Factors Associated With Improvement in Activities of Daily Living in Patients With Cardiovascular Disease in a Japanese Community Rehabilitation Ward Authors: Tomohiro Matsuo, Kenichiro Maekawa, Satoshi Yamamoto, Masato Sakaguchi, Takeshi Ito, Hiromasa Masuda, Ryosuke Matsuki, Akira Hotta, Michitaka Kato, Kentaro Iwata, Tetsuya Takahashi, Tomoyuki Morisawa Published: 2026 ABSTRACT BackgroundPatients with cardiovascular disease (CVD) admitted to community rehabilitation wards after acute care for cardiac rehabilitation (CR) continuation show improvements in activities of daily living (ADL). However, because the determinants of ADL improvement remain unclear, we aimed to identify the factors related to ADL recovery during post-acute CR.Methods and ResultsThis interim analysis of the multicenter J-CREW CR registry included 198 patients with CVD (mean age: 78.1±8.8 years; 49.5% female). Data on demographics, ADL and frailty status before hospitalization, primary and comorbid conditions, nutritional risk, admission ADL, physical and cognitive function, length of stay (LOS), and rehabilitation dose were collected. The primary outcome was factors associated with Functional Independence Measure (FIM) gain, and the secondary outcome was FIM effectiveness. In the regression model (R2=0.435, P<0.001), admission Short Physical Performance Battery (SPPB; B=1.177, P=0.004), Mini-Mental State Examination (MMSE; B=1.272, P<0.001), and LOS (B=0.117, P<0.001) were significantly associated with FIM gain, whereas admission FIM was negatively associated (B=−0.623, P<0.001). Another model (R2=0.360, P<0.001) revealed comparable associations with FIM effectiveness.ConclusionsBetter physical and cognitive function at admission, together with sufficient LOS, contributed to greater ADL recovery. These indices may serve as practical indicators of effective post-acute CR strategies. Central Figure Japan’s rapidly aging population has led to a significant increase in the number of patients with cardiovascular disease (CVD) presenting with preexisting declines in activities of daily living (ADL), frailty, sarcopenia, and multimorbidity. These conditions often result in complex disabilities that often delay the achievement of acute-phase rehabilitation goals. Furthermore, approximately 20% of older adults experience hospitalization-associated disability (HAD) following acute care,1 and HAD has been reported in 10–20% of older patients with CVD.2–4 According to the Japanese J-Proof HF registry, 37.1% of older patients with heart failure developed HAD, and the presence of HAD was a major prognostic factor associated with an increased risk of 1-year death.5,6 In addition, in patients with heart failure or acute coronary syndrome and in those who have undergone cardiovascular surgery, poor ADL status at discharge from acute-care hospitals has been identified as a predictor of death and unplanned readmission.7–11 Therefore, preventing further ADL decline in patients with CVD is important, but within the limited length of hospital stay (LOS) in acute-care hospitals, sufficient recovery of ADL is not always achieved. For these patients, community rehabilitation wards, which can provide sustained and intensive cardiac rehabilitation (CR) during the post-acute phase, have attracted increasing attention. Our previous domestic studies demonstrated that patients with CVD admitted to rehabilitation hospitals achieved significant improvements not only in ADL but also in exercise tolerance, physical function, respiratory muscle strength, and health-related quality of life.12–14 These findings suggest that CR in community rehabilitation wards significantly affects ADL recovery, but the factors contributing to these improvements remain unclear. In particular, identifying the determinants of ADL recovery during hospitalization is a clinically important issue, as it may translate to better post-discharge outcomes. Thus, in the present study we aimed to clarify the factors associated with in-hospital ADL recovery by patients with CVD admitted to community rehabilitation wards and receiving intensive CR. Methods Study Design and Study Population This study was a prospective, multicenter registry-based interim analysis of the Japanese Community Rehabilitation Ward multicenter registry of Cardiac Rehabilitation (J-CREW CR study). Eligible participants were consecutive patients with a primary diagnosis of CVD who were prospectively enrolled upon admission to 7 community rehabilitation wards between May 2023 and December 2024. The patients were considered unable to return directly to community living and required inpatient continuation of CR after acute care. The exclusion criteria were as follows: age <18 years; severe disturbance of consciousness or cognitive impairment that precluded communication; bedridden status prior to acute hospitalization; inability to provide informed consent; and death during hospitalization. The study protocol was approved by the Ethics Committee of Juntendo University (approval number: 23-008). Written informed consent was obtained from all patients by site investigators. The study was conducted in accordance with the principles of the Declaration of Helsinki. Data Collection We collected the following data from medical records and patient interviews: baseline demographics (age, sex, body mass index [BMI]), ADL prior to acute hospitalization using the Barthel Index, functional status prior to acute hospitalization based on the Kihon Checklist (KCL), cohabitation status, long-term care level before and at admission, primary and comorbid conditions during acute hospitalization, and LOS in the acute-care hospital. The KCL was used to assess frailty severity,15 with a score ≥8 points indicating frailty, as previously reported.16 Disease severity at admission to rehabilitation wards was assessed using the New York Heart Association functional class, echocardiographic parameters, and laboratory data. Additionally, we recorded the CR dose, medications at discharge, rehabilitation performance index, LOS in community rehabilitation wards, and discharge disposition. Almost all patients (99%) had started rehabilitation during the acute-care phase before being transferred. However, detailed information about the duration, frequency, and content of the acute phase of rehabilitation was not consistently available, so it was not included in the analyses. Assessments of ADL, Physical and Cognitive Function, and Exercise Tolerance ADL was assessed using the Functional Independence Measure (FIM) at both admission and discharge,17 administered by trained physical or occupational therapists. The FIM evaluates independence and required assistance across 18 items (13 motor and 5 cognitive), each scored on a 7-point Likert scale from 1 (“total assistance”) to 7 (“complete independence”), with lower scores indicative of poorer ADL. Physical function was measured using the Short Physical Performance Battery (SPPB) test, handgrip strength, and comfortable gait speed at admission and discharge. The SPPB was assessed as described by Guralnik et al.,18 consisting of a balance test, a 4-m walk test, and a 5-times chair-to-stand test, with total scores calculated. Handgrip strength was measured twice on each side using a digital hand dynamometer, and the maximum value was recorded. Comfortable gait speed was assessed over 10 m, with a 3-m approach distance, and expressed in meters per second (m/s). Cognitive function was assessed using the Japanese version of the Mini-Mental State Examination (MMSE),19 with scores ranging from 0 to 30; higher scores indicate better cognition. Exercise tolerance was evaluated using the 6-minute walk test at both admission and discharge, performed according to American Thoracic Society guidelines,20 and expressed as 6-minute walking distance (6MD). CR CR programs were prescribed by attending physicians and delivered by physical and occupational therapists. The interventions followed the Japanese Circulation Society guidelines,21 supplemented with individualized training based on each patient’s physical and ADL status. The programs included ADL practice, balance training, instrumental ADL training, cognitive tasks, movement instruction, and patient education. The CR programs were conducted daily for up to 3 h. Physical therapy was provided to all patients. Occupational therapy was also routinely considered. However, when physical function and ADL were relatively preserved and exercise therapy (including aerobic and resistance training) was the main component of the rehabilitation program, patients were treated with physical therapy alone after consultation with the attending physician. Patients with swallowing dysfunction received speech therapy (ST) as necessary, following the prescription of a physician. ADL Outcome The primary outcome was the change in the FIM score during hospitalization (FIM gain), defined as the difference between the discharge and admission FIM scores.22 Because ceiling and floor effects may affect FIM gain, especially in patients with very high or low baseline FIM scores, FIM effectiveness was used as a secondary outcome to account for these limitations. FIM effectiveness was calculated using the following formula: FIM gain/(126 − admission FIM score), representing the proportion of actual improvement relative to the maximum possible.23 These indicators have been frequently used in systematic reviews of rehabilitation research.24 Statistical Analysis Changes in ADL and related measures between admission and discharge were analyzed using paired t-tests or Wilcoxon signed-rank tests. Multiple regression analyses were conducted to identify factors associated with FIM gain and effectiveness. The independent variables were selected based on clinical relevance and previous literature rather than automated statistical procedures. Because the proportion of missing data was <5% for all variables, patients with missing values for key variables were excluded from multivariate analyses using a complete-case approach. The variables included demographic information (e.g., age and sex), information prior to acute hospitalization, primary CVD, comorbidities, nutritional risk (Geriatric Nutritional Risk Index), admission FIM, physical function (SPPB), cognitive function (MMSE), LOS, and rehabilitation dose. As sensitivity analyses, the multivariable models additionally adjusted for ST intervention and medical vs. surgical status. Additional exploratory analyses were performed using FIM efficiency and FIM gain at 30 days as alternative indicators of recovery speed. FIM efficiency was calculated by dividing FIM gain by LOS in the rehabilitation ward. FIM gain at 30 days was defined as the difference in between the admission FIM score and the score after 30 days. Patients discharged within 30 days were excluded from this analysis. In a subgroup analysis, patients were included in the ADL improvement (FIM effectiveness ≥0.5) or no improvement groups (<0.5). Receiver operating characteristic (ROC) curve analyses were performed using significant predictors from the regression models to evaluate the ability to discriminate between ADL improvement and no improvement. Optimal cutoff values were determined using the Youden index. Logistic regression models combining clinically relevant predictors were constructed, and the predicted probabilities were used in the ROC analyses. Exploratory analyses included testing logistic regression models with interaction terms between physical and cognitive function. Continuous variables were mean-centered to avoid multicollinearity, and their product was used as the interaction term. Two models were tested: a simple model including only physical and cognitive function and their interaction (Model A), and an adjusted model including major covariates (Model B). Model fit was evaluated, and all analyses were performed using SPSS Statistics version 28.0 (IBM Corp., Armonk, NY), with statistical significance set at P<0.05. Results Patients’ Demographics A total of 266 consecutive patients admitted to community rehabilitation wards following acute cardiovascular treatment were recruited for the study. After excluding in-hospital deaths, those unable to ambulate, and those without informed consent, the final analysis included 198 patients. Figure 1 illustrates the study flow, and Table 1 summarizes the baseline characteristics of the patients. Figure 1.Flowchart of patient selection and inclusion in the study. Table 1.Baseline Characteristics of Study Patients n=198Avg.SDBaseline data Age, years78.18.8 Female, n (%)98 (49.5) Body mass index, kg/m220.93.7 Barthel Index before hospitalization93.714.5 Kihon Checklist before hospitalization, points9.15.1 Living alone, n (%)65 (32.8) Length of hospital stay in acute-care hospital, days39.325.1Principal disease before transfer, n (%) Heart failure56 (28.3) Ischemic heart disease33 (16.7) Cardiac surgery57 (28.8) Aortic disease41 (20.7) Combined surgery (cardiac and aortic surgery)7 (3.5) Other4 (2.0)Information for admission to rehabilitation hospital NYHA class, n (%) Class 118 (9.1) Class 261 (30.8) Class 3111 (56.1) Class 48 (4.0) Comorbidities, n (%) Diabetes mellitus63 (31.8) Chronic kidney disease50 (25.3) Chronic heart failure91 (46.0) Respiratory disease53 (15.2) Musculoskeletal disorders83 (41.9) Cerebrovascular disease68 (34.3) Peripheral arterial disease18 (9.1) Atrial fibrillation68 (34.3)Echocardiography Left ventricular ejection fraction, %53.412.6Laboratory data Hemoglobin, g/dL11.21.6 Albumin, g/mL3.30.5 Creatinine, mg/dL1.10.6 eGFR, mL/min/1.73 m254.322.0 C-reactive protein, mg/dL1.482.31 BNP, pg/mL248.5326.2 NT-proBNP, pg/mL2,954.24,593.3 GNRI88.010.8 No risk, n (%)30 (15.1) Low risk, n (%)38 (19.2) Moderate risk, n (%)71 (35.9) High risk, n (%)59 (29.8)Others Total CR time, min/day140.623.0 Length of hospital stay, days67.536.0 Rehabilitation performance index56.839.7 Temporary transfer, n (%)30 (15.2) Home discharge, n (%)154 (77.8)Avg, average; BNP, B-type natriuretic peptide; CR cardiac rehabilitation; eGFR, estimated glomerular filtration rate; GNRI, Geriatric Nutritional Risk Index; NT-proBNP, N-terminal pro-B-type natriuretic peptide; NYHA class, New York Heart Association functional classification; SD, standard deviation. Flowchart of patient selection and inclusion in the study. Table 1.Baseline Characteristics of Study Patients n=198Avg.SDBaseline data Age, years78.18.8 Female, n (%)98 (49.5) Body mass index, kg/m220.93.7 Barthel Index before hospitalization93.714.5 Kihon Checklist before hospitalization, points9.15.1 Living alone, n (%)65 (32.8) Length of hospital stay in acute-care hospital, days39.325.1Principal disease before transfer, n (%) Heart failure56 (28.3) Ischemic heart disease33 (16.7) Cardiac surgery57 (28.8) Aortic disease41 (20.7) Combined surgery (cardiac and aortic surgery)7 (3.5) Other4 (2.0)Information for admission to rehabilitation hospital NYHA class, n (%) Class 118 (9.1) Class 261 (30.8) Class 3111 (56.1) Class 48 (4.0) Comorbidities, n (%) Diabetes mellitus63 (31.8) Chronic kidney disease50 (25.3) Chronic heart failure91 (46.0) Respiratory disease53 (15.2) Musculoskeletal disorders83 (41.9) Cerebrovascular disease68 (34.3) Peripheral arterial disease18 (9.1) Atrial fibrillation68 (34.3)Echocardiography Left ventricular ejection fraction, %53.412.6Laboratory data Hemoglobin, g/dL11.21.6 Albumin, g/mL3.30.5 Creatinine, mg/dL1.10.6 eGFR, mL/min/1.73 m254.322.0 C-reactive protein, mg/dL1.482.31 BNP, pg/mL248.5326.2 NT-proBNP, pg/mL2,954.24,593.3 GNRI88.010.8 No risk, n (%)30 (15.1) Low risk, n (%)38 (19.2) Moderate risk, n (%)71 (35.9) High risk, n (%)59 (29.8)Others Total CR time, min/day140.623.0 Length of hospital stay, days67.536.0 Rehabilitation performance index56.839.7 Temporary transfer, n (%)30 (15.2) Home discharge, n (%)154 (77.8)Avg, average; BNP, B-type natriuretic peptide; CR cardiac rehabilitation; eGFR, estimated glomerular filtration rate; GNRI, Geriatric Nutritional Risk Index; NT-proBNP, N-terminal pro-B-type natriuretic peptide; NYHA class, New York Heart Association functional classification; SD, standard deviation. Baseline Characteristics of Study Patients Avg, average; BNP, B-type natriuretic peptide; CR cardiac rehabilitation; eGFR, estimated glomerular filtration rate; GNRI, Geriatric Nutritional Risk Index; NT-proBNP, N-terminal pro-B-type natriuretic peptide; NYHA class, New York Heart Association functional classification; SD, standard deviation. Rehabilitation in Community Rehabilitation Wards All patients received physical therapy; 92.4% additionally received occupational therapy, and 52.5% received ST. The mean daily duration of CR was 140.6±23.0 min. The mean LOS in community rehabilitation wards was 67.5±36.0 days. A temporary transfer back to an acute-care hospital for additional treatment was required for 15.2% of patients. At discharge, 77.8% of patients could return home (Table 1). Of the 8 patients classified as NYHA class IV at admission, none required an interruption in rehabilitation or an unplanned transfer back to an acute-care hospital. In total, 7 were discharged to their homes, and 1 was discharged to a long-term care facility. There were no serious, rehabilitation-related adverse events. Changes in ADL, Physical Function, Exercise Tolerance, and Cognitive Function As presented in Table 2, the total FIM, motor function FIM, and cognitive function FIM scores at discharge were significantly higher than those at admission (all P<0.001). There was a significant improvement in total SPPB score and its subcomponents, gait speed and handgrip strength, from admission to discharge (all P<0.001). Furthermore, both the 6MD and MMSE improved significantly after CR (both P<0.001). Table 2.Changes in ADL, Physical Function, Exercise Tolerance, and Cognitive Function Before rehabilitationAfter rehabilitationP valueAvg.SDAvg.SDFIM, points67.622.2104.521.6<0.001 FIM-motor, points44.316.474.716.8<0.001 FIM-cognitive, points23.37.729.86.6<0.001SPPB, points5.33.88.03.8<0.001 Balance, points2.01.52.91.4<0.001 4-m gait speed, points2.21.43.01.3<0.001 5 chair–stand, points1.11.42.11.6<0.001Gait speed, m/s0.730.300.930.32<0.001Hand grip strength, kg16.48.017.78.1<0.0016MD, m155.9127.8242.3138.8<0.001MMSE, points24.54.725.94.9<0.0016MD, 6-min walking distance; ADL, activities of daily living; FIM, Functional Independence Measure; MMSE, Mini-Mental State Examination; SPPB, Short Physical Performance Battery. Other abbreviations as in Table 1. Table 2.Changes in ADL, Physical Function, Exercise Tolerance, and Cognitive Function Before rehabilitationAfter rehabilitationP valueAvg.SDAvg.SDFIM, points67.622.2104.521.6<0.001 FIM-motor, points44.316.474.716.8<0.001 FIM-cognitive, points23.37.729.86.6<0.001SPPB, points5.33.88.03.8<0.001 Balance, points2.01.52.91.4<0.001 4-m gait speed, points2.21.43.01.3<0.001 5 chair–stand, points1.11.42.11.6<0.001Gait speed, m/s0.730.300.930.32<0.001Hand grip strength, kg16.48.017.78.1<0.0016MD, m155.9127.8242.3138.8<0.001MMSE, points24.54.725.94.9<0.0016MD, 6-min walking distance; ADL, activities of daily living; FIM, Functional Independence Measure; MMSE, Mini-Mental State Examination; SPPB, Short Physical Performance Battery. Other abbreviations as in Table 1. Changes in ADL, Physical Function, Exercise Tolerance, and Cognitive Function 6MD, 6-min walking distance; ADL, activities of daily living; FIM, Functional Independence Measure; MMSE, Mini-Mental State Examination; SPPB, Short Physical Performance Battery. Other abbreviations as in Table 1. Factors Related to ADL Improvement Multiple regression analyses were performed with FIM gain and effectiveness as the dependent variables (Table 3). For the model with FIM gain (R2=0.435, P<0.001), the significant factors associated with FIM gain were admission SPPB (B=1.177, P=0.002), admission MMSE (B=1.272, P<0.001), and LOS (B=0.117, P<0.001). Admission FIM was negatively associated with FIM gain (B=−0.630, P<0.001). An increase of 1 point in either the SPPB or MMSE corresponded to an increase of approximately 1.3 points in FIM gain. Each additional day of hospitalization was associated with an increase of 0.12 points in FIM gain. For the model with FIM effectiveness (R2=0.360, P<0.001), admission SPPB (B=0.021, P<0.001), admission MMSE (B=0.018, P<0.001), and LOS (B=0.002, P=0.002) were significant factors. All variance inflation factors were <2.5, suggesting no substantial multicollinearity. In the sensitivity analyses adjusting for ST intervention and medical vs. surgical status (Supplementary Tables 1,2), neither variable was significantly associated with FIM gain or FIM effectiveness. The associations of SPPB, MMSE, and LOS with ADL recovery remained unchanged. Additional exploratory analyses of FIM efficiency and FIM gain at 30 days (Supplementary Table 3) revealed that admission SPPB remained significantly associated with both outcomes. Table 3.Summary of Multiple Regression Analysis With FIM Gain and FIM Effectiveness as Dependent Variables (n=198) FIM gainFIM effectivenessB95.0% CIP valueVIFB95.0% CIP valueVIFIntercept16.069[−21.139, 53.278]0.395–−0.081[−0.639, 0.477]0.775–Basic demographics Age−0.067[−0.361, 0.227]0.6531.540−0.003[−0.007, 0.001]0.1881.540 Sex2.555[−2.281, 7.391]0.2991.3400.069[−0.004, 0.141]0.0641.340Information before hospitalization Barthel Index0.163[−0.021, 0.346]0.0821.3360.002[−0.001, 0.005]0.1171.336 Frailty−1.195[−6.062, 3.671]0.6281.309−0.026[−0.099, 0.047]0.4801.309Information at acute-care hospital Primary disease* Ischemic heart disease−2.940[−9.749, 3.869]0.3951.512−0.045[−0.147, 0.058]0.3911.512 Cardiac surgery−2.053[−7.948, 3.843]0.4931.628−0.020[−0.109, 0.068]0.6501.628 Aortic disease−1.812[−8.468, 4.844]0.5921.612−0.026[−0.126, 0.074]0.6131.612 Combined surgery10.650[−1.549, 22.850]0.0871.1980.175[−0.009, 0.358]0.0621.198 Others−12.076[−27.485, 3.333]0.1241.110−0.062[−0.293, 0.169]0.5981.110 Musculoskeletal comorbidities1.203[−3.481, 5.886]0.6131.2260.008[−0.062, 0.079]0.8121.226 LOS at acute-care hospital0.004[−0.090, 0.099]0.9291.3060.000[−0.001, 0.002]0.7671.306Clinical factors at the time of admission to the rehabilitation ward GNRI−1.113[−3.519, 1.293]0.3631.412−0.014[−0.050, 0.022]0.4381.412 FIM−0.623[−0.761, −0.486]<0.0012.1190.001[−0.001, 0.003]0.5592.119 SPPB1.177[0.377, 1.978]0.0042.1220.021[0.009, 0.033]<0.0012.122 MMSE1.272[0.722, 1.822]<0.0011.5230.018[0.010, 0.026]<0.0011.523 LOS at rehabilitation ward0.117[0.049, 0.184]<0.0011.3490.002[0.001, 0.003]0.0031.349 Daily rehabilitation time1.408[−0.538, 3.355]0.1551.1600.012[−0.017, 0.041]0.4131.160 R2=0.435R2=0.360 F=10.544F=7.711*Heart failure was used as the reference category. Data were complete for most variables. Missing data were only observed for Frailty (n=4) and MMSE (n=2). These cases were excluded from the multivariate analyses using a complete-case approach. B, unstandardized coefficient; CI, confidence interval; F, overall model F-statistic (P<0.001); LOS, length of hospital stay; R2, coefficient of determination; VIF, variance inflation factor. Other abbreviations as in Tables 1,2. Table 3.Summary of Multiple Regression Analysis With FIM Gain and FIM Effectiveness as Dependent Variables (n=198) FIM gainFIM effectivenessB95.0% CIP valueVIFB95.0% CIP valueVIFIntercept16.069[−21.139, 53.278]0.395–−0.081[−0.639, 0.477]0.775–Basic demographics Age−0.067[−0.361, 0.227]0.6531.540−0.003[−0.007, 0.001]0.1881.540 Sex2.555[−2.281, 7.391]0.2991.3400.069[−0.004, 0.141]0.0641.340Information before hospitalization Barthel Index0.163[−0.021, 0.346]0.0821.3360.002[−0.001, 0.005]0.1171.336 Frailty−1.195[−6.062, 3.671]0.6281.309−0.026[−0.099, 0.047]0.4801.309Information at acute-care hospital Primary disease* Ischemic heart disease−2.940[−9.749, 3.869]0.3951.512−0.045[−0.147, 0.058]0.3911.512 Cardiac surgery−2.053[−7.948, 3.843]0.4931.628−0.020[−0.109, 0.068]0.6501.628 Aortic disease−1.812[−8.468, 4.844]0.5921.612−0.026[−0.126, 0.074]0.6131.612 Combined surgery10.650[−1.549, 22.850]0.0871.1980.175[−0.009, 0.358]0.0621.198 Others−12.076[−27.485, 3.333]0.1241.110−0.062[−0.293, 0.169]0.5981.110 Musculoskeletal comorbidities1.203[−3.481, 5.886]0.6131.2260.008[−0.062, 0.079]0.8121.226 LOS at acute-care hospital0.004[−0.090, 0.099]0.9291.3060.000[−0.001, 0.002]0.7671.306Clinical factors at the time of admission to the rehabilitation ward GNRI−1.113[−3.519, 1.293]0.3631.412−0.014[−0.050, 0.022]0.4381.412 FIM−0.623[−0.761, −0.486]<0.0012.1190.001[−0.001, 0.003]0.5592.119 SPPB1.177[0.377, 1.978]0.0042.1220.021[0.009, 0.033]<0.0012.122 MMSE1.272[0.722, 1.822]<0.0011.5230.018[0.010, 0.026]<0.0011.523 LOS at rehabilitation ward0.117[0.049, 0.184]<0.0011.3490.002[0.001, 0.003]0.0031.349 Daily rehabilitation time1.408[−0.538, 3.355]0.1551.1600.012[−0.017, 0.041]0.4131.160 R2=0.435R2=0.360 F=10.544F=7.711*Heart failure was used as the reference category. Data were complete for most variables. Missing data were only observed for Frailty (n=4) and MMSE (n=2). These cases were excluded from the multivariate analyses using a complete-case approach. B, unstandardized coefficient; CI, confidence interval; F, overall model F-statistic (P<0.001); LOS, length of hospital stay; R2, coefficient of determination; VIF, variance inflation factor. Other abbreviations as in Tables 1,2. Summary of Multiple Regression Analysis With FIM Gain and FIM Effectiveness as Dependent Variables (n=198) *Heart failure was used as the reference category. Data were complete for most variables. Missing data were only observed for Frailty (n=4) and MMSE (n=2). These cases were excluded from the multivariate analyses using a complete-case approach. B, unstandardized coefficient; CI, confidence interval; F, overall model F-statistic (P<0.001); LOS, length of hospital stay; R2, coefficient of determination; VIF, variance inflation factor. Other abbreviations as in Tables 1,2. Predictive Performance of Factors Related to ADL Improvement ROC curve analyses demonstrated that the area under the curve (AUC) for predicting ADL improvement was 0.672 (P<0.001) for SPPB and 0.695 (P<0.001) for MMSE, indicating modest predictive ability for both indices (Figure 2). The optimal cutoff values were 5.5 for SPPB (sensitivity 0.542, specificity 0.778) and 23.5 for MMSE (sensitivity 0.781, specificity 0.644). Figure 2.Receiver operating characteristic (ROC) curve analysis for predicting ADL improvement. The predictive performance of 3 models was compared: SPPB alone, MMSE alone, and a logistic regression model including SPPB, MMSE, and their interaction terms, adjusted for major covariates. SPPB: AUC=0.672 (95% CI, 0.587–0.757). MMSE: AUC=0.695 (95% CI, 0.594–0.796). SPPB×MMSE interaction model: AUC=0.799 (95% CI, 0.720–0.879). ADL, activities of daily living; AUC, area under the curve; CI, confidence interval; MMSE, Mini-Mental State Examination; SPPB, Short Physical Performance Battery. Receiver operating characteristic (ROC) curve analysis for predicting ADL improvement. The predictive performance of 3 models was compared: SPPB alone, MMSE alone, and a logistic regression model including SPPB, MMSE, and their interaction terms, adjusted for major covariates. SPPB: AUC=0.672 (95% CI, 0.587–0.757). MMSE: AUC=0.695 (95% CI, 0.594–0.796). SPPB×MMSE interaction model: AUC=0.799 (95% CI, 0.720–0.879). ADL, activities of daily living; AUC, area under the curve; CI, confidence interval; MMSE, Mini-Mental State Examination; SPPB, Short Physical Performance Battery. Although LOS was identified as a significant factor in the regression analyses, it was excluded from ROC analyses because it cannot be determined at the time of admission. Logistic regression models incorporating both variables showed that SPPB and MMSE remained significant predictors (both P<0.001 in Models A and B), with significant interaction effects (Model A: P=0.004; Model B: P<0.001). The combined models demonstrated improved predictive performance, with AUCs of 0.747 for Model A and 0.799 for Model B, compared to SPPB (AUC=0.672) or MMSE (AUC=0.695) alone. Discussion This is the first multicenter study to investigate the factors associated with ADL improvement among patients admitted to community rehabilitation wards for intensive CR following acute cardiovascular treatment. Admission physical function (SPPB), cognitive function (MMSE), and LOS were identified as factors associated with FIM gain and effectiveness. In addition, baseline FIM was negatively associated with FIM gain, indicating that patients with lower initial ADL had greater improvement during hospitalization. The study cohort had a mean age of 78 years and comprised a frail population with multiple comorbidities and nutritional risk, although many retained a degree of independence in ADL. Despite these vulnerabilities, significant improvements in ADL, physical function, exercise tolerance, and cognitive function were achieved through continued CR in rehabilitation wards. These improvements exceeded the previously reported minimal critically important difference values,25–27 thus supporting the effectiveness of the intensive and prolonged CR program provided in this setting, which averaged 140 min/day. Regression analyses highlighted the importance of the SPPB score at the time of admission as an indicator of ADL recovery. The SPPB comprehensively evaluates balance, gait speed, and chair–stand performance, all of which directly contribute to daily activities such as mobility and toileting. A lower SPPB score at discharge from acute care is associated with a greater risk of rehospitalization or death,28 and our findings further support its role as a factor related to recovery in the post-acute phase. Similarly, Fujii et al.29 identified the SPPB and pre-hospitalization KCL as factors associated with rehabilitation outcomes, which is consistent with the present results. Furthermore, the MMSE was an independent contributor to rehabilitation effectiveness, underscoring the role of cognitive function in rehabilitation effectiveness. Cognitive impairment negatively affects ADL outcomes,24,30 likely through reduced comprehension of instructions, adherence to therapy, and motivation. Therefore, patients with preserved cognition may experience greater functional recovery. The present analysis revealed that each 1-point increase in admission SPPB or MMSE was associated with an approximate 1.3-point increase in FIM gain, which suggests that improvements in both physical and cognitive performance contribute meaningfully to ADL recovery. This magnitude of change suggests a clinically relevant improvement in functional independence beyond mere statistical significance. Although it may seem counterintuitive that age was not associated with an improvement in ADL, this finding likely reflects the characteristics of community rehabilitation wards, where sufficient rehabilitation intensity and duration can be provided even for older or frail patients. Previous studies have shown that 80% of older patients with hospital-associated functional decline improved their ADL when receiving ≥75 min of daily rehabilitation.31 Those results imply that chronological age alone does not preclude functional recovery when adequate rehabilitation is provided. Instead, physical and cognitive function at admission may better capture functional reserve and responsiveness to rehabilitation. Therefore, they are more clinically relevant determinants of ADL recovery than age itself. However, the supplementary analysis using FIM efficiency revealed that older age was associated with a slower rate of functional recovery, which suggests that, although older patients may ultimately achieve similar overall functional gains, they may require a longer rehabilitation period to do so. These findings underscore the importance of allocating adequate rehabilitation time to account for age-related differences in recovery speed. LOS in community rehabilitation wards was another significant factor associated with ADL improvement. Patients admitted to rehabilitation wards often have residual impairments due to complications or HAD during acute care, particularly older adults requiring a longer recovery period. The mean LOS of 67.5 days observed in this study likely provided sufficient recovery time and reflects the benefits of prolonged intensive CR. However, the causal direction remains uncertain. Prolonged hospitalization may facilitate recovery by allowing for more time to intervene, or it may reflect delayed improvement in patients with greater initial impairment. As FIM gain is time-dependent, its magnitude may be influenced by LOS, potentially overestimating the contribution of LOS in regression models. To address this issue, we conducted supplementary analyses using FIM efficiency, a rate-based index of recovery, in which age and daily rehabilitation time were identified as independent factors. These results suggested that although total functional gain may depend on LOS in the rehabilitation ward, the rate of recovery may be more strongly influenced by patient characteristics and rehabilitation exposure. Recovery trajectories in post-acute CR are often nonlinear, particularly for older and frail patients, who may require an initial period of limited improvement before achieving meaningful gains. In addition, our supplementary analyses using FIM gain at 30 days demonstrated that primary disease category and prehospital ADL were associated with early functional improvement. These findings suggested that early recovery may be influenced more by baseline functional reserve and disease characteristics, whereas later recovery may depend more on LOS, further supporting the concept of nonlinear recovery trajectories. The negative association between the FIM score at admission and FIM gain likely reflects a ceiling effect. A possible floor effect cannot be ruled out for patients with severe disabilities. By applying FIM effectiveness, which adjusts for potential improvement, we confirmed that SPPB, MMSE, and LOS remained significant factors associated with ADL recovery, suggesting that these factors equally affect both absolute and relative recovery. Notably, although low baseline ADL has been reported as a barrier to recovery in stroke and hip fracture cohorts,32 our findings differ, possibly due to the unique Japanese system that allows extended hospitalization and daily intensive rehabilitation. ROC analyses showed that the SPPB and MMSE had only modest predictive ability (AUCs of 0.672 and 0.695, respectively). Thus, although SPPB and MMSE were statistically significant, they may have limited clinical utility as standalone screening tools. Notably, combining SPPB and MMSE in a logistic regression improved predictive accuracy (AUC up to 0.799), suggesting that physical and cognitive functions interact synergistically in ADL recovery. However, this interaction should be considered exploratory and requires further validation. The moderate improvement in predictive power underscores the need to incorporate additional clinical variables for practical prognostication. The lack of association between daily rehabilitation time and improvement in ADL suggests that baseline physical and cognitive function may be more critical determinants of responsiveness to rehabilitation than therapy duration alone. However, supplementary analyses using FIM efficiency identified daily rehabilitation time as an independent factor associated with faster functional recovery rates, suggesting that greater rehabilitation exposure may play a more important role in accelerating the recovery process than in determining the extent of functional improvement. For patients with low SPPB or MMSE scores, an individualized, multidisciplinary approach may be required, including task-oriented activities led by occupational therapists, speech therapists when indicated, and ward-based activity promotion. In these cases, community rehabilitation may focus not only on short-term ADL gains but also on adjusting goals, preventing further decline, and planning for early discharge with long-term care support. Moreover, because recovery trajectories and rehabilitation components may differ across patient subgroups, the primary associations remained robust after adjustment for diagnostic category and ST intervention. Taken together, the findings emphasize the importance of assessing both physical and cognitive function upon admission to rehabilitation wards. They also suggest that the SPPB and MMSE may serve as practical tools for early risk stratification upon admission. Although these factors are not directly modifiable through rehabilitation after admission, they can be used to tailor the intensity of the program and the allocation of resources, enabling the early identification of patients likely to regain ADL independence and informs personalized rehabilitation planning. The SPPB and MMSE are feasible bedside assessments that may serve as standard evaluation tools in post-acute CR. Nevertheless, interventional studies with standardized protocols and long-term outcomes are required to clarify causality and optimize rehabilitation strategies. Study Limitations First, despite being a multicenter study, the sample size was relatively small due to the limited number of patients eligible for prolonged rehabilitation. The modest sample size may have limited the statistical power to detect disease-specific or between-hospital differences. A larger cohort might identify additional factors associated with ADL recovery and enable more robust subgroup analyses. Second, the observational design precludes causal inference. Third, because it was a multicenter study, minor variability in functional assessments between raters could not be excluded. In addition, although all participating hospitals implemented CR programs based on the guidelines of the Japanese Circulation Society, the specific content and intensity of each program may have varied slightly across facilities. Additionally, although most patients received acute-phase rehabilitation before transfer, detailed information on the duration, intensity, and content was not consistently available, limiting adjustment for its potential influence. Fourth, data on prehospital physical and cognitive function were limited, making it unclear to what extent the admission values reflected premorbid status. Fifth, as no control group was included, any improvements cannot be attributed solely to CR. Finally, this study was limited to patients admitted to community rehabilitation wards, which are unique to the Japanese healthcare system. Compared with Western rehabilitation settings, these wards are characterized by a longer LOS and an older, more multimorbid patient population. Therefore, the present findings should be generalized with caution to countries with different healthcare systems. Nevertheless, the results emphasize that physical and cognitive function are key factors in ADL recovery among older patients with CVD, offering valuable insights for similar populations worldwide. Conclusions This study demonstrated that continuing CR in community rehabilitation wards following acute cardiovascular treatment improves ADL. In particular, better physical and cognitive function at admission, together with sufficient LOS, were identified as factors associated with ADL recovery. The findings highlight the importance of comprehensive admission assessments in guiding effective rehabilitation strategies for patients with CVD during the post-acute phase. Future analyses of this multicenter registry are planned to include 1-year prognostic outcomes, which may help clarify the long-term impact of post-acute cardiac rehabilitation. Disclosures T.T. is a member of Circulation Reports’ Editorial Team. The other authors declare no conflicts of interest. IRB Information This study was approved by the Ethics Committee of Juntendo University (Approval No: 23-008). Informed Consent All patients provided written informed consent. Supplementary Files Supplementary Table 1. Supplementary Table 2. Supplementary Table 3.