Rethinking EDSS-based ambulation assessment in multiple sclerosis using continuous variable monitoring Authors: Noah M. Werner, Melanie Schuette, Ramona Hagler, Jan Voth, Balázs Danajka, Patricia Kirschner, Markus Heibel, Tjalf Ziemssen, Uwe K. Zettl, Sven G. Meuth, Marc Pawlitzki, Lars Masanneck Published: 2026 ABSTRACT BackgroundMultiple sclerosis (MS) is a chronic inflammatory disease of the central nervous system (CNS) characterized by relapses and progressive disability. The Expanded Disability Status Scale (EDSS), used to quantify disability, is based on single, discretely assessed and potentially inaccurate patient-estimated walking ability, whereas digital health technologies (DHTs) enable continuous activity monitoring and more objective assessment of real-world functional performance.MethodsIn this prospective observational study conducted at two German centers, patients with relapsing-remitting MS (RRMS) underwent clinical assessments at baseline (V1) and study completion (V2). Walking distance and step counts were measured using a measuring wheel and pedometer, while continuous physical activity was assessed via smartwatch-derived metrics.ResultsSixteen patients with RRMS were included (median age 57.5 years [interquartile range (IQR) 49.25–63.25]; median EDSS 4.5 [IQR 3.5–6]). Patient-estimated walking distance at V1 showed moderate correlation with clinically measured distance (Spearman‘s ρ = 0.60, p = 0.013), with 12 of 16 patients misjudging distances relative to EDSS thresholds. Walking distance showed intra-individual variability between V1 and V2 (median absolute difference: 113.6 m). Median daily walking distance (ρ = −0.61, p = 0.0123), step count (ρ = −0.64, p = 0.0082), and peak steps (ρ = −0.69, p = 0.0032) correlated negatively with EDSS.ConclusionPatient-estimated maximum walking distance demonstrated moderate agreement with clinical performance and frequently crossed EDSS thresholds, while clinical assessments varied substantially within individuals over the short study duration, underscoring the limitations of single evaluations. In contrast, smartwatch-derived metrics aligned with clinical measures, reflected EDSS scores, and captured real-world mobility.Supplementary InformationThe online version contains supplementary material available at 10.1186/s42466-026-00501-8. Introduction Multiple Sclerosis (MS) is a chronic inflammatory disease of the central nervous system (CNS) [1, 2], characterized by both relapses and progressive disability accumulation [1, 3, 4, 5]. Despite substantial progress in reducing relapse rates in MS, many patients continue to accumulate disability, largely driven by progression independent of relapse activity (PIRA), which has emerged as a central determinant of long-term outcomes [6]. Various strategies have been proposed to characterize PIRA [8–10]. Nonetheless, the Expanded Disability Status Scale (EDSS) remains the primary clinical instrument to operationalize progression in MS in real-world practice [9, 11]. Despite its widespread use, the EDSS has been increasingly questioned [12]. It is a non-linear [13], predominantly motor-focused [12] scale, in which walking distance becomes a decisive determinant, particularly at scores between 4 and 6 [14]. Importantly, walking ability in MS can be conceptualized at various levels, including patients’ subjective perception of their walking capacity, the clinically assessed maximum walking distance under standardized conditions, and actual walking performance in everyday life. The reliance on patient-estimated walking capacity in routine clinical practice is often problematic, as many individuals are unable to accurately estimate their walking distance, thereby introducing considerable variability into the assessment [15]. In clinical practice, however, assessing maximum walking distance is time-consuming and is therefore often estimated based on patient report. Consequently, it remains uncertain to what extent EDSS values obtained during routine clinical visits truly reflect disease progression, given that everyday factors can also substantially influence the score [16]. To overcome these gaps, digital health technologies (DHTs), such as wearable devices, are becoming increasingly prevalent [17, 18] and are being used more frequently in neurological clinical trials [19], including studies in Duchenne muscular dystrophy [20], Parkinson’s disease [21], and Chronic Inflammatory Demyelinating Polyneuropathy [22, 23]. Recent studies in MS [24] further highlight the potential of DHTs, demonstrating that wearable-based monitoring can provide novel insights, with metrics such as average daily step count serving as additional indicators of disability severity and patient-specific characteristics [25]. Moreover, DHTs represent a promising solution for overcoming existing limitations by continuously collecting real-life data, including physical activity, sleep, and cognition [19]. Given the unprecedented longitudinal monitoring of physical activity that is made possible by these approaches, these DHTs may enable better detection of subtle fluctuations, particularly in early MS or mild progression, by continuously tracking day-to-day activity. Accordingly, this study aimed to examine three key aspects of walking ability assessment in MS. First, we wanted to quantify the discrepancies between patient-estimated and clinically measured maximum walking distance and step count at the start of the study, including misclassification across EDSS-relevant walking thresholds. Second, we assessed short-term within-person variability of clinically measured maximum walking distance over a period of seven days. Finally, we investigated the relationships between EDSS, clinical walking performance, and smartwatch-derived real-world mobility metrics, with a particular focus on bout-based measures as indicators of walking ability in everyday life. Methods Study design This prospective, two-center observational study was conducted at the Department of Neurology, University Hospital Düsseldorf, and Sauerlandklinik Hachen, Germany. Key inclusion criteria were age ≥ 18 years, a diagnosis of relapsing-remitting MS (RRMS) according to the revised 2017 McDonald criteria [26]. A clinically confirmed relapse within the preceding 90 days served as an exclusion criterion. After providing written informed consent, patients were equipped with a smartwatch (ScanWatch, Withings, Issy-les-Moulineaux, France) and instructed to wear it continuously, day and night, for the entire study period, enabling uninterrupted 24/7 data collection. Patients completed two scheduled study visits, including a walking distance test, at baseline (V1) and during the second study visit (V2) (for the study design, see Fig. 1). To better reflect everyday-life conditions, the interval between V1 and V2 was, whenever feasible, arranged to include a weekend. Fig. 1Study and Analysis Overview. This figure illustrates the design and process of the study, including patient recruitment at two centers, screening with inclusion and exclusion criteria, and subsequent study visits. The timeline depicts the study interval with two main assessment points. At each visit, clinical examinations and patient-reported outcome measures (questionnaires) were performed, including neurological disability scoring. The diagram further highlights the number of patients screened, excluded, and finally included in the analysis. Abbreviations: UKD, University Hospital Düsseldorf; SC, Sauerlandklinik Hachen; EDSS, Expanded Disability Status Scale Study and Analysis Overview. This figure illustrates the design and process of the study, including patient recruitment at two centers, screening with inclusion and exclusion criteria, and subsequent study visits. The timeline depicts the study interval with two main assessment points. At each visit, clinical examinations and patient-reported outcome measures (questionnaires) were performed, including neurological disability scoring. The diagram further highlights the number of patients screened, excluded, and finally included in the analysis. Abbreviations: UKD, University Hospital Düsseldorf; SC, Sauerlandklinik Hachen; EDSS, Expanded Disability Status Scale Before performing the walking distance test, patients completed a questionnaire estimating the maximum distance they could walk before requiring a first break and the approximate number of steps this would involve (see Supplementary Fig. 1 in Online Resource 1). After completing the questionnaires and estimating their maximum walking distance and step count, patients were instructed to walk as far as possible. The exact start and end times were documented. Walking distance and steps were assessed using a digital measuring wheel (DigiRollerPlus 2, Calculated Industries) and a manual pedometer, while the smartwatch simultaneously recorded steps and distance digitally. The test was continued until the patients could no longer walk and required a break to define the maximum walking distance. During V2, the procedure was repeated without smartwatch recording. Outcome parameters Disability status was evaluated using the EDSS [11]. The scale ranges from 0 (no disability) to 10 (death due to MS) [11]. In this study, we used the EDSS scores from previous medical examinations, which were not older than 30 days, for baseline characteristics, and did not perform a full EDSS assessment at V1 and V2. These days, we conducted the walking tests independently. All physicians in this study, as well as those who performed the EDSS, were certified EDSS raters. During both study visits, the distance covered in meters and the number of steps were counted and documented on the test sheet provided. For the following analysis, the walking distances from V1 and V2 were evaluated comparatively. Digital measures The smartwatch was used to record the following longitudinal activity parameters: daily steps as well as the estimated duration of soft (light), moderate, and intense (vigorous) activity (in seconds), according to the Physical Activity Guidelines for Americans [27] and the Compendium of Physical Activities [28]. In addition to the specified steps, we also recorded the algorithmically estimated distances covered in meters, as calculated by the smartwatch. In order to approximate the maximum uninterrupted walking distance in everyday life, as assessed in the EDSS, we derived time-based activity intervals for steps and distance from the continuous smartwatch data. For the analysis, a walking interval was selected that was preceded by at least 15 min of inactivity and followed by another 15 min of inactivity. This ensured that the segment under consideration represented a complete, uninterrupted walking phase. Across the defined intervals, the longest step and distance streaks were extracted and summarized using minimum, maximum, and median values. In addition, we used daily peak-step data from the smartwatch, which represents the highest step intensity achieved in a one-hour time window per day. The peak value is calculated by forming a rolling sum of the step count over a time window and then determining the maximum of this rolling sum. The returned peak-steps are the daily maximum step counts in such a 1-hour window during the observation period. Adherence calculation and criteria An adherence of at least 75% of wearing time during waking hours (6:00 a.m. to 11:00 p.m.) was defined for the activity analysis. If there were days that were below this threshold, or if the individual adherence of the respective patients was not met on more than 25% of the days, they were excluded from the analysis. Adherence per patient is visualized in Supplementary Fig. 2 (in Online Resource 1). These criteria are comparable to the adherence criteria of our previous work [22]. Software Data analysis and visualization were computed using R (version 4.4.1) [29] in RStudio (version 2024.09.1 + 394), including the Base and Stats packages and ggplot2 [30]. Data handling and visualization were conducted using the tidyverse package [31], including dplyr and tidyr. Additional statistical analyses were supported by the Hmisc package [32]. A web application to synchronize, organize, and visualize data was set up based on the Dash framework version 2.6.1. Descriptive statistics and calculation of distribution characteristics All analyses were exploratory in nature due to the sample size (n = 16) and the number of assessed correlations. Distributional assumptions were evaluated using the Shapiro–Wilk tests, in conjunction with inspection of summary distributions. Continuous variables are reported as mean (standard deviation, SD) when approximately normally distributed, and as median (interquartile range, IQR) otherwise. Normality was assessed for all variables. Correlations were then examined using Pearson’s r for normally distributed continuous data, and Spearman’s rank correlation for non-normally distributed or ordinal data. We treated the analyses as exploratory and performed them without correction for multiple testing. As a secondary sensitivity analysis for analyses assessing associations between EDSS scores and corresponding patient-reported, smartwatch-derived, and V1 walking measures, FDR correction was applied using the Benjamini–Hochberg procedure to account for multiple comparisons. For within-subject comparisons (V1 vs. V2) and comparisons of observed vs. estimated values at V1 and V2, paired t-tests were applied to normally distributed data, whereas Wilcoxon tests were used for non-normally distributed (non-parametric) data. Absolute differences between estimated and observed walking distances at V1 and between observed walking distances at V1 and V2 were summarized descriptively and tested against zero using one-sample t-tests or Wilcoxon signed-rank tests, depending on normality assessed by the Shapiro–Wilk test. Given the exploratory nature of the analyses, results were primarily interpreted based on effect sizes and direction of associations, with p-values considered descriptive rather than confirmatory. Data availability statement The data that support the findings of this study are available from the corresponding author upon reasonable request. Access may be subject to applicable ethical, legal, and institutional restrictions. Results Cohort characteristics 16 patients with RRMS were recruited across both centres. The median age at enrolment was 57.5 (IQR: 49.25–63.25) years, with a median disease duration of 12.5 (IQR: 3.50–22.00) years. Most patients (68.75%) were female. The baseline clinical values were a median EDSS of 4.5 (IQR: 3.0–6.0) and a median Body-Mass-Index (BMI) of 26.37 (IQR: 22.95–29.96). Of the sixteen patients, eleven received anti-CD20 treatment, one patient received teriflunomide, and four patients did not receive any specific disease-modifying therapy (see Supplementary Table 1 in Online Resource 1). Analysis of absolute deviations between patient-estimated and clinically measured walking distance and step counts at V1 The median patient-estimated number of steps during the walking distance test in V1 was 700 (IQR: 137.50–1400), and the median patient-estimated walking distance was 400 m (IQR: 62.50–600). The median number of steps measured by the pedometer during the walking distance test in V1 was 443 steps (IQR: 229.50–1255.50). The median walking distance measured by the measuring wheel during the same test was 214.33 m (IQR: 111.80–694.58). The patient-estimated steps showed a moderate positive, but not significant, correlation with the steps measured by the pedometer in the V1 walking distance test (Spearman’s ρ = 0.48, p = 0.059). The patient-estimated walking distance correlated moderately and significantly with the distance measured using the measuring wheel in the V1 walking distance test (ρ = 0.60, p = 0.013; Fig. 2) and remained statistically significant after FDR correction (Supplementary Table 2; Supplementary Fig. 3 in Online Resource 1). Fig. 2Correlation Heatmaps of observed Walking Performance, Digital Measurements, and Disability Scores. (A) Correlation between observed walking performance (actual distance and steps during the walking test) and patient-reported estimates, as well as digitally measured steps and distance. (B) Correlation between observed walking performance and EDSS scores with smartwatch-derived activity parameters, including median steps, median distance, longest step and distance streaks, and peak steps. Red indicates positive correlations, while blue indicates negative correlations. Asterisks denote statistically significant correlations (*p < 0.05). Pearson or Spearman correlation coefficients were applied depending on data distribution. Abbreviations: EDSS, Expanded Disability Status Scale Correlation Heatmaps of observed Walking Performance, Digital Measurements, and Disability Scores. (A) Correlation between observed walking performance (actual distance and steps during the walking test) and patient-reported estimates, as well as digitally measured steps and distance. (B) Correlation between observed walking performance and EDSS scores with smartwatch-derived activity parameters, including median steps, median distance, longest step and distance streaks, and peak steps. Red indicates positive correlations, while blue indicates negative correlations. Asterisks denote statistically significant correlations (*p < 0.05). Pearson or Spearman correlation coefficients were applied depending on data distribution. Abbreviations: EDSS, Expanded Disability Status Scale Of note, 12 of 16 patients misclassified their walking distance relative to EDSS-defined thresholds (Fig. 3; Supplementary Table 3 in Online Resource 1). Fig. 3Comparison of Estimated vs. Actual Walking Distance. Scatter plots illustrate individual patient performance in the walking test. Estimated values (blue) before attending and actual values (orange) of the performed walking distance are shown. The x-axis depicts individual study participants, while the y-axis represents the actual walking distance in meters achieved. Red dashed lines indicate clinically relevant thresholds of walking performance (100 m, 200 m, 500 m). Abbreviations: EDSS, Expanded Disability Status Scale Comparison of Estimated vs. Actual Walking Distance. Scatter plots illustrate individual patient performance in the walking test. Estimated values (blue) before attending and actual values (orange) of the performed walking distance are shown. The x-axis depicts individual study participants, while the y-axis represents the actual walking distance in meters achieved. Red dashed lines indicate clinically relevant thresholds of walking performance (100 m, 200 m, 500 m). Abbreviations: EDSS, Expanded Disability Status Scale Analysis of clinically measured walking distance at V1 and V2 Over a median interval of seven days (IQR: 7–10), clinically measured maximum walking distance showed substantial within-person variability (median absolute difference 113.6 m; Fig. 4; Supplementary Tables 4–5 in Online Resource 1), with changes occurring in both directions. Comparing the group median of the measured walking distance in V1, the measured distance in V2 was slightly higher at 250.69 m (IQR: 177.63–369.81). Fig. 4Comparison of walking distance between Visit 1 and Visit 2. Boxplots display the distribution of walking distances at visit 1 (blue) and visit 2 (green). Individual patient trajectories are connected with grey lines. Red bars indicate the mean value for each group. Substantial within-person variability is observed, with changes occurring in both directions. Absolute differences were significantly greater than zero (p < 0.001), indicating pronounced intraindividual variability Comparison of walking distance between Visit 1 and Visit 2. Boxplots display the distribution of walking distances at visit 1 (blue) and visit 2 (green). Individual patient trajectories are connected with grey lines. Red bars indicate the mean value for each group. Substantial within-person variability is observed, with changes occurring in both directions. Absolute differences were significantly greater than zero (p < 0.001), indicating pronounced intraindividual variability In V1, the Spearman correlation between patient-estimated walking distances and EDSS was ρ = -0.90 (p < 0.001), while clinically measured walking distance correlated moderately with EDSS (ρ = -0.47, p = 0.064). In V2, the Spearman correlation between patient-estimated walking distances and EDSS was ρ = -0.54 (p = 0.029), while clinically measured walking distance was more strongly associated with EDSS (ρ = -0.61, p = 0.013). Smartwatch analysis Overall, smartwatch adherence during predefined waking hours was high, with most participants meeting the prespecified threshold for inclusion in activity analyses (see Supplementary Fig. 2 in Online Resource 1). Daily physical activity was assessed based on the number of steps taken and distance walked, measured by the smartwatch. Across recorded days, median daily step count was 2210 (IQR 796.50–4118.50) and median daily smartwatch-derived walking distance was 1779.8 m (IQR 590.30–3136.80). The distribution of daily steps is shown in Supplementary Fig. 4 (in Online Resource 1). Correlation of smartwatch-derived activity metrics with EDSS scores Smartwatch-derived step counts and distances during the V1 walking distance test correlated strongly with clinically measured values (steps: ρ = 0.83, p < 0.001; distance: ρ = 0.86, p < 0.001; Fig. 2), remaining statistically significant after FDR correction (Supplementary Table 2; Supplementary Fig. 3 in Online Resource 1). The median values for the longest steps and distance streak in everyday life per patient, derived from the smartwatch recordings, correlated moderately and significantly with the clinical measurements at V1(steps: ρ = 0.51, p = 0.0464; distance: ρ = 0.53, p = 0.0373; Fig. 2), although these associations did not remain significant after FDR correction (Supplementary Table 2; Supplementary Fig. 3 in Online Resource 1). A graph of the distance streak derived by the smartwatch, together with the patient-estimated distances and those measured clinically, is shown in Fig. 5. Fig. 5Comparison of Estimated, Clinical Measured and Median longest Distance Streak. Scatter plots illustrate individual patient walking performance. Estimated walking distances (blue) prior to the study visit, actual walking distances (orange), and median longest distance streaks measured via digital monitoring (green) are displayed for each participant. The x-axis represents individual study participants, while the y-axis shows walking distance in meters. Red dashed lines indicate clinically relevant thresholds of walking performance (5 m, 100 m, 200 m, 500 m). Abbreviations: EDSS, Expanded Disability Status Scale Comparison of Estimated, Clinical Measured and Median longest Distance Streak. Scatter plots illustrate individual patient walking performance. Estimated walking distances (blue) prior to the study visit, actual walking distances (orange), and median longest distance streaks measured via digital monitoring (green) are displayed for each participant. The x-axis represents individual study participants, while the y-axis shows walking distance in meters. Red dashed lines indicate clinically relevant thresholds of walking performance (5 m, 100 m, 200 m, 500 m). Abbreviations: EDSS, Expanded Disability Status Scale With regard to EDSS scores, both the median walking distance and the median number of steps per subject show a moderate negative correlation when using smartwatch-derived everyday data (ρ = -0.61, p = 0.0123 for walking distance; ρ = -0.64, p = 0.0082 for steps), as does the median peak-steps per subject (ρ = -0.69, p = 0.0032) during the entire period in which the smartwatch was worn (Fig. 2), remaining statistically significant after FDR correction (Supplementary Table 2; Supplementary Fig. 3 in Online Resource 1). Discussion In this proof-of-concept study, we connected patient-estimated, clinically assessed, and smartwatch-derived walking measures in people with MS. Three observations stand out: self-estimated walking distance showed only moderate agreement with clinical performance and frequently crossed EDSS-relevant ambulation thresholds; clinically measured walking distance varied substantially within individuals over just seven days; and smartwatch-derived measures correlated closely with reference measurements during in-clinic testing and likely captured complementary aspects of real-world mobility over a 7-day period. This study demonstrated a strong correlation between clinically assessed and digitally measured walking distance and step count. This close agreement provided the basis for linking everyday activity parameters with functional clinical measures. At the same time, many patients substantially misclassified their own physical performance. Despite the overall correlation between in-clinic testing and patient-estimated maximum walking distance, discrepancies in patients‘ estimates were large enough in several cases to result in EDSS-relevant shifts, affecting 12 of 16 patients in our cohort. These findings are consistent with previous reports showing that both patients and physicians frequently misjudge walking distance [15, 33]. Consistent with these observations, studies in which daily activity was objectively measured using smartwatches have shown that patient-estimated walking distance correlates only weakly with objectively measured performance, with most patients underestimating their walking ability, leading to clinically relevant EDSS misclassification [34]. This discrepancy between subjective perception and objectively measured performance has previously been described [35–38] and is also reflected in patient reports of perceived disability despite stable EDSS scores [39]. Since motor impairment plays a central role in MS disability assessment, especially in the mid-range of the scale (EDSS 4.0–6.0) [14], even small variations in walking distance can lead to a change in disability grade. A second important observation from our study was the short-term variability in clinically measured walking distances. Even between two clinical assessments separated by only a median of seven days, maximum walking distances showed considerable fluctuations. Although the group median changed only slightly, the individual fluctuations showed that even frequent clinical tests capture only a snapshot of walking performance and may not take into account short-term fluctuations in motor function. Even subtle deterioration of the disease can contribute to this variability. It has been shown that insidious progression occurs even at lower EDSS scores and often goes unnoticed in patients who are officially classified as having RRMS [7]. Such gradual changes are easily overlooked in sporadic clinical examinations, underscoring the need for complementary tools that can more reliably capture fluctuations in practice. In consideration of these limitations, DHTs offer a promising alternative. Previous studies have shown that accelerometer-based activity tracking correlates with EDSS [40, 41] and several groups have investigated the relationship between step counts and clinical scores [42]. The median values of the peak-steps also showed a strong correlation with the EDSS, offering the possibility of approaching as closely as possible the individual limit of the maximum possible walking distance before a resting period and thus showing the potential as a short-term marker. By analyzing predefined walking-distance intervals in our data, we likewise observed the expected day-to-day fluctuations but were able to derive daily or weekly medians. These aggregated measurements correlated with clinical walking tests and provided a more stable representation of typical walking ability than isolated clinical assessments, again underscoring the potential of objective measurements in real-world conditions. Even when walking distance is measured according to strict standardized procedures for EDSS assessments, performance in MS patients is likely influenced by prior daily activity. Fatigue severity in MS increases throughout the day, typically peaking in the late afternoon [43], and physical exertion further exacerbates fatigue [44], potentially impairing walking performance in later assessments, highlighting the need for more frequent monitoring, which DHTs may facilitate. The high adherence to smartwatch use observed in this study, consistent with our previous long-term trials [22], indicates that patients are willing to wear such devices on a daily basis. This study has several limitations. EDSS ratings were performed by different physicians, potentially introducing inter-rater variability; however, this reflects usual clinical practice in clinical settings, and in addition, all raters were certified EDSS raters with extensive clinical expertise. Furthermore, we did not incorporate the patient-reported walking distance from V1 into the current EDSS assessment because the EDSS was not recalculated if the prior score was less than 30 days old. Overall, patient-reported walking distance and EDSS-derived ambulatory status were closely aligned, suggesting good concordance with the contemporaneous EDSS estimate even without updating the score based on V1. Minor discrepancies between V1 walking-distance estimates and those documented at the time of EDSS determination were observed in a small number of cases, further underscoring that self-estimated walking distance can fluctuate even over short time intervals. Furthermore, the walking distance assessments were not standardized for the time of day, which may have led to deviations due to daily fluctuations. However, this is in line with standard clinical practice and highlights a limitation that often occurs with single-point assessments. Additionally, due to the proof-of-concept design, our cohort included patients with comparatively higher EDSS scores. Future studies should evaluate larger and more heterogeneous cohorts, including individuals with lower disability levels and longer follow-up, to validate our findings and further explore the potential of digital health technologies to detect subtle disease progression at earlier stages of MS. Given the small cohort size, the present findings should be interpreted as exploratory. Nevertheless, we applied FDR correction to correlations between real-world parameters and EDSS, as well as to clinical assessments of the V1 walking test. The sensitivity analysis using FDR correction showed overall similar trends, with the main findings remaining consistent after adjustment. However, larger cohorts are needed to confirm these observations. Most consumer wearable devices use accelerometer-based technology to detect walking activity [45], as does the Withings Scanwatch. Established frameworks for validating wearable-based biometric monitoring technologies, such as the V3 framework (verification, analytical validation, clinical validation), have been proposed [46]. However, to our knowledge, no peer-reviewed validation study has specifically evaluated the step count or walking distance algorithms of the Withings ScanWatch. While older Withings models have been validated, these findings cannot be directly extrapolated to the ScanWatch [47–49]. In the present study, smartwatch-recorded data showed good correlation with manual measurements (measuring wheel, pedometer) during V1. This provides preliminary analytical verification; however, confirmation in larger cohorts is warranted. There is wide variability in step count accuracy based on device model, movement speed, and testing environment, and the gold standard remains video-recorded step count [45]. External influences such as infections or mobility aid use can confound activity signals [50, 51]. In our analysis, a 15-minute period of inactivity before and after each episode was used to define self-initiated walking bouts that may represent a complete, uninterrupted walking phase. To contextualize this choice, alternative interval lengths (5, 10, 20, and 30 min) were explored. Shorter intervals may be more susceptible to short-term fluctuations and therefore less representative, whereas longer intervals may be too prolonged to adequately capture spontaneous daily mobility. However, these alternative intervals were not included in the final analysis and were therefore not systematically evaluated. Consequently, no definitive conclusions can be drawn regarding the optimal duration of inactivity intervals, as this was not the primary focus of the study. Nevertheless, our findings suggest that bout-based measures provide a feasible approach for assessing real-world daily activity. Future studies with larger cohorts and longer monitoring periods are needed to systematically evaluate different interval thresholds and to determine the most appropriate definition of real-world walking bouts. In contrast to the six-minute walking test, which requires rest without specifying a fixed duration [52, 53], the 15-minute inactivity threshold was pragmatically chosen to distinguish isolated walking from continuous activity. However, bout-detection algorithms in free-living accelerometry studies vary considerably [54], and the sensitivity of our results to this threshold was not formally evaluated. In conclusion, patient-estimated walking distance reflected EDSS-related disability only partially and may be unreliable around clinically relevant ambulation thresholds. Short-term variability in clinical walking distance tests further limits the interpretation of single snapshot assessments. Therefore, even if the maximum walking distance would be measured in everyday clinical practice despite the high time expenditure, the values would be of limited usefulness. Smartwatch-derived real-world mobility metrics, particularly bout-based measures such as peak-steps and longest streaks, offer a feasible complementary perspective on ambulatory function and may help contextualize EDSS-focused assessments in routine care and research. Nevertheless, smartwatches, such as those owned by patients, are of course not regulated and also have various differences. This makes it all the more important to always check all parameters in terms of quality. In our study setting, it was easy to track patient adherence over 7 days using the same smartwatch, and adherence was consistently high. Over a longer period, it may be advisable to check these parameters regularly to avoid snapshots and ensure continuous monitoring, as well as to take into account the different manufacturer standards. Integrating validated digital mobility parameters into routine care may improve detection of disease progression, offer the potential to reduce reliance on time-consuming walking tests, and enable a more differentiated, patient-centered assessment of disability in multiple sclerosis. Electronic Supplementary Material Supplementary Material 1. Supplementary Material 1.