Background: Physiotherapy in the clinic is highly recommended for improving gait, balance and falls risk in people with Parkinson's disease. In addition, technology may help boost unsupervised exercise hours at home. Strolll is an augmented-reality (AR) neurorehabilitation platform for delivering gait-and-balance exercises onto AR glasses that can be performed under direct supervision of the therapist in the clinic, but also independently at home. Strolll AR also has the option to integrate AR cueing in gait-and-balance exercises to assist people with more severe mobility impairments in performing the exercises. The objective of this pragmatic randomized controlled trial (RCT) on Strolll AR is to examine its clinical feasibility and effectiveness for improving indicators of gait, balance and falls risk. A secondary objective is to evaluate procedures for tailoring assistive AR cues.

Methods: 80-100 people with Parkinson's disease (Hoehn and Yahr stages 1-3) with gait and/or balance impairments will participate in this study. This study is a pragmatic RCT in which all participants follow the same procedure. After a baseline assessment (T0), participants will start with a 6-week usual care control period, followed by a midterm assessment (T1). Subsequently, participants will undergo two weeks of in-clinic familiarization with Strolll AR. Then, participants will start with the 6-week Strolll AR intervention at home, followed by a final in-clinic assessment (T2). The primary study parameters are feasibility (i.e., safety, adherence, performance and user experience) and effectiveness for improving indicators of gait, balance and falls risk. For the statistical analyses on effectiveness, participants will be allocated to control (using T0-T1 change data) or intervention (using T1-T2 change data) groups using multiple (n=20) randomizations. Recruitment started in May 2024 and the last T2 assessment is expected in January 2025.

Discussion: The design of this particular pragmatic RCT will demonstrate feasibility and effectiveness in a real-world setting and in a representative population. Strolll AR may facilitate the transition from supervised care in the clinic to independent care at home, providing a platform for delivering individualized treatment, assisted with AR cues when deemed beneficial, for improving gait, balance and falls risk in people with Parkinson's disease.

Study design This study is a fully blinded two-armed pragmatic RCT, which will be executed in the Netherlands implemented at multiple physical or exercise therapy clinics. The study consists of three in-clinic assessments performed by the participant's own therapist. After a baseline assessment (T0), participants will start with a 6-week usual care control period, followed by a midterm assessment (T1). Subsequently, participants will undergo two weeks of in-clinic familiarization with Strolll AR supervised by their therapist. Then, participants will start with the 6-week Strolll AR intervention at home integrated with usual care, followed by the final in-clinic assessment (T2).

The definition of a pragmatic RCT is that the intervention is tested in a real-world population in a real-world setting (i.e., the clinic), with an appropriate comparison arm (i.e., usual care alone in the current study) and relevant outcomes to determine its effectiveness. Special trial designs can be considered. In this particular pragmatic RCT all participants follow the same T0-control-T1-intervention-T2 procedure, suitable for both between-groups and within-subjects comparisons of the Strolll AR intervention effects as well as for evaluating its clinical feasibility. With regard to the between-groups comparison, participants will be allocated to control (using T0-T1 change data) or intervention (using T1-T2 change data) groups using multiple (n=20) two-armed randomizations. The randomization will be predefined by an independent person using a Matlab script, locked with a date and timestamp, and kept concealed from participants, therapists and researchers. After the final T2 assessment of the last participant, the 20 randomizations of participants over control and intervention groups will be revealed to the researchers and applied in the data analysis. Participants, therapists, and researchers will therefore all be blind to group allocation during the study. This design, in which all participants receive the Strolll AR intervention, also allows for a complementary within-subjects evaluation of the intervention effect (i.e., being less susceptible to between-subjects variation than aforementioned between-groups evaluation), whilst at the same time providing ample information on the clinical feasibility of the Strolll AR platform implemented in the clinical pathway in terms of its safety, adherence, performance and user experience (i.e., from both participants' and therapists' perspectives). Moreover, a design with the same procedure for all participants makes implementation of this study in the clinical pathway easier (hence the term pragmatic RCT) and potentially reduces drop-out rate because no participants are withheld from receiving the intervention due to the randomization.

Procedure and outcomes After enrollment, the study consists of three in-clinic assessments spread out over 14 weeks, demarcating a usual-care control period and a Strolll AR intervention period. Below we describe the procedure for each of these parts of the design.

Participant enrollment: enrollment and informed consent The therapist will inform eligible participants about the study. If people are interested, they will receive a flyer with instructions on how to send their contact details to the researchers. Upon receival of contact details of a potential participant, the researcher will call this person to provide more information about the study. Hereafter, potential participants will receive the participant information letter by mail or email. A week after receiving the participant information letter, the researcher will again call the potential participant to answer any questions and to ask if he or she wants to participate. If so, in- and exclusion criteria will be checked by the researcher (to the extent possible, otherwise criteria will be evaluated during the baseline assessment session T0). Furthermore, the potential participant will be informed that the medication should remain stable during the study. If the potential participant is eligible, he or she will be asked to sign the informed consent form, after which the researcher also signs the informed consent form and notifies the therapist that the person can start with the pragmatic RCT. Written informed consent must be given before commencing with the baseline assessment (T0).

Baseline assessment (T0): characterization and effectiveness The baseline assessment serves to characterize the study sample and to provide baseline values for evaluating the effectiveness of Strolll AR. Study parameters used to characterize the participants are demographics (i.e., age, gender, disease duration) and current medication use (i.e., the Levodopa Equivalent Daily Dose [LEDD]). Questionnaires to describe the participant population are the Montreal Cognitive Assessment [MoCA], Movement Disorders Society Unified Parkinson Disease Rating Scale [MDS-UPDRS], New Freezing of Gait Questionnaire [NFOGQ], Physical Activity Scale for the Elderly [PASE], Cumulative Illness Rating Scale [CIRS], and 12-month falls history. All these outcomes will be collected using online questionnaires, except for the MoCA and MDS-UPDRS part Ia, III and IV, which will be administered in the clinic by the therapist. To be able to evaluate the potential effects of Strolll AR for improving indicators of gait, balance and falls risk, the following standard clinical tests will be administered by the therapists: Timed Up-and-Go test (TUG), Five Times Sit-to-Stand test (FTSTS), 10-meter walk test (10MWT) and Mini Balance Evaluation Scale Test (Mini-BESTest). In addition, participants will fill out the Falls Efficacy Scale International (FES-I).

Usual care (6-weeks) During the usual-care control period, participants will not receive any additional training or instructions related to the Strolll AR intervention.

Midterm assessment (T1): effectiveness For the midterm assessment, the standard clinical tests of gait, balance and falls-risk indicators (TUG, FTSTS, 10MWT, Mini-BESTest) will again be administered by the therapist. In addition, the participants are asked to fill out the FES-I and 6-week falls history (to assess falls during the usual-care period) questionnaires at home.

Practice with Strolll AR intervention supervised in clinic (2 weeks) To make the participant familiar with the Strolll AR platform technology and procedures and to allow the therapist to evaluate safety aspects, the Strolll AR intervention starts with 1 to 3 supervised in-clinic sessions of (cueing-assisted) personalized AR exercises.

In these supervised sessions in the clinic, the therapist will also test and validate procedures to select and tailor effective assistive AR cues from a library of different AR cues. This is needed for two reasons: 1) to test the premise that a one-size-fits-all cue does not exist as there is heterogeneity in effect of cues among participants, 2) to ensure that the Strolll AR gait-and-balance exercises can be effectively assisted with personalized cues for those who need it (e.g., people experiencing freezing of gait or people with small steps and shuffling gait). The so-obtained best cues for supporting gait, as evaluated subjectively using both the participant's experience and the therapist's impression, will be used during the remainder of the Strolll AR intervention for those who are deemed to benefit from it. We will document if AR cues are added to Strolll AR gait-and-balance exercises, and when applicable, specify the type of AR cues as well as their personalized parameters (intercue distance, height, width, audiofeedback, et cetera).

To deliver the Strolll AR intervention, participants in this study will by default be equipped with Magic Leap 2 AR glasses for its unmatched vertical AR field of view. Special insert lenses from a complementary lens kit (supports 8 prescriptions from -2.0 to +3.0) made available at each clinical site will be used to correct vision for people wearing spectacles. In case vision cannot be corrected with these insert lenses (e.g., due to a special prescription not part of the lens kit), either tailored insert lenses will be ordered for Magic Leap 2 or Microsoft HoloLens 2 with a slightly smaller vertical AR field of view will be used which can be worn over spectacles to deliver the intervention.

Once the participant is deemed sufficiently capable of using the Strolll AR platform technology independently at home in a safe manner (after training in the clinic with the therapist for one to maximally three sessions of 1 hour in two weeks), the participant will take the AR glasses home together with a complementary SIM-card equipped WiFi router to continue the Strolll AR intervention independently at home. If the Strolll AR intervention at home is not deemed safe or suitable by the therapist, the intervention will stop and the reason for this will be documented by the therapist as part of the clinical feasibility evaluation.

Training with Strolll AR intervention independently at home (6 weeks) The Strolll AR intervention, integrated with usual care, comprises of a 6-week home-based individualized AR gait-and-balance exercise program assisted with tailored AR cues when deemed beneficial. The AR exercises are designed to train aspects associated with gait, balance and falls risk in a gamified manner to maximize training compliance. Strolll AR comprises seven AR exercises, namely Basketballl, Smash!, Mole Patrolll, Hot Buttons, Puzzle Walk, Wobbly Waiter and Cue Challenge, intended to improve gait, balance and falls risk in people with Parkinson's disease. Exercises that require walking have optional complementary AR cueing (Smash!, Cue Challenge), integrated cueing (speed cue in Wobbly Waiter) or AR-mediated goal-directedness (the moles in Mole Patrolll, the puzzle pieces in Puzzle Walk), to various degrees enabling and assisting people with more severe mobility impairments to participate. For each AR exercise, feedback of performance in terms of game-play performance metrics and functional performance metrics are given during the exercises. For Basketballl, for example, metrics associated with the number of balls thrown in the basket (game-play performance) and the number of sit-to-stands or squats made (functional performance) are given, derived from AR glasses data.

Participants will be invited by their therapist to use Strolll AR minimally 5 times a week for 30 active minutes per day, resulting in the recommended 150 minutes per week. They can perform the exercises anytime during the day, in one bout or in multiple chunks during the day, with complementary AR cueing when deemed beneficial. The precise exercises, exercise duration, difficulty level and order will be prescribed by the therapist via the Strolll web portal. This will be evaluated and adjusted on a weekly basis in telephone calls, e-consults or during an onsite therapy session using information of the participant in combination with the feedback from exercises that are reported in the Strolll web portal, such as adherence and game-play and functional performance metrics. Adherence will be defined as the percentage of performed over prescribed gait-and-balance exercises (i.e., both frequency and session-duration adherence). The data on game-play level and performance will be used to evaluate whether the exercise program was both progressive (increase in level) and achievable (consistently high game-play performance scores) as intended. The weekly contact moments with the therapist will also be used to ask if participants experienced any adverse events (safety flags). Specifically, safety will be evaluated as the number of adverse events due to the Strolll AR intervention, such as falls, dizziness, eye strain and headache. Potential technical issues will be documented by the researcher, who will be available to provide technical support during the study based on email or telephone contacts with the participants.

Final assessment (T2): effectiveness and clinical feasibility To be able to evaluate the effectiveness of the Strolll AR intervention, the therapist will again administer standard clinical tests of gait, balance and falls risk (TUG, FTSTS, 10MWT, Mini-BESTest). Participants will fill out the FES-I questionnaire together with the researcher during a telephone call.

Clinical feasibility of the Strolll AR platform according to the participants will be assessed in terms of user experience (i.e., the User Experience Questionnaire [UEQ]), technology acceptance and use (i.e., Unified Theory of Acceptance and Use of Technology informed questions), top-3 barriers and facilitators of Strolll AR, and intervention-specific questions regarding Strolll AR, including cueing if used during the intervention (see the Supplementary Material for the Strolll AR evaluation questionnaire). Participants will also report their falls history during the intervention period. The questionnaires of the final assessment are completed through a telephone call with the participant to ensure comprehensibility, which was first assessed in think-aloud sessions with a number of participants. In light of the potential future implementation in the clinical pathway, therapists will also be asked about their experience with Strolll AR through a purpose-made questionnaire, including a top-3 barriers and facilitators (assessed before training their first participants and after training their last participant) and the UEQ (assessed after training their last participant). Furthermore, a focus group with the therapist will be organized after the study for an in-depth evaluation of the top-3 barriers and facilitators.

To be eligible to participate, a person must meet the following criteria:

• 21 years or older
• have command of the Dutch language
• diagnosed with Parkinson's disease according to the UK Parkinson's Disease Brain Bank criteria (stages 1-3 on the Hoehn and Yahr scale)
• bothersome gait or balance impairments (i.e., negatively affecting their ability to perform their usual daily activities as indicated by the person with Parkinson's disease and/or their therapist).

A potential participant who meets any of the following criteria will be excluded from participation:

• inability to comply with the protocol, i.e. additional neurological diseases and/or orthopedic problems seriously interfering with gait function
• insufficient physical capacity to safely perform the intervention (e.g., frequent faller) or severe cognitive impairments (as observed by the therapist)
• visual or hearing impairments (after corrective aids)
• inability to walk independently for 30 minutes (in bouts of 5-10 minutes)
• severe visual hallucinations or illusions
• no stable dosage of medication.

Finally, in case potential participants are susceptible to hallucinations, have a history of seizures, have deep brain stimulation, or a pacemaker, it should be discussed with their therapist if they can participate, as per Strolll's Instructions For Use (DOC-IFU-00021 (NL), Revision 01, Date of Issue 06 2024).