{"id":27654,"date":"2022-06-17T09:25:16","date_gmt":"2022-06-17T14:25:16","guid":{"rendered":"https:\/\/saluddigital.com\/?p=27654"},"modified":"2025-10-20T11:45:23","modified_gmt":"2025-10-20T17:45:23","slug":"modelos-de-ia-capaces-de-predecir-riesgos-clinicos-en-hospitales","status":"publish","type":"post","link":"https:\/\/saluddigital.com\/en\/big-data\/modelos-de-ia-capaces-de-predecir-riesgos-clinicos-en-hospitales\/","title":{"rendered":"AI models capable of predicting clinical risks in hospitals"},"content":{"rendered":"
\n\t\t\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t

A recent study shows the evaluation of an Artificial Intelligence model capable of predicting different clinical risks in different hospitals in real time.<\/h2>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t

Recently the study \u201cPrediction models based on machine learning for different clinical risks in different hospitals: Evaluation of in vivo performance\u201d, was published in the Journal of Medical Internet<\/em> (JMIR). Its main objective was to evaluate clinical risk prediction models in live workflows and thus be able to compare their performance in that environment with their performance when using retrospective data.<\/p>

The importance of this study lies in the fact that they attempted a generalization of the results by applying the same research to three different use cases in three hospitals. Furthermore, the use of machine learning to develop clinical risk models is often limited to evaluations with retrospective data. This study shows the evaluation of the model through the use of data and clinical workflow in real time.<\/p>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t

\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\"\"\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t

The prediction models used were trained for the prediction of clinical risk of delirium, sepsis and acute kidney injury, in three different hospitals and with retrospective data. Likewise, these models of machine learning models, specifically deep learning, were used to train a tool called transformer model<\/em>.<\/p>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t

\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t

\u201cThe models were trained using a calibration tool that is common to all hospitals and use cases. The models had a common design, but were calibrated using data specific to each hospital. The models were implanted in these three hospitals and used in daily clinical practice. The predictions made by these models were recorded and correlated with the diagnosis at discharge. Its performance was compared with evaluations on retrospective data and interhospital evaluations were carried out, \u201dexplains the study.<\/p>

The results showed that the performance of the models using clinical workflow data was similar to using retrospective data. The average value of the Receiver Operating Characteristic Curve - ROC or receiver operating characteristic curve (AUROC), had a decrease value of 0.6% from 94.8 to 94.2%.<\/p>

\u201cCross-hospital assessments showed very poor performance: the mean AUROC decreased by 8 percentage points (from 94.2% to 86.3% at discharge), indicating the importance of model calibration with deployment hospital data\u201d , shows the study.<\/p>

Thus, the authors concluded that calibrating the model with data from the various hospitals achieves better results and model performance in live workflows. Check out the full study at the following link:<\/p>

https:\/\/www.jmir.org\/2022\/6\/e34295<\/a><\/p>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t

\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t
\n\t\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t\t\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/i><\/span>\n\t\t\t\t\t\t\t\t<\/i><\/span>\n\t\t\t\t\t\t\t\t\t\t\t\t\t<\/span>\n\t\t\t\t\t\t\t\t\t\t\t\t BIBLIOGRAPHY<\/a>\n\t\t\t\t\t<\/div>\n\n\t\t\t\t\t

JMIR<\/strong><\/p>

https:\/\/www.jmir.org\/2022\/6\/e34295<\/a><\/p><\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t<\/div>","protected":false},"excerpt":{"rendered":"

Estudio reciente muestra la evaluaci\u00f3n de un modelo de Inteligencia Artificial capaz de predecir diferentes riesgos cl\u00ednicos en distintos hospitales en tiempo real. Recientemente el estudio \u201cModelos de predicci\u00f3n basados en el aprendizaje autom\u00e1tico para diferentes riesgos cl\u00ednicos en distintos hospitales: Evaluaci\u00f3n del rendimiento en vivo\u201d, fue publicado en el Journal of Medical Internet (JMIR). Su objetivo principal fue evaluar modelos de predicci\u00f3n de riesgo cl\u00ednico en flujos de trabajo en vivo y as\u00ed poder comparar su rendimiento en ese entorno con su rendimiento cuando se utilizan datos retrospectivos. La importancia de este estudio radica en que intentaron una generalizaci\u00f3n de los resultados aplicando la misma investigaci\u00f3n a tres casos de uso diferentes en tres hospitales. Adem\u00e1s, el uso de aprendizaje autom\u00e1tico para desarrollar modelos de riesgo cl\u00ednico, suele limitarse a evaluaciones con datos retrospectivo. Este estudio muestra la evaluaci\u00f3n del modelo mediante el uso de datos y flujo de trabajo cl\u00ednico en tiempo real. Los modelos de predicci\u00f3n utilizados, fueron entrenados para la predicci\u00f3n de riesgo cl\u00ednico de delirio, sepsis y lesi\u00f3n renal aguda, en tres hospitales distintos y con datos retrospectivos. Asimismo, dichos modelos de modelos de aprendizaje autom\u00e1tico, en espec\u00edfico de aprendizaje profundo fueron utilizados para entrenar una herramienta llamada Transformer model. \u201cLos modelos se entrenaron utilizando una herramienta de calibraci\u00f3n que es com\u00fan para todos los hospitales y casos de uso. Los modelos ten\u00edan un dise\u00f1o com\u00fan, pero se calibraron utilizando los datos espec\u00edficos de cada hospital. Los modelos se implantaron en estos tres hospitales y se utilizaron en la pr\u00e1ctica cl\u00ednica diaria. Las predicciones realizadas por estos modelos se registraron y correlacionaron con el diagn\u00f3stico al alta. Se compar\u00f3 su rendimiento con evaluaciones sobre datos retrospectivos y se realizaron evaluaciones interhospitalarias\u201d, explica el estudio. Los resultados mostraron que, el rendimiento de los modelos con datos de flujos de trabajo cl\u00ednicos fue similar a la utilizaci\u00f3n de datos retrospectivos. El promedio del valor de la Curva Caracter\u00edstica Operativa del Receptor – ROC o receiver operating characteristic curve (AUROC), tuvo un valor decrecimiento de 0,6% de 94,8 a 94,2%. \u201cLas evaluaciones entre hospitales mostraron un rendimiento muy reducido: el AUROC medio disminuy\u00f3 en 8 puntos porcentuales (del 94,2% al 86,3% al alta), lo que indica la importancia de la calibraci\u00f3n del modelo con los datos del hospital de despliegue\u201d, muestra el estudio. De esta forma los autores concluyeron que la calibraci\u00f3n el modelo con datos de los diversos hospitales logra mejores resultados y rendimiento del modelo en flujos de trabajo en vivo. Consulta el estudio completo en el siguiente enlace: https:\/\/www.jmir.org\/2022\/6\/e34295 BIBLIOGRAF\u00cdA JMIR https:\/\/www.jmir.org\/2022\/6\/e34295<\/p>","protected":false},"author":1,"featured_media":27656,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[3399,156,160],"tags":[145],"class_list":["post-27654","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-ai-analitica","category-big-data","category-noticias","tag-noticias"],"aioseo_notices":[],"_links":{"self":[{"href":"https:\/\/saluddigital.com\/en\/wp-json\/wp\/v2\/posts\/27654","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/saluddigital.com\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/saluddigital.com\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/saluddigital.com\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/saluddigital.com\/en\/wp-json\/wp\/v2\/comments?post=27654"}],"version-history":[{"count":0,"href":"https:\/\/saluddigital.com\/en\/wp-json\/wp\/v2\/posts\/27654\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/saluddigital.com\/en\/wp-json\/wp\/v2\/media\/27656"}],"wp:attachment":[{"href":"https:\/\/saluddigital.com\/en\/wp-json\/wp\/v2\/media?parent=27654"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/saluddigital.com\/en\/wp-json\/wp\/v2\/categories?post=27654"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/saluddigital.com\/en\/wp-json\/wp\/v2\/tags?post=27654"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}