Interpreting Patient-Reported Outcomes After Ischemic Stroke: Defining Meaningful Change in EQ-5D Across Recovery Phases

Interpreting Patient-Reported Outcomes After Ischemic Stroke: Defining Meaningful Change in EQ-5D Across Recovery Phases | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Interpreting Patient-Reported Outcomes After Ischemic Stroke: Defining Meaningful Change in EQ-5D Across Recovery Phases Pingping Li, Min Zhao, Yining Huang, Weiping Sun, Xuejing Jin, and 2 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7989766/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 21 Mar, 2026 Read the published version in Health and Quality of Life Outcomes → Version 1 posted 9 You are reading this latest preprint version Abstract Background: The EQ‑5D is widely applied to measure patient-reported outcomes, yet its minimally important difference (MID) has not been clearly established across distinct recovery phases after stroke. This study aimed to determine phase‑specific MIDs in EQ‑5D following stroke and to explore heterogeneity by estimation method, direction of change, and stroke etiology. Methods: A total of 9 978 adults with neuroimaging‑confirmed acute ischemic stroke were included in a prospective longitudinal cohort study. EQ‑5D and modified Rankin Scale (mRS) scores were recorded at admission (V1), hospital discharge (V2), 3-month (V3), and 1-year since admission (V4). Anchor-based MIDs were estimated at both group and individual levels and triangulated by distribution-based and instrument-defined approaches. Changes during the recovery phases (V1-V2, V2-V3, and V3-V4) were grouped into 3 categories: improved, no change, and deteriorated. Subgroup analyses were conducted according to the TOAST classification. Results: Phase-specific group-level MIDs for improvement decreased over time: anchor-based estimates were 0.19 at V2, 0.14 at V3, and 0.11 at V4, while deterioration MIDs were smaller (0.11 to 0.09). Distribution-based and instrument-defined estimates fluctuated slightly around the anchor-based values but followed a similar downward trend over time. Individual‑level analyses yielded valid MIDs only for improvement at V2 (0.10) and V3 (0.01). Cardioembolic strokes had higher MIDs than large-artery atherosclerosis and small-artery occlusion, while baseline utilities showed the reverse. Conclusions: This study provides the phase‑specific MIDs for utility measures after ischemic stroke, showing a declining trend from acute to chronic recovery and confirming robustness across multiple estimation methods. These MID values may assist in the interpretation of patient‑reported outcome changes related to different healthcare interventions in stroke patients at different recovery phases. EQ-5D Meaningful important difference Stroke Recovery phase Longitudinal study Figures Figure 1 Figure 2 Figure 3 1. Background Stroke remains the second leading cause of death and the third leading cause of disability globally(1). Ischemic stroke (IS) accounts for approximately 85% of all strokes and is associated with substantial mortality and economic burden (2). In China, the incidence and recurrence of stroke continue to rise over recent decades(3-5). Hospitalization cost for IS rose nearly 40% from 2010 to 2020(6), and rehabilitation now exceeds ¥63,000 per patient with limited insurance coverage(7). Stroke recovery has long been evaluated using clinician-reported instruments such as the National Institutes of Health Stroke Scale (NIHSS), Barthel Index (BI), and the modified Rankin Scale (mRS)(8). While these instruments effectively capture neurological deficits and functional status, they may not fully reflect patients’ own perceptions of recovery. The growing emphasis on patient-reported outcomes (PROs) has highlighted health utility measures such as the EuroQol five-dimensional questionnaire (EQ-5D), which are crucial for evaluating quality of life and health technology assessment (9-11). Interpreting the clinical relevance of observed changes for EQ-5D requires the minimally important difference (MID), defined as the smallest score change that patients perceive as meaningful(12). Although MIDs for EQ-5D have been established in various chronic conditions, evidence in stroke remains limited(13, 14). Existing studies are often constrained by small sample sizes, reliance on single anchors or one distribution-based index(15-17). Furthermore, no previous study has investigated how MIDs vary across distinct recovery phases or stroke etiological subtypes classified by the Trial of Org 10172 in Acute Stroke Treatment (TOAST). Given that health utility preferences are population-specific, MIDs for utility also depend on local population value sets(13, 18, 19). Therefore, this study aimed to estimate phase-specific MIDs for EQ-5D-3L in patients after acute ischemic stroke (AIS) using real-world, longitudinal registry data. Secondary objectives were to compare MID estimates across three commonly used methods, namely, anchor-based, distribution-based, and instrument-defined approaches; to evaluate MIDs by TOAST subtypes; and to assess MID using a credibility instrument specifically designed for anchor-based MID(20). 2. Methods 2.1 Study Design and Participants We used the data from the Chinese Acute Ischemic Stroke Treatment Outcome Registry (CASTOR; ClinicalTrials.gov Identifier: NCT02470624). CASTOR is a prospective, longitudinal, multicenter observational registry designed to evaluate outcomes among patients with AIS. Between March 2015 and December 2018, 40 tertiary hospitals across 19 provinces in China enrolled eligible patients. The protocol was approved by the institutional review board of the Peking University First Hospital (Approval No. 2015[922]) and corresponding ethics committees at each participating center before study initiation. All participants provided written informed consent prior to enrolment. Eligibility criteria (detailed in the published protocol(21)) required patients to be adults (≥18 years) with neuroimaging-confirmed AIS. 2.2 Measures and Scales Used Health-related quality of life (HRQOL) was assessed in this study using the EQ-5D-3L and functional outcome using the mRS at admission (V1), hospital discharge (V2), 90 ±14 days since admission (V3), and 360 ±28 days since admission (V4). In the present analysis, MIDs for EQ-5D-3L health utility were calculated separately for three study periods: V1-V2, V2-V3, and V3-V4. These phases were selected because they correspond to the acute, subacute and chronic phases of IS, which enhances clinical relevance. The EQ-5D-3L, a validated multi-attribute utility instrument to assess HRQOL, consists of two core components: a descriptive system measuring health state and a visual analogue scale (EQ-VAS)(22). The descriptive system includes five dimensions: mobility, self-care, usual activities, pain/discomfort, and anxiety/depression. Each dimension has 3 response categories (1 = no problems, 2 = some problems, 3 = extreme problems). Two EQ-5D-3L value sets have been published for China(23, 24). The 2014 version was used to calculate health utility in this study(24). The EQ-5D-3L was the target instrument for estimating MIDs, as it allows quantification of patients’ perceived changes in health status into utility values relevant for both clinical and economic evaluations. In stroke research, MIDs have been established for the BI and mRS, in addition to the NIHSS(17). As mRS was collected in the CASTOR study, it was selected as the anchor in the present analysis. The mRS is a clinician-reported outcome tool designed to evaluate global disability in stroke. It categorizes functional status into seven hierarchical grades (0-6), where 0 denotes no symptoms and 6 indicates death. The mRS has shown good reliability, validity, and responsiveness(8). Large-scale prospective studies have shown that a ≥1-point change reflects meaningful functional progression(25), a finding further corroborated by a 2017 Delphi consensus of 122 academic stroke neurologists(26), which established any single-grade shift to represent the MID in stroke populations 2.3 Statistical Analysis To ensure data quality and sample integrity, if the proportion of missing data for EQ-5D-3L health state items and mRS levels is less than 5% (as in this study), cases with missing values were directly excluded; if the missing rate falls between 5% and 30%, multiple imputation was applied. All analyses were performed in the R statistical programming language. 2.3.1 Interpretation of Meaningful Change The mRS served as the anchor for determining the MID of EQ-5D-3L in this study. The health state level transitions of mRS were classified as: (1) no change (non-responders), (2) minimal improvement or deterioration (responders, defined as a change of -1 or +1 grade)(26). Patients with either minimal change or no change were included in the anchor-based analysis. Four anchor-based statistical methods were employed to calculate the MID(27): (1) average change (AC), defined as the mean EQ-5D utility change of responders (within-group difference); (2) change difference (CD), the difference in the average change score of respondents and non-responders (between-group difference); (3) linear regression of EQ-5D change on mRS change, with the regression coefficient of the mRS shift interpreted as the MID; (4) receiver operating characteristic (ROC) curve analysis, which required an area under the curve (AUC) > 0.7 for acceptable discriminative ability, using EQ-5D score change as the test measure and mRS score change as the reference, with the threshold defined by the maximum Youden’s index. The ROC-based MID represents an individual-level threshold, whereas the other three methods provide group-level estimates. Instrument-defined method calculates the average health utility difference between baseline health state and single-level transitions to adjacent states, taking the specific clinical profile of IS into account(18). Participants with baseline profiles of 11111 were excluded, as no better health state was possible. We used the 0.2SD, 0.5SD, standard error of measurement (SEM), and 0.5 effect size (ES) in the distribution-based analysis. In addition, we calculated the minimal detectable change (MDC) which represents a statistical threshold for the smallest change detectable beyond measurement error(27, 28). We used the MDC to ascertain whether our MID estimates exceeded measurement error. Specifically, we calculated the ratio of MID to MDC at two confidence levels (90% and 95%), and at both individual and group levels. Ratios greater than one indicated that the MID exceeded measurement errors(29). Detailed calculation methods for all distribution‑based indices are provided in Supplementary Material A. Credibility refers to how well the study design and implementation minimize the risk of inaccurate MID estimates. Guyatt et al. (2020) developed a credibility assessment tool for anchor-based MID studies(20), which, along with its recent extension(30), provides a framework for evaluating methodological rigor. The tool includes five core criteria and four additional criteria for transition rating anchors. 2.3.2 Subgroup Analysis TOAST classification is internationally recognized as the preferred standard for etiological diagnosis of IS(2). This criterion categorizes the causes of AIS patients into five types: large-artery atherosclerosis (LAA), cardioembolic infarction (CE), small-artery occlusion (SA), stroke of other explicit etiology (SOE), and stroke of undetermined etiology (SUE). Subtype-level analyses were conducted because pathophysiology, baseline severity, and recovery trajectories vary across subtypes. The Kruskal-Wallis (K-W) test was used to compare whether baseline and follow-up scores of EQ-5D differed significantly across TOAST subgroups. For outcomes with significant post-hoc differences ( P <.05), subgroup-specific MIDs were calculated. In addition, a linear mixed effects model with random intercepts and random slopes was created. 3. Results 3.1 Sample Characteristics From 10 029 enrolled patients, 10,002 were eligible. The analytical cohorts included 9 978 (99.8%), 9 944 (99.4%), 9 162 (91.6%), and 8 396 (83.9%) patients at V1 to V4, respectively (Figure 1). Table 1 presents the baseline characteristics of the study population. Detailed characteristics for this cohort have been previously reported by Wang et al.(31). At admission, 47.1% of the Chinese patient cohort presented with mild stroke severity (NIHSS score 1-4)(31). Table 1. Baseline patient characteristics. Participants, No. (%) Characteristic V1 (N = 9978) V2 (N = 9944) V3 (N = 9162) Age, mean (SD), y 64.0 (11.9) 64.0 (11.9) 63.9 (11.9) Male sex 6564 (65.8) 6564 (65.8) 6564 (65.8) BMI, mean (SD) a 24.6 (3.40) 24.56 (3.42) 24.56 (3.42) TOAST classifications b LAA (n) 2890 (64.6) 2880 (64.6) 2731 (64.2) CE (n) 190 (4.2) 189 (4.2) 167 (3.9) SA (n) 1131 (25.3) 1129 (25.3) 1107 (26.0) Other determined cause (n) 105 (2.3) 105 (2.4) 100 (2.3) Undetermined cause (n) 159 (3.6) 158 (3.5) 151 (3.5) NIHSS Scores, mean (SD) 5.3 (4.98) 3.6(4.24) NA Treatments specific for AIS IV thrombolysis 859 (8.6) 859 (8.6) 801 (8.7) Antiplatelet 9536 (95.7) 9511 (95.6) 8764 (95.7) Anticoagulant 1460 (14.6) 1455 (14.6) 1318 (14.4) Volume expanders 1384 (13.9) 1384 (13.9) 1316 (14.4) Neuroprotective agents 9380 (94.1) 9355 (94.1) 8614 (94.0) Cerebral perfusion-enhancing agents 8718 (87.5) 8694 (87.4) 7999 (87.3) TCM 7591 (76.1) 7578 (76.2) 6992 (76.3) Others 348 (3.5) 348 (3.5) 321 (3.5) Abbreviations: TOAST, Trial of Org 10172 in Acute Stroke Treatment; LAA, large-artery atherosclerosis; CE, cardioembolic stroke; SA, small-artery occlusion; NIHSS, National Institution of Health Stroke Scale; NA, not applicable; AIS, acute ischemic stroke; IV, intravenous injection; TCM, traditional Chinese medicine. a BMI data available for 2487, 2481, 2318 patients for V1, V2, and V3 respectively. b TOAST classifications data available for 4475, 4461, and 4256 patients for V1, V2, and V3 respectively. 3.2 Longitudinal HRQOL Trends Significant improvements were observed in health utility (greatest from V1-V2; 1-year AUC: 0.802 [95% CI, 0.797 to 0.807]) and EQ-VAS (1-year AUC: 83.33 [95%CI, 83.02 to 83.64]) (both P <.001). Concurrently, moderate-to-severe disability (mRS 3-5) declined substantially from 43% to 15% (See Supplementary Table 1). Based on mRS-defined meaningful functional progress (See Supplementary Figure 1), the proportion of patients with functional improvement from V1 rose across visits: 47% at V1-V2, 61% at V1-V3, and 68% at V1-V4, while the proportion with unchanged status declined correspondingly from 46% to 31% and 25%. When analysed across three consecutive recovery phases, the proportion showing meaningful improvement decreased stepwise (V1-V2: 47%; V2-V3: 34%; V3-V4: 25%), whereas the proportion with unchanged status increased in parallel (46%, 57%, and 65%, respectively). Only a small proportion of patients (7-9%) demonstrated meaningful deterioration at any time. 3.3 MID EQ-5D-3L correlated strongly with mRS across visits (all r >0.75, P 0.33, P <.001; See Supplementary Figure 2, right panel). Although score changes were skewed (See Supplementary Figure 3, P < .001), the large sample size justified using the mean rather than the median to estimate the MID. Within each timepoint (V2, V3, V4), patients with minimal improvement had greater mean utility score changes than those with no meaningful change or minimal deterioration ( P <.001; Table 2, horizontal comparison). When compared across timepoints, significant differences were observed between the minimal improvement and no meaningful change groups, but not for the minimal deterioration group ( P <.001; Table 2, vertical comparison). Table 2. AC‑driven MIDs and proportion of patients above thresholds by recovery phases. Analysis type Timepoints Minimal improvement a No meaningful change a Minimal deterioration a ANCOVA P value Within-timepoint V2 from V1 0.27±0.19 0.08±0.15 -0.09±0.24 <.0001 3477 (35.0%) 5685 (57.2%) 782 (7.8%) V3 from V2 0.15±0.18 0.04±0.13 -0.08±0.17 <.0001 2454 (26.8%) 5660 (61.8%) 1048 (11.4%) V4 from V3 0.12±0.16 0.02±0.11 -0.09±0.16 <.0001 1917 (22.8%) 5544 (66.0%) 935 (11.1%) Across-timepoint V2 vs V3 vs V4 <.001 <.001 .954 Abbreviation: AC, average change; MID, minimally important difference; ANCOVA, analysis of covariance. a Minimal improvement: 1‑point decrease in mRS; no meaningful change: 0‑point change in mRS; minimal deterioration: 1‑point increase in mRS. Note. AC‑driven MID was defined as the mean change in EQ‑5D utility among responders. n (%) indicates the number (percentage) of patients whose utility change exceeded this mean change. See Supplementary Table 2 and Figure 2 summarise MID estimates across phases, with anchor‑based results as primary. The mean change for improvement in responders declined over time: acute phase 0.22 (95% CI, 0.21 to 0.22), subacute 0.15 (95% CI, 0.15 to 0.16), and chronic 0.12 (95% CI, 0.11 to 0.13). In contrast, for patients who experienced minimal clinically important deterioration, the MIDs did not differ significantly across the three recovery phases. Regression-derived MIDs approximated the mean change results. Linear regression diagnostics supported the model validity (See Supplementary Figure 4 and 5). Table 2 also shows the number of patients exceeding the AC‑driven MID. The proportion of patients exceeding the improvement threshold decreased across phases, with most showing no meaningful change and few deteriorating. To triangulate anchor‑based estimates, MIDs were also derived using distribution‑based and instrument‑defined approaches (Figure 2). Distribution‑based values (ranged 0.05-0.29) closely matched anchor‑based estimates (0.10-0.22). Instrument‑defined analyses yielded differences of 0.11 and 0.12 (See Supplementary Figure 6). ROC analyses provide individual‑level thresholds. However, because of the small number of patients with deterioration and the resulting imbalance between positive (deterioration) and negative (no change) cases (e.g., V2: 430 vs 4 570; See Supplementary Table 3), valid ROC estimates were obtained only for the improvement direction at V2 and V3 (AUC 0.70-0.72; Figure 3). These optimal cut‑points were 0.104 at V2 and 0.007 at V3. All group-level anchor-based MIDs were greater than the MDC 95 and MDC 90 (See Supplementary Figure 7). In contrast, the ROC‑derived MIDs were lower than the corresponding individual‑level MDC values. Sensitivity, specificity, and predictive values for both improvement and deterioration, together with the individual‑level MDCs, are showed in See Supplementary Table 4. The result of credibility assessment showed 4 out of 5 criteria fully met, confirming high trustworthiness of the established MIDs (rank of 2 out of 11; higher rank means higher credibility; See Supplementary Table 5). 3.4 Subtype-specific MIDs The K-W test revealed significant differences in baseline and follow-up health utility among TOAST subtypes ( P LAA > CE across Visits 2-4 ( P .05; See Supplementary Table 7). As shown in See Supplementary Figure 8, the baseline health utility distributions of SUE, SOE, and LAA overlapped. Based on prior analyses, mixed‑effects regression results (See Supplementary Table 8) and the heterogeneity observed within the SUE and SOE subtypes, SA, CE, and LAA were selected as subgroups for subtype‑specific MID estimation using anchor‑ and distribution‑based methods. For reporting, we focused on the overall direction of change. Across all visits, MIDs consistently followed the order of CE > LAA > SAO (See Supplementary Figure 9), a finding that was supported by the mixed-effects regression model (See Supplementary Table 8 and 9). 4. Discussion This study establishes, for the first time, phase-specific MIDs for the EQ-5D-3L after AIS. MIDs demonstrated a consistent decline from acute to chronic phases, showing an inverse relationship with baseline utility values. Subgroup analyses revealed the highest MIDs in cardioembolic stroke. These findings collectively indicate that patients with lower baseline health status required a greater magnitude of change to perceive a clinically meaningful improvement; this highlights the essential role of baseline utility in interpreting MIDs for both clinical and research applications(13). Consistent with prior MID research, our estimates demonstrated variability across directions of change(32, 33). Using anchor-based and distribution‑based methods, MID estimates for deterioration were smaller than those for improvement or overall changes (See Supplementary Table 2). Existing reviews of EQ-5D MID have generally reported smaller thresholds for deterioration than improvement(13, 34, 35), although some studies have identified the opposite(14, 32, 36). These differences may be attributable to patient sensitivity to health decline or variations among disease categories. The AC method quantifies within-group differences and may overestimate change due to natural recovery and other non-specific effects, thus representing a minimal important change (MIC)(37). In contrast, the CD method captures between-group differences by accounting for both responders and non-responders, providing a more conservative MID. In our study, AC-based values were consistently higher than CD-based values across all time points and directions of change, echoing the findings of Qin et al(37). The gap between AC and CD estimates narrowed over time, reflecting increasing utility gains in patients with unchanged mRS scores. Compared to previous MID of 0.10 at 3-4 weeks reported in a Taiwanese study(15), our estimate (CD: 0.14) at V2 (acute phase: median 13.0±8.5 days) was slightly larger, likely due to our population’s lower baseline utility (0.55 vs. 0.72), different value sets and anchor choice. Similarly, our MID for improvement over the 9-month interval (V3 to V4, 0.10) exceeded the 0.08 reported in a Korean study over a comparable 10-month period, with explanations including value set and our cohort’s higher proportion of patients with no baseline disability (36.2% vs. 12.5%); we also calculated a MID of 0.10 for the V1-V4 interval (11 months), which was approximately equal to that for V3-V4, although our primary analysis focused on the more clinically relevant comparison betweenV4 and V3. Our pooled anchor-based MIDs across all recovery phases (0.08-0.17) fell within the interquartile range (0.05-0.17) reported in a systematic review of EQ-5D-3L utility MIDs(13). The distribution-based (0.10-0.16) and instrument-defined estimates (0.11-0.12) in our study exceeded the review’s median pooled distribution-based value (0.08) and the upper range (0.08) of instrument-defined MIDs reported across four countries, likely reflecting our larger sample size and the use of a different value set, respectively. Furthermore, MID values in this study were pooled using mean values, in line with common practice. It should be noted, however, that methodologies for integrating MID estimates or selecting an optimal value within a single study remain undefined(38, 39), and future research could seek to identify pooled MID guided by clinical expertise. 4.1. Strengths and Limitations Strengths of this study include its large sample size and the phase‑specific assessment across acute, subacute, and chronic stages after AIS. This design better captures the heterogeneity of recovery potential and treatment strategies, with acute phase management focused on life‑saving and clinical stabilization through urgent time‑window interventions; the subacute phase emphasizing neurological recovery and rehabilitation training, supported by neurotrophic or neuroprotective agents; and the chronic phase prioritizing functional maintenance and secondary prevention of recurrence, alongside attention to psychological and social adaptation. In addition, we employed multiple methodological approaches to triangulate MID estimates and further tested their robustness and validity using bootstrap resampling, a credibility instrument, and MDC. Several limitations should be noted. First, we used EQ-5D-3L instead of the more sensitive 5L version, as the 5L was not available when the study was designed. Many patients with mRS changes already had one or more EQ-5D dimensions at an extreme level by V1-V3, consequently constraining the measurement of further change (See Supplementary Figure 10). Second, reliance on the mRS as a single anchor may be a limitation. Although multiple anchors are generally recommended, a single, well-validated anchor can be sufficient(40). The mRS was selected due to its established use in stroke MID studies and strong correlation with the EQ-5D, whereas the available EQ-VAS possesses weaker validity as an anchor(15, 17). 4.2. Implications The application of phase-specific MIDs should differ by recovery stage—a principle established in guidelines for assessing meaningful change in quality of life in cancer patients(41). Although derived from oncology, this principle holds equal weight in stroke populations, where distinguishing between early and late disease stages is essential. Specifically, acute‑phase values are most relevant for assessing short‑term treatment effects during admission, whereas subacute and chronic values are better suited for monitoring medium‑ to long‑term prognosis. We report MIDs at both group and individual levels. Group‑level estimates are most informative for trial design, such as using the AC‑derived MID of 0.15 in the subacute phase to interpret the mean change in scores in a real-world treatment group and to calculate sample size. CD‑based estimates provide an alternative approach that can be directly applied to between‑group comparisons. Individual‑level estimates may help classify responders, although ROC‑derived thresholds fell below corresponding MDC values and therefore carry some risk of misclassification. Depending on clinical priorities, stricter cut‑points may increase specificity but reduce sensitivity (See Supplementary Table 4). AC-, CD-, and regression-based results were generally consistent and reliable, while 0.5SD and SEM—which are the most widely used distribution-based metrics for MID estimation—produced values closest to anchor-based estimates. 5. Conclusion This study establishes phase-specific MIDs for the EQ-5D after IS. In clinical trials, change in EQ‑5D scores should be interpreted against these MIDs, with the most robust evidence obtained when changes are both statistically significant and clinically meaningful. The use of multiple methods to evaluate MID robustness in this analysis provides a reference for future research. Abbreviations AC: average change; AIS: acute ischemic stroke; AUC: area under the curve; CASTOR: Chinese Acute Ischemic Stroke Treatment Outcome Registry; CD: change difference; CE: cardioembolic infarction; EQ-5D-3L: EuroQol five-dimensional three-level questionnaire; EQ-VAS: EQ-5D visual analogue scale; HRQOL: health-related quality of life; LAA: large-artery atherosclerosis; MDC: minimal detectable change; MID: minimally important difference; mRS: modified Rankin Scale; NIHSS: National Institutes of Health Stroke Scale; ROC: receiver operating characteristic; SA: small-artery occlusion; SEM: standard error of measurement; SOE: stroke of other determined etiology; SUE: stroke of undetermined etiology; TOAST: Trial of Org 10172 in Acute Stroke Treatment. Declarations Ethics approval and consent to participate: This observational trial has been approved by the institutional review board of the Peking University First Hospital (approval No. 2015[922]) and that participants have signed written informed consent. Consent for publication: Not applicable Availability of data and materials: The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request. Competing interests: The authors declare that they have no competing interest. Funding: The authors declare that no funds, grants, or other support were received during the preparation of this manuscript. Authors’ contributions: PL: Conceptualization, Study Design, Data Analysis, Project Administration, Writing–Original Draft. XJ, FX, HL: Supervision, Methodology, Critical Revision of Manuscript; Joint corresponding authors and guarantors. YH, WS: Data Acquisition. MZ: Statistical Analysis. All authors contributed to data interpretation, critically revised the manuscript for important intellectual content, and approved the final submitted version. Acknowledgements: We thank LW and XG for their assistance with data cleaning. References Global, regional, and national burden of stroke and its risk factors, 1990-2019: a systematic analysis for the Global Burden of Disease Study 2019. The Lancet Neurology. 2021;20(10):795-820. Murphy SJ, Werring DJ. Stroke: causes and clinical features. Medicine (Abingdon, England : UK ed). 2020;48(9):561-6. Ma Q, Li R, Wang L, Yin P, Wang Y, Yan C, et al. Temporal trend and attributable risk factors of stroke burden in China, 1990-2019: an analysis for the Global Burden of Disease Study 2019. The Lancet Public health. 2021;6(12):e897-e906. Tu WJ, Zhao Z, Yin P, Cao L, Zeng J, Chen H, et al. Estimated Burden of Stroke in China in 2020. JAMA Netw Open. 2023;6(3):e231455. Xu J, Zhang X, Jin A, Pan Y, Li Z, Meng X, et al. Trends and Risk Factors Associated With Stroke Recurrence in China, 2007-2018. JAMA Netw Open. 2022;5(6):e2216341. 2022 WGoCSCR. Summary of China Stroke Center Report 2022. Chin J Cerebrovasc Dis. 2024;21(8):565-76. Boehringer Ingelheim I. White Paper on the Economic Burden of Stroke Rehabilitation Patients in China 2023: Yiyaomofang (ByDrug); 2023 [Available from: https://bydrug.pharmcube.com/news/detail/e6bb7c6237e3c0d28156cd1c3e56da1b. Akkara Y, Afreen R, Lemonick M, Paz SG, Rifi Z, Tosto J, et al. Standardizing Domains and Metrics of Stroke Recovery: A Systematic Review. Brain sciences. 2024;14(12). FDA. Guidance for industry: patient-reported outcome measures: use in medical product development to support labeling claims. 2024. Horsman J, Furlong W, Feeny D, Torrance G. The Health Utilities Index (HUI): concepts, measurement properties and applications. Health and quality of life outcomes. 2003;1:54. Guyatt GH, Feeny DH, Patrick DL. Measuring health-related quality of life. Ann Intern Med. 1993;118(8):622-9. Jaeschke R, Singer J, Guyatt GH. Measurement of health status. Ascertaining the minimal clinically important difference. Controlled clinical trials. 1989;10(4):407-15. Al Sayah F, Jin X, Short H, McClure NS, Ohinmaa A, Johnson JA. A Systematic Literature Review of Important and Meaningful Differences in the EQ-5D Index and Visual Analog Scale Scores. Value Health. 2025;28(3):470-6. Siriwardana AN, Hoffman AT, Morton RL, Smyth B, Brown MA. Estimating a Minimal Important Difference for the EQ-5D-5L Utility Index in Dialysis Patients. Value Health. 2024;27(4):469-77. Chen P, Lin KC, Liing RJ, Wu CY, Chen CL, Chang KC. Validity, responsiveness, and minimal clinically important difference of EQ-5D-5L in stroke patients undergoing rehabilitation. Qual Life Res. 2016;25(6):1585-96. Kim SK, Kim SH, Jo MW, Lee SI. Estimation of minimally important differences in the EQ-5D and SF-6D indices and their utility in stroke. Health and quality of life outcomes. 2015;13:32. Mishra B, Sudheer P, Agarwal A, Nilima N, Srivastava MVP, Vishnu VY. Minimal Clinically Important Difference of Scales Reported in Stroke Trials: A Review. Brain sciences. 2024;14(1). McClure NS, Sayah FA, Xie F, Luo N, Johnson JA. Instrument-Defined Estimates of the Minimally Important Difference for EQ-5D-5L Index Scores. Value Health. 2017;20(4):644-50. Jayadevappa R, Cook R, Chhatre S. Minimal important difference to infer changes in health-related quality of life-a systematic review. J Clin Epidemiol. 2017;89:188-98. Devji T, Carrasco-Labra A, Qasim A, Phillips M, Johnston BC, Devasenapathy N, et al. Evaluating the credibility of anchor based estimates of minimal important differences for patient reported outcomes: instrument development and reliability study. Bmj. 2020;369:m1714. Sun W, Ou Q, Zhang Z, Qu J, Huang Y. Chinese acute ischemic stroke treatment outcome registry (CASTOR): protocol for a prospective registry study on patterns of real-world treatment of acute ischemic stroke in China. BMC complementary and alternative medicine. 2017;17(1):357. EUROQOL. EQ-5D sample/demos https://euroqol.org/register/obtain-eq-5d/eq-5d-sample-demos/ [updated April 11th, 2025. Available from: https://euroqol.org/register/obtain-eq-5d/eq-5d-sample-demos/. Zhuo L, Xu L, Ye J, Sun S, Zhang Y, Burstrom K, et al. Time Trade-Off Value Set for EQ-5D-3L Based on a Nationally Representative Chinese Population Survey. Value Health. 2018;21(11):1330-7. Liu GG, Wu H, Li M, Gao C, Luo N. Chinese time trade-off values for EQ-5D health states. Value Health. 2014;17(5):597-604. Banks JL, Marotta CA. Outcomes validity and reliability of the modified Rankin scale: implications for stroke clinical trials: a literature review and synthesis. Stroke. 2007;38(3):1091-6. Cranston JS, Kaplan BD, Saver JL. Minimal Clinically Important Difference for Safe and Simple Novel Acute Ischemic Stroke Therapies. Stroke. 2017;48(11):2946-51. Mouelhi Y, Jouve E, Castelli C, Gentile S. How is the minimal clinically important difference established in health-related quality of life instruments? Review of anchors and methods. Health and quality of life outcomes. 2020;18(1):136. de Vet HC, Terwee CB. The minimal detectable change should not replace the minimal important difference. J Clin Epidemiol. 2010;63(7):804-5; author reply 6. Las Hayas C, Quintana JM, Padierna JA, Bilbao A, Muñoz P, Francis Cook E. Health-Related Quality of Life for Eating Disorders questionnaire version-2 was responsive 1-year after initial assessment. J Clin Epidemiol. 2007;60(8):825-33. Wang Y, Devji T, Carrasco-Labra A, Qasim A, Hao Q, Kum E, et al. An extension minimal important difference credibility item addressing construct proximity is a reliable alternative to the correlation item. J Clin Epidemiol. 2023;157:46-52. Wang L, Guan X, Zhou J, Hu H, Liu W, Wei Q, et al. Measuring the health outcomes of Chinese ischemic stroke patients based on the data from a longitudinal multi-center study. Qual Life Res. 2025;34(7):1967-77. McClure NS, Sayah FA, Ohinmaa A, Johnson JA. Minimally Important Difference of the EQ-5D-5L Index Score in Adults with Type 2 Diabetes. Value Health. 2018;21(9):1090-7. Franceschini M, Boffa A, Pignotti E, Andriolo L, Zaffagnini S, Filardo G. The Minimal Clinically Important Difference Changes Greatly Based on the Different Calculation Methods. The American journal of sports medicine. 2023;51(4):1067-73. Musoro JZ, Coens C, Sprangers MAG, Brandberg Y, Groenvold M, Flechtner HH, et al. Minimally important differences for interpreting EORTC QLQ-C30 change scores over time: A synthesis across 21 clinical trials involving nine different cancer types. European journal of cancer (Oxford, England : 1990). 2023;188:171-82. Bourke S, Bennett B, Oluboyede Y, Li T, Longworth L, O'Sullivan SB, et al. Estimating the minimally important difference for the EQ-5D-5L and EORTC QLQ-C30 in cancer. Health and quality of life outcomes. 2024;22(1):81. Cella D, Hahn EA, Dineen K. Meaningful change in cancer-specific quality of life scores: differences between improvement and worsening. Qual Life Res. 2002;11(3):207-21. Qin Z, Zhu Y, Shi DD, Chen R, Li S, Wu J. The gap between statistical and clinical significance: time to pay attention to clinical relevance in patient-reported outcome measures of insomnia. BMC medical research methodology. 2024;24(1):177. Trigg A, Griffiths P. Triangulation of multiple meaningful change thresholds for patient-reported outcome scores. Qual Life Res. 2021;30(10):2755-64. Oliveira A, Machado A, Marques A. Minimal Important and Detectable Differences of Respiratory Measures in Outpatients with AECOPD(†). Copd. 2018;15(5):479-88. Wang Y, Devji T, Qasim A, Hao Q, Wong V, Bhatt M, et al. A systematic survey identified methodological issues in studies estimating anchor-based minimal important differences in patient-reported outcomes. J Clin Epidemiol. 2022;142:144-51. Sprangers MA, Moinpour CM, Moynihan TJ, Patrick DL, Revicki DA. Assessing meaningful change in quality of life over time: a users' guide for clinicians. Mayo Clinic proceedings. 2002;77(6):561-71. Additional Declarations No competing interests reported. Supplementary Files Appendix1.MIDSupplementaryTablesandFigures.docx Additional file 1 MID Supplementary Tables and Figures Appendix2.eMethods.docx Additional file 2. eMethods Cite Share Download PDF Status: Published Journal Publication published 21 Mar, 2026 Read the published version in Health and Quality of Life Outcomes → Version 1 posted Editorial decision: Revision requested 14 Jan, 2026 Reviews received at journal 14 Dec, 2025 Reviews received at journal 30 Nov, 2025 Reviewers agreed at journal 16 Nov, 2025 Reviewers agreed at journal 16 Nov, 2025 Reviewers invited by journal 10 Nov, 2025 Editor assigned by journal 02 Nov, 2025 Submission checks completed at journal 02 Nov, 2025 First submitted to journal 30 Oct, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7989766","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":546863942,"identity":"89d8aeb0-02f0-4882-a76e-e660a9395b45","order_by":0,"name":"Pingping Li","email":"","orcid":"","institution":"China Pharmaceutical University","correspondingAuthor":false,"prefix":"","firstName":"Pingping","middleName":"","lastName":"Li","suffix":""},{"id":546863943,"identity":"185345c6-bbd4-44d3-8690-77720b0e9977","order_by":1,"name":"Min Zhao","email":"","orcid":"","institution":"China Pharmaceutical University","correspondingAuthor":false,"prefix":"","firstName":"Min","middleName":"","lastName":"Zhao","suffix":""},{"id":546863944,"identity":"5a4b9980-a2e8-4262-a839-65616ab482d4","order_by":2,"name":"Yining Huang","email":"","orcid":"","institution":"Peking University First Hospital","correspondingAuthor":false,"prefix":"","firstName":"Yining","middleName":"","lastName":"Huang","suffix":""},{"id":546863945,"identity":"8863be9f-c6d7-4841-80cc-9f4473e6ad98","order_by":3,"name":"Weiping Sun","email":"","orcid":"","institution":"Peking University First Hospital","correspondingAuthor":false,"prefix":"","firstName":"Weiping","middleName":"","lastName":"Sun","suffix":""},{"id":546863946,"identity":"c99fab05-d92b-4ec3-8dd5-f85824a99f16","order_by":4,"name":"Xuejing Jin","email":"","orcid":"","institution":"Beijing University of Chinese Medicine","correspondingAuthor":false,"prefix":"","firstName":"Xuejing","middleName":"","lastName":"Jin","suffix":""},{"id":546863947,"identity":"1d937364-a69f-471c-81bb-b06ea44d1bbc","order_by":5,"name":"Feng Xie","email":"","orcid":"","institution":"McMaster University","correspondingAuthor":false,"prefix":"","firstName":"Feng","middleName":"","lastName":"Xie","suffix":""},{"id":546863948,"identity":"7179f302-8973-4e45-b37f-2bdddc592ebb","order_by":6,"name":"Hongchao Li","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA3klEQVRIiWNgGAWjYBACPmYeMJ3AwMB8ACJ0gIAWNoQWtgQitTDAtfAYEKmFnffg44Jfdnn87T0fPxe2Mcjx3Uhg/FyA12F8ycYz+5KLJc6c3Sw9s43BWPJGArP0DPx+MZPm7TmQ2HAjdxszbxtD4oYbCTAPEtAy/0bOM5CWeuK08Pw4ADQ8hw2kJcGACC3GxrwNyYkbzxwzlp5xTsJw5pmHzdL4tPDznzF8zPPHLnHe8eaHnwvKbOT5jicf/IxPCxgwtkFoZgYGCRC3gZAGIPgD1zIKRsEoGAWjABMAAD7NRV9nkpyLAAAAAElFTkSuQmCC","orcid":"","institution":"China Pharmaceutical University","correspondingAuthor":true,"prefix":"","firstName":"Hongchao","middleName":"","lastName":"Li","suffix":""}],"badges":[],"createdAt":"2025-10-30 13:08:37","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7989766/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7989766/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1186/s12955-026-02524-w","type":"published","date":"2026-03-21T15:57:51+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":96364098,"identity":"3dcf1add-b0cb-463d-abc8-f5329d15af08","added_by":"auto","created_at":"2025-11-20 10:08:53","extension":"tiff","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":508020,"visible":true,"origin":"","legend":"","description":"","filename":"Figure2.tiff","url":"https://assets-eu.researchsquare.com/files/rs-7989766/v1/f204cb2e015371c471f9b5e8.tiff"},{"id":96286737,"identity":"ee4bf0ed-968b-494c-8fd6-52075c7206bf","added_by":"auto","created_at":"2025-11-19 12:05:50","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":162604,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.docx","url":"https://assets-eu.researchsquare.com/files/rs-7989766/v1/6c419dc507cafbf164442ef7.docx"},{"id":96286742,"identity":"1c0844ae-5f6e-43f2-82f5-2a042d8b9c75","added_by":"auto","created_at":"2025-11-19 12:05:50","extension":"tiff","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":2667960,"visible":true,"origin":"","legend":"","description":"","filename":"Figure3.tiff","url":"https://assets-eu.researchsquare.com/files/rs-7989766/v1/130c449eab869984c266ff74.tiff"},{"id":96286743,"identity":"96ae614b-e4bc-4656-a6f1-bc7ce11eb26b","added_by":"auto","created_at":"2025-11-19 12:05:51","extension":"tif","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1826044,"visible":true,"origin":"","legend":"","description":"","filename":"Figure1.tif","url":"https://assets-eu.researchsquare.com/files/rs-7989766/v1/a5c2f231a6ca3e9bd43564f8.tif"},{"id":96286735,"identity":"a4364a42-ba49-4b03-bb9e-d30a950162c5","added_by":"auto","created_at":"2025-11-19 12:05:50","extension":"json","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":8772,"visible":true,"origin":"","legend":"","description":"","filename":"30c3325effc24f14ab41c1bb19b53611.json","url":"https://assets-eu.researchsquare.com/files/rs-7989766/v1/bd19bf7dbc9666152a46ed60.json"},{"id":96286751,"identity":"4757b329-df24-4989-9234-0c101739e295","added_by":"auto","created_at":"2025-11-19 12:05:51","extension":"docx","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":58455738,"visible":true,"origin":"","legend":"","description":"","filename":"Appendix1.MIDSupplementaryTablesandFigures.docx","url":"https://assets-eu.researchsquare.com/files/rs-7989766/v1/a0b2c548357a47150e970b73.docx"},{"id":96286740,"identity":"1a072212-d30d-4b95-8ea6-5b6530010e06","added_by":"auto","created_at":"2025-11-19 12:05:50","extension":"docx","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":47069,"visible":true,"origin":"","legend":"","description":"","filename":"Appendix2.eMethods.docx","url":"https://assets-eu.researchsquare.com/files/rs-7989766/v1/0a7fb96bd293770b4c218a1b.docx"},{"id":96286747,"identity":"d9075766-a80c-4ced-940c-30edc72a6958","added_by":"auto","created_at":"2025-11-19 12:05:51","extension":"tiff","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":508020,"visible":true,"origin":"","legend":"","description":"","filename":"Figure2.tiff","url":"https://assets-eu.researchsquare.com/files/rs-7989766/v1/6129dc27a43d9213d99c9f46.tiff"},{"id":96286750,"identity":"535f6632-575d-45ce-9457-42a346ad367f","added_by":"auto","created_at":"2025-11-19 12:05:51","extension":"tiff","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":2667960,"visible":true,"origin":"","legend":"","description":"","filename":"Figure3.tiff","url":"https://assets-eu.researchsquare.com/files/rs-7989766/v1/bd3a3d5a89d3685d8bf2c07c.tiff"},{"id":96286749,"identity":"cca96b89-d12c-416e-ae2f-f7be67fb31e2","added_by":"auto","created_at":"2025-11-19 12:05:51","extension":"tif","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1826044,"visible":true,"origin":"","legend":"","description":"","filename":"Figure1.tif","url":"https://assets-eu.researchsquare.com/files/rs-7989766/v1/69e224e8f01b56e18ead5271.tif"},{"id":96286746,"identity":"b7f70567-e0be-4a19-9af8-0e54eec64e58","added_by":"auto","created_at":"2025-11-19 12:05:51","extension":"png","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":77623,"visible":true,"origin":"","legend":"","description":"","filename":"OnlineFigure2.png","url":"https://assets-eu.researchsquare.com/files/rs-7989766/v1/937f5f04bfcfa121dbce9a44.png"},{"id":96286744,"identity":"41c00f72-269b-493d-8d10-bd46f468a3a1","added_by":"auto","created_at":"2025-11-19 12:05:51","extension":"png","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":21738,"visible":true,"origin":"","legend":"","description":"","filename":"OnlineFigure3.png","url":"https://assets-eu.researchsquare.com/files/rs-7989766/v1/b9d368ca531dba190295727a.png"},{"id":96286748,"identity":"7e6b3372-afcd-484f-91af-563525242aaa","added_by":"auto","created_at":"2025-11-19 12:05:51","extension":"png","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":283291,"visible":true,"origin":"","legend":"","description":"","filename":"OnlineFigure1.png","url":"https://assets-eu.researchsquare.com/files/rs-7989766/v1/95ff55e8da75cd054c6452f9.png"},{"id":96286734,"identity":"15c47ba8-fe8f-4e24-8239-289655577d92","added_by":"auto","created_at":"2025-11-19 12:05:50","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":672157,"visible":true,"origin":"","legend":"\u003cp\u003eFlow chart of patients\u003c/p\u003e","description":"","filename":"Figure1.png","url":"https://assets-eu.researchsquare.com/files/rs-7989766/v1/ea0d755aa8cf8d20eff00d84.png"},{"id":96364810,"identity":"d45a6ec0-e30d-45e5-8e06-5763b8d02be5","added_by":"auto","created_at":"2025-11-20 10:09:40","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":85828,"visible":true,"origin":"","legend":"\u003cp\u003eDistribution of minimal importance differences (MIDs) across three recovery phases.\u003c/p\u003e","description":"","filename":"Figure2.png","url":"https://assets-eu.researchsquare.com/files/rs-7989766/v1/14189365956169c5c4b1454c.png"},{"id":96364797,"identity":"6bf8147f-c767-4329-b804-ef190f1bca91","added_by":"auto","created_at":"2025-11-20 10:09:39","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":40295,"visible":true,"origin":"","legend":"\u003cp\u003eReceiver operating characteristic (ROC) curves for improvement.\u003c/p\u003e","description":"","filename":"Figure3.png","url":"https://assets-eu.researchsquare.com/files/rs-7989766/v1/36f67c8cb9693670ebc26b67.png"},{"id":105223306,"identity":"4ffc9e7e-3e73-47cc-9567-fe16b06157f9","added_by":"auto","created_at":"2026-03-23 16:03:28","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1490286,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7989766/v1/fd7bd41c-7a0b-4031-8925-0fcab5f498bc.pdf"},{"id":96286753,"identity":"d8369b95-08f7-4a38-942d-cf8a74060ba4","added_by":"auto","created_at":"2025-11-19 12:05:53","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":58455738,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eAdditional file 1 \u003c/strong\u003eMID Supplementary Tables and Figures\u003c/p\u003e","description":"","filename":"Appendix1.MIDSupplementaryTablesandFigures.docx","url":"https://assets-eu.researchsquare.com/files/rs-7989766/v1/7c7e8d8c58b4e27999ec1a18.docx"},{"id":96363482,"identity":"6b54d329-c0f2-4ad0-80ad-afd3cdcc14a6","added_by":"auto","created_at":"2025-11-20 10:07:07","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":47069,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eAdditional file 2. \u003c/strong\u003eeMethods\u003c/p\u003e","description":"","filename":"Appendix2.eMethods.docx","url":"https://assets-eu.researchsquare.com/files/rs-7989766/v1/9fa2c3daa44e3e2a6342031f.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Interpreting Patient-Reported Outcomes After Ischemic Stroke: Defining Meaningful Change in EQ-5D Across Recovery Phases","fulltext":[{"header":"1. Background","content":"\u003cp\u003eStroke remains the second leading cause of death and the third leading cause of disability globally(1). Ischemic stroke (IS) accounts for approximately 85% of all strokes and is associated with substantial mortality and economic burden (2). In China, the incidence and recurrence of stroke continue to rise over recent decades(3-5). Hospitalization cost for IS rose nearly 40% from 2010 to 2020(6), and rehabilitation now exceeds ¥63,000 per patient with limited insurance coverage(7). Stroke recovery has long been evaluated using clinician-reported instruments such as the National Institutes of Health Stroke Scale (NIHSS), Barthel Index (BI), and the modified Rankin Scale (mRS)(8). While these instruments effectively capture neurological deficits and functional status, they may not fully reflect patients’ own perceptions of recovery. The growing emphasis on patient-reported outcomes (PROs) has highlighted health utility measures such as the EuroQol five-dimensional questionnaire (EQ-5D), which are crucial for evaluating quality of life and health technology assessment (9-11). Interpreting the clinical relevance of observed changes for EQ-5D requires the minimally important difference (MID), defined as the smallest score change that patients perceive as meaningful(12).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAlthough MIDs for EQ-5D have been established in various chronic conditions, evidence in stroke remains limited(13, 14). Existing studies are often constrained by small sample sizes, reliance on single anchors or one distribution-based index(15-17). Furthermore, no previous study has investigated how MIDs vary across distinct recovery phases or stroke etiological subtypes classified by the Trial of Org 10172 in Acute Stroke Treatment (TOAST). Given that health utility preferences are population-specific, MIDs for utility also depend on local population value sets(13, 18, 19).\u003c/p\u003e\n\u003cp\u003eTherefore, this study aimed to estimate phase-specific MIDs for EQ-5D-3L in patients after acute ischemic stroke (AIS) using real-world, longitudinal registry data. Secondary objectives were to compare MID estimates across three commonly used methods, namely, anchor-based, distribution-based, and instrument-defined approaches; to evaluate MIDs by TOAST subtypes; and to assess MID using a credibility instrument specifically designed for anchor-based MID(20).\u003c/p\u003e"},{"header":"2. Methods","content":"\u003ch2\u003e2.1 Study Design and Participants\u003c/h2\u003e\n\u003cp\u003eWe used the data from the Chinese Acute Ischemic Stroke Treatment Outcome Registry (CASTOR; ClinicalTrials.gov Identifier: NCT02470624). CASTOR is a prospective, longitudinal, multicenter observational registry designed to evaluate outcomes among patients with AIS. Between March 2015 and December 2018, 40 tertiary hospitals across 19 provinces in China enrolled eligible patients. The protocol was approved by the institutional review board of the Peking University First Hospital (Approval No. 2015[922]) and corresponding ethics committees at each participating center before study initiation. All participants provided written informed consent prior to enrolment. Eligibility criteria (detailed in the published protocol(21)) required patients to be adults (≥18 years) with neuroimaging-confirmed AIS.\u003c/p\u003e\n\u003ch2\u003e2.2 Measures and Scales Used\u003c/h2\u003e\n\u003cp\u003eHealth-related quality of life (HRQOL) was assessed in this study using the EQ-5D-3L and functional outcome using the mRS at admission (V1), hospital discharge (V2), 90 ±14 days since admission (V3), and 360 ±28 days since admission (V4). In the present analysis, MIDs for EQ-5D-3L health utility were calculated separately for three study periods: V1-V2, V2-V3, and V3-V4. These phases were selected because they correspond to the acute, subacute and chronic phases of IS, which enhances clinical relevance.\u003c/p\u003e\n\u003cp\u003eThe EQ-5D-3L, a validated multi-attribute utility instrument to assess HRQOL, consists of two core components: a descriptive system measuring health state and a visual analogue scale (EQ-VAS)(22). The descriptive system includes five dimensions: mobility, self-care, usual activities, pain/discomfort, and anxiety/depression. Each dimension has 3 response categories (1 = no problems, 2 = some problems, 3 = extreme problems). Two EQ-5D-3L value sets have been published for China(23, 24). The 2014 version was used to calculate health utility in this study(24). The EQ-5D-3L was the target instrument for estimating MIDs, as it allows quantification of patients’ perceived changes in health status into utility values relevant for both clinical and economic evaluations.\u003c/p\u003e\n\u003cp\u003eIn stroke research, MIDs have been established for the BI and mRS, in addition to the NIHSS(17). As mRS was collected in the CASTOR study, it was selected as the anchor in the present analysis. The mRS is a clinician-reported outcome tool designed to evaluate global disability in stroke. It categorizes functional status into seven hierarchical grades (0-6), where 0 denotes no symptoms and 6 indicates death. The mRS has shown good reliability, validity, and responsiveness(8). Large-scale prospective studies have shown that a ≥1-point change reflects meaningful functional progression(25), a finding further corroborated by a 2017 Delphi consensus of 122 academic stroke neurologists(26), which established any single-grade shift to represent the MID in stroke populations\u003c/p\u003e\n\u003ch2\u003e2.3 Statistical Analysis\u003c/h2\u003e\n\u003cp\u003eTo ensure data quality and sample integrity, if the proportion of missing data for EQ-5D-3L health state items and mRS levels is less than 5% (as in this study), cases with missing values were directly excluded; if the missing rate falls between 5% and 30%, multiple imputation was applied. All analyses were performed in the R statistical programming language.\u003c/p\u003e\n\u003ch3\u003e2.3.1 Interpretation of Meaningful Change\u003c/h3\u003e\n\u003cp\u003eThe\u0026nbsp;mRS\u0026nbsp;served as the anchor for determining the MID of EQ-5D-3L in this study. The health state level transitions of\u0026nbsp;mRS\u0026nbsp;were classified as: (1) no change (non-responders), (2) minimal improvement or deterioration (responders, defined as a change of -1 or +1 grade)(26). Patients with either minimal change or no change were included in the anchor-based analysis. Four anchor-based statistical methods were employed to calculate the MID(27): (1) average change (AC), defined as the mean EQ-5D utility change of responders (within-group difference); (2) change difference (CD), the difference in the average change score of respondents and non-responders (between-group difference); (3) linear regression of EQ-5D change on mRS change, with the regression coefficient of the mRS shift interpreted as the MID; (4) receiver operating characteristic (ROC) curve analysis, which required an area under the curve (AUC) \u0026gt; 0.7 for acceptable discriminative ability, using EQ-5D score change as the test measure and mRS score change as the reference, with the threshold defined by the maximum Youden’s index. The ROC-based MID represents an individual-level threshold, whereas the other three methods provide group-level estimates.\u003c/p\u003e\n\u003cp\u003eInstrument-defined method calculates the average health utility difference between baseline health state and single-level transitions to adjacent states, taking the specific clinical profile of IS into account(18). Participants with baseline profiles of 11111 were excluded, as no better health state was possible.\u003c/p\u003e\n\u003cp\u003eWe used the 0.2SD, 0.5SD, standard error of measurement (SEM), and 0.5 effect size (ES) in the distribution-based analysis. In addition, we calculated the minimal detectable change (MDC) which represents a statistical threshold for the smallest change detectable beyond measurement error(27, 28). We used the MDC to ascertain whether our MID estimates exceeded measurement error. Specifically, we calculated the ratio of MID to MDC at two confidence levels (90% and 95%), and at both individual and group levels. Ratios greater than one indicated that the MID exceeded measurement errors(29). Detailed calculation methods for all distribution‑based indices are provided in Supplementary Material A.\u003c/p\u003e\n\u003cp\u003eCredibility refers to how well the study design and implementation minimize the risk of inaccurate MID estimates. Guyatt et al. (2020) developed a credibility assessment tool for anchor-based MID studies(20), which, along with its recent extension(30), provides a framework for evaluating methodological rigor. The tool includes five core criteria and four additional criteria for transition rating anchors.\u003c/p\u003e\n\u003ch3\u003e2.3.2 Subgroup Analysis\u003c/h3\u003e\n\u003cp\u003eTOAST classification is internationally recognized as the preferred standard for etiological diagnosis of IS(2). This criterion categorizes the causes of AIS patients into five types: large-artery atherosclerosis (LAA), cardioembolic infarction (CE), small-artery occlusion (SA), stroke of other explicit etiology (SOE), and stroke of undetermined etiology (SUE). Subtype-level analyses were conducted because pathophysiology, baseline severity, and recovery trajectories vary across subtypes. The Kruskal-Wallis (K-W) test was used to compare whether baseline and follow-up scores of EQ-5D differed significantly across TOAST subgroups. For outcomes with significant post-hoc differences (\u003cem\u003eP\u003c/em\u003e\u0026lt;.05), subgroup-specific MIDs were calculated. In addition, a linear mixed effects model with random intercepts and random slopes was created.\u003c/p\u003e"},{"header":"3. Results","content":"\u003ch2\u003e3.1 Sample Characteristics\u003c/h2\u003e\n\u003cp\u003eFrom 10 029 enrolled patients, 10,002 were eligible. The analytical cohorts included 9 978 (99.8%), 9 944 (99.4%), 9 162 (91.6%), and 8 396 (83.9%) patients at V1 to V4, respectively (Figure 1). Table 1 presents the baseline characteristics of the study population. Detailed characteristics for this cohort have been previously reported by Wang et al.(31). At admission, 47.1% of the Chinese patient cohort presented with mild stroke severity (NIHSS score 1-4)(31).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 1.\u003c/strong\u003e Baseline patient characteristics.\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\" width=\"98%\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd colspan=\"3\"\u003e\n \u003cp\u003e\u003cstrong\u003eParticipants, No. (%)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eCharacteristic\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eV1 (N = 9978)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eV2 (N = 9944)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eV3 (N = 9162)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eAge, mean (SD), y\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e64.0 (11.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e64.0 (11.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e63.9 (11.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eMale sex\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e6564 (65.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e6564 (65.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e6564 (65.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eBMI, mean (SD)\u003csup\u003ea\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e24.6 (3.40)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e24.56 (3.42)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e24.56 (3.42)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eTOAST classifications\u003csup\u003eb\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eLAA (n)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e\u003cstrong\u003e2890 (64.6)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e2880 (64.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e2731 (64.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eCE (n)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e\u003cstrong\u003e190 (4.2)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e189 (4.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e167 (3.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eSA (n)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e\u003cstrong\u003e1131 (25.3)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e1129 (25.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e1107 (26.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eOther determined cause (n)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e\u003cstrong\u003e105 (2.3)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e105 (2.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e100 (2.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eUndetermined cause (n)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e\u003cstrong\u003e159 (3.6)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e158 (3.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e151 (3.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eNIHSS Scores, mean (SD)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e5.3 (4.98)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e3.6(4.24)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eNA\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eTreatments specific for AIS\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eIV thrombolysis\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e\u003cstrong\u003e859 (8.6)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e859 (8.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e801 (8.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eAntiplatelet\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e\u003cstrong\u003e9536 (95.7)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e9511 (95.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e8764 (95.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eAnticoagulant\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e\u003cstrong\u003e1460 (14.6)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e1455 (14.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e1318 (14.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eVolume expanders\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e\u003cstrong\u003e1384 (13.9)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e1384 (13.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e1316 (14.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eNeuroprotective agents\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e\u003cstrong\u003e9380 (94.1)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e9355 (94.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e8614 (94.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eCerebral perfusion-enhancing agents\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e\u003cstrong\u003e8718 (87.5)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e8694 (87.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e7999 (87.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eTCM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e\u003cstrong\u003e7591 (76.1)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e7578 (76.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e6992 (76.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eOthers\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e\u003cstrong\u003e348 (3.5)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e348 (3.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e321 (3.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"4\"\u003e\n \u003cp\u003eAbbreviations: TOAST, Trial of Org 10172 in Acute Stroke Treatment; LAA, large-artery atherosclerosis; CE, cardioembolic stroke; SA, small-artery occlusion; NIHSS, National Institution of Health Stroke Scale; NA, not applicable; AIS, acute ischemic stroke; IV, intravenous injection; TCM, traditional Chinese medicine.\u003c/p\u003e\n \u003cp\u003e\u003csup\u003ea\u003c/sup\u003eBMI data available for 2487, 2481, 2318 patients for V1, V2, and V3 respectively.\u003c/p\u003e\n \u003cp\u003e\u003csup\u003eb\u003c/sup\u003e TOAST classifications data available for 4475, 4461, and 4256 patients for V1, V2, and V3 respectively.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003ch2\u003e3.2 Longitudinal HRQOL Trends\u003c/h2\u003e\n\u003cp\u003eSignificant improvements were observed in health utility (greatest from V1-V2; 1-year AUC: 0.802 [95% CI, 0.797 to 0.807]) and EQ-VAS (1-year AUC: 83.33 [95%CI, 83.02 to 83.64]) (both \u003cem\u003eP\u003c/em\u003e\u0026lt;.001). Concurrently, moderate-to-severe disability (mRS 3-5) declined substantially from 43% to 15% (See Supplementary Table 1).\u0026nbsp;Based on mRS-defined meaningful functional progress (See Supplementary Figure\u0026nbsp;1), the proportion of patients with functional improvement from V1 rose across visits: 47% at V1-V2, 61% at V1-V3, and 68% at V1-V4, while the proportion with unchanged status declined correspondingly from 46% to 31% and 25%. When analysed across three consecutive recovery phases, the proportion showing meaningful improvement decreased stepwise (V1-V2: 47%; V2-V3: 34%; V3-V4: 25%), whereas the proportion with unchanged status increased in parallel (46%, 57%, and 65%, respectively). Only a small proportion of patients (7-9%) demonstrated meaningful deterioration at any time.\u003c/p\u003e\n\u003ch2\u003e3.3 MID\u003c/h2\u003e\n\u003cp\u003eEQ-5D-3L correlated strongly with mRS across visits (all r \u0026gt;0.75, \u003cem\u003eP\u003c/em\u003e\u0026lt;.001; See Supplementary Figure 2, left panel). For score changes, correlations were moderate to high (all r \u0026gt;0.33, \u003cem\u003eP\u003c/em\u003e\u0026lt;.001; See Supplementary Figure 2, right panel). Although score changes were skewed (See Supplementary Figure 3, \u003cem\u003eP\u003c/em\u003e \u0026lt; .001), the large sample size justified using the mean rather than the median to estimate the MID. Within each timepoint (V2, V3, V4), patients with minimal improvement had greater mean utility score changes than those with no meaningful change or minimal deterioration (\u003cem\u003eP\u003c/em\u003e\u0026lt;.001; Table 2, horizontal comparison). When compared across timepoints, significant differences were observed between the minimal improvement and no meaningful change groups, but not for the minimal deterioration group (\u003cem\u003eP\u003c/em\u003e\u0026lt;.001; Table 2, vertical comparison).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 2.\u0026nbsp;\u003c/strong\u003eAC‑driven MIDs and proportion of patients above thresholds by recovery phases.\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\" width=\"100%\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eAnalysis type\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eTimepoints\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eMinimal improvement \u003csup\u003ea\u003c/sup\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eNo meaningful change\u003csup\u003e\u0026nbsp;a\u003c/sup\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eMinimal deterioration\u003csup\u003e\u0026nbsp;a\u003c/sup\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eANCOVA\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eP\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;value\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"6\"\u003e\n \u003cp\u003eWithin-timepoint\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\"\u003e\n \u003cp\u003eV2 from V1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e0.27\u0026plusmn;0.19\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e0.08\u0026plusmn;0.15\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e-0.09\u0026plusmn;0.24\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\"\u003e\n \u003cp\u003e\u0026lt;.0001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e3477 (35.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e5685 (57.2%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e782 (7.8%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\"\u003e\n \u003cp\u003eV3 from V2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e0.15\u0026plusmn;0.18\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e0.04\u0026plusmn;0.13\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e-0.08\u0026plusmn;0.17\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\"\u003e\n \u003cp\u003e\u0026lt;.0001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e2454 (26.8%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e5660 (61.8%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e1048 (11.4%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\"\u003e\n \u003cp\u003eV4 from V3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e0.12\u0026plusmn;0.16\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e0.02\u0026plusmn;0.11\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e-0.09\u0026plusmn;0.16\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\"\u003e\n \u003cp\u003e\u0026lt;.0001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e1917 (22.8%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e5544 (66.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e935 (11.1%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eAcross-timepoint\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eV2 vs V3 vs V4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026lt;.001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u0026lt;.001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e.954\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"6\"\u003e\n \u003cp\u003eAbbreviation: AC, average change; MID, minimally important difference; ANCOVA, analysis of covariance.\u003c/p\u003e\n \u003cp\u003e\u003csup\u003ea\u0026nbsp;\u003c/sup\u003eMinimal improvement: 1‑point decrease in mRS; no meaningful change: 0‑point change in mRS; minimal deterioration: 1‑point increase in mRS.\u003c/p\u003e\n \u003cp\u003e\u003cem\u003eNote.\u003c/em\u003e AC‑driven MID was defined as the mean change in EQ‑5D utility among responders. \u003cem\u003en\u003c/em\u003e (%) indicates the number (percentage) of patients whose utility change exceeded this mean change.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003eSee Supplementary Table 2 and Figure 2 summarise MID estimates across phases, with anchor‑based results as primary. The mean change for improvement in responders declined over time: acute phase 0.22 (95% CI, 0.21 to 0.22), subacute 0.15 (95% CI, 0.15 to 0.16), and chronic 0.12 (95% CI, 0.11 to 0.13). In contrast, for patients who experienced minimal clinically important deterioration, the MIDs did not differ significantly across the three recovery phases. Regression-derived MIDs approximated the mean change results. Linear regression diagnostics supported the model validity (See Supplementary Figure 4 and 5). Table 2 also shows the number of patients exceeding the AC‑driven MID. The proportion of patients exceeding the improvement threshold decreased across phases, with most showing no meaningful change and few deteriorating. To triangulate anchor‑based estimates, MIDs were also derived using distribution‑based and instrument‑defined approaches (Figure 2). Distribution‑based values (ranged 0.05-0.29) closely matched anchor‑based estimates (0.10-0.22). Instrument‑defined analyses yielded differences of 0.11 and 0.12 (See Supplementary Figure 6).\u003c/p\u003e\n\u003cp\u003eROC analyses provide individual‑level thresholds. However, because of the small number of patients with deterioration and the resulting imbalance between positive (deterioration) and negative (no change) cases (e.g., V2: 430 vs 4 570; See Supplementary Table 3), valid ROC estimates were obtained only for the improvement direction at V2 and V3 (AUC 0.70-0.72; Figure 3). These optimal cut‑points were 0.104 at V2 and 0.007 at V3.\u003c/p\u003e\n\u003cp\u003eAll group-level anchor-based MIDs were greater than the MDC\u003csub\u003e95\u003c/sub\u003e and MDC\u003csub\u003e90\u0026nbsp;\u003c/sub\u003e(See Supplementary Figure 7). In contrast, the ROC‑derived MIDs were lower than the corresponding individual‑level MDC values. Sensitivity, specificity, and predictive values for both improvement and deterioration, together with the individual‑level MDCs, are showed in See Supplementary Table 4. The result of credibility assessment showed 4 out of 5 criteria fully met, confirming high trustworthiness of the established MIDs (rank of 2 out of 11; higher rank means higher credibility; See Supplementary Table 5).\u003c/p\u003e\n\u003ch2\u003e3.4 Subtype-specific MIDs\u003c/h2\u003e\n\u003cp\u003eThe K-W test revealed significant differences in baseline and follow-up health utility among TOAST subtypes (\u003cem\u003eP\u003c/em\u003e\u0026lt;.001; See Supplementary Table 6). Post-hoc analysis consistently showed SA \u0026gt; LAA \u0026gt; CE across Visits 2-4 (\u003cem\u003eP\u003c/em\u003e\u0026lt;.001; See Supplementary Table 7). Comparisons between SUE and SOE, and between these subtypes and SA, LAA, or CE, were mostly non‑significant (\u003cem\u003eP\u003c/em\u003e\u0026gt;.05; See Supplementary Table 7). As shown in See Supplementary Figure 8, the baseline health utility distributions of SUE, SOE, and LAA overlapped. Based on prior analyses, mixed‑effects regression results (See Supplementary Table 8) and the heterogeneity observed within the SUE and SOE subtypes, SA, CE, and LAA were selected as subgroups for subtype‑specific MID estimation using anchor‑ and distribution‑based methods. For reporting, we focused on the overall direction of change. Across all visits, MIDs consistently followed the order of CE \u0026gt; LAA \u0026gt; SAO (See Supplementary Figure 9), a finding that was supported by the mixed-effects regression model (See Supplementary Table 8 and 9).\u003c/p\u003e"},{"header":"4. Discussion","content":"\u003cp\u003eThis study establishes, for the first time, phase-specific MIDs for the EQ-5D-3L after AIS. MIDs demonstrated a consistent decline from acute to chronic phases, showing an inverse relationship with baseline utility values. Subgroup analyses revealed the highest MIDs in cardioembolic stroke. These findings collectively indicate that patients with lower baseline health status required a greater magnitude of change to perceive a clinically meaningful improvement; this highlights the essential role of baseline utility in interpreting MIDs for both clinical and research applications(13).\u003c/p\u003e\n\u003cp\u003eConsistent with prior MID research, our estimates demonstrated variability across directions of change(32, 33). Using anchor-based and distribution‑based methods, MID estimates for deterioration were smaller than those for improvement or overall changes (See Supplementary Table 2). Existing reviews of EQ-5D MID have generally reported smaller thresholds for deterioration than improvement(13, 34, 35), although some studies have identified the opposite(14, 32, 36). These differences may be attributable to patient sensitivity to health decline or variations among disease categories.\u003c/p\u003e\n\u003cp\u003eThe AC method quantifies within-group differences and may overestimate change due to natural recovery and other non-specific effects, thus representing a minimal important change (MIC)(37). In contrast, the CD method captures between-group differences by accounting for both responders and non-responders, providing a more conservative MID. In our study, AC-based values were consistently higher than CD-based values across all time points and directions of change, echoing the findings of Qin et al(37). The gap between AC and CD estimates narrowed over time, reflecting increasing utility gains in patients with unchanged mRS scores.\u003c/p\u003e\n\u003cp\u003eCompared to previous MID of 0.10 at 3-4 weeks reported in a Taiwanese study(15), our estimate (CD: 0.14) at V2 (acute phase: median 13.0±8.5 days) was slightly larger, likely due to our population’s lower baseline utility (0.55 vs. 0.72), different value sets and anchor choice. Similarly, our MID for improvement over the 9-month interval (V3 to V4, 0.10) exceeded the 0.08 reported in a Korean study over a comparable 10-month period, with explanations including value set and our cohort’s higher proportion of patients with no baseline disability (36.2% vs. 12.5%); we also calculated a MID of 0.10 for the V1-V4 interval (11 months), which was approximately equal to that for V3-V4, although our primary analysis focused on the more clinically relevant comparison betweenV4 and V3.\u0026nbsp;Our pooled anchor-based MIDs across all recovery phases (0.08-0.17) fell within the interquartile range (0.05-0.17) reported in a systematic review of EQ-5D-3L utility MIDs(13).\u0026nbsp;The distribution-based (0.10-0.16) and instrument-defined estimates (0.11-0.12) in our study exceeded the review’s median pooled distribution-based value (0.08) and the upper range (0.08) of instrument-defined MIDs reported across four countries, likely reflecting our larger sample size and the use of a different value set, respectively.\u0026nbsp;Furthermore, MID values in this study were pooled using mean values, in line with common practice. It should be noted, however, that methodologies for integrating MID estimates or selecting an optimal value within a single study remain undefined(38, 39), and future research could seek to identify pooled MID guided by clinical expertise.\u003c/p\u003e\n\u003ch2\u003e4.1. Strengths and Limitations\u003c/h2\u003e\n\u003cp\u003eStrengths of this study include its large sample size and the phase‑specific assessment across acute, subacute, and chronic stages after AIS. This design better captures the heterogeneity of recovery potential and treatment strategies, with acute phase management focused on life‑saving and clinical stabilization through urgent time‑window interventions; the subacute phase emphasizing neurological recovery and rehabilitation training, supported by neurotrophic or neuroprotective agents; and the chronic phase prioritizing functional maintenance and secondary prevention of recurrence, alongside attention to psychological and social adaptation. In addition, we employed multiple methodological approaches to triangulate MID estimates and further tested their robustness and validity using bootstrap resampling, a credibility instrument, and MDC.\u003c/p\u003e\n\u003cp\u003eSeveral limitations should be noted. First, we used EQ-5D-3L instead of the more sensitive 5L version, as the 5L was not available when the study was designed.\u0026nbsp;Many patients with mRS changes already had one or more EQ-5D dimensions at an extreme level by V1-V3, consequently constraining the measurement of further change (See Supplementary Figure\u0026nbsp;10).\u0026nbsp;Second, reliance on the mRS as a single anchor may be a limitation. Although multiple anchors are generally recommended, a single, well-validated anchor can be sufficient(40). The mRS was selected due to its established use in stroke MID studies and strong correlation with the EQ-5D, whereas the available EQ-VAS possesses weaker validity as an anchor(15, 17).\u0026nbsp;\u003c/p\u003e\n\u003ch2\u003e4.2. Implications\u003c/h2\u003e\n\u003cp\u003eThe application of phase-specific MIDs should differ by recovery stage—a principle established in guidelines for assessing meaningful change in quality of life in cancer patients(41). Although derived from oncology, this principle holds equal weight in stroke populations, where distinguishing between early and late disease stages is essential. Specifically, acute‑phase values are most relevant for assessing short‑term treatment effects during admission, whereas subacute and chronic values are better suited for monitoring medium‑ to long‑term prognosis. We report MIDs at both group and individual levels. Group‑level estimates are most informative for trial design, such as using the AC‑derived MID of 0.15 in the subacute phase to interpret the mean change in scores in a real-world treatment group and to calculate sample size. CD‑based estimates provide an alternative approach that can be directly applied to between‑group comparisons. Individual‑level estimates may help classify responders, although ROC‑derived thresholds fell below corresponding MDC values and therefore carry some risk of misclassification. Depending on clinical priorities, stricter cut‑points may increase specificity but reduce sensitivity (See Supplementary Table 4). AC-, CD-, and regression-based results were generally consistent and reliable, while 0.5SD and SEM—which are the most widely used distribution-based metrics for MID estimation—produced values closest to anchor-based estimates.\u003c/p\u003e"},{"header":"5. Conclusion","content":"\u003cp\u003eThis study establishes phase-specific MIDs for the EQ-5D after IS. In clinical trials, change in EQ‑5D scores should be interpreted against these MIDs, with the most robust evidence obtained when changes are both statistically significant and clinically meaningful. The use of multiple methods to evaluate MID robustness in this analysis provides a reference for future research.\u003c/p\u003e\n\u003cp\u003e\u003cbr\u003e\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003cp\u003eAC: average change; AIS: acute ischemic stroke; AUC: area under the curve; CASTOR: Chinese Acute Ischemic Stroke Treatment Outcome Registry; CD: change difference; CE: cardioembolic infarction; EQ-5D-3L: EuroQol five-dimensional three-level questionnaire; EQ-VAS: EQ-5D visual analogue scale; HRQOL: health-related quality of life; LAA: large-artery atherosclerosis; MDC: minimal detectable change; MID: minimally important difference; mRS: modified Rankin Scale; NIHSS: National Institutes of Health Stroke Scale; ROC: receiver operating characteristic; SA: small-artery occlusion; SEM: standard error of measurement; SOE: stroke of other determined etiology; SUE: stroke of undetermined etiology; TOAST: Trial of Org 10172 in Acute Stroke Treatment.\u003c/p\u003e\n\u003cp\u003e\u003cbr\u003e\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate:\u0026nbsp;\u003c/strong\u003eThis observational trial has been approved by the institutional review board of the Peking University First Hospital (approval No. 2015[922]) and that participants have signed written informed consent.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication:\u0026nbsp;\u003c/strong\u003eNot applicable\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and materials:\u003c/strong\u003e The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests:\u003c/strong\u003e The authors declare that they have no competing interest.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding:\u0026nbsp;\u003c/strong\u003eThe authors declare that no funds, grants, or other support were received during the preparation of this manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors’ contributions:\u003c/strong\u003e PL: Conceptualization, Study Design, Data Analysis, Project Administration, Writing–Original Draft.\u003c/p\u003e\n\u003cp\u003eXJ, FX, HL: Supervision, Methodology, Critical Revision of Manuscript; Joint corresponding authors and guarantors.\u003c/p\u003e\n\u003cp\u003eYH, WS: Data Acquisition.\u003c/p\u003e\n\u003cp\u003eMZ: Statistical Analysis.\u003c/p\u003e\n\u003cp\u003eAll authors contributed to data interpretation, critically revised the manuscript for important intellectual content, and approved the final submitted version.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements:\u0026nbsp;\u003c/strong\u003eWe thank LW and XG for their assistance with data cleaning.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eGlobal, regional, and national burden of stroke and its risk factors, 1990-2019: a systematic analysis for the Global Burden of Disease Study 2019. The Lancet Neurology. 2021;20(10):795-820.\u003c/li\u003e\n\u003cli\u003eMurphy SJ, Werring DJ. Stroke: causes and clinical features. Medicine (Abingdon, England : UK ed). 2020;48(9):561-6.\u003c/li\u003e\n\u003cli\u003eMa Q, Li R, Wang L, Yin P, Wang Y, Yan C, et al. Temporal trend and attributable risk factors of stroke burden in China, 1990-2019: an analysis for the Global Burden of Disease Study 2019. The Lancet Public health. 2021;6(12):e897-e906.\u003c/li\u003e\n\u003cli\u003eTu WJ, Zhao Z, Yin P, Cao L, Zeng J, Chen H, et al. Estimated Burden of Stroke in China in 2020. JAMA Netw Open. 2023;6(3):e231455.\u003c/li\u003e\n\u003cli\u003eXu J, Zhang X, Jin A, Pan Y, Li Z, Meng X, et al. Trends and Risk Factors Associated With Stroke Recurrence in China, 2007-2018. JAMA Netw Open. 2022;5(6):e2216341.\u003c/li\u003e\n\u003cli\u003e2022 WGoCSCR. Summary of China Stroke Center Report 2022. Chin J Cerebrovasc Dis. 2024;21(8):565-76.\u003c/li\u003e\n\u003cli\u003eBoehringer Ingelheim I. White Paper on the Economic Burden of Stroke Rehabilitation Patients in China 2023: Yiyaomofang (ByDrug); 2023 [Available from: https://bydrug.pharmcube.com/news/detail/e6bb7c6237e3c0d28156cd1c3e56da1b.\u003c/li\u003e\n\u003cli\u003eAkkara Y, Afreen R, Lemonick M, Paz SG, Rifi Z, Tosto J, et al. Standardizing Domains and Metrics of Stroke Recovery: A Systematic Review. Brain sciences. 2024;14(12).\u003c/li\u003e\n\u003cli\u003eFDA. Guidance for industry: patient-reported outcome measures: use in medical product development to support labeling claims. 2024.\u003c/li\u003e\n\u003cli\u003eHorsman J, Furlong W, Feeny D, Torrance G. The Health Utilities Index (HUI): concepts, measurement properties and applications. Health and quality of life outcomes. 2003;1:54.\u003c/li\u003e\n\u003cli\u003eGuyatt GH, Feeny DH, Patrick DL. Measuring health-related quality of life. Ann Intern Med. 1993;118(8):622-9.\u003c/li\u003e\n\u003cli\u003eJaeschke R, Singer J, Guyatt GH. Measurement of health status. Ascertaining the minimal clinically important difference. Controlled clinical trials. 1989;10(4):407-15.\u003c/li\u003e\n\u003cli\u003eAl Sayah F, Jin X, Short H, McClure NS, Ohinmaa A, Johnson JA. A Systematic Literature Review of Important and Meaningful Differences in the EQ-5D Index and Visual Analog Scale Scores. Value Health. 2025;28(3):470-6.\u003c/li\u003e\n\u003cli\u003eSiriwardana AN, Hoffman AT, Morton RL, Smyth B, Brown MA. Estimating a Minimal Important Difference for the EQ-5D-5L Utility Index in Dialysis Patients. Value Health. 2024;27(4):469-77.\u003c/li\u003e\n\u003cli\u003eChen P, Lin KC, Liing RJ, Wu CY, Chen CL, Chang KC. Validity, responsiveness, and minimal clinically important difference of EQ-5D-5L in stroke patients undergoing rehabilitation. Qual Life Res. 2016;25(6):1585-96.\u003c/li\u003e\n\u003cli\u003eKim SK, Kim SH, Jo MW, Lee SI. Estimation of minimally important differences in the EQ-5D and SF-6D indices and their utility in stroke. Health and quality of life outcomes. 2015;13:32.\u003c/li\u003e\n\u003cli\u003eMishra B, Sudheer P, Agarwal A, Nilima N, Srivastava MVP, Vishnu VY. Minimal Clinically Important Difference of Scales Reported in Stroke Trials: A Review. Brain sciences. 2024;14(1).\u003c/li\u003e\n\u003cli\u003eMcClure NS, Sayah FA, Xie F, Luo N, Johnson JA. Instrument-Defined Estimates of the Minimally Important Difference for EQ-5D-5L Index Scores. Value Health. 2017;20(4):644-50.\u003c/li\u003e\n\u003cli\u003eJayadevappa R, Cook R, Chhatre S. Minimal important difference to infer changes in health-related quality of\u0026nbsp;life-a systematic review. J Clin Epidemiol. 2017;89:188-98.\u003c/li\u003e\n\u003cli\u003eDevji T, Carrasco-Labra A, Qasim A, Phillips M, Johnston BC, Devasenapathy N, et al. Evaluating the credibility of anchor based estimates of minimal important differences for patient reported outcomes: instrument development and reliability study. Bmj. 2020;369:m1714.\u003c/li\u003e\n\u003cli\u003eSun W, Ou Q, Zhang Z, Qu J, Huang Y. Chinese acute ischemic stroke treatment outcome registry (CASTOR): protocol for a prospective registry study on patterns of real-world treatment of acute ischemic stroke in China. BMC complementary and alternative medicine. 2017;17(1):357.\u003c/li\u003e\n\u003cli\u003eEUROQOL. EQ-5D sample/demos https://euroqol.org/register/obtain-eq-5d/eq-5d-sample-demos/ [updated April 11th, 2025. Available from: https://euroqol.org/register/obtain-eq-5d/eq-5d-sample-demos/.\u003c/li\u003e\n\u003cli\u003eZhuo L, Xu L, Ye J, Sun S, Zhang Y, Burstrom K, et al. Time Trade-Off Value Set for EQ-5D-3L Based on a Nationally Representative Chinese Population Survey. Value Health. 2018;21(11):1330-7.\u003c/li\u003e\n\u003cli\u003eLiu GG, Wu H, Li M, Gao C, Luo N. Chinese time trade-off values for EQ-5D health states. Value Health. 2014;17(5):597-604.\u003c/li\u003e\n\u003cli\u003eBanks JL, Marotta CA. Outcomes validity and reliability of the modified Rankin scale: implications for stroke clinical trials: a literature review and synthesis. Stroke. 2007;38(3):1091-6.\u003c/li\u003e\n\u003cli\u003eCranston JS, Kaplan BD, Saver JL. Minimal Clinically Important Difference for Safe and Simple Novel Acute Ischemic Stroke Therapies. Stroke. 2017;48(11):2946-51.\u003c/li\u003e\n\u003cli\u003eMouelhi Y, Jouve E, Castelli C, Gentile S. How is the minimal clinically important difference established in health-related quality of life instruments? Review of anchors and methods. Health and quality of life outcomes. 2020;18(1):136.\u003c/li\u003e\n\u003cli\u003ede Vet HC, Terwee CB. The minimal detectable change should not replace the minimal important difference. J Clin Epidemiol. 2010;63(7):804-5; author reply 6.\u003c/li\u003e\n\u003cli\u003eLas Hayas C, Quintana JM, Padierna JA, Bilbao A, Mu\u0026ntilde;oz P, Francis Cook E. Health-Related Quality of Life for Eating Disorders questionnaire version-2 was responsive 1-year after initial assessment. J Clin Epidemiol. 2007;60(8):825-33.\u003c/li\u003e\n\u003cli\u003eWang Y, Devji T, Carrasco-Labra A, Qasim A, Hao Q, Kum E, et al. An extension minimal important difference credibility item addressing construct proximity is a reliable alternative to the correlation item. J Clin Epidemiol. 2023;157:46-52.\u003c/li\u003e\n\u003cli\u003eWang L, Guan X, Zhou J, Hu H, Liu W, Wei Q, et al. Measuring the health outcomes of Chinese ischemic stroke patients based on the data from a longitudinal multi-center study. Qual Life Res. 2025;34(7):1967-77.\u003c/li\u003e\n\u003cli\u003eMcClure NS, Sayah FA, Ohinmaa A, Johnson JA. Minimally Important Difference of the EQ-5D-5L Index Score in Adults with Type 2 Diabetes. Value Health. 2018;21(9):1090-7.\u003c/li\u003e\n\u003cli\u003eFranceschini M, Boffa A, Pignotti E, Andriolo L, Zaffagnini S, Filardo G. The Minimal Clinically Important Difference Changes Greatly Based on the Different Calculation Methods. The American journal of sports medicine. 2023;51(4):1067-73.\u003c/li\u003e\n\u003cli\u003eMusoro JZ, Coens C, Sprangers MAG, Brandberg Y, Groenvold M, Flechtner HH, et al. Minimally important differences for interpreting EORTC QLQ-C30 change scores over time: A synthesis across 21 clinical trials involving nine different cancer types. European journal of cancer (Oxford, England : 1990). 2023;188:171-82.\u003c/li\u003e\n\u003cli\u003eBourke S, Bennett B, Oluboyede Y, Li T, Longworth L, O'Sullivan SB, et al. Estimating the minimally important difference for the EQ-5D-5L and EORTC QLQ-C30 in cancer. Health and quality of life outcomes. 2024;22(1):81.\u003c/li\u003e\n\u003cli\u003eCella D, Hahn EA, Dineen K. Meaningful change in cancer-specific quality of life scores: differences between improvement and worsening. Qual Life Res. 2002;11(3):207-21.\u003c/li\u003e\n\u003cli\u003eQin Z, Zhu Y, Shi DD, Chen R, Li S, Wu J. The gap between statistical and clinical significance: time to pay attention to clinical relevance in patient-reported outcome measures of insomnia. BMC medical research methodology. 2024;24(1):177.\u003c/li\u003e\n\u003cli\u003eTrigg A, Griffiths P. Triangulation of multiple meaningful change thresholds for patient-reported outcome scores. Qual Life Res. 2021;30(10):2755-64.\u003c/li\u003e\n\u003cli\u003eOliveira A, Machado A, Marques A. Minimal Important and Detectable Differences of Respiratory Measures in Outpatients with AECOPD(\u0026dagger;). Copd. 2018;15(5):479-88.\u003c/li\u003e\n\u003cli\u003eWang Y, Devji T, Qasim A, Hao Q, Wong V, Bhatt M, et al. A systematic survey identified methodological issues in studies estimating anchor-based minimal important differences in patient-reported outcomes. J Clin Epidemiol. 2022;142:144-51.\u003c/li\u003e\n\u003cli\u003eSprangers MA, Moinpour CM, Moynihan TJ, Patrick DL, Revicki DA. Assessing meaningful change in quality of life over time: a users' guide for clinicians. Mayo Clinic proceedings. 2002;77(6):561-71.\u003c/li\u003e\n\u003c/ol\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"health-and-quality-of-life-outcomes","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"hqlo","sideBox":"Learn more about [Health and Quality of Life Outcomes](http://hqlo.biomedcentral.com)","snPcode":"12955","submissionUrl":"https://submission.nature.com/new-submission/12955/3","title":"Health and Quality of Life Outcomes","twitterHandle":"@BioMedCentral","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"EQ-5D, Meaningful important difference, Stroke, Recovery phase, Longitudinal study","lastPublishedDoi":"10.21203/rs.3.rs-7989766/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7989766/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eBackground: \u003c/strong\u003eThe EQ‑5D is widely applied to measure patient-reported outcomes, yet its minimally important difference (MID) has not been clearly established across distinct recovery phases after stroke. This study aimed to determine phase‑specific MIDs in EQ‑5D following stroke and to explore heterogeneity by estimation method, direction of change, and stroke etiology.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods: \u003c/strong\u003eA total of 9 978 adults with neuroimaging‑confirmed acute ischemic stroke were included in a prospective longitudinal cohort study. EQ‑5D and modified Rankin Scale (mRS) scores were recorded at admission (V1), hospital discharge (V2), 3-month (V3), and 1-year since admission (V4). Anchor-based MIDs were estimated at both group and individual levels and triangulated by distribution-based and instrument-defined approaches. Changes during the recovery phases (V1-V2, V2-V3, and V3-V4) were grouped into 3 categories: improved, no change, and deteriorated. Subgroup analyses were conducted according to the TOAST classification.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults: \u003c/strong\u003ePhase-specific group-level MIDs for improvement decreased over time: anchor-based estimates were 0.19 at V2, 0.14 at V3, and 0.11 at V4, while deterioration MIDs were smaller (0.11 to 0.09). Distribution-based and instrument-defined estimates fluctuated slightly around the anchor-based values but followed a similar downward trend over time. Individual‑level analyses yielded valid MIDs only for improvement at V2 (0.10) and V3 (0.01). Cardioembolic strokes had higher MIDs than large-artery atherosclerosis and small-artery occlusion, while baseline utilities showed the reverse.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusions: \u003c/strong\u003eThis study provides the phase‑specific MIDs for utility measures after ischemic stroke, showing a declining trend from acute to chronic recovery and confirming robustness across multiple estimation methods. These MID values may assist in the interpretation of patient‑reported outcome changes related to different healthcare interventions in stroke patients at different recovery phases.\u003c/p\u003e","manuscriptTitle":"Interpreting Patient-Reported Outcomes After Ischemic Stroke: Defining Meaningful Change in EQ-5D Across Recovery Phases","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-11-19 12:05:46","doi":"10.21203/rs.3.rs-7989766/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2026-01-14T07:05:19+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-12-14T06:53:35+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-11-30T05:20:01+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"140486221216033624630481853030506157701","date":"2025-11-16T17:24:16+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"153831482735877746953655875002747268967","date":"2025-11-16T17:09:01+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-11-10T09:37:00+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-11-02T23:13:44+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-11-02T23:13:11+00:00","index":"","fulltext":""},{"type":"submitted","content":"Health and Quality of Life Outcomes","date":"2025-10-30T13:07:02+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"health-and-quality-of-life-outcomes","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"hqlo","sideBox":"Learn more about [Health and Quality of Life Outcomes](http://hqlo.biomedcentral.com)","snPcode":"12955","submissionUrl":"https://submission.nature.com/new-submission/12955/3","title":"Health and Quality of Life Outcomes","twitterHandle":"@BioMedCentral","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"329c5e91-c025-4e57-bed3-1b4c323d72f2","owner":[],"postedDate":"November 19th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2026-03-23T16:00:53+00:00","versionOfRecord":{"articleIdentity":"rs-7989766","link":"https://doi.org/10.1186/s12955-026-02524-w","journal":{"identity":"health-and-quality-of-life-outcomes","isVorOnly":false,"title":"Health and Quality of Life Outcomes"},"publishedOn":"2026-03-21 15:57:51","publishedOnDateReadable":"March 21st, 2026"},"versionCreatedAt":"2025-11-19 12:05:46","video":"","vorDoi":"10.1186/s12955-026-02524-w","vorDoiUrl":"https://doi.org/10.1186/s12955-026-02524-w","workflowStages":[]},"version":"v1","identity":"rs-7989766","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7989766","identity":"rs-7989766","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}