Using data mining for prediction of hospital length of stay: an application of the CRISP-DM Methodology

Hospitals are nowadays collecting vast amounts of data related with patient records. All this data hold valuable knowledge that can be used to improve hospital decision making. Data mining techniques aim precisely at the extraction of useful knowledge from raw data. This work describes an implementation of a medical data mining project approach based on the CRISP-DM methodology. Recent real-world data, from 2000 to 2013, were collected from a Portuguese hospital and related with inpatient hospitalization. The goal was to predict generic hospital Length Of Stay based on indicators that are commonly available at the hospitalization process (e.g., gender, age, episode type, medical specialty). At the data preparation stage, the data were cleaned and variables were selected and transformed, leading to 14 inputs. Next, at the modeling stage, a regression approach was adopted, where six learning methods were compared: Average Prediction, Multiple Regression, Decision Tree, Artificial Neural Network ensemble, Support Vector Machine and Random Forest. The best learning model was obtained by the Random Forest method, which presents a high quality coefficient of determination value (0.81). This model was then opened by using a sensitivity analysis procedure that revealed three influential input attributes: the hospital episode type, the physical service where the patient is hospitalized and the associated medical specialty. Such extracted knowledge confirmed that the obtained predictive model is credible and with potential value for supporting decisions of hospital managers.

Blood transfusion in cardiac surgery is a risk factor for increased hospital length of stay in adult patients

Abstract Background Allogeneic red blood cell (RBC) transfusion has been proposed as a negative indicator of quality in cardiac surgery. Hospital length of stay (LOS) may be a surrogate of poor outcome in transfused patients. Methods Data from 502 patients included in Transfusion Requirements After Cardiac Surgery (TRACS) study were analyzed to assess the relationship between RBC transfusion and hospital LOS in patients undergoing cardiac surgery and enrolled in the TRACS study. Results According to the status of RBC transfusion, patients were categorized into the following three groups: 1) 199 patients (40%) who did not receive RBC, 2) 241 patients (48%) who received 3 RBC units or fewer (low transfusion requirement group), and 3) 62 patients (12%) who received more than 3 RBC units (high transfusion requirement group). In a multivariable Cox proportional hazards model, the following factors were predictive of a prolonged hospital length of stay: age higher than 65 years, EuroSCORE, valvular surgery, combined procedure, LVEF lower than 40% and RBC transfusion of > 3 units. Conclusion RBC transfusion is an independent risk factor for increased LOS in patients undergoing cardiac surgery. This finding highlights the adequacy of a restrictive transfusion therapy in patients undergoing cardiac surgery. Trial registration Clinicaltrials.gov identifier: http://NCT01021631.

Finite mixture regression model with random effects: application to neonatal hospital length of stay

A two-component mixture regression model that allows simultaneously for heterogeneity and dependency among observations is proposed. By specifying random effects explicitly in the linear predictor of the mixture probability and the mixture components, parameter estimation is achieved by maximising the corresponding best linear unbiased prediction type log-likelihood. Approximate residual maximum likelihood estimates are obtained via an EM algorithm in the manner of generalised linear mixed model (GLMM). The method can be extended to a g-component mixture regression model with the component density from the exponential family, leading to the development of the class of finite mixture GLMM. For illustration, the method is applied to analyse neonatal length of stay (LOS). It is shown that identification of pertinent factors that influence hospital LOS can provide important information for health care planning and resource allocation. (C) 2002 Elsevier Science B.V. All rights reserved.

Pressure ulcers and nutritional deficits in elderly long-term care patients: Effects of a comprehensive nutritional program on pressure ulcer healing, length of hospital stay and charges to patients

The elderly are at the highest risk of developing pressure ulcers that result in prolonged hospitalization, high health care costs, increased mortality, and decreased quality of life. The burden of pressure ulcers will intensify because of a rapidly increasing elderly population in the United States (US). Poor nutrition is a major predictor of pressure ulcer formation. The purpose of this study was to examine the effects of a comprehensive, interdisciplinary nutritional protocol on: (1) pressure ulcer wound healing (2) length of hospital stays, and (3) charges for pressure ulcer management. Using a pre-intervention/post intervention quasi-experimental design the study sample was composed of 100 patients 60 years or older, admitted with or acquiring a pressure ulcer. A pre-intervention group (n= 50) received routine pressure ulcer care (standard diet, dressing changes, and equipment). A post-intervention group received routine care plus an interdisciplinary nutrition intervention (physical therapy, speech therapy, occupational therapy, added protein and calories to the diet). Research questions were analyzed using descriptive statistics, frequencies, Chi-Square Tests, and T-tests. Findings indicated that the comprehensive, interdisciplinary nutritional protocol had a significant effect on the rate of wound healing in Week3 and Week4, total hospital length of stay (pre-intervention M= 43.2 days, SD=31.70 versus M=31.77, SID-12.02 post-intervention), and pressure ulcer length of stay (pre-intervention 25.28 days, SD5.60 versus 18.40 days, SD 5.27 post-intervention). Although there was no significant difference in total charges for the pre-intervention group ($727,245.00) compared to the post-intervention group ($702,065.00), charges for speech (m=$5885.12, SD=$332.55), pre albumin (m=$808.52,SD= $332.55), and albumin($278 .88, SD=55.00) were higher in the pre-intervention group and charges for PT ($5721.26, SD$3655.24) and OT($2544 .64, SD=1712.863) were higher in the post-intervention group. Study findings indicate that this comprehensive nutritional intervention was effective in improving pressure ulcer wound healing, decreasing both hospital length of stay for treatment of pressure ulcer and total hospital length of stay while showing no significant additional charges for treatment of pressure ulcers.

Pressure ulcers and nutritional deficits in elderly long-term care patients : effects of a comprehensive nutritional program on pressure ulcer healing, length of hospital stay and charges to patients

The elderly are at the highest risk of developing pressure ulcers that result in prolonged hospitalization, high health care costs, increased mortality, and decreased quality of life. The burden of pressure ulcers will intensify because of a rapidly increasing elderly population in the United States (US). Poor nutrition is a major predictor of pressure ulcer formation. The purpose of this study was to examine the effects of a comprehensive, interdisciplinary nutritional protocol on: 1) pressure ulcer wound healing 2) length of hospital stays, and 3) charges for pressure ulcer management. Using a pre-intervention/post intervention quasi-experimental design the study sample was composed of 100 patients 60 years or older, admitted with or acquiring a pressure ulcer. A pre-intervention group (n= 50) received routine pressure ulcer care (standard diet, dressing changes, and equipment). A post-intervention group received routine care plus an interdisciplinary nutrition intervention (physical therapy, speech therapy, occupational therapy, added protein and calories to the diet). Research questions were analyzed using descriptive statistics, frequencies, Chi-Square Tests, and T-tests. Findings indicated that the comprehensive, interdisciplinary nutritional protocol had a significant effect on the rate of wound healing in Week3 and Week4, total hospital length of stay (pre-intervention M= 43.2 days, SD=31.70 versus M=31.77, SD=12.02 post-intervention), and pressure ulcer length of stay (pre-intervention 25.28 days, SD5.60 versus 18.40 days, SD 5.27 post-intervention). Although there was no significant difference in total charges for the pre-intervention group ($727,245.00) compared to the post-intervention group ($702,065.00), charges for speech (m=$5885.12, SD=$332.55), pre albumin (m=$808.52,SD= $332.55), and albumin($278 .88, SD=55.00) were higher in the pre-intervention group and charges for PT ($5721.26, SD$3655.24) and OT($2544 .64, SD=1712.863) were higher in the post-intervention group. Study findings indicate that this comprehensive nutritional intervention was effective in improving pressure ulcer wound healing, decreasing both hospital length of stay for treatment of pressure ulcer and total hospital length of stay while showing no significant additional charges for treatment of pressure ulcers.

Walking in hospital is associated with a shorter length of stay in older medical inpatients

Evidence suggests that inactivity during a hospital stay is associated with poor health outcomes in older medical inpatients. We aimed to estimate the associations of average daily step-count (walking) in hospital with physical performance and length of stay in this population. Medical in-patients aged ⩾65 years, premorbidly mobile, with an anticipated length of stay ⩾3 d, were recruited. Measurements included average daily step-count, continuously recorded until discharge, or for a maximum of 7 d (Stepwatch Activity Monitor); co-morbidity (CIRS-G); frailty (SHARE F-I); and baseline and end-of-study physical performance (short physical performance battery). Linear regression models were used to estimate associations between step-count and end-of-study physical performance or length of stay. Length of stay was log transformed in the first model, and step-count was log transformed in both models. Similar models were used to adjust for potential confounders. Data from 154 patients (mean 77 years, SD 7.4) were analysed. The unadjusted models estimated for each unit increase in the natural log of stepcount, the natural log of length of stay decreased by 0.18 (95% CI −0.27 to −0.09). After adjustment of potential confounders, while the strength of the inverse association was attenuated, it remained significant (βlog(steps) = −0.15, 95%CI −0.26 to −0.03). The back-transformed result suggested that a 50% increase in step-count was associated with a 6% shorter length of stay. There was no apparent association between step-count and end-of-study physical performance once baseline physical performance was adjusted for. The results indicate that step-count is independently associated with hospital length of stay, and merits further investigation.

Atrial Fibrillation in Decompensated Heart Failure: Associated Factors and In-Hospital Outcome

Background: Studies on atrial fibrillation (AF) in decompensated heart failure (DHF) are scarce in Brazil. Objectives: To determine AF prevalence, its types and associated factors in patients hospitalized due to DHF; to assess their thromboembolic risk profile and anticoagulation rate; and to assess the impact of AF on in-hospital mortality and hospital length of stay. Methods: Retrospective, observational, cross-sectional study of incident cases including 659 consecutive hospitalizations due to DHF, from 01/01/2006 to 12/31/2011. The thromboembolic risk was assessed by using CHADSVASc score. On univariate analysis, the chi-square, Student t and Mann Whitney tests were used. On multivariate analysis, logistic regression was used. Results: The prevalence of AF was 40%, and the permanent type predominated (73.5%). On multivariate model, AF associated with advanced age (p < 0.0001), non-ischemic etiology (p = 0.02), right ventricular dysfunction (p = 0.03), lower systolic blood pressure (SBP) (p = 0.02), higher ejection fraction (EF) (p < 0.0001) and enlarged left atrium (LA) (p < 0.0001). The median CHADSVASc score was 4, and 90% of the cases had it ≥ 2. The anticoagulation rate was 52.8% on admission and 66.8% on discharge, being lower for higher scores. The group with AF had higher in-hospital mortality (11.0% versus 8.1%, p = 0.21) and longer hospital length of stay (20.5 ± 16 versus 16.3 ± 12, p = 0.001). Conclusions: Atrial fibrillation is frequent in DHF, the most prevalent type being permanent AF. Atrial fibrillation is associated with more advanced age, non-ischemic etiology, right ventricular dysfunction, lower SBP, higher EF and enlarged LA. Despite the high thromboembolic risk profile, anticoagulation is underutilized. The presence of AF is associated with longer hospital length of stay and high mortality.

Experience in the Surgical Management of Abdominal Compartment Syndrome in the Hospital Occidente de Kennedy

The abdominal compartment syndrome (ACS) is the result of various physiological alterations produced by an abnormal increase of the intra-abdominal pressure. Some of these patients will undergo a surgical procedure for its management. Methods: This is a retrospective case series of 28 patients with ACS who required surgical treatment at the Hospital Occidente de Kennedy between 1999 and 2003. We assessed retrospectively the behavior of McNelis’s equation for prediction of the development of the ACS. Results: The leading cause of ACS in our study was intraabadominal infection (n=6 21,4%). Time elapsed between diagnosis and surgical decompression was less than 4 hours in 75% (n=21) of the cases. The variables that improved significantly after the surgical decompression were CVP (T: 4,0 p: 0,0001), PIM (T: 2,7; p: 0,004), PIA (T1,8; p:0,034) and Urine Output (T:-2,4; p:0,02). The values of BUN, Creatinine and the cardiovascular instability did not show improvement. The ICU and hospital length of stay were 11 days (SD: 9) and 18 days (SD13) respectively. Global mortality was 67,9% (n=19) and mortality directly attributable to the syndrome was 30% (n=8). The behavior of the McNelis’s equation was erratic. Conclusions: The demographic characteristics as well as disease processes associated with ACS are consistent with the literature. The association between physiological variables and ACS is heterogeneous between patients. Mortality rates attributable to ACS in our institution are within the range described world-wide. The behavior of the McNelis’s equation seems to depend greatly upon fluid balance.

Derrame Pleural Parapneumônico: perfil e evolução de crianças internadas no Hospital Universitário João de Barros Barreto

Introdução: A pneumonia é uma das infecções mais comuns em crianças, sendo o derrame pleural uma complicação potencial, principalmente nos países em desenvolvimento, onde existem sérias limitações de recursos diagnóstico-terapêuticos. Objetivos: Conhecer o perfil das crianças portadoras de derrame pleural parapneumônico, analisando sua evolução a partir da terapêutica clínica e cirúrgica que são submetidas. Método: Estudo descritivo, transversal, de coleta prospectiva, no qual foram estudadas todas as crianças internadas em um hospital de referência de doenças infecciosas na Região Norte-Brasil com diagnóstico de derrame pleural parapneumônico, submetidas a procedimento cirúrgico para abordagem do mesmo, no período de outubro de 2010 a outubro de 2011. Resultados: A amostra foi composta por 46 crianças, com distribuição igual entre os sexos, predominância de menores de 3 anos (74%) e procedentes do interior do estado (54,3%). Parcela considerável (28%) possuía algum grau de inadequação do estado nutricional. A média de tempo de doença até à admissão na referida instituição foi de 16,9 dias, e todos os indivíduos eram provenientes de outro hospital. A duração média da internação hospitalar foi de 26,0 dias, e a do estado febril, de 9,8 dias. Foram utilizados 2,2 esquemas antimicrobianos, em média, por paciente e o Ceftriaxone foi o antibiótico mais comum. Diagnóstico etiológico só foi alcançado em um único caso. Em 22 crianças, (47,8%), havia empiema pleural, e elas apresentaram maior tempo de drenagem. Foi encontrada associação entre crianças com nanismo e cirurgias múltiplas (Teste G = 8,40; p= 0,040). A grande maioria das criança foi operada uma única vez (80,4%), e a drenagem torácica fechada foi a cirurgia mais realizada. A drenagem torácica aberta foi instalada em 24,0% dos pacientes. A toracotomia com descorticação foi realizada em 2 pacientes (4,0%). Todas as crianças da amostra obtiveram recuperação clínica e radiológica em até quatro meses após a alta hospitalar, e não houve óbito na amostra. Conclusão: A população estudada possui doença em estágio avançado e distúrbios nutricionais prevalentes, que podem influenciar na evolução. É necessária padronização da antibioticoterapia, e reconsiderar a drenagem torácica aberta no empiema pleural, a partir de novos estudos sobre o tema, principalmente na indisponibilidade da videotoracoscopia.

Aplicação do escore MELD em pacientes submetidos a transplante de fígado: análise retrospectiva da sobrevida e dos fatores preditivos a curto e longo prazo

BACKGROUND: The model for end-stage liver disease (MELD) was developed to predict short-term mortality in patients with cirrhosis. There are few reports studying the correlation between MELD and long-term posttransplantation survival. AIM: To assess the value of pretransplant MELD in the prediction of posttransplant survival. METHODS: The adult patients (age >18 years) who underwent liver transplantation were examined in a retrospective longitudinal cohort of patients, through the prospective data base. We excluded acute liver failure, retransplantation and reduced or split-livers. The liver donors were evaluated according to: age, sex, weight, creatinine, bilirubin, sodium, aspartate aminotransferase, personal antecedents, brain death cause, steatosis, expanded criteria donor number and index donor risk. The recipients' data were: sex, age, weight, chronic hepatic disease, Child-Turcotte-Pugh points, pretransplant and initial MELD score, pretransplant creatinine clearance, sodium, cold and warm ischemia times, hospital length of stay, blood requirements, and alanine aminotransferase (ALT >1,000 UI/L = liver dysfunction). The Kaplan-Meier method with the log-rank test was used for the univariable analyses of posttransplant patient survival. For the multivariable analyses the Cox proportional hazard regression method with the stepwise procedure was used with stratifying sodium and MELD as variables. ROC curve was used to define area under the curve for MELD and Child-Turcotte-Pugh. RESULTS: A total of 232 patients with 10 years follow up were available. The MELD cutoff was 20 and Child-Turcotte-Pugh cutoff was 11.5. For MELD score > 20, the risk factors for death were: red cell requirements, liver dysfunction and donor's sodium. For the patients with hyponatremia the risk factors were: negative delta-MELD score, red cell requirements, liver dysfunction and donor's sodium. The regression univariated analyses came up with the following risk factors for death: score MELD > 25, blood requirements, recipient creatinine clearance pretransplant and age donor >50. After stepwise analyses, only red cell requirement was predictive. Patients with MELD score < 25 had a 68.86%, 50,44% and 41,50% chance for 1, 5 and 10-year survival and > 25 were 39.13%, 29.81% and 22.36% respectively. Patients without hyponatremia were 65.16%, 50.28% and 41,98% and with hyponatremia 44.44%, 34.28% and 28.57% respectively. Patients with IDR > 1.7 showed 53.7%, 27.71% and 13.85% and index donor risk <1.7 was 63.62%, 51.4% and 44.08%, respectively. Age donor > 50 years showed 38.4%, 26.21% and 13.1% and age donor <50 years showed 65.58%, 26.21% and 13.1%. Association with delta-MELD score did not show any significant difference. Expanded criteria donors were associated with primary non-function and severe liver dysfunction. Predictive factors for death were blood requirements, hyponatremia, liver dysfunction and donor's sodium. CONCLUSION: In conclusion MELD over 25, recipient's hyponatremia, blood requirements, donor's sodium were associated with poor survival.

A randomized comparative study of patients undergoing myocardial revascularization with or without cardiopulmonary bypass surgery: The MASS III Trial

The MASS III Trial is a large project from a single institution, The Heart Institute of the University of Sao Paulo, Brazil (InCor), enrolling patients with coronary artery disease and preserved ventricular function. The aim of the MASS III Trial is to compare medical effectiveness, cerebral injury, quality of life, and the cost-effectiveness of coronary surgery with and without of cardiopulmonary bypass in patients with multivessel coronary disease referred for both strategies. The primary endpoint should be a composite of cardiovascular mortality, cerebrovascular accident, nonfatal myocardial infarction, and refractory angina requiring revascularization. The secondary end points in this trial include noncardiac mortality, presence and severity of angina, quality of life based on the SF-36 Questionnaire, and cost-effectiveness at discharge and at 5-year follow-up. In this scenario, we will analyze the cost of the initial procedure, hospital length of stay, resource utilization, repeat hospitalization, and repeat revascularization events during the follow-up. Exercise capacity will be assessed at 6-months, 12-months, and the end of follow-up. A neurocognitive evaluation will be assessed in a subset of subjects using the Brain Resource Center computerized neurocognitive battery. Furthermore, magnetic resonance imaging will be made to detect any cerebral injury before and after procedures in patients who undergo coronary artery surgery with and without cardiopulmonary bypass.

A decision support system to facilitate management of patients with acute gastrointestinal bleeding

Objective: To develop a model to predict the bleeding source and identify the cohort amongst patients with acute gastrointestinal bleeding (GIB) who require urgent intervention, including endoscopy. Patients with acute GIB, an unpredictable event, are most commonly evaluated and managed by non-gastroenterologists. Rapid and consistently reliable risk stratification of patients with acute GIB for urgent endoscopy may potentially improve outcomes amongst such patients by targeting scarce health-care resources to those who need it the most. Design and methods: Using ICD-9 codes for acute GIB, 189 patients with acute GIB and all. available data variables required to develop and test models were identified from a hospital medical records database. Data on 122 patients was utilized for development of the model and on 67 patients utilized to perform comparative analysis of the models. Clinical data such as presenting signs and symptoms, demographic data, presence of co-morbidities, laboratory data and corresponding endoscopic diagnosis and outcomes were collected. Clinical data and endoscopic diagnosis collected for each patient was utilized to retrospectively ascertain optimal management for each patient. Clinical presentations and corresponding treatment was utilized as training examples. Eight mathematical models including artificial neural network (ANN), support vector machine (SVM), k-nearest neighbor, linear discriminant analysis (LDA), shrunken centroid (SC), random forest (RF), logistic regression, and boosting were trained and tested. The performance of these models was compared using standard statistical analysis and ROC curves. Results: Overall the random forest model best predicted the source, need for resuscitation, and disposition with accuracies of approximately 80% or higher (accuracy for endoscopy was greater than 75%). The area under ROC curve for RF was greater than 0.85, indicating excellent performance by the random forest model Conclusion: While most mathematical models are effective as a decision support system for evaluation and management of patients with acute GIB, in our testing, the RF model consistently demonstrated the best performance. Amongst patients presenting with acute GIB, mathematical models may facilitate the identification of the source of GIB, need for intervention and allow optimization of care and healthcare resource allocation; these however require further validation. (c) 2007 Elsevier B.V. All rights reserved.

Variação clínica indesejada no parto no SNS português : o papel da oferta

RESUMO - A variação clínica indesejada é um problema global que atinge os diversos sistemas de saúde. Vários autores relacionaram as instituições de saúde e suas características (a oferta) com as decisões clínicas, originando variação clínica entre prestadores de cuidados de saúde. Este estudo procurou identificar a existência de variação clínica indesejada nos nascimentos assistidos no SNS entre 2002 e 2009. É conhecido que uma taxa elevada de cesarianas é prejudicial para as mães e crianças. Neste sentido, procurou analisar-se a variação na percentagem de cesarianas realizadas por hospital do SNS e a influência do número de profissionais nestes valores. A metodologia utilizada foi a análise de fontes de informação que incluíram a caracterização dos internamentos e o número e especialidade dos profissionais de saúde no SNS português. Os resultados permitem afirmar que existe de variação clínica indesejada nos nascimentos no SNS, nomeadamente: (1) a percentagem de cesarianas realizada por hospital varia entre 19,78% e 40,09%; (2) o número de médicos obstetras varia entre os hospitais do SNS, entre 2,1 e 31,1 por 1000 partos; (3) o número de enfermeiros obstetras varia entre 3,8 e 50,7 por 1000 partos; (4) o número médio de dias internamento da mulher é 1,54 dias mais curto nos partos vaginais, que nos partos por cesariana, e 1 dia para o tempo mediano; (5) o tempo mediano de internamento da mulher submetida a cesariana é mais curto nos hospitais que realizam mais este procedimento; (6) não existe relação entre a idade da mãe e a percentagem de cesarianas; (7) nem do número de profissionais de saúde ajustado por 1000 partos; (8) não é possível identificar alterações significativas na percentagem de cesarianas entre hospitais universitários e não universitários.

Use of corticosteroids and the outcome of infants with bronchopulmonary dysplasia

Ventilator-dependent premature infants are often treated with dexamethasone. Several trials showed that steroids while improve pulmonary compliance and facilitate extubation, some treated infants may have adverse effects, such as alterations of growth curves. We conducted this retrospective study to evaluate the effects of steroids on mechanical ventilation, oxygen therapy, hospital length stay and mortality, in ventilator-dependent infants with bronchopulmonary dysplasia (BPD) (defined as the need of oxygen supplementation at 28 days of life). Twenty-six newborns with BPD were evaluated during 9 -- 42 days postpartum (mean = 31 days) and were divided into two groups: Group I - 14 newborns that did not receive dexamethasone, and Group II - 12 newborns that received dexamethasone at 14 --21 days of life. Dexamethasone was given at a dose of 0.25 mg per kilogram of body weight twice daily intravenously for 3 days, after which the dose was tapered. RESULTS: There were no statistically significant differences in the mean length of mechanical ventilation (Group I - 37 days, Group II - 35 days); oxygen supplementation (Group I - 16 days, Group II - 29 days); hospital stay (Group I - 72 days, Group II - 113 days); mortality (Group I - 35.7%, Group II - 41.6%). At birth, Group II was lighter (BW: Group I - 1154 grams ± 302, Group II - 791 grams ± 165; p < 0.05) and smaller (height: Group I - 37.22 cm ± 3.3, Group II - 33.5 ± 2.4; p< 0.05) than Group I. At 40 weeks, there were no statistically significant differences between groups in relation to anthropometric measurements. CONCLUSIONS: The use of corticosteroids in bronchopulmonary dysplasic infants may influence the somatic growth during its use. However, after its suspension, a recovery seems to occur, suggesting that its influence could be transitory.

Colchicine to Reduce Atrial Fibrillation in the Postoperative Period of Myocardial Revascularization

Abstract Background: The high prevalence of atrial fibrillation (AF) in the postoperative period of myocardial revascularization surgery increases morbidity and mortality. Objective: To assess the efficacy of colchicine to prevent AF in the postoperative period of myocardial revascularization surgery, the impact of AF on hospital length of stay and death, and to identify its risk factors. Methods: Between May 2012 and November 2013, 140 patients submitted to myocardial revascularization surgery were randomized, 69 to the control group and 71 to the colchicine group. Colchicine was used at the dose of 1 mg orally, twice daily, preoperatively, and of 0.5 mg, twice daily, until hospital discharge. A single dose of 1 mg was administered to those admitted 12 hours or less before surgery. Results: The primary endpoint was AF rate in the postoperative period of myocardial revascularization surgery. Colchicine group patients showed no reduction in AF incidence as compared to control group patients (7.04% versus 13.04%, respectively; p = 0.271). There was no statistically significant difference between the groups regarding death from any cause rate (5.6% versus 10.1%; p = 0,363) and hospital length of stay (14.5 ± 11.5 versus 13.3 ± 9.4 days; p = 0.490). However, colchicine group patients had a higher infection rate (26.8% versus 8.7%; p = 0.007). Conclusion: The use of colchicine to prevent AF after myocardial revascularization surgery was not effective in the present study. Brazilian Registry of Clinical Trials number RBR-556dhr.