Anterior cruciate ligament (ACL) injury can lead to pain, knee instability, and variable disability level, from reduced participation in sports to more severe limitations in activities of daily living [1,2]. ACL reconstruction (ACL-R) is the sixth most common orthopedic surgical procedure in the United States [3]. The goals for ACL-R include restored knee stability and return to prior activities. Recent studies evaluating insurance databases have noted that pediatric and adolescent athletes (young athletes) are the largest demography of patients per capita undergoing ACL-R [4,5]. Young athletes also demonstrate higher reinjury rates following ACL-R, partly because of their high level of functional demands following surgery. Biomechanical studies note that persistent muscular weakness or dysfunction is a significant risk factor for reinjury after the patient has returned to preprocedure activities and that postoperative rehabilitation is critical in reducing these negative outcomes [6,7].

Rehabilitation—specifically restoration of motion, optimizing muscular strength, and improving neuromuscular control—is demonstrably important in optimizing the quality of life of young athletes and function following ACL-R [6-8]. Adequate rehabilitation requires patients to complete prescribed exercises daily, primarily in the form of home exercise programs (HEPs) [9]. Unfortunately, due to lack of education, communication, and oversight, many patients do not fully understand or appreciate the importance of their HEP [10]. Particularly, in a young athlete population, low rehabilitation adherence is a common and critical obstacle in ACL-R recovery [8,11]. Inadequate rehabilitation can result in poorer clinical outcomes, including reduced quality of life, increased pain, and increased chances of reinjury [12-14]. Commonly reported barriers to rehabilitation adherence in the ACL-R population include fear of pain or reinjury, low self-motivation (self-efficacy) or exercise level, and a feeling of low patient control have been cited as common barriers faced by patients following ACL-R [13].

In response to these obstacles, our group developed an interactive health technology (IHT) called interACTION (iA; Elizur), to prescribe and monitor home exercises for knee-specific rehabilitation. iA is a web-based rehabilitation platform that incorporates wearable motion sensors with a mobile device app, using real-time motion tracking to provide exercise biofeedback to the patient and log HEP performance metrics for remote clinicians (Figure 1). Through these mechanisms, iA helps to bridge the gaps in current physical therapy practices by allowing remote monitoring of HEP compliance and performance. iA has been evaluated in both healthy controls and adult patients with total knee replacement, which established a proof of concept of the technology [15].

Similar to many IHTs, iA in its current form is not grounded in models of behavioral change that may translate to a prolonged influence on patient behavior. IHTs founded on theoretical models of behavioral intervention, supported by technology, have been shown to promote exercise adherence more effectively [16-18]. The Wheel of Sukr is a proposed mechanism to introduce behavioral intervention concepts into IHTs. The Wheel of Sukr (Figure 2), which was initially described in its application to diabetes self-management, provides a structure for the incorporation of numerous behavioral elements (ie, motivation, self-management, fun, and sustainability) into a game-like experience of an IHT [19]. This process of IHT gamification introduces behavioral intervention concepts through techniques typically used in game development [20]. It seeks to insert concepts such as engagement, reward, and incentive into the IHT platform to better engage patients and affect behavioral change [21]. Adolescent athletes typically note prolonged (6-12 months) recovery following ACL-R as a hurdle to rehabilitation adherence, so the focus of Wheel of Sukr on sustainability was felt to be of relevance to this patient cohort [11].

The objective of this study is to use a user-centered design process to incorporate behavioral change strategies derived from self-management theory into the iA platform using the Wheel of Sukr gamification framework [19]. With the implementation of gamification on the iA platform, we hope to better engage young athlete ACL-R patients to improve their adherence to postoperative rehabilitation programs and subsequently improve their recovery. Self-reported questionnaires focus on patient adherence and barriers to adherence are also recorded to help characterize this patient population and to inform the development of behavioral change strategies.

Young athletes aged 10-18 years with a history of ACL injury and surgical reconstruction were included in this study. Patients were identified for study inclusion by reviewing the electronic medical records at our institution. All patients were between 4 weeks and 1 year from the ACL-R before their consideration for study participation. Concomitant procedures, such as meniscal repairs and chondroplasties, performed at the time of ACL-R did not exclude participation but patients undergoing multiple-ligament reconstructions were excluded. Eligible patients were contacted by a research coordinator, at which time a thorough description of the study was presented, and informed consent was obtained from the patient or guardian.

After the participants were enrolled and informed consent was obtained, their age, sex, site of surgery (right or left), surgical procedures performed (including concomitant procedures), and date of surgery were recorded from the electronic medical record. On their interview date, participants were presented with a set of self-reported questionnaires, which were completed by either the participant or their proxy (parent or legal guardian). The questionnaire included demographic questions as well as the Exercise Adherence Rating Scale 10 items (EARS-10), General Self-Efficacy (GSE) scale, Tampa Scale for Kinesiophobia (TSK), and Numeric Pain Rating Scale (NPRS). The EARS-10 assesses adherence to prescribed HEPs by asking participants to answer questions about their exercise adherence and attitudes toward exercise.

GSE measures a participant’s perceived self-efficacy. Scores predict the self-perceptions of self-efficacy. The participants rated their perceived self-efficacy, and higher self-efficacy was correlated with emotion, optimism, and better health outcomes. The TSK assesses fear of movement, fear of physical activity, and fear avoidance. A score <37 is classified as low fear, and a score >37 is classified as high fear [22].

Descriptive statistics were used to summarize patient demographics and the results for the EARS-10 GSE, TSK, and NPRS. The Pearson correlation coefficient (r) was calculated between these patient-reported outcome measures and was interpreted as follows: 0-0.1, negligible; 0.1-0.3, weak; 0.3-0.7, moderate; 0.7-1.0, strong .

The patient interviews were performed in private rooms in a clinical research suite, which served as a controlled environment to simulate outpatient medical encounters. The participants took part in a single individual interview with three distinct phases, as described below. The audio of participants was recorded throughout the entirety of their interviews.

Phase 1 followed a standardized, semistructured script, with open-ended questions designed to facilitate conversation with a specific focus on: (1) identifying critical elements of existing gaming experiences (ie, sports, video games) that lead to high levels of participation (adherence) and (2) eliciting experiential information regarding their knee surgery and subsequent physical therapy.

In phase 2, participants were presented with a series of cards, each presenting an aspect of game play designed to promote continued interest and participation. Each card displayed an individual component and a description of the Wheel of Sukr methodology. The language was altered to the appropriate reading level of the participants. The participants were prompted to think aloud as they reviewed these cards and were prompted to rank the cards in order of importance in maintaining interest in a game. The final card rankings of the participants were recorded. The card sort was followed by a brief semistructured survey to elucidate the motivations behind the card sort results.

Finally, for phase 3, participants were presented with the current version of the iA. Participants were again prompted to think aloud as they had experienced the device. The participants were then presented with a brief semistructured survey to determine their overall opinion of the current device, elicit suggestions for improvement, and identify any gamification changes that should be implemented.

Phase 1 of the interviews was performed by an experienced facilitator with extensive background in qualitative research methods from the Qualitative Research Core at the University of Pittsburgh Center for Research on Healthcare Data. Phases 2 and 3 of the interviews were performed by a study research assistant, as they did not require the same level of experience with qualitative interview facilitation as the first phase. The interviews were audio recorded, transcribed, checked for accuracy, and transferred to ATLAS.ti (ATLAS.ti Scientific Software Development, GmBH) for coding.

A qualitative codebook was developed by the phase 1 interviewer using an inductive approach and shared with the study team to ensure that it captured key concepts that were of interest to the investigators. Once the codebook was finalized, 2 Qualitative Research Core coders used ATLAS.ti to independently code the full data set and to reconcile any coding differences. The interviewer or primary coder then conducted a content analysis using the coded data [23]. The resulting content analysis was confirmed by the secondary coder in the form of investigator triangulation.

The research team assisted in developing the best ways to display, represent, interpret, triangulate, and summarize the qualitative findings to achieve a fully integrated description [24]. These elements were categorized according to the Wheel of Sukr framework and translated into elements relevant to improving the design of iA to increase ACL-R patient engagement.

A total of 10 interviews were conducted during the fall of 2019. Participants ranged in age from 11 to 15 years (mean 12.8, SD 1.32 years). Participant demographics are presented in Table 1. All 10 participants completed all three phases of the interviews.

Results of the EARS-10, GSE, and TSK questionnaires are presented in Table 2. The median value for the EARS-10 questionnaire was 26, with values ranging from 6 to 36. The median self-efficacy rating was 31.5, participants reported values ranging from 21 to 40, with 40 being the maximum value on the scale. Participants were divided evenly between low and high fear on the kinesiophobia scale. The median rating was 37.5 with ratings ranging from 19 to 44.

NPRS scores showed moderate rates of postoperative pain, which improved with the start of physical therapy and study enrollment (Table 3). The median pain rating postoperatively was 4, with postoperative pain ranging from 0 to 7. Pain at the start of physical therapy ranged from 0 to 6, with a median value of 2. Pain in the 24 hours before the interview was low, with a median score of 0 and a range from 0 to 2.

Pearson correlation coefficients were calculated for patient-reported outcomes. The correlation between EARS-10 and GSE was 0.79, indicating a strong correlation. Another strong correlation was found between the NPRS surgery outcome and the TSK, with a value of 0.74. Both these correlations had a significant P value. All other P values were not significant. The r value for NPRS at the beginning of PT and EARS-10 was calculated as −0.46, a moderate correlation (Table 4).

In the Wheel of Sukr card sort exercise, the collective top three themes were motivation, self-management, and esteem (when using the median). When using the mean, motivation was first; self-management was second; and growth, esteem, and fun tied for the third. The top three themes of each participant are presented in Table 5, along with their reported favorite games, and the reasons and motivations for playing their favorite game. We presented the data individually by participants because the participants generally did not describe the cards any further than what it already said on the cards themselves—that is, most did not provide information on why they liked a concept beyond how it was already described on the card, and thus it was not possible to derive themes from their descriptions.

When talking about what elements or features from the cards the participants would like to see in a game, many referenced their top three. Few participants mentioned a feature of a card, not in the top three. Textbox 1 lists the example recommendations for each card theme provided by the participants.

This study is the first in a stepwise, user-centered design process to integrate behavioral theory concepts into the design of iA, a mobile IHT to prescribe and monitor home exercises for knee-specific rehabilitation following ACL-R in young athletes. Participants reported that their favorite games were sports, with many participants reporting playing multiple sports and most participants becoming injured through their sports. Participants generally reported positive experiences with PT following ACL-R, yet most thought iA and its remote monitoring would have been a helpful addition to their overall PT experience and recovery.

When considering gameplay and mobile apps, the Wheel of Sukr domains of motivation, self-management, and confidence were deemed of highest importance, by median. The top three domains were motivation, self-management, and growth, esteem, and fun tied for third. Participants’ impressions of a device that tracked their knee PT exercises were overwhelmingly positive. The participants want the app to include a feature that tracks their exercise movement for accuracy and to include features to encourage or motivate them.

On the basis of these interviews, future versions of iA will strive to incorporate the following components of gameplay:

Having a connection to favorite games or sports would help establish interest in the app from a preexisting interest. Incorporation of sport-specific activities later in the rehabilitation process may be helpful to promote longitudinal usage of the app:

The participants wanted to know that they were doing their home exercises correctly. This feature is already present in iA, through real-time biofeedback provided by the wireless sensors, but in its current form, it acts as a simple counting mechanism with no gamified structure. By improving this interface, the app can show how their knees are moving and provide more motivational feedback during workouts. The participants also wanted a better sense of how their recovery progressed. Incorporating performance thresholds or milestones and rewards, allowing users to interact with peers via avatars, and integrating regular feedback from their remote clinicians will be explored in future studies. This would improve self-management capabilities and provide extrinsic sources of motivation.

A strong significant correlation was calculated between the EARS-10 survey measuring exercise adherence and the GSE survey measuring self-efficacy. This strong correlation indicates that individuals with low self-efficacy are more likely to have lower adherence to HEPs. As the average score of the GSE was 31, with a value of 40 being the maximum score and indicating the highest level of self-efficacy, the participants interviewed overall had a fairly high self-efficacy score. Confidence and self-management were both highly rated components of the Wheel of Sukr and ideas related to a high level of self-efficacy. Designing the next iteration of iA and taking into account the highly ranked ideals of confidence and self-management can keep those with high self-efficacy adherent to their maximum potential and engage young athletes with naturally lower self-efficacy.

A strong significant correlation between a high NPRS value postoperatively and a high rating on the TSK was calculated, suggesting that higher pain postsurgery is related to a fear of movement. This finding is consistent with those of previous studies on chronic musculoskeletal pain and kinesiophobia using the Tampa Scale [25]. Goal and incentive-based feedback, both intrinsic and extrinsic, will be implemented in future iterations of iA to keep users engaged in the early aspects of their rehabilitation despite high levels of pain. Starting the rehabilitation process out quickly postoperatively and having patients adhere to their HEP over time should decrease pain levels and kinesiophobia as the patient progresses through treatment, as patients will gain more confidence as exercises and movement become easier and less painful. In addition, direct communication from remote clinician monitoring progress can ease patient hesitation and build confidence by answering questions and providing encouragement during the rehabilitation process.

A moderate negative correlation was found between the EARS-10 and NPRS-PT ratings, indicating that higher pain during physical therapy is related to lower home exercise adherence. This correlation was not statistically significant; therefore, future work is needed to completely understand this relationship. The patients do not want to perform their assigned exercises on their own if they are in pain. Increasing patient adherence through motivational tools to draw users to the app daily will increase the frequency of home exercise performance and decrease pain over time as the knee begins to return to its previous level of function.

iA is an IHT designed to bridge the gap between patients and clinicians during rehabilitation for knee injuries and surgeries. The use of IHTs to promote health and manage illness has led to fundamental changes in health practices and transformed health care [26]. Qualitative studies indicate that patients have positive perceptions of IHTs and their benefits, including awareness of progress, feeling of being in control, empowered self-management, and improved patient-clinician communication [27-30]. IHTs founded on theoretical models of behavioral intervention supported by technology have been shown to promote exercise adherence more effectively than those devoid of behavioral modeling [16-18]. The Wheel of Sukr method was applied to iA as it is designed to affect behavioral changes for chronic conditions, and many adolescent athletes view their ACL-R rehabilitation as a long-term event and cite fatigue as a prime driver of noncompliance [11,21,31].

This study has several limitations. Given that qualitative studies emphasize exploration of participant perspectives in great depth, sample sizes tend to be small, potentially limiting generalizability. In addition, there may be bias in the sample, insofar as participants may have agreed to be interviewed because of a particular interest in the topic. However, thematic saturation was reached by the conclusion of our 10 interviews, suggesting that the sample was sufficient to obtain a representative range of perspectives and experiences. The interview responses were monitored by a quantitative data expert who determined when thematic saturation was achieved. Furthermore, our sample was diverse in terms of gender, age, and surgical procedures (concomitant procedures with ACL-R).

Gamification of iA was explored in this study to incorporate the behavioral theory concepts of gaming to improve engagement with and adherence to the iA rehabilitation platform, which can be applicable to other IHTs as well [19,21]. Moving forward, we will focus on improving the domains of motivation, self-management, and growth or esteem or fun to the iA user experience. In subsequent studies, young athlete ACL-R patients will be asked to complete typical clinical scenarios using progressively more developed prototypes of the gamified iA in a controlled laboratory setting. The usability of each gamification aspect and its ability to induce continued user engagement will be assessed objectively and subjectively. On the basis of these metrics, iA will undergo an iterative improvement process to address its current limitations.