RESUMEN
BACKGROUND: This study sought to predict time to patient hemodynamic stabilization during trauma resuscitations of hypotensive patient encounters using electronic medical records (EMR) data. METHODS: This observational cohort study leveraged EMR data from a nine-hospital academic system composed of Level I, Level II and non-trauma centers. Injured, hemodynamically unstable (initial systolic blood pressure < 90 mmHg) emergency encounters from 2015-2020 were identified. Stabilization was defined as documented subsequent systolic blood pressure > 90 mmHg. We predicted time to stabilization testing random forests, gradient boosting and ensembles using patient, injury, treatment, EPIC Trauma Narrator and hospital features from the first four hours of care. RESULTS: Of 177,127 encounters, 1347 (0.8%) arrived hemodynamically unstable; 168 (12.5%) presented to Level I trauma centers, 853 (63.3%) to Level II, and 326 (24.2%) to non-trauma centers. Of those, 747 (55.5%) were stabilized with a median of 50 minutes (IQR 21-101 min). Stabilization was documented in 94.6% of unstable patient encounters at Level I, 57.6% at Level II and 29.8% at non-trauma centers (p < 0.001). Time to stabilization was predicted with a C-index of 0.80. The most predictive features were EPIC Trauma Narrator measures; documented patient arrival, provider exam, and disposition decision. In-hospital mortality was highest at Level I, 3.0% vs. 1.2% at Level II, and 0.3% at non-trauma centers (p < 0.001). Importantly, non-trauma centers had the highest re-triage rate to another acute care hospital (12.0%) compared to Level II centers (4.0%, p < 0.001). CONCLUSION: Time to stabilization of unstable injured patients can be predicted with EMR data.
RESUMEN
BACKGROUND: This study aimed to use natural language processing to predict the presence of intra-abdominal injury using unstructured data from electronic medical records. METHODS: This was a random-sample retrospective observational cohort study leveraging unstructured data from injured patients taken to one of 9 acute care hospitals in an integrated health system between 2015 and 2021. Patients with International Classification of Diseases External Cause of Morbidity codes were identified. History and physical, consult, progress, and radiology report text from the first 8 hours of care were abstracted. Annotator dyads independently annotated encounters' text files to establish ground truth regarding whether intra-abdominal injury occurred. Features were extracted from text using natural language processing techniques, bag of words, and principal component analysis. We tested logistic regression, random forests, and gradient boosting machine to determine accuracy, recall, and precision of natural language processing to predict intra-abdominal injury. RESULTS: A random sample of 7,000 patient encounters of 177,127 was annotated. Only 2,951 had sufficient information to determine whether an intra-abdominal injury was present. Among those, 84 (2.9%) had an intra-abdominal injury. The concordance between annotators was 0.989. Logistic regression of features identified with bag of words and principal component analysis had the best predictive ability, with an area under the receiver operating characteristic curve of 0.9, recall of 0.73, and precision of 0.17. Text features with greatest importance included "abdomen," "pelvis," "spleen," and "hematoma." CONCLUSION: Natural language processing could be a screening decision support tool, which, if paired with human clinical assessment, can maximize precision of intra-abdominal injury identification.
Asunto(s)
Traumatismos Abdominales , Registros Electrónicos de Salud , Procesamiento de Lenguaje Natural , Humanos , Registros Electrónicos de Salud/estadística & datos numéricos , Estudios Retrospectivos , Traumatismos Abdominales/diagnóstico , Traumatismos Abdominales/epidemiología , Femenino , Masculino , Persona de Mediana Edad , Adulto , Anciano , Adulto JovenRESUMEN
Individuals with maxillary dentofacial deformities commonly have preexisting nasal airway obstruction with associated nasal septal deviation. Combination bimaxillary procedures with intraoral septoplasty after maxillary downfracture have been demonstrated to address dentofacial deformities and nasal obstruction simultaneously. Although effective, current transoral septoplasty techniques offer limited visualization of the superior septum and are challenging to perform due to septal mobility after Le Fort I downfracture. Here, we describe a novel approach in which a transoral septoplasty is performed before completion of the Le Fort I maxillary downfracture, improving septal stability during the dissection and affording better visualization of the septal L-strut for precise cartilage resection.
RESUMEN
Genioglossus advancement plays an important role in the armamentarium of the obstructive sleep apnea surgeon and has gone through many iterations over several decades. A recently described technique involves creating a box osteotomy, which is carried through the inferior border of the mandible in order to increase recruitment of the suprahyoid musculature. Here we introduce a further modification of the technique that uses virtual planning to improve the safety and accuracy of genial tubercle capture. In addition, angulation of the lateral osteotomies enhances bone to bone contact. Before the osteotomy, the surgeon has the opportunity to drill the buccal plate to reduce the chin profile. This approach is particularly valuable in the patient who is prognathic at baseline or who becomes prognathic after simultaneous maxillomandibular advancement. Here we discuss this unique approach, demonstrating how the patient profile may be balanced even as the genial tubercle is advanced.
RESUMEN
STUDY OBJECTIVE: There is scant research identifying thematic trends within medical research. This work may provide insight into how a given field values certain topics. We assessed the feasibility of using a machine learning approach to determine the most common research themes published in Gynecologic Oncology over a thirty-year period and to subsequently evaluate how interest in these topics changed over time. METHODS: We retrieved the abstracts of all original research published in Gynecologic Oncology from 1990 to 2020 using PubMed. Abstract text was processed through a natural language processing algorithm and clustered into topical themes using latent Dirichlet allocation (LDA) prior to manual labeling. Topics were investigated for temporal trends. RESULTS: We retrieved 12,586 original research articles, of which 11,217 were evaluable for subsequent analysis. Twenty-three research topics were selected at the completion of topic modeling. The topics of basic science genetics, epidemiologic methods, and chemotherapy experienced the greatest increase over the time period, while postoperative outcomes, reproductive age cancer management, and cervical dysplasia experienced the greatest decline. Interest in basic science research remained relatively constant. Topics were additionally reviewed for words indicative of either surgical or medical therapy. Both surgical and medical topics saw increasing interest, with surgical topics experiencing a greater increase and representing a higher proportion of published topics. CONCLUSIONS: Topic modeling, a type of unsupervised machine learning, was successfully used to identify trends in research themes. The application of this technique provided insight into how the field of gynecologic oncology values the components of its scope of practice and therefore how it may choose to allocate grant funding, disseminate research, and participate in the public discourse.
Asunto(s)
Neoplasias de los Genitales Femeninos , Femenino , Humanos , Neoplasias de los Genitales Femeninos/terapia , Publicaciones , Aprendizaje AutomáticoRESUMEN
SUMMARY: Many patients seeking orthognathic jaw surgery also experience nasal obstruction. Current transoral functional rhinoplasty techniques include septoplasty and inferior turbinate reduction, which are performed through the mouth after maxillary downfracture. Although powerful, these interventions do not treat dynamic nasal sidewall collapse. The authors describe a novel transoral alar batten graft. Using the maxillary vestibular approach, septal cartilage is harvested and delivered from the maxillary vestibule into the nasal alar-sidewall junction through a small tunnel. This procedure is simple and versatile, and carries minimal morbidity, enabling the orthognathic jaw surgeon to support the nasal sidewall through a minimal-access approach and improve the patient's nasal function and airway.
Asunto(s)
Obstrucción Nasal , Procedimientos Quirúrgicos Ortognáticos , Rinoplastia , Humanos , Rinoplastia/métodos , Nariz/cirugía , Obstrucción Nasal/etiología , Obstrucción Nasal/cirugía , Cartílago/trasplante , Tabique Nasal/cirugíaRESUMEN
Microsurgical advances have led to minimally invasive approaches for mandibular reconstruction. Currently, no resource compares all minimally invasive microvascular mandibular reconstruction (MIMMR) treatment options. Methods: All known cases of MIMMR were identified following the Preferred Reporting Items for Systematic Reviews, Meta-Analyses guidelines, and our own surgical experience. Patient demographics, MIMMR type [submandibular (SM), modified facelift/retroauricular (MFL/RA), or intraoral (IO)], methodology, and clinical outcomes were analyzed with the Fisher exact and Kruskal-Wallis tests. Results: Forty-seven patients underwent MIMMR. Ameloblastoma was the most common pathology treated using all approaches, and MFL/RA was the only approach used to treat squamous cell carcinoma (P = 0.0103). Reconstruction was reported for large, bilateral defects only via the SM or IO approach (P = 0.0216). The iliac crest or fibula was used as a donor site. The facial artery was the most common recipient vessel using the IO and SM approaches, whereas the superior thyroid and external carotid vessels were the most common in the MFL/RA approach (P < 0.0001). Virtual planning was used in all cases performed via an IO approach, 80.0% of cases using an SM approach, and no MFL/RA cases (P < 0.0001). Good aesthetic and functional outcomes were reported for every patient, and there was no difference in complication rates (P = 0.2880). Conclusions: Minimally invasive approaches are safe and effective treatment options for patients requiring mandibular microsurgery, usually in the setting of benign pathology. The IO and SM approaches usually rely on the facial vessels, whereas the MFL/RA approach permits access to the superior thyroid and external carotid vessels and cervical lymphadenectomy.
RESUMEN
BACKGROUND: Management of cleft lip and palate has been well characterized in pediatric patients, but limited data exist regarding the long-term functional outcomes of cleft patients once they reach adulthood. MATERIALS AND METHODS: An institutional, cross-sectional survey of adult patients with a history of cleft lip and/or palate was performed. The survey recorded patient characteristics, concerns, and barriers to care. Patient-reported outcome measures were assessed using the Nasal Obstruction Symptom Evaluation Scale, Epworth Sleepiness Scale, Mandibular Function Impairment Questionnaire, and the CLEFT-Q Speech Modules. RESULTS: A total of 63 patients (18.2%) participated in the survey. The mean patient age was 43.7 years (median: 41 y, range: 19-93 y), and the most common diagnosis was cleft lip and palate (51%) followed by isolated cleft palate (35%) and isolated cleft lip (14%). A subset of patients scored with moderate to severe dysfunction on each outcome measure including the Nasal Obstruction Symptom Evaluation Instrument (59%), Epworth Sleepiness Scale (7%), and Mandibular Function Impairment Questionnaire (13%). Respondent scores on the CLEFT-Q Speech modules demonstrated a bimodal distribution with lower scores in a significant subset of patients with cleft palate and cleft lip and palate. Many respondents (41%) were interested in clinical evaluation but cited barriers to seeking treatment including financial barriers (35%) or lack awareness of clinical options (27%). CONCLUSIONS: Many cleft patients have persistent needs or concerns in adulthood, especially regarding speech and nasal breathing. Systemic barriers pose challenges to these patients undergoing clinical evaluation.
Asunto(s)
Labio Leporino , Fisura del Paladar , Obstrucción Nasal , Humanos , Adulto , Niño , Labio Leporino/diagnóstico , Labio Leporino/cirugía , Fisura del Paladar/diagnóstico , Fisura del Paladar/cirugía , Estudios Transversales , SomnolenciaRESUMEN
Capsulorhexis in white intumescent cataract is often associated with extension leading to radial tears or biradial extension "Argentinian flag sign" and associated complications. We describe a novel technique of managing this situation. Vacuum rhexis is performed with a 24 G bent cannula attached to a 10-ml syringe. The flap is lifted and then caught using 24 G cannula. The suction is manually controlled by the surgeon, and the rhexis is completed using circumferential movement. The chamber stability is well maintained as there is no leakage, and the bent tip gives excellent fulcrum for movement. The advantage of this technique is that a single port is used for maintaining the chamber stability and it needs no special surgical instruments. Vacuum rhexis is a novel surgical technique used for performing successful continuous curvilinear capsulorhexis (CCC) of adequate size in white and intumescent cataracts, with a consistent and predictable outcome.
Asunto(s)
Catarata , Cápsula del Cristalino , Facoemulsificación , Humanos , Capsulorrexis/métodos , Cápsula del Cristalino/cirugía , Vacio , Succión , Facoemulsificación/métodosRESUMEN
OBJECTIVES: The NINDS rt-PA Stroke Study is frequently cited in support of alteplase for acute ischemic stroke within 3 h of symptom onset. Multiple post-hoc reanalyses of this trial have been published to adjust for a baseline imbalance in stroke severity. We performed a risk of selection bias assessment and reanalyzed trial data to determine if the etiology of this baseline imbalance was more likely due to random chance or randomization errors. METHODS: A risk of selection bias assessment was conducted using signaling questions from the Cochrane Risk of Bias 2 (ROB 2) tool. Four sensitivity analyses were conducted on the trial data based on the randomization process: assessment of imbalances in allocation in unique strata; adherence to a pre-specified restriction on randomization between time strata at each randomization center; assessment of differences in baseline computed tomography (CT) results in unique strata; and comparison of baseline characteristics between allocation groups within each time strata. A multivariable logistic regression model was used to compare reported treatment effects with revised treatment effects after adjustment of baseline imbalances identified in the sensitivity analyses. RESULTS: Based on criteria from the ROB 2 tool, the risk of bias arising from the randomization process was high. Sensitivity analyses found 11 of 16 unique strata deviated from the expected 1:1 allocation ratio. Three randomization centers violated an apriori rule regarding a maximum difference in allocation between the time strata. Three unique strata had imbalances in baseline CT results that prognostically favored alteplase. Four imbalances in baseline characteristics were identified in the 91-180-min time stratum that all prognostically favored alteplase and were consistent with a larger alteplase treatment effect size compared to the 0-90-min time stratum. After adjustments for baseline imbalances, all reported treatment effects were reduced. Three out of seven originally positive reported results were revised to non-significant. CONCLUSION: This risk of selection bias assessment revealed a high risk of selection bias in the NINDS rt-PA Stroke Study. Sensitivity analyses conducted based on the randomization process supported this assessment. Baseline imbalances in the trial were more likely due to randomization errors than random chance. Adjusted analyses accounting for baseline imbalances revealed a reduction in reported treatment effects supporting the presence of selection bias in the trial. Treatment decisions and guideline recommendations based on the original treatment effect reported in the NINDS rt-PA Stroke Study should be done cautiously.
Asunto(s)
Accidente Cerebrovascular Isquémico , Accidente Cerebrovascular , Humanos , National Institute of Neurological Disorders and Stroke (U.S.) , Sesgo de Selección , Accidente Cerebrovascular/diagnóstico , Accidente Cerebrovascular/tratamiento farmacológico , Factores de Tiempo , Activador de Tejido Plasminógeno/uso terapéutico , Resultado del Tratamiento , Estados UnidosRESUMEN
OBJECTIVE: Recent national guidelines recommend alteplase treatment for ischemic stroke within 4.5 h of symptom-onset based on meta-analyses of randomized controlled clinical trials (RCT). A detailed description of missing outcome data (MOD) due to participant loss to follow-up has never been published. The objective of this study was to perform a methodlogical survey on missing outcome data in an alteplase for ischemic stroke meta-analysis. MATERIALS AND METHODS: A methodological survey was performed on a chosen meta-analysis of alteplase for ischemic stroke RCTs that most closely aligns with recent national guideline recommendations. Data were collected to assess the number of participants lost to follow-up; differential lost to follow-up between allocation groups; baseline characteristics of those lost to follow-up; and the imputation methods used by individual trials and the chosen meta-analysis. The number of participants lost to follow-up was compared with the fragility index; and repeated for individually positive RCTs in the meta-analysis. RESULTS: The methodological survey revealed a substantial degree of missing information regarding MOD in the chosen meta-analysis and in individual RCTs. Single imputation was exclusively used in all RCTs and in the meta-analysis. The number of participants lost to follow-up was greater than the fragility index in the chosen meta-analysis and individually positive component RCTs suggesting that MOD may impact the direction of the reported effect or effect size. CONCLUSION: This methodological survey of an alteplase for ischemic stroke meta-analysis revealed MOD may be an important source of unrecognized bias. This survey highlights the need for sensitivity analyses using more robust methods of imputation.
Asunto(s)
Accidente Cerebrovascular Isquémico , Accidente Cerebrovascular , Fibrinolíticos/uso terapéutico , Humanos , Accidente Cerebrovascular/tratamiento farmacológico , Activador de Tejido Plasminógeno/uso terapéutico , Resultado del TratamientoRESUMEN
PURPOSE: The objective of this study was to test the feasibility of implementing a postoperative monitoring program for women with gynecologic cancers composed of patient-reported outcomes (PROs) and a wearable activity monitor. METHODS: We prospectively enrolled patients undergoing gynecologic cancer surgery to this single-arm study. Enrolled patients completed PROs (Patient-Reported Outcomes Measurement Information System physical function, sleep disturbance, anxiety, fatigue, and pain intensity) at baseline and one-week intervals for 4 weeks. They also wore a wearable accelerometer device that measured steps, heart rate, and intensity of physical activity. The primary outcome was feasibility. The secondary outcome was prediction of unscheduled contacts with the health care system on a given postoperative day. RESULTS: We enrolled 34 women. Three patients were unevaluable. The mean age was 58 years. The mean body mass index was 31 kg/m2; 17 patients were White (54.8%), 12 patients were Black (38.7%), and two patients (6.5%) were Asian. The overall wear time was 83.8%, and patients responded to 80.4% of the PRO instruments. Twenty-two patients (71%) had an unscheduled contact with the health care system postoperatively (median 1.5, 0.0-8.0). The day of an unscheduled health care utilization event was predicted with acceptable discrimination (area under the receiver operating characteristic curve 0.75; 95% CI, 0.67 to 0.81). PROs of fatigue and physical function were most predictive followed by wearable device outputs of lightly active minutes and average daily heart rate. CONCLUSION: Implementation of a postoperative monitoring program of patient-reported outcomes and a wearable device was feasible. The specific day of an unscheduled contact with the health care system was predicted with acceptable discrimination.
Asunto(s)
Neoplasias de los Genitales Femeninos , Dispositivos Electrónicos Vestibles , Fatiga , Estudios de Factibilidad , Femenino , Neoplasias de los Genitales Femeninos/cirugía , Humanos , Masculino , Persona de Mediana Edad , Medición de Resultados Informados por el PacienteRESUMEN
BACKGROUND: The Watch Tower Society, the main governing organization of the Jehovah's Witness (JW) faith, introduced the doctrine to refuse blood in 1945 and has been enforcing it since 1961. A member can be expelled for accepting prohibited blood components. Many reconstructive surgeries place patients at an increased risk for blood loss. There have been attempts at reducing the rate of transfusions in craniofacial surgery, even in patients not opposed to it. PRESENTATION: A 15-year-old female patient, who refused blood transfusion due to her faith, presented with a class III malocclusion, transverse maxillary constriction, and a lateral open bite. Surgery was deferred until the patient reached 18 years of age and had undergone preoperative orthodontics. A two-piece Le Fort I osteotomy was performed. Erythropoietin, oral iron, and tranexamic acid were used to minimize intraoperative blood loss. CONCLUSION: Here we discuss preoperative and intraoperative management strategies to ensure a transfusion-free environment for patients with religious objections to blood transfusions undergoing operations with increased bleeding risk.
RESUMEN
An 11-year-old male patient presented with a large, highly vascular, destructive mandibular mass. An intraoral biopsy showed pleomorphic spindle cells arranged in intersecting fascicles, with scattered atypia. A diagnosis of low-grade myofibroblastic sarcoma was made. The lesion was treated with preoperative arterial embolization followed by surgical resection and reconstruction with a fibular osteomyocutaneous free flap. To our knowledge, no reports of highly vascular mandibular low-grade myofibroblastic sarcoma are available in the literature.
RESUMEN
BACKGROUND: Although pathology in the maxillary and mandibular bones is rare in young patients, the differential diagnosis is broad. The World Health Organization (WHO) updated its classification of maxillofacial bone pathology in 2017. Using these updated guidelines, a systematic review of common maxillofacial bone lesions in the pediatric population was performed. METHODS: A PubMed search was conducted capturing English language articles from inception to July 2020. Thirty-one articles were identified that described the frequency of maxillofacial bone pathology. Data were extracted and organized using the WHO 2017 classification of odontogenic and maxillofacial bone tumors. Prevalence data were analyzed among diagnostic categories and geographical regions. The SAS version 9.4 was used to complete statistical analyses. RESULTS: The articles included patients from birth to a maximum age of 14 to 19âyears. The most common odontogenic cysts included radicular cyst (42.7%) and dentigerous cyst (39.0%) followed by odontogenic keratocyst (15.0%). Among odontogenic bone tumors, odontoma (49.3%) was most common followed by ameloblastoma (29.1%). The most common nonodontogenic bone tumor was fibrous dysplasia (42.4%), and the most common malignant bone tumor was osteosarcoma (75.0%). Significant variations were found by geographic region, with dentigerous cyst more common than radicular cyst, and ameloblastoma more common than odontoma in African and Asian countries (Pâ<â0.0001). CONCLUSIONS: This systematic review uses the WHO 2017 guidelines to classify common odontogenic and nonodontogenic maxillofacial bone lesions around the world. Pathogenesis, presentation, and available treatment options for the most common maxillofacial bone lesions are reviewed.
Asunto(s)
Ameloblastoma , Quiste Dentígero , Quistes Odontogénicos , Tumores Odontogénicos , Odontoma , Quiste Radicular , Adolescente , Adulto , Ameloblastoma/epidemiología , Niño , Quiste Dentígero/diagnóstico , Humanos , Quistes Odontogénicos/patología , Tumores Odontogénicos/diagnóstico , Quiste Radicular/diagnóstico , Adulto JovenRESUMEN
Vision-based localization is the problem of inferring the pose of the camera given a single image. One commonly used approach relies on image retrieval where the query input is compared against a database of localized support examples and its pose is inferred with the help of the retrieved items. This assumes that images taken from the same places consist of the same landmarks and thus would have similar feature representations. These representations can learn to be robust to different variations in capture conditions like time of the day or weather. In this work, we introduce a framework which aims at enhancing the performance of such retrieval-based localization methods. It consists in taking into account additional information available, such as GPS coordinates or temporal proximity in the acquisition of the images. More precisely, our method consists in constructing a graph based on this additional information that is later used to improve reliability of the retrieval process by filtering the feature representations of support and/or query images. We show that the proposed method is able to significantly improve the localization accuracy on two large scale datasets, as well as the mean average precision in classical image retrieval scenarios.
RESUMEN
BACKGROUND AND PURPOSE: Early identification of large vessel occlusions (LVO) and timely recanalization are paramount to improved clinical outcomes in acute ischemic stroke. A stroke assessment that maximizes sensitivity and specificity for LVOs is needed to identify these cases and not overburden the health system with unnecessary transfers. Machine learning techniques are being used for predictive modeling in many aspects of stroke care and may have potential in predicting LVO presence and mechanical thrombectomy (MT) candidacy. METHODS: Ischemic stroke patients treated at Loyola University Medical Center from July 2018 to June 2019 (N = 286) were included. Thirty-five clinical and demographic variables were analyzed using machine learning algorithms, including logistic regression, extreme gradient boosting, random forest (RF), and decision trees to build models predictive of LVO presence and MT candidacy by area of the curve (AUC) analysis. The best performing model was compared with prior stroke scales. RESULTS: When using all 35 variables, RF best predicted LVO presence (AUC = 0.907 ± 0.856-0.957) while logistic regression best predicted MT candidacy (AUC = 0.930 ± 0.886-0.974). When compact models were evaluated, a 10-feature RF model best predicted LVO (AUC = 0.841 ± 0.778-0.904) and an 8-feature RF model best predicted MT candidacy (AUC = 0.862 ± 0.782-0.942). The compact RF models had sensitivity, specificity, negative predictive value and positive predictive value of 0.81, 0.87, 0.92, 0.72 for LVO and 0.87, 0.97, 0.97, 0.86 for MT, respectively. The 10-feature RF model was superior at predicting LVO to all previous stroke scales (AUC 0.944 vs 0.759-0.878) and the 8-feature RF model was superior at predicting MT (AUC 0.970 vs 0.746-0.834). CONCLUSION: Random forest machine learning models utilizing clinical and demographic variables predicts LVO presence and MT candidacy with a high degree of accuracy in an ischemic stroke cohort. Further validation of this strategy for triage of stroke patients requires prospective and external validation.
Asunto(s)
Isquemia Encefálica , Accidente Cerebrovascular Isquémico , Accidente Cerebrovascular , Isquemia Encefálica/diagnóstico , Isquemia Encefálica/cirugía , Humanos , Aprendizaje Automático , Masculino , Estudios Prospectivos , Estudios Retrospectivos , Accidente Cerebrovascular/diagnóstico , Accidente Cerebrovascular/cirugía , Trombectomía , Activador de Tejido PlasminógenoRESUMEN
Background and Purpose: This study aims to determine whether machine learning (ML) and natural language processing (NLP) from electronic health records (EHR) improve the prediction of 30-day readmission after stroke. Methods: Among index stroke admissions between 2011 and 2016 at an academic medical center, we abstracted discrete data from the EHR on demographics, risk factors, medications, hospital complications, and discharge destination and unstructured textual data from clinician notes. Readmission was defined as any unplanned hospital admission within 30 days of discharge. We developed models to predict two separate outcomes, as follows: (1) 30-day all-cause readmission and (2) 30-day stroke readmission. We compared the performance of logistic regression with advanced ML algorithms. We used several NLP methods to generate additional features from unstructured textual reports. We evaluated the performance of prediction models using a five-fold validation and tested the best model in a held-out test dataset. Areas under the curve (AUCs) were used to compare discrimination of each model. Results: In a held-out test dataset, advanced ML methods along with NLP features out performed logistic regression for all-cause readmission (AUC, 0.64 vs. 0.58; p < 0.001) and stroke readmission prediction (AUC, 0.62 vs. 0.52; p < 0.001). Conclusion: NLP-enhanced machine learning models potentially advance our ability to predict readmission after stroke. However, further improvement is necessary before being implemented in clinical practice given the weak discrimination.