|
Action | 0 (0) | 1 (8) | 2 (15) | 1 (8) | 0 (0) | |||||
|
Maintenance | 6 (40) | 2 (17) | 3 (23) | 3 (23) | 2 (17) | |||||
|
|||||||||||
|
Total score (12-60 points) | 36.6 (11.3) | 39.3 (13.4) | 38.7 (11.4) | 43.7 (6.9) | 39.4 (11.9) | |||||
|
Pointse | 3.1 (1.0) | 3.4 (1.6) | 3.3 (1.1) | 3.7 (0.6) | 3.3 (1.0) |
aOnly tests in the nutrition section were performed to keep contact times and numbers low in the context of increasing COVID-19 incidence in the region.
bDifferent sample sizes within nutrition and mobility owing to the patient’s choice option to participate only in one main theme or in both main themes.
cTTM: transtheoretical model of behavior change.
dTC: technology commitment (Neyer et al [
eAverage Likert scale points per item.
Navigation tasks required fewer hints (14.1%-26.5%) per task than comprehension tasks (20%-30.5%) in all iterative test phases (
Participants’ performance in the different navigation and comprehension tasks in the iterative test phases.
Tasks | Iterative phase 1 | Iterative phase 2 | Iterative phase 3 | |||||||||
|
Total tasksa, N | Hintsb, n (%) | Failc, n (%) | Total tasks, N | Hints, n (%) | Fail, n (%) | Total tasks, N | Hints, n (%) | Fail, n (%) | |||
|
177 | 25 (14) | 0 (0) | 204 | 54 (27) | 31 (15) | 93 | 16 (17) | 4 (4) | |||
|
Next screen (1 screen) | 105 | 15 (14) | 0 (0) | 49 | 10 (20) | 8 (16) | 29 | 4 (14) | 1 (4) | ||
|
Further screen (≥2 screens) | 21 | 3 (14) | 0 (0) | 54 | 20 (37) | 10 (19) | 37 | 8 (22) | 3 (8) | ||
|
Use of back button | 32 | 6 (19) | 0 (0) | 55 | 10 (18) | 6 (11) | 18 | 1 (6) | 0 (0) | ||
|
Use of tab layout (text elements) | 12 | 1 (8) | 0 (0) | 9 | 2 (22) | 2 (22) | 0 | 0 (0) | 0 (0) | ||
|
Use of the help button | 0 | 0 (0) | 0 (0) | 28 | 10 (36) | 5 (18) | 9 | 3 (33) | 1 (11) | ||
|
Use of quizzes | 7 | 0 (0) | 0 (0) | 0 | 0 (0) | 0 (0) | 0 | 0 (0) | 0 (0) | ||
|
Use of exercise diary | 0 | 0 (0) | 0 (0) | 9 | 2 (22) | 1 (11) | 0 | 0 (0) | 0 (0) | ||
|
105 | 32 (31) | 2 (2) | 74 | 16 (22) | 0 (0) | 40 | 8 (20) | 1 (3) | |||
|
Purpose of screen | 33 | 11 (33) | 0 (0) | 26 | 5 (19) | 0 (0) | 9 | 1 (11) | 0 (0) | ||
|
Foresight of content | 15 | 6 (40) | 0 (0) | 0 | 0 (0) | 0 (0) | 12 | 2 (17) | 0 (0) | ||
|
Nutrition diary | 8 | 6 (75) | 2 (25) | 0 | 0 (0) | 0 (0) | 0 | 0 (0) | 0 (0) | ||
|
Interpretation of content | 38 | 7 (18) | 0 (0) | 72 | 11 (15) | 0 (0) | 19 | 5 (26) | 1 (5) | ||
|
Comprehension of quizzes | 11 | 2 (19) | 0 (0) | 0 | 0 (0) | 0 (0) | 0 | 0 (0) | 0 (0) |
aTotal number of tasks performed by all participants per iterative phase.
bSummed up the number and percentage of required hints for all participants for this task type in the iterative phase.
cTotal number and percentage of unsolved tasks in all participants who performed this task type in the iterative phase.
Most participants understood how to use the buttons correctly. In 86.9% of all tests, participants consistently selected the arrow button for navigation to the next screen as intended. When navigating back to previous screens, this was done completely correctly in 84.7% of all tasks.
Many participants were able to successfully interpret the active input options such as the drinking protocol (11/15, 73%), the diagram of consumed food groups (9/10, 90%), and the exercise diary (7/8, 88%).
Some participants had difficulties in interpreting the content they would expect to find based on the names of the modules or the themes. The contents that needed to be optimized are shown in
Many design elements on the overview screens and the educational features were rated positively by older adults. For educational content with texts, the font sizes, readability, and length of the text were positively rated in almost all corresponding tasks (92.3%).
Moreover, more than half of the participants (25/49, 51%) were able to successfully solve more than 90% of the tasks. Exploratory analysis of group differences suggested that those who solved more than 90% of the tasks had significantly higher technology affinity (
Content and structures optimized following iterative testing with participants.
Domain | Adaptations |
Navigation |
On the main overview screens for the nutrition and mobility modules, details on the content of modules were added. Checkboxes for the confirmation of exercise execution and the labeling of the elements were enlarged. For screens that guide different topics in a module, a question or more guidance about the content was added in addition to the title (eg, increasing activity: How can I become more active in everyday life?). |
Comprehension |
Keywords were supplemented with further information (eg, The wording An instruction for the action was added below the screen heading (eg, “Please select one of the following topics to get more information.”). Information for food groups was added ( |
Design |
The symbols for nutrients (the molecule symbol was changed to a magnifying glass), interesting information (the light bulb was changed to a book with light bulb on it), and the nutrition diary (the booklet was changed to a book) were replaced. Photographs for text elements were exchanged for symbols or drawings. Exercise photos were used instead of exercise drawings; a white background was added to the exercise photos. Any other elements besides diagrams were removed from the evaluation screens. Feedback on reaching the training goal using flowers instead of stars was added (flowers contain additional information about the number of exercises performed). |
The evaluation of the SUS showed a continuous improvement in the usability of the e-coach (
Because the normality assumptions of the ANOVA were violated, a 1-way Welch-ANOVA was performed to determine whether the SUS score was significantly different among the test phases. The improvement in the SUS score was not statistically significant among the 3 test phases (Welch
System Usability Scale (SUS) score for each iterative test phase.
Phase | Values | ||
|
n (%) | Mean (SD; range) | Median (IQR) |
Iterative phase 1 | 21 (43) | 69.3 (16.3; 42.5-97.5) | 65.0 (57.5-83.8) |
Iterative phase 2 | 16 (33) | 70.3 (18.7; 20.0-95.0) | 77.5 (60.0-84.4) |
Iterative phase 3 | 12 (25) | 78.1 (11.8; 50.0-92.5) | 82.5 (70.6-86.9) |
We showed that it is possible to conduct iterative test phases and improve the usability of a health app, even for older adults with health restrictions who are undergoing inpatient rehabilitation. After 3 iterative test phases, the average SUS increased from 69.3 (SD 16.3; median 65) to 78.1 (SD 11.8; median 82.5), indicating good usability of the e-coach and is comparable with the SUS results on eHealth application use in other studies with older adults [
Comparing the completion of tasks among the 3 iterations, it is noticeable that in the first iterative test phase, hints were needed, especially for comprehension questions (30.5%), and a few tasks were not fulfilled by all participants in this test (1.9%). In this iterative phase, hints were often needed for comprehension questions, such as what can be done on certain screens and also concerning the interpretation of what was meant by certain titles (eg, an overview of the recommended amount of physical activity). These difficulties were solved by adding more information to the descriptive texts and partly by changing the wording of individual terms.
In the second iterative test phase, the proportion of required hints in the navigation tasks (26.5%) increased noticeably compared with the first iterative test phase (14.1%), and 15.2% of all navigation tasks could not be solved by the participants in the second iterative test phase. In the first iterative test phase, many navigation tasks included easier functions, such as navigation to the next screen, and only a few tasks with more complex navigation steps. In contrast, in the second iterative test phase, more complex tasks were added, requiring, for example, 2 navigation steps, and the proportion of simpler navigation tasks decreased accordingly (
In the third iterative test phase, hints were also required more often for navigation tasks that required several navigation steps. Although there were some tasks in the last iteration phase that required more than one navigation step, only 17% (2/12) of all participants needed help with all navigation tasks, and 4.3% of all navigation tasks in the third iterative test phase were not successfully completed. However, in the comprehension tasks, the need for hints was almost similar in the last iterative phase (20%) compared with the second iterative phase (21.6%), but more tasks could not be solved (2.5%) than in the second iterative phase. Compared with the first iterative test phase, almost all participants (10/12, 83%) in the last iterative test phase were able to derive the following content from the labeling of the modules. In addition, the question about what can be done on the overview screens was also answered correctly by all participants. Nevertheless, some participants found it difficult to interpret diagrams, meaning that they required hints, and one participant was unable to solve the task.
Other aspects should be considered as possible reasons for the different performance in the tasks, in addition to the increase in complexity. There were fewer tasks and fewer participants in the third iterative test phase because increasing COVID-19 case numbers forced us to reduce contact times with the participants. In addition, all the comprehension questions that could not be answered successfully in the last test were questions about protein and energy intake shown in a diagram. This topic could be difficult for people with little existing knowledge of the subject.
Furthermore, in our exploratory analysis of group differences between half of the participants who solved more than 90% of the tasks correctly and those who solved less than 90% of the tasks correctly, we found indications that those who solved more than 90% of the tasks correctly had a significantly higher affinity for technology (
Technology commitment in our study population was at a median of 3.5 (SD 0.6) points in the group of participants who correctly solved more than 90% of the tasks and at a median of 3.0 (0.6) in the group of people who correctly solved less than 90% of the tasks. This value seems to correspond approximately to the technology readiness of people in this age group in Germany. Rasche et al [
To increase the usability of e-coaches in health care, as well as for less technically inclined older persons, detailed instructions on how to use the e-coach could be applied. Older people seem to have greater benefits from step-by-step instructions when learning to use new technologies. To address this problem in a real-world setting in terms of use of the e-coach by older adults, guidance (eg, in the form of a printed manual) should be additionally offered [
The data on the TTM phases of the participants demonstrate that geriatric patients in rehabilitation are at different phases of the behavioral change process. In the iterative tests with nutrition content as well as with physical activity content, the proportion of participants in the first 2 phases of TTM (precontemplation and contemplation) and the remaining phases (preparation, action, and maintenance) was quite balanced. As such, the readiness and implementation of behavior change in the areas of nutrition and physical activity appear to be heterogeneous among the target group, requiring the use of different strategies in the e-coach to support behavior changes. No indications of significant group differences between the half of the participants who were able to solve more than 90% of their tasks and those who were able to solve less than 90% of the tasks were found for the TTM phase distribution in the area of nutrition. For the TTM phase in the physical activity domain, no testing for group differences was conducted because of the small sample size resulting from missing data from the last iterative test. A correlation between certain phases of behavior change and the ability to successfully solve tasks could be explained only to a limited extent. In the case of tasks for operating the app, such as finding specific screens or using buttons, personal readiness to change one’s nutritional or physical activity behavior should have little influence. It could be possible that the interpretation of specific content with terms such as
When developing the e-coach, design recommendations for the target group [
In a few tests, it was noticed that a user wanted to select the correct button, but the pressure on the button was not recognized by the device. Besides a possible malfunction of the touchscreen, there are 2 points that have been described in further studies [
It is likely that more people who already had a general interest in nutrition, physical activity, or technology participated in the study. Therefore, a selection bias cannot be ruled out even if at least the measured technology commitment also corresponds to the figures from another study with older people in Germany with varying degrees of experience in the use of technology [
Data collection was conducted by study team members who were involved in the development of the e-coach. An uninvolved person would have had to conduct usability tests to increase the objectivity of these tests. This was not possible because of the financial resources of the project.
Moreover, the test situation itself may have biased the results. On the one hand, it is likely that older adults were more nervous and made more mistakes than if they had used the e-coach unobserved. On the other hand, the possibility of receiving hints and help from the examiner could also have led to a situation in which help was requested more quickly and, if necessary, tasks would have been solved after some time without the help of the study team member. In addition, it cannot be ruled out that unintentional nonverbal responses were given by the study team (eg, nodding) that were not recorded in the documentation.
In the context of this study, the time spent on the task was not determined. The time it takes a participant to complete a task can be an indicator of usability, as quick performance can also indicate ease of use. However, there is also evidence from a study by Sonderegger et al [
This study only evaluated whether e-coach elements are generally usable for older adults. It provides an indication of the e-coach’s usability for first-time users but not for a longer period of use.
This study involved older people undergoing inpatient rehabilitation in iterative optimization and usability testing for an e-coach according to the German International Organization for Standardization 9241-210:2019
As the target group is particularly vulnerable, and an individual’s willingness to continuously use the e-coach may impose an additional burden on them, it is essential to evaluate the acceptance, willingness, and adherence for long-term use of the system in a further study. Previous studies have shown that it can be very difficult to recruit patients for such long-term use of technical devices in this target group, and that good usability, as well as subjective benefits for patients, must necessarily be present [
Examples of tasks in iterative phase 2 physical activity.
Examples of tasks in iterative phase 3 nutrition.
Unsolved tasks and optimizations.
Mini Nutritional Assessment-Short Form
Physical Activity Scale for the Elderly
System Usability Scale
transtheoretical model of behavior change
We thank all participants for participating in the study and for providing us with valuable inputs. Special thanks are due to Daniel Küppers for programing the e-coach and Patrick Elfert for his valuable input and work in developing the nutrition diary. We would also like to thank the geriatric and cardiologic rehabilitation departments of the Rehabilitation Center of Oldenburg for giving us the opportunity to conduct our study.
This research was funded by an intramural grant from Carl von Ossietzky University Oldenburg (School of Medicine and Health Sciences; Research Pool 2018-018). The funder had no role in the study design, data collection, analysis and interpretation, decision to publish, or preparation of the manuscript.
LH, AH, and RD led to the conception and design of the study. LH, MS, AH, and RD contributed to their expertise in developing the app content. LH and MS coordinated and performed usability testing and evaluation. LH wrote the manuscript with input from MS, AH, and RD. AH and RD obtained funding and supervised the study. All authors read and approved the final manuscript.
None declared.