Predicting Colonic Neoplasia Surgical Complications: A Machine Learning Approach.

BACKGROUND: A range of statistical approaches have been used to help predict outcomes associated with colectomy. The multifactorial nature of complications suggests that machine learning algorithms may be more accurate in determining postoperative outcomes by detecting nonlinear associations, which are not readily measured by traditional statistics. OBJECTIVE: The aim of this study was to investigate the utility of machine learning algorithms to predict complications in patients undergoing colectomy for colonic neoplasia. DESIGN: Retrospective analysis using decision tree, random forest, and artificial neural network classifiers to predict postoperative outcomes. SETTINGS: National Inpatient Sample database (2003-2017). PATIENTS: Adult patients who underwent elective colectomy with anastomosis for neoplasia. MAIN OUTCOME MEASURES: Performance was quantified using sensitivity, specificity, accuracy, and area under the receiver operating characteristic curve to predict the incidence of anastomotic leak, prolonged length of stay, and inpatient mortality. RESULTS: A total of 14,935 patients (4731 laparoscopic, 10,204 open) were included. They had an average age of 67 ± 12.2 years, and 53% of patients were women. The 3 machine learning models successfully identified patients who developed the measured complications. Although differences between model performances were largely insignificant, the neural network scored highest for most outcomes: predicting anastomotic leak, area under the receiver operating characteristic curve 0.88/0.93 (open/laparoscopic, 95% CI, 0.73-0.92/0.80-0.96); prolonged length of stay, area under the receiver operating characteristic curve 0.84/0.88 (open/laparoscopic, 95% CI, 0.82-0.85/0.85-0.91); and inpatient mortality, area under the receiver operating characteristic curve 0.90/0.92 (open/laparoscopic, 95% CI, 0.85-0.96/0.86-0.98). LIMITATIONS: The patients from the National Inpatient Sample database may not be an accurate sample of the population of all patients undergoing colectomy for colonic neoplasia and does not account for specific institutional and patient factors. CONCLUSIONS: Machine learning predicted postoperative complications in patients with colonic neoplasia undergoing colectomy with good performance. Although validation using external data and optimization of data quality will be required, these machine learning tools show great promise in assisting surgeons with risk-stratification of perioperative care to improve postoperative outcomes. See Video Abstract . PREDICCIN DE LAS COMPLICACIONES QUIRRGICAS DE LA NEOPLASIA DE COLON UN ENFOQUE DE MODELO DE APRENDIZAJE AUTOMTICO: ANTECEDENTES:Se han utilizado una variedad de enfoques estadísticos para ayudar a predecir los resultados asociados con la colectomía. La naturaleza multifactorial de las complicaciones sugiere que los algoritmos de aprendizaje automático pueden ser más precisos en determinar los resultados posoperatorios al detectar asociaciones no lineales, que generalmente no se miden en las estadísticas tradicionales.OBJETIVO:El objetivo de este estudio fue investigar la utilidad de los algoritmos de aprendizaje automático para predecir complicaciones en pacientes sometidos a colectomía por neoplasia de colon.DISEÑO:Análisis retrospectivo utilizando clasificadores de árboles de decisión, bosques aleatorios y redes neuronales artificiales para predecir los resultados posoperatorios.AJUSTE:Base de datos de la Muestra Nacional de Pacientes Hospitalizados (2003-2017).PACIENTES:Pacientes adultos sometidos a colectomía electiva con anastomosis por neoplasia.INTERVENCIONES:N/A.PRINCIPALES MEDIDAS DE RESULTADO:El rendimiento se cuantificó utilizando la sensibilidad, especificidad, precisión y la característica operativa del receptor del área bajo la curva para predecir la incidencia de fuga anastomótica, duración prolongada de la estancia hospitalaria y mortalidad de los pacientes hospitalizados.RESULTADOS:Se incluyeron un total de 14.935 pacientes (4.731 laparoscópicos, 10.204 abiertos). Presentaron una edad promedio de 67 ± 12,2 años y el 53% eran mujeres. Los tres modelos de aprendizaje automático identificaron con éxito a los pacientes que desarrollaron las complicaciones medidas. Aunque las diferencias entre el rendimiento del modelo fueron en gran medida insignificantes, la red neuronal obtuvo la puntuación más alta para la mayoría de los resultados: predicción de fuga anastomótica, característica operativa del receptor del área bajo la curva 0,88/0,93 (abierta/laparoscópica, IC del 95%: 0,73-0,92/0,80-0,96); duración prolongada de la estancia hospitalaria, característica operativa del receptor del área bajo la curva 0,84/0,88 (abierta/laparoscópica, IC del 95%: 0,82-0,85/0,85-0,91); y mortalidad de pacientes hospitalizados, característica operativa del receptor del área bajo la curva 0,90/0,92 (abierto/laparoscópico, IC del 95%: 0,85-0,96/0,86-0,98).LIMITACIONES:Los pacientes de la base de datos de la Muestra Nacional de Pacientes Hospitalizados pueden no ser una muestra precisa de la población de todos los pacientes sometidos a colectomía por neoplasia de colon y no tienen en cuenta factores institucionales y específicos del paciente.CONCLUSIONES:El aprendizaje automático predijo con buen rendimiento las complicaciones postoperatorias en pacientes con neoplasia de colon sometidos a colectomía. Aunque será necesaria la validación mediante datos externos y la optimización de la calidad de los datos, estas herramientas de aprendizaje automático son muy prometedoras para ayudar a los cirujanos con la estratificación de riesgos de la atención perioperatoria para mejorar los resultados posoperatorios. (Traducción-Dr. Fidel Ruiz Healy ).

Association between oral antimalarial medication administration and mortality among patients with Ebola virus disease: a multisite cohort study.

BACKGROUND: Empiric antimalarial treatment is a component of protocol-based management of Ebola virus disease (EVD), yet this approach has limited clinical evidence for patient-centered benefits. METHODS: This retrospective cohort study evaluated the association between antimalarial treatment and mortality among patients with confirmed EVD. The data was collected from five International Medical Corps operated Ebola Treatment Units (ETUs) in Sierra Leone and Liberia from 2014 through 2015. The standardized protocol used for patient care included empiric oral treatment with combination artemether and lumefantrine, twice daily for three days; however, only a subset of patients received treatment due to resource variability. The outcome of interest was mortality, comparing patients treated with oral antimalarials within 48-h of admission to those not treated. Analysis was conducted with logistic regression to generate adjusted odds ratios (aORs). Multivariable analyses controlled for ETU country, malaria rapid diagnostic test result, age, EVD cycle threshold value, symptoms of bleeding, diarrhea, dysphagia and dyspnea, and additional standard clinical treatments. RESULTS: Among the 424 cases analyzed, 376 (88.7%) received early oral antimalarials. Across all cases, mortality occurred in 57.5% (244). In comparing unadjusted mortality prevalence, early antimalarial treated cases yielded 55.1% mortality versus 77.1% mortality for those untreated (p = 0.005). Multivariable analysis demonstrated evidence of reduced aOR for mortality with early oral antimalarial treatment versus non-treatment (aOR = 0.34, 95% Confidence Interval: 0.12, 0.92, p = 0.039). CONCLUSION: Early oral antimalarial treatment in an EVD outbreak was associated with reduced mortality. Further study is warranted to investigate this association between early oral antimalarial treatment and mortality in EVD patients.

Association between treatment with oral third-generation cephalosporin antibiotics and mortality outcomes in Ebola virus disease: a multinational retrospective cohort study.

OBJECTIVE: To evaluate the association between oral third-generation cephalosporin antibiotic treatment and mortality in Ebola virus disease (EVD). METHODS: This retrospective cohort studied EVD-infected patients admitted to five Ebola Treatment Units in Sierra Leone and Liberia during 2014-15. Empiric treatment with cefixime 400 mg once daily for five days was the clinical protocol; however, due to resource variability, only a subset of patients received treatment. Data on sociodemographics, clinical characteristics, malaria status and Ebola viral loads were collected. The primary outcome was mortality compared between cases treated with cefixime within 48 h of admission to those not treated within 48 h. Propensity scores were derived using clinical covariates. Mortality between treated and untreated cases was compared using propensity-matched conditional logistic regression and bootstrapped log-linear regression analyses to calculate an odds ratio (OR) and relative risk (RR), respectively, with associated 95% confidence intervals (CI). RESULTS: Of 424 cases analysed, 360 (84.9%) met the cefixime treatment definition. The mean age was 30.5 years and 40.3% were male. Median cefixime treatment duration was 4 days (IQR: 3, 5). Among cefixime-treated patients, mortality was 54.7% (95% CI: 49.6-59.8%) vs. 73.4% (95% CI: 61.5-82.7%) in untreated patients. In conditional logistic regression, mortality likelihood was significantly lower among cases receiving cefixime (OR = 0.48, 95% CI: 0.32-0.71; P = 0.01). In the bootstrap analysis, a non-significant risk reduction was found with cefixime treatment (RR = 0.82, 95% CI: 0.64-1.16, P = 0.11). CONCLUSION: Early oral cefixime may be associated with reduced mortality in EVD and warrants further investigation.

OBJECTIF: Evaluer l'association entre le traitement antibiotique oral avec des céphalosporine de troisième génération et la mortalité dans la maladie au virus Ebola (MVE). MÉTHODES: Cette étude de cohorte rétrospective a été menée chez des patients infectés par la maladie au virus Ebola admis dans cinq unités de traitement Ebola en Sierra Leone et au Libéria en 2014-2015. Le traitement empirique avec Cefixime 400 mg une fois par jour pendant cinq jours était le protocole clinique. Cependant, en raison de la variabilité des ressources, seul un sous-ensemble de patients a reçu un traitement. Des données sur la sociodémographie, les caractéristiques cliniques, le statut du paludisme et les charges virales d'Ebola ont été collectées. Le critère principal était la mortalité comparée entre les cas traités au céfixime dans les 48 heures suivant l'admission et ceux non traités dans les 48 heures. Les scores de propension ont été dérivés à l'aide de covariables cliniques. La mortalité entre les cas traités et non traités a été comparée à l'aide d'analyses de régression logistique conditionnelle et de régression log-linéaire bootstrapées pour calculer respectivement un rapport de cotes (OR) et un risque relatif (RR), avec des intervalles de confiance (IC) à 95% associés. RÉSULTATS: Sur 424 cas analysés, 360 (84,9%) répondaient à la définition du traitement au céfixime. L'âge moyen était de 30,5 ans et 40,3% étaient des hommes. La durée médiane du traitement par le céfixime était de 4 jours (IQR: 3, 5). Parmi les patients traités au Cefixime, la mortalité était de 54,7% (IC95%: 49,6 à 59,8%) vs 73,4% (IC95%: 61,5 à 82,7%) chez les patients non traités. Dans la régression logistique conditionnelle, la probabilité de mortalité était significativement plus faible parmi les cas recevant du céfixime (OR = 0,48 ; IC95%: 0,32 à 0,71; P = 0,01). Dans l'analyse bootstrap, une réduction du risque non significative a été trouvée avec le traitement au céfixime (RR = 0,82, IC95%: 0,64 à 1,16 ; P = 0,11). CONCLUSION: Le céfixime par voie orale rapide peut être associé à une mortalité réduite dans la MVE et mérite une investigation plus approfondie.

Vitamin A Supplementation Was Associated with Reduced Mortality in Patients with Ebola Virus Disease during the West African Outbreak.

BACKGROUND: Micronutrient supplementation is recommended in Ebola virus disease (EVD); however, there are limited data on therapeutic impacts of specific micronutrients. OBJECTIVE: To evaluate the association between vitamin A supplementation and mortality in EVD. METHODS: This retrospective cohort included patients with EVD admitted to 5 International Medical Corps Ebola Treatment Units (ETUs) in 2 countries during 2014-2015. Protocolized treatments with micronutrients were used at all ETUs: however, because of resource constraints, only a subset of patients received vitamin A. Standardized data on demographics, clinical characteristics, malaria status, and Ebola viral loads (cycle threshold values) were collected. The outcome of interest was mortality between cases treated with 200,000 IU of vitamin A on care days 1 and/or 2, and those not. Propensity scores based on the first 48 h of care were derived using covariates of age, ETU duration, malaria status, cycle threshold values, and clinical symptoms. Patients were matched 1:1 using nearest neighbors with replacement. Mortality between cases treated and not treated with vitamin A was compared using generalized estimating equations to calculate RR with associated 95% CI. RESULTS: There were 424 cases analyzed, of which 330 (77.8%) were treated with vitamin A. The mean age was 30.5 y and 40.3% were men. The most common symptoms were diarrhea (85.6%), anorexia (80.7%), and abdominal pain (76.9%). Mortality proportions among cases treated and not treated with vitamin A were 55.0% and 71.9%, respectively. In the propensity-matched analysis, mortality was significantly lower among cases receiving vitamin A (RR = 0.77, 95% CI: 0.59, 0.99; P = 0.041). In a subgroup analysis of patients treated with multivitamins already containing vitamin A, additional vitamin A supplementation did not impact mortality. CONCLUSION: Early vitamin A supplementation was associated with reduced mortality in patients with EVD, and should be further studied and considered for use in future epidemics.