Admission Interval Research Articles

Does Early Rehabilitation Admission Improve Functional Outcomes of Stroke Patients after a Short Hospital Stay? A Thai Retrospective Study on 596 Patients

Background: It remains unknown whether inpatient rehabilitation lasting three weeks or less, following one or two weeks of acutephase treatment for stroke patients at Siriraj Hospital, Mahidol University (SH-MU), Thailand, is effective, especially when the beginning is deferred. Objective: To examine functional outcomes (FO) of short hospitalized rehabilitation (SHR) for stroke patients with three different stroke onsets to rehabilitation admission intervals (SOTRAI), and to determine whether there are any differences in FO regarding SOTRAI. Materials and Methods: The investigators conducted a retrospective cohort study of medical records on inpatient rehabilitated at the Department of Rehabilitation Medicine, SH-MU. The investigators enrolled a cohort of stroke patients who underwent SHR at a Thai university-based, quaternarycare medical center between 2015 and 2019. The predictor variable was SOTRAI, categorized as early, which is 30 days or less, early-late, which is 31 to 90 days, and late, which is more than 90 days. The FO variables were the modified Barthel Index (MBI) score change, rehabilitation efficiency, and effectiveness, and percentages of responders. Appropriate statistics were computerized with the significant level at p-value less than 0.05. Results: Five hundred ninety-six patients, with 44.8% females and 45.8% early SOTRAI, with the mean age of 65.3±13.4 years (range of 19 to 94) were included. Early SOTRAI was associated with the highest FO parameters (p<0.001). Conclusion: Stroke survivors should be apt to SHR within 30 days after the symptoms begin. In other words, early SOTRAI could help improve FO, compared with late SOTRAI. Keywords: Stroke; Rehabilitation; Hospital admission; Outcome; Inpatient

A nomogram for predicting the transition from recurrent acute pancreatitis to chronic pancreatitis

Abstract Objectives: Acute pancreatitis, recurrent acute pancreatitis, and chronic pancreatitis are recognized as a continuum of pancreatic diseases. Recurrence increases the risk of progression to chronic pancreatitis. The aim of this study was to search for clinical features that may promote the progression of chronic pancreatitis in patients with recurrent acute pancreatitis. Methods: We retrospectively reviewed patients with recurrent acute pancreatitis from Medical Information Mart for Intensive Care-IV database. They were divided into a training cohort and a validation cohort. A nomogram was constructed based on clinical features during the second hospitalization. The discrimination and calibration of the nomogram were evaluated using the concordance index, area under the time-dependent receiver operating characteristic curve, and calibration plots. Results: A total of 432 recurrent acute pancreatitis patients were evaluated, of which 93 (21.53%) were diagnosed with chronic pancreatitis later. Age, biliary pancreatitis, admission interval, alcohol dependence, lipase, and platelet were selected. The concordance index was 0.717 (95% CI: 0.691-0.743) for the training cohort and 0.718 (95% CI: 0.662-0.774) for the validation cohort. The area under the time-dependent receiver operating characteristic curve was >0.7 over 1000 days. Conclusions: A nomogram was developed and validated to evaluate the transition from recurrent acute pancreatitis to chronic pancreatitis.

Effects of integrated traditional Chinese and Western medicine for acute pancreatitis: A real-world study in a tertiary teaching hospital.

This study aimed to evaluate whether integrated traditional Chinese medicine (TCM) and Western medicine (WM) is more effective than WM for acute pancreatitis (AP). Patients with AP were enrolled and divided into the TCM and WM (TCM&WM) and WM groups according to the therapeutic protocol in real clinical settings. We applied 1:3 propensity score matching, which was to adjust confounding factors. The primary outcome was mortality, whereas the secondary outcomes were organ failure, organ supportive therapies, local complications, hospitalization cost, and length of hospital stay. Sensitivity and subgroup analyses were also performed. Of 5442 patients with AP, 4691 and 751 were included in the TCM&WM and WM groups, respectively. After PSM, patient baseline characteristics were well balanced. Compared with the WM group (n=734), the TCM&WM group (n=2096) had lower overall mortality rate (1.7%vs. 3.4%; risk ratio, 0.482; 95% confidence interval, 0.286-0.810; p=0.005). The TCM&WM group was associated with lower risk of persistent renal failure, multiple organ failure, and infection, lower utilization of organ supportive therapies, shortened lengths of hospital and intensive care unit stay, and lower hospital costs. Sensitivity analyses showed similar results. Subgroup analysis favored TCM&WM treatment for patients aged<60 years, with hypertriglyceridic etiology, and with admission interval between 24 and 48h. TCM&WM treatment can achieve lower risks of mortality and organ failure and better economic effectiveness in patients with AP than WM treatment. This study provides a promising alternative of TCM&WM treatment for AP in the real-world setting.

Fast algorithms of computing admissible intervals for discrete distributions with single parameter

It is of great interest to compute optimal exact confidence intervals for the success probability (p) in a binomial distribution, the number of subjects with a certain attribute (M) or the total number of subjects (N) in a hypergeometric distribution, and the mean λ of a Poisson distribution. In this paper, efficient algorithms are proposed to compute an admissible exact interval for each of the four parameters when the sample size (n) or the random observation X is large. The algorithms are utilized in four practical examples: evaluating the relationship between two diseases, certifying companies, estimating the proportion of drug users, and analyzing earthquake frequency. The intervals computed by the algorithms are shorter, and the calculations are faster, demonstrating the accuracy of the results and the time efficiency of the proposed algorithms.

Impact of index admission cholecystectomy vs interval cholecystectomy on readmission rate in acute cholangitis: National Readmission Database survey.

Elective cholecystectomy (CCY) is recommended for patients with gallstone-related acute cholangitis (AC) following endoscopic decompression to prevent recurrent biliary events. However, the optimal timing and implications of CCY remain unclear. To examine the impact of same-admission CCY compared to interval CCY on patients with gallstone-related AC using the National Readmission Database (NRD). We queried the NRD to identify all gallstone-related AC hospitalizations in adult patients with and without the same admission CCY between 2016 and 2020. Our primary outcome was all-cause 30-d readmission rates, and secondary outcomes included in-hospital mortality, length of stay (LOS), and hospitalization cost. Among the 124964 gallstone-related AC hospitalizations, only 14.67% underwent the same admission CCY. The all-cause 30-d readmissions in the same admission CCY group were almost half that of the non-CCY group (5.56% vs 11.50%). Patients in the same admission CCY group had a longer mean LOS and higher hospitalization costs attributable to surgery. Although the most common reason for readmission was sepsis in both groups, the second most common reason was AC in the interval CCY group. Our study suggests that patients with gallstone-related AC who do not undergo the same admission CCY have twice the risk of readmission compared to those who undergo CCY during the same admission. These readmissions can potentially be prevented by performing same-admission CCY in appropriate patients, which may reduce subsequent hospitalization costs secondary to readmissions.

Input-to-state stability of impulsive switched systems under mode-dependent event-triggered impulsive control

This paper aims to investigate input-to-state stability (ISS) of nonlinear impulsive switched systems under mode-dependent event-triggered impulsive control (MDETIC) method. A novel mode-dependent event-triggered mechanism (MDETM) is devised, where triggering impulsive instants are determined by some predesigned event conditions. Different from existing MDETM, there are two types of impulsive behavior involved, including the impulsive behavior of system itself and event-triggered one. The novel MDETM not only reduces communication cost and energy consumption but also excludes “Zeno behavior.” Then, the ISS criteria of the closed-loop system are established based on MDETIC strategy, admissible edge-dependent average dwell time (AED-ADT), and admissible edge-dependent average impulsive interval (AED-AII) methods. Finally, two examples are given to verify the effectiveness of the main results.

Input-to-state stability of discrete-time impulsive switched delayed systems with delay-dependent impulse under two asynchronous cases

This paper studies a class of discrete-time impulsive switched delayed systems with delay-dependent impulses, aiming to solving the input-to-state stability (ISS) problem in the case of asynchronous switching signal between subsystem and controller and asynchrony of switching signals and impulsive signals. Among them, the delay effect is handled by a new Lyapunov-Krasovskii function divided into a delay-dependent part and a delay-independent part. In contrast to admissible edge-dependent average dwell time (AED-ADT) methods that address switching signals, we propose admissible edge-dependent average impulsive interval (AED-AII) to deal with delayed impulses, thereby establishing the relationship between these methods and the decay rate. Meanwhile, a stabilizing controller for discrete impulsive switched linear delayed systems is proposed and the minimum average dwell time and controller gain are obtained. Compared with the previous work, the parameters of Lyapunov functional relation depend on the directed edges; the operation of subsystem is divided into asynchronous time and synchronous time, which are less conservative; the combination of AED-ADT and AED-AII can be applied to a wider range of impulsive switched systems. Finally, two examples are given to show the effectiveness of the results.

Association between the initial physical activity and functional recovery after 1 month of inpatient rehabilitation for subacute stroke: stratified analysis by nutritional status.

Objectively measured physical activity volume serves as a predictive factor for functional recovery in patients with stroke. Malnutrition, a frequent complication of stroke, may influence the relationship between physical activity and functional recovery. This study aimed to examine the association between physical activity volume and functional recovery in patients with stroke, stratified by their nutritional status. This multicenter prospective observational study included 209 patients with stroke admitted to two Japanese convalescent rehabilitation hospitals. Participants were categorized based on the geriatric nutritional risk index (GNRI) at admission [≥92, high GNRI group ( n = 133); <92, low GNRI group ( n = 76)]. Physical activity levels were measured as the duration of total physical activity (TPA), which is the sum of light-intensity physical activity and moderate-to-vigorous physical activity, using a triaxial accelerometer during the first 7 days after admission. Outcome measures are represented as the relative gain of the motor score on functional independence measure (M-FIM effectiveness) during the first month after admission. The multiple regression analysis, adjusting for age, sex, comorbidity, onset to admission intervals, motor paralysis, initial M-FIM, and cognitive FIM, showed that the duration of TPA in the first 7 days was significantly associated with the M-FIM effectiveness over the first month in both low GNRI [ B = 0.12, 95% confidential intervals (CI) = 0.01; 0.24, P = 0.049] and high GNRI group ( B = 0.11, 95% CI = 0.01; 0.21, P = 0.027). This study demonstrates a positive predictive association between early TPA level and functional recovery in stroke patients, irrespective of their nutritional status.

Simple coiling of middle meningeal artery embolization for chronic subdural hematoma: An inverse probability of treatment weighting matched cohort study.

Middle meningeal artery embolization (MMAE) for chronic subdural hematoma (CSDH) has gained much attention in recent years. However, unintended embolization may occur when employing liquid embolic agents or particles. We present our clinical experience in simple coiling of MMAE to manage CSDH. Patients underwent either surgical evacuation or MMAE with simple coiling for CSDH were reviewed. Clinical and radiographic outcomes were assessed at admission, 1-month, and 6-month intervals. Two treatment groups were matched with inverse probability of treatment weighting. One hundred twelve patients were included, with 27 patients in MMAE group and 87 patients in surgery group. In MMAE group, significant reductions were observed in hematoma width (admission vs. 1-month, 2.04 [1.44-2.60] cm vs. 0.62 [0.37-0.95] cm, p < 0.001). The adjusted odds ratio (aOR) of surgical rescue rate (0.77 95%CI 0.13-4.47, p = 0.77), hematoma reduction (>50%) (0.21 95%CI 0.04-1.07, p = 0.06), and midline shift improvement rate (3.22, 95%CI 0.84-12.4, p = 0.09) had no substantial disparities between two groups at 1-month follow-up. In addition, no significant difference was noted between two groups in terms of hematoma reduction (>50%) at 6-month follow-up (aOR 1.09 95%CI 0.32-3.70, p = 0.89). No procedure-related complications were found in MMA embolization group. Simple coiling for MMA had comparable outcomes with surgical evacuation for CSDH. Our findings suggest that simple coiling can be an alternative choice for liquid agents or particles in MMA embolization for CSDH with acceptable safety.

Stability of nonlinear impulsive systems with non-uniformly distributed packet loss

This study focuses mainly on the stability of a class of nonlinear systems with impulsive disturbance. In particular, we consider the possible non-uniformly distributed packet loss phenomenon in the controller-to-actuator channel. This packet loss is described by a characterization scenario in the average sense, known as the average packet loss interval (APLI). In this scenario, an average allocation strategy between the number of impulsive sampling intervals and the length of packet loss intervals is established. This implies that the limitations on the proportion or the upper bound of packet loss intervals can be relaxed. A sampled-data control is put forward based on impulsive disturbance, which not only brings better control performance but also reduces conservatism. In view of the Halanay inequality, the number of subsystem switchings caused by packet loss is allowed to be arbitrary. Based on the elastic constraint relationships established between system parameters and APLI, a series of conditions guaranteeing the globally uniformly exponential stability (GUES) are provided. Further, the admissible packet loss intervals can be estimated on an average basis by solving an algebraic inequality. Finally, a numerical example is discussed to demonstrate the validity of our study.

Unified stability criteria of asynchronous discrete-time impulsive switched delayed systems: bounded admissible edge-dependent average dwell time method

This article investigates input-to-state stability (ISS) problem of asynchronous discrete-time impulsive switched delayed systems, in which switching and impulse may be asynchronous. A novel bounded admissible edge-dependent average dwell time (BAED-ADT) switching signal is designed. Based on this division rule of switching subsystems, we adopt the slow switching method for the stable subsystems and the fast switching method for the unstable subsystems, so that stable subsystems can compensate unstable subsystems. Combined with admissible edge-dependent average impulsive interval (AED-AII) impulsive signal, a new unified ISS stability result is established, which could be applied to impulsive switched systems with an arbitrary combination of unstable and stable subsystems. Without solving the linear matrix inequality, it is allowed to readjust the AED-ADT boundary of the newly proposed stability condition according to the actual impulsive and switching signal. Further, compared with the existing fruits, the new stability condition is an extension and improvement of the previous one, which is less conservative. Several numerical examples are provided to show the validity of the conditions and the advantages of the theoretic results.

Uniportal versus multiportal nonintubated thoracoscopic anatomical resection for lung cancer: A propensity-matched analysis

No studies have compared between uniportal and multiportal nonintubated thoracoscopic anatomical resection for non-small cell lung cancer (NSCLC). We aimed to compare short- and long-term postoperative outcomes concerning these two methods.Our retrospective dataset comprised patients with NSCLC who underwent uniportal or multiportal nonintubated thoracoscopic anatomical resection between January 2011 and December 2019. The primary outcome was recurrence-free survival. Propensity scores were matched according to age, sex, body mass index, pulmonary function, tumor size, cancer stage, and surgical method.In total, 1130 such patients underwent nonintubated video-assisted thoracoscopic surgery (VATS), and 490 consecutive patients with stage I-III NSCLC underwent nonintubated anatomical resection, including lobectomy and segmentectomy (uniportal, n = 158 [32.3%]; multiportal, n = 331 [67.7%]). The uniportal group had fewer dissected lymph nodes and lymph node stations. In paired group analysis, the uniportal group had shorter operation durations (99.8 vs. 138.2 min; P < 0.001), lower intensive care unit (ICU) admission rates and ICU admission intervals (7.0% vs. 27.8%; P < 0.001), and shorter postoperative hospital stays (4.1 days vs. 5.2 days; P < 0.001). The most common postoperative complication was prolonged air leaks. No surgical mortality was observed. The multiportal group had higher complication rates for grades ≥ II NSCLC; however, this difference was not significant (4.4% vs. 1.3%, respectively; P = 0.09).Nonintubated uniportal VATS for anatomical resection had better results for some perioperative outcomes than multiportal VATS. Oncological outcomes such as recurrence-free and overall survival remained uncompromised, despite fewer dissected lymph nodes.

Infection with SARS-CoV-2 variant Gamma (P.1) in Chile increased ICU admission risk three to five-fold.

The 2021 wave of SARS-CoV-2 infection in Chile was characterized by an explosive increase in ICU admissions, which disproportionately affected individuals younger than 60 years. This second wave was also accompanied by an explosive increase in Gamma (P.1) variant detections and the massive vaccine rollout. We unveil the role the Gamma variant played in stressing the use of critical care, by developing and calibrating a queueing model that uses data on new onset cases and actual ICU occupancy, symptom's onset to ICU admission interval, ICU length-of-stay, genomic surveillance, and vaccine effectiveness. Our model shows that infection with the Gamma (P.1) variant led to a 3.5-4.7-fold increase in ICU admission for people younger than 60 years. This situation occurred on top of the already reported higher infection rate of the Gamma variant. Importantly, our results also strongly suggest that the vaccines used in Chile (inactivated mostly, but also an mRNA), were able to curb Gamma variant ICU admission over infections.

Stability analysis for hybrid deterministic system under delay-dependent impulses uniting properties of edges

In this summary, for a kind of hybrid deterministic system, we emphasize the input-to-state stability (ISS) property. Meanwhile, the delay-dependent impulse and switching behavior are permitted to arise asynchronously. In other words, the impulses can arise many times or not once within a period of switching interval. For impulsive signals, A fresh idea called admissible edge-dependent average impulsive interval (AED-AII) is addressed according to properties of edges. By developing the multiple Lyapunov-Krasovskii functionals (MLKFs), merged with the AED-ADT switching scheme and AED-AII method, some ameliorated criteria in accordance with the ISS of hybrid delayed systems are established with regard to stabilizing and destabilizing impulses. Compared with previous work, our setup enjoys the important features: the candidate functions fully taps the effect of the delay existing in the system dynamics and impulse on the function part of the MLKFs; the AED-AII method combined with the AED-ADT switching scheme in our paper contains a larger group of the impulsive and switching signals; and for the case in which all the subsystems are unstable, the stability criterion is suitable for the arbitrary time delay. Finally, an example shows that the results of this paper are effective.

Early prediction of 30- and 14-day all-cause unplanned readmissions.

An unplanned readmission is a dual metric for both the cost and quality of medical care. We employed the random forest (RF) method to build a prediction model using a large dataset from patients' electronic health records (EHRs) from a medical center in Taiwan. The discrimination abilities between the RF and regression-based models were compared using the areas under the ROC curves (AUROC). When compared with standardized risk prediction tools, the RF constructed using data readily available at admission had a marginally yet significantly better ability to identify high-risk readmissions within 30 and 14days without compromising sensitivity and specificity. The most important predictor for 30-day readmissions was directly related to the representing factors of index hospitalization, whereas for 14-day readmissions the most important predictor was associated with a higher chronic illness burden. Identifying dominant risk factors based on index admission and different readmission time intervals is crucial for healthcare planning.

Clinic-epidemiological profile and acute kidney injury among pediatric snake bite cases admitted in a tertiary care center of South India

Background: Snake bite is a neglected public health problem in the tropical countries, particularly in rural areas. Snake bite is more common in children and young adults. This study intends to bring out the burden of the issue in rural children of South India. Aims and Objectives: The objectives of the study are as follows: (1) To study the epidemiological profile, the clinical profile, complications, and outcome of pediatric snake bite cases and (2) to study the incidence, predictors, and outcome of acute kidney injury (AKI). Materials and Methods: A prospective and observational study carried out for 1½ years among 0–12-year-old snake bite children. The epidemiological details, details related to snake bite event, pre hospital events, clinical and laboratory parameters on admission, outcome, complications, incidence of AKI, and their possible predictive factors among the snake bite children are analyzed. Statistical analysis was done using SPSS software (IBM, USA) version 20. Results: Male children above 5 years of age in rural areas during outdoor activities are more commonly affected during the monsoon period and during night hours. Hemotoxic snake bites are more common. Incidence of AKI following snake bite is 10%. Possible predictive factors are long bite to admission interval, tender lymph nodes, anemia, increased total count, elevated serum bilirubin on admission, and increased requirement of Anti Snake Venom. Conclusion: This study proves the burden of snake bite among rural children and their morbidity and mortality following snake bite-induced AKI. A high index of suspicion should be kept in mind based on possible predictive factors for development of AKI in snake bite children.

Effect of Late-Onset Stroke Rehabilitation on Medical Morbidities and Functional Recovery: A Single-Center Observational Study.

Background The concept of focused rehabilitative care and a dedicated rehabilitation setup is fairly less known among the people of developing countries. The main objective of the study is to assess the effect of late-onset stroke rehabilitation on its overall prognosis and to see whether and how the late initiation of rehabilitation would pose asignificant effect on functional recovery and adverse medical outcomes Methods A single-center, prospective observational study was conducted in a tertiary rehabilitation center for a duration of one and a half years. The subjects admitted to the Department of Physical Medicine and Rehabilitation were divided into three groups with respect to the onset to admission interval (OAI). Patients who got admitted within 30 days (OAI≤ 30 days) were consideredthe early rehabilitation group. OAI between 31 and 150 days was the late rehabilitation group and OAI of 151 days or more was the very late rehabilitation group. Barthel Index (BI), Modified Rankin Score (MRS), and 6-minute walk test (6MWT) were used as functional outcome measures at admission and discharge from indoors. Results Maximum complications were observed in the late rehabilitation group, i.e. 2.75±2.74, with an overall mean of 2.02±2.04 (p=0.003). Functional recovery assessed using BI shows a significant difference between the early (35.79±8.86) and late (40.62±10.5) rehabilitation groups (p= 0.0005). 6MWT at discharge showed significant improvement in the early rehabilitation group (p=0.005). Conclusion Early onset of rehabilitation showed better functional recovery and fewer adverse medical outcomes in stroke patients. A longer length of hospital stay (LOHS), with a mean duration of 88 ±2 days was needed for patients with higher OAIs.

EGS P22 Management of acute cholecystitis during the Covid-19 pandemic - A complete cycle

Abstract Background Acute cholecystitis is a common surgical emergency. NICE guidelines recommend early laparoscopic cholecystectomy should be performed within 1 week of diagnosis. Emergency and elective surgical provision was affected during the first wave of the covid pandemic due to high postoperative mortality and pulmonary complications. The objectives of this audit were to assess and improve the local management of acute cholecystitis during the Covid-19 pandemic. Methods A retrospective review of all patients who presented with acute cholecystitis over 2 months was performed in April 2020. Several interventions were introduced following the first cycle to improve our adherence with NICE guidelines, including ‘green elective surgical pathway’, clinical priority coding for elective surgeries, a hot gallbladder pathway, and a dedicated weekend and evening list. A re-audit was performed over 2 months in July 2020. These findings were presented at the departmental meeting. Results 34 patients were reviewed in the first audit and 37 in the re-audit. A higher number of patients with acute cholecystitis in cycle 2 had admission covid swabs (97.3% vs 64.7%) and interval covid swabs (54.1% vs 0%) compared to cycle 1. One patient was tested positive for Covid-19 in cycle 2 and was listed for elective operation. Cycle 2 revealed a higher rate of hot gallbladder (from 8.8% to 10.8%) and a shorter median length from admission to operation (from 4 to 1.5 days). A higher rate of patients was readmitted due to complications of cholecystitis in cycle 2 (29.7% vs 20.6%) compared to cycle 1. Conclusions The overall performance is still behind the NICE guideline recommendation. We have identified factors such as pressure on theatre facilities and resources and cancer fast-track priority surgeries during the Covid-19 pandemic. Ongoing audit and optimisation of the hot gallbladder pathway, elective surgical pathway, extra lists, and clinical priority of operations are important to improve the quality of care for patients with acute cholecystitis.

Integral input-to-state stability of switched delayed systems with delay-dependent impulses under generalized impulsive and switching scheme

This paper studies integral input-to-state stability (iISS) of a class of hybrid delayed systems with impulse and switching behavior, in which the delay-dependent impulse and switching behavior are allowed for asynchronous occurrence. Relative to admissible edge-dependent average dwell time (AED-ADT) concept of the switching signal, a fresh concept of admissible edge-dependent average impulsive interval (AED-AII) is proposed for the impulse signal. Several sufficient criteria for iISS are achieved by using multiple Lyapunov–Krasovskii functional (MLKF) tool, AED-ADT switching, and AED-AII schemes. An improved inequality that describing the relation among Lyapunov function continuous switching signal and discrete impulse is constructed, where the jump amplitude at discrete switching points and impulsive moments including delay-dependent part and delay-independent part are involved. Compared with existing results, our conclusions have the following novelties: (1) the influence of time delay existing in continuous dynamics and discrete impulse on the function part of L-K functional is dug out, which is less conservative; (2) the AED-AII scheme proposed in this paper is combined with the AED-ADT switching method, which contains a larger set of the impulsive and switching signal. Moreover, the stability criterion is suitable for the arbitrary time delay in situation that all the subsystems are destabilizing. Finally, an example is given to verify the validity of the results.

Long-Term Rehabilitation Outcomes of Neurological Patients: A Multicenter Study

Objective: To evaluate functional outcomes in patients three years after rehabilitation and to identify factors significantly associated with improvement.Materials and Methods: This prospective cohort study was carried out in nine tertiary centers to compare functional outcomes (Barthel Index; BI) at discharge with scores at three-year follow-up among various diagnoses and types of admissions. Related factors were evaluated for association with improvement in functional score.Results: Three hundred and eighteen patients (mean age: 54 years; 60% male) were included. More than half of all patients suffered from a spinal cord injury. After three years, 35% of patients were still receiving physical therapy. Only those who were admitted for intensive rehabilitation showed significant improvement after three years. One hundred and ten patients or 35.8% showed significant improvement over time. A univariate analysis showed type of diagnosis, type of admission, onset to admission interval, BI at discharge, and presence of depression and complications at follow-up to be significantly associated with improvements in functional score in the follow-up period. Using a multivariate analysis, only the type of diagnosis, low BI at discharge, and absence of depression and complications at follow-up related to functional improvement.Conclusion: One-third of patients had sustained functional improvements from rehabilitation three years after discharge. Participants admitted into intensive rehabilitation showed significant improvements in functional scores between discharge and follow-up. TBI diagnosis, low BI at discharge, absence of depression and complications at follow-up related to long-term functional improvement at the three-year mark.