This prospective study was carried out between January and February 2022. A convenience sample of fifteen children with CP was recruited from a pool of patients of four private pediatric physiotherapy clinics in Greece. Children were eligible for the study when they met the following inclusion criteria: (1) age 6 to 12 years old, (2) diagnosed with unilateral or bilateral type of spastic CP, (3) functional mobility in categories I or II according to Gross Motor Function Classification System (GMFCS), (4) ability to understand verbal instructions. The exclusion criteria were: (1) orthopaedic surgery in the last year, (2) botulinumtoxin-A (Bont-A) injections within four months before the assessment, (3) categories III to IV on GMFCS, (4) difficulties in following verbal instructions, and (5) absence of electronic devices (laptop or smartphone). Parents of all participants were fully informed about the aims of the study, and they provided written informed consent. The study protocol followed the ethical standards of the World Medical Association (Declaration of Helsinki) and was approved by the Ethics Committee of the Physiotherapy Department, University of Thessaly (Protocol ID-719/23-9-2021).

To test the hypothesis, the study employed a within-subjects design in which children completed the assessments in two environments: remotely at home and in-person at the private physiotherapy clinic. These sessions were scheduled with 24 to 48 hours apart. For the remote evaluations, the VSee platform was utilized to facilitate real-time, two-way communication between children, parents, and the physiotherapist at home. VSee was selected for its high-resolution video quality, secure interface, and user-friendly accessibility, as it functions without requiring software installation on laptops or smartphones. To minimize potential learning effect, the functional tests order in both experimental conditions was randomized using the Random Sequence Generator software (https://www.random.org/sequences/). However, the order of experimental conditions remained the same for all participants, starting with the in-person assessment, followed 24 to 48 hours later by the remote assessment. To avoid fatigue, a resting period of 10 minutes between tests was provided in both experimental conditions. All children were instructed to avoid any vigorous activities (such as outdoor activities) or physiotherapy sessions between the two experimental conditions (remote and in-person). The study design is presented in Figure 1. To ensure thorough preparation and clarity, all parents were provided with comprehensive written guidelines. This documentation included instructions for equipment setup, navigation of the VSee platform, and explicit safety protocols for emergencies, complemented by instructional videos and illustrated guides for each specific test. Furthermore, supportive materials (ropes, meter tapes, goniometers, adhesive tapes, and a wooden step) were given to all children for the remote assessment. In both remote and in-person assessments, the same parent was present. All measurements were performed by the same physiotherapist (GK). The entire physiotherapy team possessed prior clinical expertise in conducting remote evaluations and delivering telerehabilitation services [21–23].

One day before the remote evaluation, each participating parent attended a 1-hour virtual orientation with the same physiotherapist (GK). This session focused on organizing the home environment and mastering the VSee platform. The physiotherapist provided practical advice, such as optimizing internet stability by positioning devices near the router, and delivered a verbal walkthrough on the correct use of testing materials. Furthermore, the session dedicated a specific time to emergency preparedness, ensuring parents felt confident managing potential technical failures or safety concerns, such as a child losing their balance.

Prior to study enrollment, each child underwent a thorough clinical intake at a private physiotherapy center to document their medical history, anthropometric data (height and weight), and specific physical impairments related to CP. The functional evaluations were conducted across two consecutive days, with 24 to 48 hours apart, separating the in-person assessment and the remote one performed at the child’s home. All children were asked to wear the same comfortable clothes and footwear for both experimental conditions. At the in-person assessment, children rested for 30–40 minutes before testing to avoid any fatigue due to their transportation to the physiotherapy clinic.

During the remote assessment, the children remained at home, maintaining no physical contact with the physiotherapist, who conducted the session from the clinic. Parents were only observed the assessment procedure and received strict instructions not to be involved unless an adverse event occurred. To ensure a stable and standardized viewing angle, all parents were supplied with a 50 × 50 cm wooden box to serve as a platform for their laptop or smartphone. Additionally, a wooden step that simulated a 17 cm stair step was provided to all participants for the remote assessment of the LSU. To conduct the TUG test, the laptop or smartphone was placed on the box at an angle, roughly 100 to 150 cm from the starting point. This specific positioning ensured the physiotherapist had a clear, unobstructed view of the entire 3-meter path, including the critical moments when the child initiated the movement from the chair and returned to a seated position. For the LSU, the laptop/smartphone was placed on the box, 100–150 cm away, at the back side of the child, and at an angle to fully capture the hip extension of the child while standing up and down from the wooden step. Similarly, for the FTSTS, the laptop/smartphone was placed beside each child to fully capture their performance during the five repetitions.

For PBS, the laptop/smartphone was placed at a distance that allowed the physiotherapist an unobstructed, full-body view of the child in both standing and seated postures.

Children were given a chance of two practice trials for each test before starting measurements. All the aforementioned functional tests were performed under the verbal guidance of the physiotherapist (GK) and in line with the recommended guidelines for each test [9, 10, 24, 25]

The physiotherapist was primed to record any adverse events for both the in-person and the remote assessment. Adverse events were defined as any physical or emotional incidents related to the assessment, including loss of balance resulting in a fall, dizziness, or significant distress. It was established that the occurrence of any such physical event would result in the immediate termination of the session. Additionally, the physiotherapist documented any factors that necessitated stopping a remote assessment, such as poor internet connectivity or a total loss of communication with the participant.

The normality of the data distribution was evaluated using the Shapiro-Wilk test. Descriptive statistics for categorical variables were presented as frequencies and percentages, while continuous variables were expressed as either means (with standard deviation, SD) or medians (with interquartile ranges, IQR), based on their distribution. The mean difference (MD) was expressed as in-person minus remote measurement. The required sample size was determined using the G*Power 3.1.9.4 software (University of Düsseldorf, Germany). Based on an “exact” bivariate normal model test to detect a moderate effect size (f = 0.7) with 80% power and a significance level of a = 0.05, a minimum of thirteen children was required. To account for a potential 20% attrition rate [28], common in studies of this nature, the final recruitment target was set at fifteen patients.

Differences between the two experimental conditions were calculated using the paired sample t-test. Statistical agreement for all continuous variables was evaluated using the ICC, specifically employing a two-way random-effects model to assess absolute agreement. Following the criteria established by Koo & Li, reliability was categorized as poor (< 0.5), moderate (0.5 to 0.75), good (0.75 to 0.9), or excellent (> 0.9) [29]. To further examine the level of agreement between the remote and in-person sessions and to detect possible outliers, Bland-Altman plots were generated [30]. Additionally, the standard error of measurement (SEM) was determined using the following formula to quantify the precision of the assessment [31]:

S E M = S D × 1 - I C C

This value subsequently informed the calculation of the smallest detectable change (SDC95%), representing the threshold for true change at a 95% confidence interval (CI) [32]:

S D C 95 % = S E M × 1.96 × 2

The SDC95% serves as the threshold for determining whether an observed difference reflects a “true” clinical improvement rather than a measurement error. As such, a lower SDC indicates a more sensitive assessment tool [31]. All statistical computations were conducted using SPSS for Mac, version 29.0 (SPSS Inc, Chicago, IL, USA), with the level of significance established at p < 0.05.

There are several limitations to acknowledge in this study. Firstly, the small number of participants, the small age group (children from 6 to 12 years old), and their classification as GMFCS levels I and II, limit the generalizability of the findings in older children and in the broader pediatric CP populations, such as those with more severe limitations on motor skills (GMFCS levels III or higher). Although the initial power analysis showed almost fifteen children to be included in the study, functional, behavioral variations, and environmental factors may affect children’s performance, and therefore, our findings should be considered as preliminary. It is important to note that because our primary objective was to determine the agreement between in-person and remote sessions, other forms of reliability, such as intra-rater or test-retest reliability within a specific setting, were not analyzed [38]. While this study focused on inter-modal consistency, test-retest reliability remains a vital metric for establishing the precision of individual measurements, which is necessary for monitoring clinical progress over time.

It is worth noting that in this study, parents acted strictly as passive witnesses to the assessment procedure and did not provide any physical assistance. However, even the purely observational presence of a parent in the home environment can psychologically influence the child’s performance dynamics. This passive parental oversight may introduce assessment bias by subtly altering the child’s motivation, compliance, or performance anxiety, which can consequently impact execution speed and balancing strategy compared to traditional clinic-based environments [39].

Despite the prior experience of the research team in conducting remote assessments utilizing functional tests among adult populations [22, 23], the administration of the tests by the same physiotherapist in both experimental conditions may have increased the risk of observer bias. However, this approach was intentionally adopted to reflect routine clinical practice, where the same clinician typically conducts repeated assessments and monitors patient progress over time.

An additional consideration when interpreting the present findings is the potential influence of a learning effect associated with repeated test administration. To minimize this possibility, the tests’ order was randomized across participants in the two experimental conditions. Moreover, children did not demonstrate superior performance in the remote condition, suggesting that prior exposure to the tasks at the in-person assessment was unlikely to have substantially affected the results. Nevertheless, the potential contribution of familiarity with the assessment procedures should be considered in future studies examining agreement between in-person and remote testing modalities.

Furthermore, in our effort to minimize confounding factors of measurement variability, we standardized the home environment through parent training, equipment provisioning, and strict positioning guidelines. This rigorous experimental control was necessary to establish baseline agreement, as there were no previous data for this specific pediatric population. Therefore, our study does not truly reflect routine telehealth practice in a typical home setting, limiting the external validity. Overall, due to the absence of sensitivity analysis, the results of this study should be interpreted with caution as they may be influenced by individual variances.

Taking these limitations into account, future research should include higher sample sizes to assess these tests on different age and GMFCS level groups by incorporating test-retest and intra-rater reliability to obtain a broader perspective on their use in telehealth.

In conclusion, the preliminary findings demonstrate that the TUG, LSU, FTSTS, and PBS can produce similar results with the in-person assessment when performed remotely at home in children aged 6 to 12 years old with CP. The incorporation of these functional tests into remote assessment holds greater significance, emphasizing the opportunity of reaching more populations with limited access to in-person assessment due to multiple reasons. However, the results of this pilot study should be treated consciously, as at an individual level, there were some considerable variations. Clinicians wishing to implement these tests for remote assessment at home, when in-person assessment is not achievable, may need to consider greater standardization and provide training and facilitation to parents/caregivers. Future studies with larger sample sizes are needed to ensure the accuracy of these results.