Daily Medical News – Machine Learning Delivers Personalized Oxygen for Patients on Ventilators – Critical Care

Machine learning delivers personalized oxygen for patients on ventilators.

WhatsApp Group Join Now
Telegram Group Join Now
Instagram Group Join Now

By HospiMedica International Staff Writers
Posted on June 10, 2024

Supplemental oxygen is one of the most commonly prescribed treatments worldwide, with 13 to 20 million patients requiring mechanical ventilation each year. Mechanical ventilation is an important life support technology that facilitates the movement of air in and out of the lungs, acting like a bellows. Modern ventilators are a significant advance from the old “iron lung” machines commonly thought of. Today's devices are sophisticated, compact digital machines that administer oxygen through a small plastic tube inserted into the throat. Despite these technical improvements, determining the appropriate oxygen level for each patient is largely based on guesswork. While clinicians determine oxygen levels using devices that measure SpO2 saturation, which indicates the saturation of oxygen in a patient's blood, previous studies have not conclusively determined this. whether higher or lower SpO2 targets are more beneficial for patients;

To remove the guesswork involved in ventilation, a team at the University of Chicago Medicine (Chicago, IL, USA) used a machine learning model to determine how different levels of oxygenation can be adjusted based on individual patient characteristics. But how can it affect the results? Their findings suggest that personalized oxygen goals can significantly reduce mortality rates, potentially revolutionizing critical care practices. Early studies by various research groups attempted to determine whether oxygen levels were high or low, but in general, these studies were inconclusive. The UChicago Medicine researchers suggested that the neutral results of these trials may not mean that oxygen levels are unrelated to patient outcomes, but rather that the effects of different oxygen levels vary from patient to patient. Randomized trials have an average effect of zero.

Image: Personal oxygenation can improve outcomes for patients on ventilators (Image courtesy of 123RF)

As personalized medicine grows in popularity, there is increasing interest in leveraging machine learning to predict the best treatment for individual patients. In the field of mechanical ventilation, these predictive models can potentially determine ideal oxygen saturation for a patient based on specific characteristics such as age, gender, heart rate, body temperature, and reason for admission to the ICU. are The team and their colleagues used data from previous randomized trials to develop and refine their machine learning model. After initial development with US data, the model was applied to patient data from Australia and New Zealand. According to their findings, for patients whose oxygen levels were optimal according to the model, overall mortality decreased by 6.4 percent. It's important to note that outcomes cannot be universally predicted based on any single characteristic—for example, not all brain injury patients will benefit from low oxygen levels, statistical trends show. Having said that – this creates the need for a comprehensive tool like machine learning models. which integrates diverse patient data.

Despite the algorithm's complexity, the input variables are common clinical parameters, making it easy for healthcare teams to use such devices in the future. At UChicago Medicine, algorithms are already integrated directly into electronic health record (EHR) systems to support various clinical decisions. The researchers believe that mechanical ventilation can be administered in a similar way. For hospitals lacking the resources to integrate such advanced machine learning tools into EHRs, there is also the possibility of developing a web-based application that allows clinicians to enter patient characteristics and act like an online calculator. Will allow to get predictions. These applications require extensive validation, testing, and refinement before they can be clinically implemented, but the potential benefits justify investment in further research.

“If the results are valid and generalizable, the results are astounding,” said Derek Ingus, MD, a critical care specialist. “If one could immediately assign each patient to the appropriate group of their predicted benefit or harm and assign their oxygen target accordingly, this intervention could theoretically be the most severe disease in the history of the field. will make the greatest improvement in lives saved.”

Related Links:
University of Chicago Medicine

WhatsApp Group Join Now
Telegram Group Join Now
Instagram Group Join Now

Leave a Comment