SCCM Resource Library-The Future of Mechanical Ventilation

Video Transcription

Welcome to this lecture on the future of mechanical ventilation. This is my conflict of interest slide.  Let's begin. First of all, anyone foolish enough to make predictions about the future technology should at least have the humility to admit how difficult it is to be accurate.  Lord Kelvin was one of the most respected scientists of his time, yet his predictions about x-rays could not have been more wrong.  Medical doctors of the 19th century had limited understanding of physiology, but that did not make them humble enough to avoid ridiculous predictions.  At the dawn of any new technology, there will be naysayers, even those people who should know better.  I like this one most of all. Imagine if that chairman of IBM from 1943 could see the present day when IBM is working on the first medical quantum computer.  Even though Albert Einstein saw no evidence of practical nuclear energy, you would think he could imagine it.  This brings us to the subject of artificial intelligence and how it is affecting our world.  I recall this event when it was inconceivable that a ventilator could defeat a human world chess champion.  But the interesting thing is that this event did not terminate chess competitions. Now the tournaments are held between teams of humans cooperating with computers.  Before I can make predictions about the future of mechanical ventilation, I want to present a brief history to set the contextual background.  Men have tried to resuscitate their colleagues probably as far back as the condition of respiratory failure was noticed.  The lower illustration shows two men resuscitation by extending the arms above the head for inspiration and pressing on the chest for expiration.  Believe it or not, I was taught this method in Boy Scouts in the early 1960s.  I remember when mouth-to-mouth resuscitation was invented. It was such a novelty that it was on display at the county fair.  This is perhaps the most bizarre resuscitation method, and I sure hope there was a one-way valve in that tube.  It appears that the earliest ventilator technology was based on some form of tank apparatus for negative pressure ventilation, the precursor of the modern iron lung.  Here is perhaps the simplest ventilator ever invented, just a handheld pump.  The Draeger company was one of the very first to develop resuscitation equipment because they were a mining company and mining accidents frequently required resuscitation.  This device was actually patented in Germany, with the ideal of creating perfect patient-ventilator synchrony.  About this time, the iron lung was improved and commercialized in the U.S. by legendary Jack Emerson, a layperson with a keen inventive mind.  Perhaps the first book describing mechanical ventilators was published by a group of anesthesiologists in the U.K.  It is a rich source of both history and technology, and I have used many illustrations from this book in this talk.  Another important innovation of the MUSHAM book was including actual pressure, volume, and flow waveforms generated by the different kinds of ventilators.  Nowadays, of course, ventilators display waveforms on their screens.  In the U.S., Ray Bennett was the first to pioneer innovative positive pressure ventilators, starting with the breathing assist device for the Air Force pilots.  He designed the Bennett PR series of ventilators and IPPB machines that most of us in my generation have used.  Also about this time, Forrest Berg also designed breathing assist machines for the Air Force.  When I entered the field in the late 1970s, the Bennett MA-1 was the most popular adult ventilator.  It had only one mode, volume control, continuous mandatory ventilation, or what is called volume assist control. Again, we see an invention by Jack Emerson.  The key feature of Emerson ventilators was that they were built to be very sturdy, often using actual plumbing supplies.  I also used a pediatric version of this early Engstrom ventilator from Denmark.  The Engstrom ventilators lasted up to the present day, where they have evolved into a ventilator made by General Electric called the Carescape R860.  One of the early popular ICU ventilators was the Servo 900 series.  It evolved from anesthesia machines, and I believe it was the first adult ventilator to offer pressure control modes.  Forrest Berg also invented the first infant ventilator that became very popular in the 1970s.  This was one of the few ventilators with a switch that allowed the operator to change the mode.  I started in respiratory therapy school in 1976. The next year, the profession got its first equipment book.  Out of 31 ventilators described, it only mentioned three modes of mechanical ventilation.  The latest book on respiratory care equipment lists almost 500 different names for modes of ventilation.  This slide shows the rise in ventilator complexity using a pictorial history of the evolution of ventilators.  Early ventilators were controlled by analog circuit with knobs and switches. The advent of digital electronics in the 1990s stimulated a rapid evolution of technology.  Today, ventilators are complex computers with many digital displays and virtual controls.  In the last 45 years or so, the technological complexity of mechanical ventilators has grown exponentially.  Unfortunately, the educational resources dedicated to teaching mechanical ventilation have not kept pace.  As a result, there is a growing knowledge gap on the part of clinicians charged with managing mechanical ventilation.  Coping with the complexity of mechanical ventilation requires the acquisition of a hierarchy of skills.  This hierarchy is the basis of a standardized approach to teaching mechanical ventilation. As with any new skill, one must first memorize some key terminology.  Next is to understand the fundamental physical and mathematical principles of patient-ventilator interaction.  The next level is to combine these concepts into a practical taxonomy for classifying modes of ventilation.  Being able to identify modes, we are in a position to compare them in terms of their ability to serve the three goals of mechanical ventilation.  Finally, all of this knowledge is used to select and manage the most appropriate mode for a particular patient at a specific point in time.  So, how do we deal with this kind of technological complexity? We begin at the most general level, that of defining the basic goals of mechanical ventilation.  Despite a multitude of indications for ventilation, there are only three basic goals for the outcome. These are safety, comfort, and liberation.  Safety means first to provide adequate gas exchange. In fact, you could view a ventilator as simply a device to manipulate the V-Q ratio of the lungs.  But two decades of research has shown that we must accomplish gas exchange in a way that protects the lungs.  That means selection of optimum values of PEEP to avoid atelectrauma and tidal volume dosage to avoid volutrauma.  Comfort means to promote patient-ventilator synchrony and the balance of the work of breathing between machine and patient.  Finally, liberation means to minimize the duration of ventilation and the risk of adverse events.  Modes of ventilation have evolved in complexity simply because engineers have developed multiple different ways to serve these goals.  The power of recognizing the goals of ventilation lies in the fact that there is only one predominant goal at any point in time.  At the start of ventilation, safety is the main goal. After the patient has stabilized and is recovering, comfort becomes the predominant goal.  Finally, liberation becomes the dominant goal until extubation. At each point in time, our assessment of the goal guides our selection of the best mode to serve that goal.  This greatly simplifies and rationalizes the mode selection process. Seen in this way, the goals of ventilation drive the selection of modes.  As the goals change, we select modes that better serve these goals. Engineers understand this and try to design more and more robust modes with multiple technological capabilities  to serve as many goals as possible without the need for humans to make selections for multiple modes or even mode settings.  The great variety of modes of ventilation makes a taxonomy necessary to distinguish among them.  This is the same reason we have a taxonomy for drugs, to distinguish brand names from generic names.  To begin with, we define a mode as a predetermined pattern of patient-ventilator interaction.  The important thing is to distinguish between arbitrary mode names coined by manufacturers for marketing purposes and mode tags or classifications based on a formal taxonomy.  A mode taxonomy we are referring to here has three main components. Any mode can be classified in terms of these three characteristics.  First is the control variable, that is, pressure control or volume control.  Second is the breath sequence, that of continuous mandatory ventilation, intermittent mandatory ventilation, or continuous spontaneous ventilation.  Third is the targeting scheme, which is basically the feedback control mechanism that makes the mode unique.  There are currently seven main targeting schemes used on ventilators today, and some modes have more than one.  The evolution of modes is essentially the evolution of the software targeting schemes, moving from manual to automatic to fully autonomous.  With that as a background, we can predict the future will contain artificial intelligence. Currently, the vast majority of modes require human intervention.  The ventilator displays information about the patient-ventilator interaction, and then clinicians combine this with other observations to manage ventilator settings.  Another approach described in the literature is the application of artificial intelligence in the form of decision support systems.  Here, the ventilator both gathers and analyzes information about patient-ventilator interactions, and then provides the clinician with some type of recommendations for ventilator management.  But the clinician makes the final decisions. There are some modes, however, that use artificial intelligence for total control of ventilator management.  No human intervention is needed, at least for some portion of the duration of ventilation.  This should not be surprising, given the degree to which artificial intelligence already touches every aspect of our lives.  Machine intelligence is dominating many fields and applications, including the military, business, stock market trading, gaming, and home automation.  Anyone who carries a smartphone is using advanced artificial intelligence daily. Take, for example, the advent of self-driving cars.  The argument is that when fully implemented, transportation will be quicker, safer, and cheaper.  The interesting thing to note is that decision-making skills will shift from individual operators to teams of software designers.  What does that have to do with mechanical ventilation? Well, the same argument can be made.  With full automation using artificial intelligence, mechanical ventilation will be more compliant with treatment protocols, more responsive to changes in patient conditions, and ultimately safer and cheaper.  What I am talking about is the expansion of what is called intelligent targeting systems.  Targeting systems are essentially feedback control programs that decide what optimum settings should be during mechanical ventilation.  This includes things like selecting pressure control versus volume control, mandatory versus spontaneous breaths, tidal volume dosage, PEEP, and FiO2 settings.  Of course, more sophisticated software requires more sophisticated computers. Quantum computers promise almost unlimited computing power for medical applications in the near future.  The power of medical computing is being distributed through the Internet of Things, and in particular, the Internet of Medical Things.  The Internet of Things describes physical objects that are embedded with sensors, processing ability, software, and other technologies that connect and exchange data with other devices and systems over the Internet or other communication networks.  As I see it, the future evolution of mechanical ventilation will depend on three key technologies.  First is sensor technology. We need to identify which signals from the patient are most appropriate for guiding the targeting systems and how to reliably measure those signals.  Second is artificial intelligence-based ventilator control software.  Third is the power of networked instruments to gather data to allow AI systems to learn as well as to monitor and control.  Ultimately, in the Internet of Medical Things, ventilator hardware will be little more than flow control modules that are linked both to a central network and to other ventilators.  The purpose will be central control and ongoing evolution of learning. AI will be self-programming using data from networked individual patients.  This programming will be, at least initially, augmented with the knowledge of human experts. In fact, I am working on such a ventilator in my spare time.

Video Summary

In this lecture on the future of mechanical ventilation, the speaker acknowledges the difficulty of making accurate predictions about future technology. The history of mechanical ventilation is briefly discussed, starting from ancient methods of resuscitation to the development of early ventilator technology. The speaker emphasizes the growing complexity of mechanical ventilators and the need for improved education and understanding among clinicians. The three goals of mechanical ventilation are identified as safety, comfort, and liberation. Different modes of ventilation are introduced and classified based on control variables, breath sequences, and targeting schemes. The speaker also highlights the potential for artificial intelligence (AI) in the future of mechanical ventilation, discussing the use of AI for decision support systems and total control of ventilator management. The role of sensor technology, AI-based control software, and networked instruments in the evolution of mechanical ventilation is highlighted. The speaker suggests that in the future, ventilator hardware will primarily serve as flow control modules, with AI systems self-programming and continuously evolving based on data from networked patients.

Asset Subtitle

Procedures, 2022

Asset Caption

This session will present a perspective on training our teams on mechanical ventilation. Speakers will review the curriculum needed to prepare future critical care professionals, how to prepare trainees for the emergence of artificial intelligence and automation in mechanical ventilation, and the role of simulation in training on mechanical ventilation.

Meta Tag

Content Type Presentation

Knowledge Area Procedures

Knowledge Level Intermediate

Knowledge Level Advanced

Membership Level Select

Tag Mechanical Ventilation

Year 2022

Keywords

future of mechanical ventilation

history of mechanical ventilation

complexity of mechanical ventilators

modes of ventilation

artificial intelligence in mechanical ventilation

	Society of Critical Care Medicine 500 Midway Drive Mount Prospect, IL 60056 USA Phone: +1 847 827-6888 Fax: +1 847 439-7226 Email: support@sccm.org	Contact Us About SCCM Newsroom Advertising & Sponsorship DONATE	MySCCM LearnICU Patients & Families Surviving Sepsis Campaign Critical Care Societies Collaborative	GET OUR NEWSLETTER
© Society of Critical Care Medicine. All rights reserved. \| Privacy Statement \| Terms & Conditions The Society of Critical Care Medicine, SCCM, and Critical Care Congress are registered trademarks of the Society of Critical Care Medicine.