Share on facebook
Share on twitter
Share on linkedin

Project Spotlight: VentOS

VentOS seeks to provide a safe and efficient ventilator operating software that is compatible with various types of open-source and commercial ventilators.

Written by: Tif Ho

The Challenge

At the global onset of COVID-19 in March, medical experts predicted that there would be a worldwide  shortage of ventilators throughout the pandemic. This prediction brought on an outpouring of highly-skilled, altruistic volunteers who sought to fulfill the shortage. In a truly global effort, medical specialists, engineers, quality assurance specialists, project managers, manufacturers, and numerous other professionals from around the world came together in order to offer their time, skills, and resources. These generous volunteers self-organized into groups for the sole purpose of producing rapid-build ventilators that could be used to treat and save the lives of COVID-19 patients. The volunteers based their projects on the premise that these ventilators could be designed, manufactured, and distributed at a speed which would quickly meet demand. 

Problematically, as the groups of volunteers began working on ventilator designs, they came across several issues. Firstly, ventilators are complex medical machines. They generally require long build times and high levels of safety assurance. Volunteers could hasten this process by relying on commonly-found, non-proprietary materials and open-source designs. However, their ventilators were still required to undergo an extensive and often lengthy regulatory process

Further complicating the issue was that many of the volunteers were designing ventilators for the first time. The overwhelming amount of information on safety and lack of deep expertise made it difficult to discern accuracy and effectiveness of safety standards.

Secondly, the sheer number of ventilator groups and pressing circumstances raised questions about the uniformity and reliability of a ramped-up regulatory process. Each ventilator group relied on different safety standards and design processes. Further complicating the issue was that many of the volunteers were designing ventilators for the first time. The overwhelming amount of information on safety and lack of deep expertise made it difficult to discern accuracy and effectiveness of safety standards. These concerns tacked more time onto an already difficult regulatory process.

Thirdly, like existing ventilators, the rapid-build designs that volunteers were working on continued to require a skilled operator. However, research reveals that even in well-resourced Intensive Care Units (ICUs), there is a wide variability in ventilator management. Inadequate training and ventilator alarm systems can result in the death of a patient. The potential for these situations increases during a pandemic, when limited resources and the highly-contagious nature of a disease means that ventilators are often monitored by a single operator, without any checks on their processes. This is particularly the case in disadvantaged communities which have few resources, including a limited number of, or inadequately trained, ventilator operators. These factors reveal the need for ventilators to include Clinical Decision Support (CDS), so as to help operators make quick and accurate decisions in emergency situations.

The Solution

Driven by the belief that we can achieve more together, anesthesiologist Dr. Erich Schulz, Dr. Robert L. Read, and Engineer Ben Coombs initiated the VentOS project, supported and encouraged by numerous volunteers in the Helpful Engineering organisation. As many ventilator groups struggled to meet regulatory standards, the VentOS team has prioritised risk management and quality assurance as the foundation of its work.

They felt that they could best support the efforts and increase the possibility of success of these groups by developing frameworks and resources that the various projects could share. In pursuance of this goal, the team decided to design software that would aim to be compatible with multiple types of ventilators.

By bringing together experts from various open-source teams, the project hopes to aid development of robust ventilator software to make ventilation as safe as possible. Ultimately, this could extend to the development of advanced features that would enhance Clinical Support Decision by helping ventilator operators to comply with complex protocols, validate data accuracy, and weigh competing risks for effective decision-making in complex situations.

The mission of VentOS is “To create a free and open-source software library and embedded operating system to enable engineering teams to develop safe and effective invasive and non-invasive ventilators for diverse contexts.” Before initiating Project VentOS, the team reviewed more than 100 open-source ventilator projects. They hope to help these groups “see the most value from the software engineering work already done within these projects.”

helpful-engineering-ventos-system-architecture

Proposed VentOS system architecture.

The VentOS team is working to address multiple challenges, including :

  1. Hardware manufacturers of ventilators want software to work in the best possible way with their hardware’s capabilities, which may include only one of the four different pressure and flow sensors that they recommend as a minimum
  2. System administrators need to be able to preload sensible default values for specific medical settings or contexts
  3. Ventilator operators need to be alerted if the ventilator fails in any way
  4. Ventilator operators need to be able to suspend alarms briefly so that they can think clearly in stressful, emergency situations
  5. Ventilator operators need to know what has triggered an alarm so as to obtain guidance on causes and solutions
  6. Ventilator operators need to be able to finely adjust alarm parameters in order to avoid alarm fatigue
  7. System developers need to be able to access ventilator logs for ongoing system refinement and quality assurance

Planning to work in an agile pattern, they have laid out requirements for a target minimal viable product (MVP). By focusing initially on a relatively simple MVP the team intends to focus the bulk of their efforts in establishing robust software engineering processes. The MVP will include life-saving alarms with settings that range from usual to extreme for:

  • Overpressure
  • Pressure sensor miscalibration
  • Pausing alarm signals for two minutes
  • Disconnection alarm when no no breathing is detected in X seconds

Rendering of initial inspiratory (orange) and expiratory (blue) flow and pressure (green) signals that will act as test cases for complex signal processing algorithms optimised to run on inexpensive low-power microcontroller units. The test here is “Can the algorithm detect that a serious problem has occured at the 50 second point?” Can it do so despite the noise, and only relying on a single sensor?

VentOS is currently in the design phase. The team remains focused on creating high-quality software. They are using both python and C in the project, using the power of Python’s data analysis tools, while developing core C modules to run on microcontrollers. They are currently working  on converting initial experimental algorithms from Python to C. Thus far, both Coombs and Schulz have found that their greatest challenge during the project has been in breaking down the silos between ventilator groups in order to make use of the work that has already been done by these groups. Nonetheless, throughout, the VentOS team has continued to document their processes in order to achieve regulatory compliance so as to provide the best possible support to ventilator groups. Schulz states that ultimately, their goal is to create “verifiable, well-engineered safety-critical software” that can be embedded into various ventilator hardwares and used in diverse medical contexts. “We cannot put a well rested, highly trained medical specialist beside every ventilator in the world, but just maybe we can train the ventilators to fill in at least some of the gaps.”

How You Can Help

VentOS is currently in a pre-concept phase. In order to continue with their work, the team is seeking volunteers for their open roles:

View the full list of roles and descriptions here.

Want to volunteer?

Join the #project-ventos channel in the Helpful Engineering slack. Please include information about your background and how you would like to help.

Want to learn more?

Learn more about this project at https://gitlab.com/project-ventos/ventos

The designs in this article are presented As-Is. The goal is to present designs that can foster further discussion and be utilized in countries that permit this product. These are not finalized designs and do not represent certification from any country. You accept responsibility and release Helpful Engineering from liability for the manufacture or use of this product. This design was created in response to the announcement on March 10, 2020, from the HHS.  Secretary of the Department of Health and Human Services (HHS) who issued a declaration pursuant to the Public Readiness and Emergency Preparedness (PREP) Act

Link to Prep act. :https://www.phe.gov/Preparedness/legal/prepact/Pages/default.aspx

ALL WARRANTIES OF ANY KIND WHATSOEVER, EXPRESS, IMPLIED AND STATUTORY, ARE HEREBY DISCLAIMED. ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE HEREBY DISCLAIMED. THIS DEVICE (INCLUDING ANY ACCESSORIES AND COMPONENTS) IS PRESENTED ‘AS IS.’

 

Related posts

helpful-engineering-uvisor

Project Spotlight: UVisor

A group of global volunteers are currently developing UVisor, a low-cost helmet with a UV chamber and fan unit that is designed to be a better alternative to protective equipment options currently on the market.

blog-header-helpful-engineering-project-spotlight-marley

Project Spotlight: MARLEy

COVID-19 caused many schools to shut down. Consequently, students now lack access to hands-on learning. MARLEy tries to solve this problem through an Augmented Reality (AR) environment that allows for interactive learning.