This article is written in response to the MIT paper calling for the rapid deployment of an open-source, low-cost ventilator based on a Breath Valve Mask system plus an external drive mechanism created by MIT.

Like many with engineering history in ventilator development, I was excited to see globe-spanning work to quickly create or adapt emergency ventilators to address the mounting COVID-19 pandemic. I read the MIT paper with their BVM system and came away thinking the system isn’t the simple answer we all want. This is a summary of my review of the MIT system, along with ideas on alternatives for a quickly deployed emergency ventilator.

Before starting the immediate manufacture of a BVM driver system, there are two important questions.

First, will a BVM (breath valve mask) system be clinically effective, no matter how it’s driven?

I believe they might, but I’m not clinically-qualified to say; better to ask an MD in respiratory care and get the opinion of experienced Respiratory Therapists. However, I think the bag valve mask system needs several modifications to be effective for ARDS patients; modifications I describe later in this article.

Secondly, do hospital supply chains see enough BVM’s to support the projected patient load?

Automating BVM’s might work, but are there enough BVM’s which form the heart of this solution? Unfortunately, I don’t have visibility into this supply chain question. I recommend querying hospital supply chain professionals to clear this hurdle.

Better option

I would look into quickly making Vortran, O-Two, Medtronic PB 560 or other ventilators for which manufacturers are sharing already-validated designs. The final assembly of a ventilator has nuance, but it’s not rocket science relative to other electronic devices. It would be easier to fast track FDA-registration and manufacturing controls starting with a validated design versus starting from a university design of unknown safety and efficacy!

The main risks with the university and crowd-sourced designs are the big question: if they will be safe and effective, and I would caution against using an unproven design from inexperienced designers. Mechanical ventilation is nuanced. It is easy for the well-intended to cause injury with a ventilator design – despite noble goals. Don’t get me wrong, I celebrate the hard work and good intent of the academic and crowd-sourced designs, but I also choose to apply a reality filter to the designer’s lack of ventilation experience.

Gotchas for quick ventilator manufacture

For validated vents like the Medtronic PB 560 and others, it is the specialized components (electronic valves, blowers, breathing circuit filters) that I predict have deep supply chain challenges.

I would look first to Vortran or O-Two ventilators (similar to the famous Bird Mark 7 ventilators, now BD/Carefusion). These are mechanical-only solutions, already validated, FDA-approved, and simple to manufacture quickly at scale compared to university-designed electro-mechanical systems.

The Vortran and similar ventilators are designs that should be fast-tracked to put into immediate high-volume manufacturing, and I’m happy to see Vortran has partnered with Xerox to produce these in high volume. I say these things having worked on and appreciating the design of the Philips/Respironics V680 and Medtronic PB 980 ventilators. These higher-end ventilators are awesome machines; sophisticated (also expensive and challenging to make quickly at mass scale)!

BVM use considerations for COVID/ARDS

To be effective, a BVM would likely need to be paired with a breath path filter, to prevent virus spread to caregivers, PEEP valve with high setting options (clinicians indicate to us this is important to ARDS patients), invasive ventilation (due to high PEEP, and the chance of leakage from the mask), and supplemental O2. These seem like reasonable off the shelf configuration items, pending supply availability!

In summary, the MIT BVM driver is interesting and similar to other crowd-source ideas out there, but I disagree with some of the benefits reported in the MIT paper. Specifically, I doubt the reported cost advantages and have concerns about basic safety and reliability for a novice-designed life-support device. I would look elsewhere first (Vortran specifically), but I’m fully prepared to be proven wrong and pitch in to move forward with the BVM designs.