Ankylosing spondylitis (AS) is one of the most severe rheumatic diseases [1]. It affects the axial skeleton and, in some cases, the peripheral joints, entheses, skin, eyes, and gastrointestinal organs. Inflammation can especially affect the lower spine and cause stiffness, leading to immobility. These symptoms can interfere with daily activities and significantly impact people’s quality of life [2]. AS involves people of various ages. About 80% of the patients develop the first symptoms at an age younger than 30 years, but the disease duration is increasing [3]. Due to its early onset, severity, and high-level chronicity, AS is associated with a substantial burden on the individual and the health care system in Germany [4]. For this reason, it is important to start early treatment to avoid progressions such as chronic pain, bone damage, and ankylosis. Nowadays, a combination of medical treatments in the form of biologicals and regular exercises is efficient and safe to reduce inflammation and receive remission, which leads to a reduction of work disability and sick leaves (from 2000 to 2012, early retirement rates decreased from 19% to 14%) [4].

Besides drug treatments, physical therapy is important to improve well-being and prevent stiffness in patients with AS [5]. Yoga, a combination of stretching and strengthening exercises, has been accepted as a popular treatment for rheumatic and musculoskeletal diseases [6]. The combination of drugs and yoga has also shown better results than treatment with drugs only [7].

Over the years, there has been considerable growth in the use of mobile eHealth apps as supporting treatment for rheumatic diseases such as rheumatoid arthritis or systemic lupus erythematosus [8,9]. Particularly during the COVID-19 pandemic, mobile eHealth apps have been used as a complementary treatment to increase the well-being of patients [10]. Especially in the field of fitness apps and exercise therapy, there has been an increase in recent years [11]. Digital apps can be beneficial in preventing a bad course of the disease and could be a cost-efficient solution in helping health insurance companies save costs [12]. It also creates more time for patients, as exercises can be performed directly from home [13]. The aim is to save as much time as possible, as appointments with a physiotherapist are often difficult to get or specialized clinics are far away. Therefore, many patients look on the internet for ways to stay active. However, there is often a lack of specific programs adapted to the disease. From our experience, many patients also feel safe knowing that experts have put the exercises together, for example, when the exercise program is recommended by their physicians.

The problem with many fitness apps is the 1-sided benefit. Often only exercises can be performed, but the patient cannot assess the effect of the exercises on the course of the disease [12]. YogiTherapy tries to combine these 2 aspects. The mobile app was developed in 2020 to provide a training section with video instructions for yoga exercises specified for patients with AS and an assessment section giving the ability to monitor the progression of the disease [14]. In the assessment section, patients can conduct tests related to disease-specific mobility, function, disease activity, and quality of life. The results are then displayed graphically on a timeline to give the patient automatic feedback. The app also provides information with pictures on individual exercises, about the disease itself, and nutrition recommendations for rheumatic diseases.

The objective of this work was to improve the existing prototype of YogiTherapy and evaluate the app. In detail, we reimplemented YogiTherapy as a cross-platform app with the open-source tools React Native [15] and Expo [16] to create a modern design and intuitive navigation and to extend its functionalities. We included the option to change the language between English and German, with the possibility to add other languages easily. Furthermore, we added search and filtering functions to the training section and enabled exporting of the test results. With the refined app, we performed a study with patients with AS to evaluate the usability and quality. The study participants first performed a task performance test (TPT) with a think-aloud protocol and completed the German version of the Mobile App Rating Scale (MARS-G) afterward. Finally, we improved YogiTherapy based on the study and published it in the iOS App [17] and Google Play stores [18].

A total of 16 participants took part in the TPT, but 4 participants (IDs 3, 7, 10, and 13) did not complete the MARS questionnaire. Table 2 summarizes the task completions of the TPT. The overall completion (84/96, 88% completed tasks) was high. A total of 9 participants completed all tasks successfully; 4 participants failed only 1 of the 6 tasks, 1 participant failed 2 tasks, and 2 participants failed half (3/6, 50%) of the tests. Task 1 (create an account) and task 2 (watch a yoga video) had a completion rate of 100% (16/16). Task 1 (create an account) was easy to complete but required, on average, 65.1 seconds to complete, which was the second-highest completion time (Figure 2). Task 4 (take a test) had the lowest completion rate (11/16, 69%) and also the highest average completion time of 73.3 seconds. Task 6 (share your progress) needed, on average, the shortest time to be completed (14.1 s).

In the think-aloud protocol, difficulties navigating the app became apparent. Especially, navigating back to a previous page could have been more intuitive for most participants, as they were unfamiliar with the double-click method. One user reclaimed the size of the text as it was too small to read. In addition, many participants could not find the filter icon since some entered the word “filter” into the search bar or did not know the filter icon. Furthermore, some participants had issues finding the assessment section, which we labeled with the word “Bewertung” in German. Even though all participants could start a yoga video (task 2), some had initial problems starting the video as they tried to click into the pictures of the associated asanas below the video.

For the MARS, the engagement, functionality, aesthetics, and information mean scores were good [21], resulting in a good app quality mean score of 3.79 (SD 0.51; Table 3). Compared to the app quality mean score, the app subjective quality score (3.52, SD 0.69) was slightly smaller, as two-thirds (8/12, 67%) of the participants were unsure whether they would pay for the app (item XXVI). Nevertheless, they would recommend the app to others (item XXIV) and use it frequently (item XXV). The evaluation also showed that YogiTherapy is interesting to use (item II); is target specific (item V); and has good performance (item VI), layout (item X), and graphics (item XI). However, customization possibilities (item III) could be improved.

This study aimed to improve YogiTherapy as a mobile eHealth app that can serve as an additional non–drug-based treatment for patients with AS. YogiTherapy provides yoga videos tailored to patients with AS and allows tracking of disease progression with standardized assessment tests. Overall, the app’s quality was good, which was demonstrated by a high task completion rate in the TPT and a high score in the MARS questionnaire.

In the TPT, the participants could complete most tasks without any previous explanation, resulting in an overall performance rate of 88% (84/96 completed tasks). Hence, the app use can be considered intuitive. Participants with the largest difficulties who could not complete 2 or more tasks were 60 years of age or older. Therefore, the differences depending on age might indicate that YogiTherapy might be harder to operate for less experienced users. However, we did not record how experienced the participants were with using smartphones and apps. This should be taken into consideration in follow-up studies.

To further improve the usability of YogiTherapy based on the study results, we analyzed task 3 (filter videos by properties) and task 4 (take a test) in more detail since they had the lowest success rates. The think-aloud protocol revealed that even participants who completed all tasks successfully had difficulties recognizing or finding the filter icon in task 3. Hence, we increased the size of the filter icon and added the label “Filter” as an explanation below the icon (Figure 1D). Furthermore, participants had issues finding the tests in task 4 since they did not associate the German translation of “assessment” (Bewertung) with the tests (Figure 1C). Based on this feedback, we renamed the section in the German translation “Test/Fortschritt.” This case demonstrates the importance of proper translations and correct wording.

The TPT was also performed on a previous version of the YogiTherapy app [14], but we used new tasks to also cover the added functionalities of the revised version and did not allow the participants to familiarize themselves with the app before the test. Nevertheless, 3 of the tasks were comparable between the studies. In both studies, all participants could successfully watch a yoga exercise video (task 1 in the previous study and task 2 in this study). However, in the previous version, the participants could start an arbitrary video while they had to search and start a specific video in this study. Hence, the level of difficulty of this task was higher in this study. The BASDAI test could be completed by 90% of the participants in the previous study, which was conducted in English, and only by 69% (11/16) of this study, which was conducted in German (task 5 in the previous study and task 4 in this study). The decrease in the completion rate of this task was most probably caused by the misleading German translation of “assessment,” as discussed earlier. In the previous study, 80% of the participants could visualize the results of the BASMI test (task 4). After the refinement of the app, the task completion rate slightly increased, as 88% (14/16) of the participants could visualize the results of the BASDAI test in this study (task 5).

The results from the MARS showed that every section scored above average from the participants’ point of view. YogiTherapy had an app quality mean score of 3.79 (SD 0.51) and thus exceeded the average score of 3.17 (SD 0.75) of 18 pain-related mobile apps [27] and the average score of 3.48 (SD 0.28) of 60 chronic condition or multimorbidity apps [28]. This implies that the app is readily acceptable and can be used by intended patients with AS [29].

The graphics item XI had one of the highest scores (4.08, SD 0.76), which can be attributed to the well-thought-out graphics used in developing YogiTherapy. We used representative images, and the layout was consistent and straightforward. The functionality mean score (3.92, SD 0.55) was the second-highest section score of the MARS. The good results in the navigation item VIII contradict the findings of the TPT, where several participants had issues navigating the app. However, this contradiction can be clarified because the study conductor explained to the participants how to navigate after the TPT. While this is a restriction of the presented study, it highlights that participants understood how the navigation works after explanation. We built the design so that the previous screen could be accessed with a finger wipe or a double-click, so the participants thought it was efficient and adequate. After the study, we added an explanation of the navigation to the welcome page to introduce this functionality. The performance item VI obtained the highest item score as we focused on the performance during implementation. For example, we bundled the pictures of the asanas into the app. Except for functionalities like watching yoga videos and sharing results, all functionalities work without an internet connection, providing less latency delay and a more pleasant experience for users.

The engagement section of the MARS gives important insights since YogiTherapy aims to be a target-specific app for patients with AS. Most of the study participants agreed that the YogiTherapy is target-specific, as the target item V reached a score of 3.92 (SD 0.49). On the other side, more customization features, such as sound or content notifications, could be added to YogiTherapy. The participants also indicated that the interactivity (item IV) of YogiTherapy is just about average. However, the participants are highly interested in the project and are very willing to use YogiTherapy as a complementary solution to their challenge.

Compared to other fitness apps (eg, Daily Yoga [30]), YogiTherapy was developed specifically for patients with AS by doctors, computer scientists, and medical professionals. Nevertheless, the participants doubted the credibility (item XVIII) of the information provided in the app. The low score could originate from most participants not seeing the well-referenced information section before participating in the MARS. Credibility could be increased by the treating physician recommending YogiTherapy.

The MARS of our app was in the middle of 9 mobile apps for patients with rheumatic diseases, for which the app quality mean scores ranged from 3.44 to 4.19 [12]. The app Rheuma Auszeit was developed with a patient organization and reached the highest MARS. Rheuma Auszeit was the only 1 of the 9 apps providing instructions for mental and physical exercises. However, it did not provide a functionality to track the disease progression [12]. Like our app, Rheuma Auszeit does not allow customization, which also led to a low score for this item (Rheuma Auszeit: 2.5; YogiTherapy: 2.9) [31]. The other 8 apps featured progression tracking, but only 3 of them used validated questionnaires, which are established in clinical practice [12]. In contrast to those apps, YogiTherapy fully accompanies patients by giving exercise instructions and allowing the disease status to be assessed with validated tests. Nevertheless, YogiTherapy could further benefit from customization functionality and patient input.

Despite the promising results, the presented usability study had several limitations. The most important limitation lies in the small study population of 16 participants for the TPT and 12 for the MARS recruited in a single hospital due to time restrictions. In addition to patients, who are the end users, domain experts, health care professionals, and physiotherapists should be included in the study population. Furthermore, participant’s typical physical activity and experience with digital technologies, especially apps in the fitness sector, should be assessed. Even though the MARS is widely established, the mHealth App Usability Questionnaire (MAUQ) [32] could be a valuable alternative, as it was designed to determine the usability of mobile health apps [22].

In conclusion, the refined YogiTherapy app exhibited promising outcomes in usability testing among most participants. In the future, additional research should also focus on how patients use YogiTherapy daily and evaluate whether the app can motivate them to be physically more active. Moreover, there remain small improvement possibilities in the implementation to further enhance the app’s intuitiveness. Overall, we demonstrated that YogiTherapy is a user-friendly app specified for patients with AS, providing yoga videos for physical exercise and disease-specific tests to monitor disease progression. We made YogiTherapy accessible in the iOS App [17] and Google Play stores [18] to make it widely available to patients.