Usability, defined as the extent to which users can achieve specific goals effectively and satisfactorily [35], was assessed using the System Usability Scale (SUS) developed by Brooke et al [36]. This standardized questionnaire contains 10 statements about the usability of a system. A 5-point Likert scale is used to answer the questions. For the classification of the SUS score, we used the adjective rating scale of Bangor et al [37]. A total of 100 points can be achieved, which can be divided into the following categories: best (100.0‐84.1 points), excellent (84.0‐80.8 points), good (80.7‐71.1 points), okay (71.0‐51.7 points), poor (51.6‐25.1 points), and worst possible (25.0‐0.0 points) [36,37].

In addition, technological commitment, according to Neyer et al [38], was recorded to reflect technical readiness, which encompasses technical acceptance, competence, and control. A 5-point scale with the response options strongly disagree (1), disagree (2), somewhat disagree (3), somewhat agree (4), and strongly agree (5) is used to assess attitudes toward and use of modern technology in general [38].

Regarding the feasibility of using the entire assistance system, this study focuses on the participants’ experiences, data collection, storage, processing, and analysis. To identify the strengths and weaknesses of the MuTS and the tablet app, participants complete a report after each use of the assistance system documenting the use of each individual integrated module (see Figure 1). Finally, a self-developed questionnaire was used to evaluate the overall system (see Multimedia Appendix 1). This questionnaire consists of 10 questions, which can be answered on a scale of 1 to 5, with 1 being “very positive” and 5 being “very negative.” The questions relate to the content and use of the tablet app, the handling of the equipment in the MuTS, the exercise, and the general process. In addition, 5 open-ended questions were included to gather feedback on the content that needs improvement, problems in the process, what participants liked best, and the extent to which they believe that the MuTS, including app, can influence their nutritional and mobility behavior. The participants’ experiences with the overall system will inform the testing and trial of the MuTS, the app (tablet-based and in the MuTS), and data collection for a planned subsequent RCT. Furthermore, analysis algorithms will be developed to evaluate the data, facilitating the RCT.

The German version of the MNA-SF (0‐14 points) was used to screen for malnutrition and risk factors. It contains 6 questions to differentiate between normal nutritional status (12‐14 points), risk of malnutrition (8‐11 points), and malnutrition (0‐7 points) [33,39]. In order to gain better insights into dietary habits, participants were asked to maintain a food diary in the tablet app for 3 consecutive days during the week preceding T0, T1, and T2. The data collected were screened by study staff members for calories, protein, and fat intake. In addition, the fluid and food intake were evaluated according to the recommendations of the German Nutrition Society, specifically for older adults [40].

The short physical performance battery (SPPB) and the TUG test were used to assess possible mobility deficits and to track their development over time, as the tests were performed at baseline, T0, T1, and T2. The SPPB consists of 3 parts to assess physical functions: static balance, walking speed, and the time it takes to get up from a chair 5 times (5-CRT). The rating scale contains 3 categories: poor (0‐4 points), moderate (5‐8 points), and good mobility (9‐12 points) [41,42]. The TUG is a test to measure the time it takes a person to get up from a chair (if possible, without the help of the arms), walk 3 meters, turn around, walk back, and sit down on the chair again. Walking speeds above 0.8 m/s were considered normal [23,43].

In addition to recording the health status using the questionnaires mentioned above, an activity sensor (Axivity AX6) was used to objectify physical activities. Therefore, the sensor was attached to the front of the participants’ thighs for 7 days, a week before T0, T1, and T2. The sensor has 6 axes and records linear acceleration and angular velocities [44]. This allowed movement data to be recorded over 7 days and analyzed according to the World Health Organization (WHO) global guidelines on physical activities [45]. Using a cut point analysis from Skjødt, the sensor data were analyzed for minutes of movement and their intensity [46]. This metric objectifies the participant’s weekly exercise duration, categorizing sedentary, light, moderate, and intense activity.

Potential participants aged 70 years and over were recruited via our register of people who had previously expressed interest in participating in the AS-Tra project. Other recruitment strategies included the distribution of flyers in places frequently visited by older adults, such as physiotherapy practices, senior centers, and care support centers, as well as public relations work through press releases and newspaper articles. Interested individuals were requested to contact the study team by telephone or email. Subsequently, a consent form to be included in the registry, along with the MNA-SF and the inclusion and exclusion criteria, was sent by mail to those who expressed interest. Based on the initial data collected, it was possible to assess whether the individual is suitable for the study. This information was also recorded in the registry, which automatically created an order in which the subjects are requested for the study. If the individuals contacted were not interested, they were asked whether they no longer wished to be included in the directory or whether they may be contacted again at a later date for possible follow-up studies. Recruitment began one and a half weeks before the start of the study and continued throughout the study until 10 participants were included. The age group above 70 years was selected because it differs significantly from the mid-60s age group, many of whom are still working [47]. The risk of health impairments continues to increase in the over-70 population [48], and digital support, which plays an important role in prevention, is often limited or entirely absent [49,50]. In addition, several components from previous projects, including the app, were already tailored to older, fit users (aged 70 y and above) or specifically designed for geriatric and rehabilitation patients (eg, the aTUG chair) [17,21-24].

In this study, eligible participants were those who met at least one of the following inclusion criteria:

Exclusion criteria included (1) inability to provide informed consent, (2) physical inability to visit the study center, (3) inadequate understanding of the study process or the German language, (4) severe mobility impairments that prevent locomotion, and (5) severe visual impairments that preclude the independent use of the system.

All participants were informed about the study by a member of the study team. The procedure was described in detail and questions could be asked.

The pilot study was carried out from October 2024 to February 2025 in accordance with the study protocol approved by the medical ethical committee of the Carl von Ossietzky University Oldenburg, Germany (2024‐116). The AS-Tra system was evaluated over 4 weeks per participant. Participants used the MuTS on their own, not simultaneously with other participants.

Figure 2 shows a schematic overview of the study’s procedures, highlighting the 4 measurement points baseline (wk 0), T0 (wk 1), T1 (wk 2), and T2 (wk 4), as well as the corresponding assessments, questionnaires, and data collection periods (digital food diary activity sensor).

All appointments (baseline-T2) were accompanied by at least one trained study staff member to observe the use of the system and to record the observations in the form of an interview format using questionnaires.

At baseline, after checking the inclusion and exclusion criteria, sociodemographic and health-related data, the MNA-SF, and the SPPB were recorded, and the participants received a tablet with the app to use in their home environment. Furthermore, an Axivity sensor was attached to the front of the participant’s thigh.

At T0 (after 1 wk), the participants returned to the MuTS, and the use of all elements in the MuTS (see Figure 1) was explained. All assessments (listed in Figure 2) were carried out together with the study staff. For the following 3 weeks, participants were encouraged to use the MuTS independently for regular assessments and to test the tablet app at home. They were recommended to do exercises twice a week on the THERA-Trainer Senso.

At T1 (after 2 wk), all assessments were carried out again. At the last appointment (T2), after 4 weeks, the final measurement took place, including all assessments. In addition, questionnaires were collected to evaluate the SUS, technological commitment, and the Evaluation Overall System (EOS) questionnaire (see Multimedia Appendix 1). Moreover, the MNA-SF and the SPPB were repeated. In week 5, participants wore the activity sensor and used the digital food diary for 3 consecutive days, but they were no longer allowed to use the MuTS.

At T0, T1, and T2, participants completed the Experience Report (ER) questionnaire (see Multimedia Appendix 1), documenting the use of the MuTS and the tablet app.

As shown in Figure 2, the activity sensor was worn for 7 days in week 1, week 3, and week 5. For each of those weeks, participants were required to maintain a food diary for 3 consecutive days. This allowed the collection of nutritional and activity data from each participant at 3 points during the study.

All collected data, including sociodemographic information, medical history, MNA-SF, SPPB, and technology-related data such as technology commitment, SUS, ER, and EOS, were collected using a combination of paper and digital methods. For older adults, questionnaires were provided on paper if preferred, while most data were collected digitally through the REDCap (Research Electronic Data Capture; version 14.5.44) data management system by the study staff members [51,52]. The paper-based records were digitized by the study staff using a dual entry method.

Participants’ characteristics, the data of the MNA-SF, the SPPB, and the TUG were analyzed descriptively using SPSS (version 29.0.0). The data were presented as total numbers, percentages, and means with SDs or medians with IQRs depending on the distribution. The data of the ER and the EOS questionnaires were also evaluated descriptively and presented as total numbers and percentages.

The mean values, including SDs of the MNA-SF and SPPB scores of all participants, were graphically processed and color-coded to indicate the scores achieved in both questionnaires at the baseline and end of the study.

Data from the activity sensor were analyzed using the procedure described above and displayed graphically using a Python script. The number of minutes of moderate activity during the weeks the sensor was worn was visualized to show its progression over the 4-week study period. In addition, these values were compared to the recommendation of the WHO guideline value of 150 minutes of moderate activity per week [45].

The study was approved by the Ethical Review Board of the Carl von Ossietzky Universität Oldenburg (2024‐116). All participants were informed of the aim of the research, the methods used, how the results would be presented, and the right to decline to participate in the study or to withdraw consent to participate at any time without further explanation. All participants provided written informed consent before data collection. All data were stored pseudonymously, which means that without a key list, it is not possible to link personal data to research data. Only study staff have access to the key list. All study participants received an expense allowance of €30 (US $35.28). If they did not participate in the study until the end, they received €5 (US $5.88) instead.

A total of 10 older adults (5 females and 5 males) were included in the pilot study. One participant dropped out after T0, after 1 week of participation due to illness, so that data from all 10 participants at baseline and 9 at T2 were included in the final analysis.

In total, there are complete datasets from 6 participants, that is, excluding the 67% dropout rate. For the remaining 3 participants, datasets from the food diary are unavailable. In one case, the diary was not filled out at all, and in the other 2 cases, it was only filled out half as often as desired. Otherwise, all data are available. In addition, all participants adhered to the MuTS usage recommendations, were present at least once a week, regularly performed the assessments, exercised, and wore the activity sensor, as shown in Figure 2.

The average age was 80 (SD 5) years, with an average BMI of 23.7 (SD 5) kg/m² among female participants and 25.0 (SD 6.3) kg/m² among male participants. The baseline characteristics and T2 data of MNA-SF and SPPB, as well as the SUS and technology commitment score, are shown in Table 1. The SUS score at T2 was good (mean 79.3, SD 13.4 points). The mean score of technology commitment, assessed at T2, was 42.2 (SD 7.6) points.

The results of the MNA-SF and SPPB recorded at times baseline and T2 of all participants are shown in Figure 3 as mean values and SDs. The color-code indicates whether malnutrition, a risk of malnutrition, or a normal nutritional status were present in the MNA-SF and whether poor mobility (0‐4 points), moderate mobility (5‐8 points), or good mobility (9‐12 points) were achieved in the SPPB. Descriptive changes in the results of these 2 questionnaires are visible at a glance.

Data from the activity sensor showed the total number of minutes recorded during the week in which the sensor was worn, including the number of minutes spent sitting or lying down. The recording duration of the Axivity sensor demonstrated high data completeness at all times. Moreover, the number of minutes of low, moderate, and high activities during that week at the time points baseline, T1, and T2 were measured in which the time spent on moderate activity is of particular interest in this study (see Table 1). Therefore, the graph (Figure 4) showing the number of minutes of moderate activity performed by a participant at the specified time points also includes the threshold value of 150 minutes per week recommended by the WHO guideline [45]. This threshold is represented by the horizontal red line for reference, making it possible to determine whether this value was achieved in one or more measurements or not. Overall, 7 participants exceeded the 150-minute threshold value in all 3 Axivity measurements during the 4-week study. In 1 out of 9 cases, this threshold value was exceeded in 2 out of 3 measurements, and 1 participant exceeded the 150 minutes once. Only 1 participant remained below the limit in every Axivity measurement.

The SUS score was good (mean 79.3, SD 13.4 points). According to the ER of the participants and their EOS (see Multimedia Appendix 1), only minor problems were encountered during the use of the MuTS. Some technical problems occurred, particularly regarding the connection between devices in the MuTS, force plates that did not calibrate correctly automatically, the training plan that did not display correctly, or occasionally the touch screen that did not respond. All of these problems occurred at maximum in 17% of MuTS sessions.

The food diary was to be maintained 3 times per participant during the 4-week study. This resulted in a total of 30 uses for 10 participants. Due to 1 dropout, there was a total of 28 uses. In 22 of the 28 cases, entering data into the food diary led to problems and errors, and in 17 cases, it resulted in crashes, which corresponds to 57% (17/30) of the cases in which users attempted to enter nutritional data. Only in 4 of the 28 cases did no problems occur. This caused significant frustration for users, who were often unable to make further entries until they restarted the app.

The rating of the EOS with regard to individual points, such as content, handling, training program, process, and use, was generally between 1 and 2 on a scale of 1 to 5, where 1 corresponds to “positive” and 5 to “negative.” With regard to the tablet app and regular visits and use, the answers tended to be between 2 and 3. All exact mean values including SDs for the individual questions, calculated by scoring all 9 participants, can be seen in Table 2. Averaged across all questions of the EOS, the system was rated 2.01 (SD 0.99).

Overall, the results show good usability (SUS, mean 79.3, SD 13.4) and also good feasibility, as only one of the 10 participants dropped out and the remaining 9 provided complete datasets. All participants adhered to the usage recommendations during the study. The only problems were with the tablet-based food diary, which crashed in 57% of cases, resulting in incomplete dietary data for 3 participants. This is also reflected in the EOS (see Table 2).

The aim of this study was to evaluate the usability for older adults (≥70 y), which can be classified as good (SUS, mean 79.3, SD 13.4), and to test the feasibility of data collection in preparation for a future RCT study, which was successful. Moreover, feedback on the use of the entire system was positive (EOS mean 2.01, SD 0.99; Table 2). The Axivity AX6 is a reliable wearable device for objectively recording activity data and classifying it into different intensity levels. Nutritional data were mainly recorded using the MNA-SF and the food diary in the tablet app, with the technical component of the tablet app proving to be the biggest problem in this study. This led to many crashes and instabilities in the app. This must be fixed before the RCT, as otherwise, there is little added value from the diary entries if the technical implementation causes several problems.

All of these main findings are discussed in more detail further, classified with the help of other studies, and interpreted.

The requirements of older adults for support services vary greatly and range from web-based solutions to personal counseling and exercise services [53]. Studies show that personal interaction and contact, for example, when using exergames, play a central role in the well-being of older people [54]. This suggests that independent exercises or measurements, where participants used the system alone, may have a more negative effect on usability compared to measurements carried out together with the study team. Previous studies with older adults showed similar or slightly higher usability scores for mostly app-based systems and, on average, younger participants [55-57]. All referenced studies included older adults aged 65 years and above, whereas in our study, the minimum age was 70 years. The average age in other studies was around the mid-70s, which is approximately 5 years below the average age in our study [55-57]. Their systems mainly focused on a specific health area [55-57]. In contrast, the AS-Tra system is a multidisciplinary, complex intervention that covers both mobility and nutrition and also includes the independent use of measuring and training devices. Due to this breadth of content and functionality, demands on the system design increase to ensure ease of use.

With regard to the feasibility of data collection, which was one of the objectives of this study, the questionnaires used proved to be meaningful and suitable for assessing the health status of older adults. The SPPB is a good predictor of physical performance and mobility in older adults, as it measures balance, walking speed, and lower extremity strength [42]. In addition, the participants in this pilot study regularly wore an activity sensor, which allowed their activity to be recorded objectively (see Figure 4). Such wearables have also been used in other studies to track physical activities [15,16]. The WHO guidelines [45] define whether a person gets enough exercise, which can be compared with the objectively measured amount of activity recorded using the activity sensor. On average, participants achieved the recommended 150 minutes of moderate activity per week (7 out of 10 participants did exceed this limit during the 4-week study period). The high SD also clearly shows that there are very large differences between individuals in terms of weekly activity (see Table 1). The observation that the achievement of the level of moderate activity recommended by the WHO can vary greatly from person to person is also reflected in the study by Jefferis et al [58]. In this cohort study, accelerometer data were collected using a GT3x accelerometer from a total of 2450 older adults aged between 70 and 93 years, of whom only 25% (n=613) achieved the recommended 150 minutes of moderate activity per week. Nevertheless, the Axivity AX6 proved to be a reliable wearable in this pilot study as in the study by Powell et al [59].

Data on nutrition and thus data collection using the MNA-SF and a tablet-based food diary could also be recorded and evaluated. As a validated screening tool, the MNA-SF is very well suited for testing older adults for the risk of malnutrition [33]. With regard to the food diary, the EOS clearly showed that using the tablet app and entering data into the digital food diary caused the most difficulties (see Table 2). The participants stated that the tablet app often crashed or took a very long time when entering food into the food diary, and that some foods could not be found or were not sufficiently specified. In addition to the technical challenges that still need to be improved to make the tablet app run smoothly, the use of such apps presented challenges for older adults. Using the app can be difficult simply due to limitations in physiological functions, which make tapping buttons or scrolling difficult. In addition, the cognitive load can be very high for older adults, as they have to keep several navigation steps in mind to make entries in the diary [12,13].

In previous iterative studies, the app design was already adapted to the needs of older adults using a user-centered design [8,12,22,53]. In this study, care was taken to provide detailed instructions on how to use the tablet in order to reduce cognitive load and provide participants with the best possible support, as many older adults do not have much experience using apps on mobile devices [21,22]. The review by Zager Kocjan et al [19] also showed that it is important to closely involve older adults in the development of the assistance system or app to counteract the lack of trust in technology. In this study, participants were therefore closely supervised and were given printed instructions for the tablet app and for entering data in the food diary to take home with them. Nevertheless, it seemed that the technical component of the tablet app was the biggest problem in the pilot study, not the execution of the entries themselves. This must also be examined in the subsequent RCT study once the tablet app has been developed further, so that there are no more crashes or delays in the entries, as food diaries also offer older adults the potential to sharpen their focus on healthy and possibly better nutrition [13].

The comparison between 2 groups, one of which uses the AS-Tra system over a longer period of 12 weeks and the other which does not, will be recorded in a planned RCT study. This study started in May 2025 and is scheduled to run for 2 years. These results will allow a more detailed examination of the effect of the complex intervention used in this study.

Nevertheless, the successful completion of the pilot study suggests that the system is likely to have a positive impact on optimizing the mobility and nutritional status of older adults. Therefore, the AS-Tra system appears to be very promising in terms of preventive and independent measures for healthy aging in older adults.

Certain limitations of this study must be taken into account. First, selection bias cannot be ruled out as it must be assumed that participants who register for the study already have a general interest in good mobility and healthy nutrition. This also applies if older adults must have a deficit in mobility and/or nutrition to be included in the study.

Second, the assessment of the exclusion criterion regarding whether participants were able to understand the study content is based on the subjective assessment of the study team. However, this assessment is supported by the objective assessment of health status using the questionnaires and tests (MNA-SF and SPPB).

Third, it should be noted that, compared to measuring mobility using automated assessments and mobile devices, assessing nutritional status is fundamentally a subjective process, as it relies solely on self-reported information using tools such as food diaries [13,60]. Although methods such as image analysis tools, photo-based diaries, and chemical sensors are increasingly available for assessing nutrition, their application in everyday life is complex and subject to limitations in terms of feasibility, data protection, and ethical concerns [13,61]. The MNA is therefore currently considered a guideline and is recommended as a survey method despite its limitations due to subjectivity. That is why researchers rely on nonautomated, subjective information, which may be distorted by memory gaps or misreporting of, for example, portion sizes [18].

Fourth, the study was neither designed nor powered to prove the positive trends observed in health status. Therefore, they should be interpreted with extreme caution and are not generalizable to the general population.

Finally, the data were collected by the members of the study team who were themselves involved in the development of the AS-Tra system, which means that the same level of objectivity cannot be guaranteed as if the study had been conducted by an uninvolved person.

Future studies could focus on the objective collection of dietary data to prevent the tablet app from crashing frequently when using the food diary. In addition, the subsequent RCT can be used to examine the effectiveness of the AS-Tra system more closely, which could provide insightful findings on healthy aging in older adults.

The main objectives of this pilot study were to test the usability and feasibility of collecting behavioral and health data using an MuTS, along with the associated app, in older adults (≥70 y) with deficits in mobility and/or nutrition over a period of 4 weeks in order to gain insights for a subsequent randomized controlled study.

Overall, the usability of the system was rated as good (SUS, mean 79.3, SD 13.4). However, the varying support needs of older people and the importance of personal interaction can influence the acceptance of the technical assistance system. Compared to simpler app-based systems, AS-Tra has a broader focus on mobility and nutrition and therefore has high requirements on design and user-friendliness, as well as active training opportunities and regular assessments to ensure that older adults can use it independently.

Data collection proved feasible using questionnaires, activity sensors, and automated assessments. The tablet-based food diary caused most of the difficulties due to technical problems (in 57% [17/30] of cases where participants wanted to make entries, crashes occurred). Despite these obstacles, participants rated the system and the overall assessment of its use positively. Improvements to the tablet app, particularly the food diary, will be essential for the RCT, especially since the report on the nutritional status is based mainly on subjective self-reports compared to automated mobility measurements.

The evaluation of potential health benefits from using the AS-Tra system will be investigated in a 12-week RCT, which started in May 2025.