This was a parallel-group, single-center, single-blinded, phase 2 randomized controlled trial to investigate the feasibility, safety, and efficacy of a telerehabilitation program in people with PD. This study was conducted following the Declaration of Helsinki, and the trial protocol was approved by the Ethical Committee of Hospital Ophir Loyola (CAAE: 42496620.9.0000.5550). Written informed consent was obtained from all participants before enrollment. This trial has been registered at Registro Brasileiro de Ensaios Clínicos (ReBEC) under the identifier: RBR-6sz837s.

From June 2021 to December 2022, people with PD diagnosed according to the United Kingdom Parkinson's Disease Society Brain Bank criteria (19) were screened from the Movement Disorders Unit of the Hospital Ophir Loyola (Belém, Brazil). The inclusion criteria were as follows: (i) age between 30 and 80 years, (ii) mild stage disease (Hoehn and Yahr stage equal to or less than 2), and (iii) stable antiparkinsonian medication dose in the previous month. The exclusion criteria were as follows: (i) secondary or atypical parkinsonism, (ii) severe psychotic symptoms (score >2 in item 1.2 of the International Parkinson's and Movement Disorder Society-Unified Parkinson's Disease Rating Scale), (iii) clinical diagnosis of dementia; (iv) diagnosis of severe systemic disease (such as infections, severe heart disease, malignant neoplasm, liver or kidney failure, and poorly controlled diabetes), (v) the presence of orthopedic diseases, other neurological diseases, or cardiac comorbidities that prevent or pose a risk for performing aerobic or stretching exercises, (vi) no access to teleconference technology (smartphone, tablet, or computer), and (vii) being unable to perform tasks on the computer or without family assistance, which is needed for remote activities.

Participants were randomly assigned (1:1 ratio) to two groups, namely, control and telerehabilitation, using a web-based system stratified for sex and disease duration. The randomization was carried out by a researcher not involved in patient recruitment or assessment or data analysis. The researchers involved in the patient assessment or data analysis were blinded to group allocation. Due to the nature of the intervention, participants and the physiotherapists who conducted the telerehabilitation sessions could not be blinded to the treatment allocation. Patients were instructed not to talk about the interventions during evaluations. Considering the exploratory aspect of the study, focused on feasibility, the sample size calculation was not performed.

Figure 1 presents a summary of the study schedule and assessment. Participants were evaluated at the baseline (T0), end of intervention (T4), and follow-up at 8 weeks after the baseline (T8) as wash-out evaluation. Participants were not allowed to receive any other type of rehabilitation during the study period. Since the randomization of participants, antiparkinsonian medications were not changed during the study period.

The telerehabilitation (TR) group underwent a 4-week telerehabilitation program at their homes, consisting of individual remote sessions supervised by a physiotherapist with real-time visual feedback and verbal cues (1 time/day, 3 days per week, for a total of 12 sessions, 60 min/session). Before sessions, participants' aerobic capacity and endurance were evaluated by the 6-min walk test (20). Only participants with good performance in the 6-min walk test would receive the exercise program. Telerehabilitation sessions were performed at home using a smartphone, tablet, or computer via free teleconference web platforms (e.g., Whatsapp^® or Google Meet^®). A caregiver supervised all the sessions in person to assist the participants if needed.

Each session included warm-up, mobility, strength, balance, and cool-down exercises based on a previous therapist-supervised exercise protocol (21) of moderate level of intensity. In the first week, the telerehabilitation sessions focused on familiarizing the patients and their caregivers with the method and ensuring a clear understanding of the exercises. The physiotherapist verbally explained the exercises, and when patients faced difficulty executing them, the physiotherapist demonstrated the movements. In the second and third weeks, additional exercises were introduced to enhance postural control on unstable surfaces and facilitate the progression toward weight transfer while also considering eliminating visual input or including head movements. During the fourth week, dynamic balance exercises were incorporated to promote postural control while simultaneously involving movements of the upper and lower limbs. This phase aimed at enhancing the patients' overall dynamic stability and functional capabilities. The therapeutic interventions were carefully designed throughout the process to target proprioception in the foot, sacroiliac joint, and cervical spine, ensuring appropriate positioning during exercise sessions and fostering improvements in postural control and stability. If participants complained about fatigue during sessions, the level of intensity was reduced.

For the control (CT) group, the physiotherapist offered participants an informative illustrated booklet with a demonstration and description of the exercises from the telerehabilitation program. Participants were instructed to consult the booklet and perform the training three times per week. The booklet is available in Supplementary Appendix S1. For both groups, the physiotherapist checked training feedback and adverse effects (AEs) once a week via a telephone call.

At the baseline (T0), we collected demographic information, including age, gender, and educational status. Moreover, we evaluated clinical data, such as the Hoehn and Yahr stage (22), the Movement Disorder Society-Unified Parkinson's Disease Rating Scale (MDS-UPDRS) (23), the Montreal Cognitive Assessment (MoCA) (24), the Time Up and Go Test (TUG) (25), the five-repetitions sit-to-stand test (5STS) (26), the activities-specific balance confidence (ABC) scale (27), the Geriatric Depression Scale (GDS-15) (28), and the 8-item Parkinson's Disease Questionnaire (PDQ-8) (29).

At revaluations (T4 and T8), participants underwent only the following tests: MDS-UPDRS Part III, TUG, 5STS, ABC scale, and the PDQ-8.

The primary outcome was the feasibility of telerehabilitation sessions (TR group) and booklet-based exercise program (CT group). It was assessed by adherence and safety. Adherence was defined as the percentage of sessions attended. Based on the percentage of the sessions attended, participants were categorized as high adherence (>80%), partial adherence (20%−80%), and non-adherence (< 20%) (30). Safety was evaluated in both groups by tracking the cumulative number of AEs and severe AEs from the baseline through the end of follow-up. AEs were defined as all-cause mortality, hospitalization for falls, or other diseases preventing exercise participation. The occurrence of AEs was compared between groups.

As secondary outcomes, gait and dynamic movements were evaluated by the TUG test, 5STS, and ABC scale in T4 and T8. The global motor status was evaluated by MDS-UPDRS Part III. Patient-reported outcomes were evaluated by the PDQ-8.

The normality of data was checked with the Kolmogorov–Smirnov test. The results were reported as numbers and percentages for categorical variables and mean and 95% confidence interval (normal distribution) or median and interquartile range (non-normal distribution) for continuous variables. We used the two-tailed t-test (normal distribution) or Mann–Whitney test (non-normal distribution) to compare the characteristics of continuous data between groups at the baseline. The chi-square test was used to test binary data between the groups at the baseline.

We performed an intention-to-treat analysis. Changes from baseline to T4 and T8 in the secondary outcomes (TUG, 5STS, ABC scale, MDS-UPDRS Part III, and PDQ-8—normal distribution) were assessed using the two-way repeated-measure analysis of variance was applied using “Time” as the within-group factor and “Group” as the between-group factor, and the Bonferroni post-hoc test was used. Effect sizes were represented by the partial eta-squared values. The level of statistical significance for all tests was set at a p-value of < 0.05. SPSS (IBM SPSS Statistics, version 23.0) was used for statistical analyses.

We invited 26 people with PD for the study. Nineteen participants (73% recruitment success) volunteered to participate and were randomized for the groups (CT group: n = 11; TR group: n = 8). All TR group participants had good performance in the 6-min walk test and completed 12 telerehabilitation sessions. One participant in the CT group did not complete the follow-up evaluation at T4. At T8, two participants of the CT group did not complete the follow-up evaluation, and one participant of the TR group did not complete the follow-up evaluation (CT group dropout rate 27%; TR group dropout rate 12.5%; and total participants dropout rate 21%).

A total of 18 participants (CT group: n = 10; TR group: n = 8) completed the T4 evaluation, and 15 (CT group: n = 8; TR group: n = 7) participants completed the T8 evaluation (Figure 2). Most participants were men (n = 10; 55.5%), aged from 47 to 75 years, and had a mean levodopa equivalent daily dose of 760 mg/day. Baseline clinical and demographic characteristics were similar across both groups, with no significant between-group differences (Table 1).

All participants randomized to the TR group completed 12 telerehabilitation sessions, with 100% of the number of participants. There were no deaths, hospitalizations, or life-threatening AEs in any group. No AEs occurred during the evaluations or any of the telerehabilitation sessions. Eleven participants (57.8%) reported at least one AE. The proportion of individuals who experienced some AEs was similar between the TR group compared to the CT group (75 vs. 45%, p = 0.35).

In the TR group, two participants reported flu symptoms, three others presented pain, and one individual presented tiredness. In the CT group, three participants reported flu symptoms, one reported pain, and one was diagnosed with COVID-19. For none of the individuals, these AEs were a reason to interrupt the intervention or result in the loss of follow-up. There were no reports of falls.

Regarding TUG, there was no significant difference between the TR and CT groups in the T4 and T8 evaluations (Figure 3A; Table 2). When the two groups were analyzed together, there was a reduction in TUG in T8 compared to the baseline (F_{2, 26} = 7.47, p = 0.004; Bonferroni post-hoc test: p = 0.004), with a partial eta-squared of 0.36, indicating a large effect size (Figure 3B; Table 2).

For 5STS, PDQ-8, and MDS-UPDRS Part III, we found that the results were not significant when compared between groups in T4 and T8 (Table 2).