How Machine Learning Can Improve Mechanical Ventilator Control?

Mechanical ventilators have been used in hospitals and the healthcare industry for decades now and have been evolving ever since because of continuous technological advancements. With the introduction of artificial intelligence and machine learning, such respiratory care techniques have now been enhanced to better support critical patients as well as make the treatment affordable for families.

Additionally, the outbreak of the covid-19 pandemic has further increased the demand for mechanical ventilators and boosted the overall market revenue. According to a detailed report published by Research Dive, the North America mechanical ventilator market is predicted to generate a revenue of $963.4 million and grow at a 6.3% CAGR during the 2019-2025 forecast timeframe.

What is a Mechanical Ventilator?

Mechanical ventilators assist patients in breathing undergoing surgeries or who can’t breathe independently due to some respiratory diseases. In some cases, the sedated patient is connected to the mechanical ventilator via a hollow tube that is inserted into their mouth and down into their trachea. This process is called invasive ventilation. In the case of non-invasive ventilation, no tube is inserted into the patient’s airway instead, a mask is provided to the patient and can also be used at home easily for any difficulties in breathing.

Penetration of Machine Learning to Enhance Mechanical Ventilation Treatment

In both ventilation cases, the mechanical ventilator takes breathes for the patient by following a clinician-prescribed breathing waveform that is obtained from the patient’s respiratory measurement. Highly trained clinicians pay their complete attention if the ventilator’s performance matches the patient’s need and doesn’t cause lung damage. However, there are always slight chances of the process’s failure.

To prevent such harm, scientists have now integrated machine learning algorithms into the treatment by utilizing signals from an artificial lung. The ML algorithm will measure the airway pressure and will bring necessary adjustments to the airflow for providing more consistently matched clinician-prescribed values. This approach minimizes manual intervention and improves the ventilator’s performance, thus reducing the risks of lung damage.

How is it Better Than the Traditional PID Method?

Mechanical ventilators are usually controlled by the traditional PID (Proportion: comparison of the measured and target pressure, Integral: previous measurement’s sum, Differential: previous measurement’s difference) that relies on the history of errors between measured and desired states. This method has been used since the 17^th century and is considered outdated by most scientists and researchers since clinicians constantly need to tune the ventilator for certain critical cases. On the contrary, a machine-learned ventilator controller performs on the training data generated by the former method and offers a more accurate data-driven alternative. Moreover, this method also reduces the clinician’s intervention, thus automating the whole treatment.

Bottom Line

The above described deep-learning approach to mechanical ventilation treatment using an artificial lung and training dataset is the beginning to make a huge impact on medical science and real ventilators. The researched approach has proved to be more efficient than the traditional PID method and makes the whole treatment cost-effective. Moreover, the ML-based approach for mechanical ventilation control also reduces manual intervention and provides an accurate data-driven solution.

About the Author(s)

 

Niti Pawar

Niti Pawar is a creative writer with almost 2 years of experience in the industry and holds a Bachelor’s degree in Information Technology from Pune University. She is extremely passionate about writing and is keen to gain significant experience and expertise in the same field. Niti aspires to become an accomplished editor and a linguist in her upcoming years. Besides work, she actively participates in recreational activities like trekking, badminton, singing, and travelling.