3D printed open source tourniquet: Rationale, failure analysis and proposed next steps of the Glia tourniquets during the Gaza protests (May 11)

Tarek Loubani
May 12, 2018 · 8 min read

Funding support: Shuttleworth Foundation; Division of Emergency Medicine, London Health Sciences Centre (EMLondon); Faculty of Medicine, University of Western Ontario

WARNING: Graphic images and content.

See part 2 of this report here: 3D printed tourniquet: Day 2 of Gaza field trials ends badly (May 14)

A s an emergency physician, I understand that there is never a happy ending to the terrible situations emergency physicians deal with, like car accidents or stabbings or gunshot wounds or heart attacks. There is only a less terrible ending. I thought this prepared me for what happened on Friday, May 11, 2018 in Gaza. It didn’t.

Why we need a 3D printed tourniquet

The Gaza Strip has faced three wars in the past decade with a total death count of at least 3,800 and 17,195 injuries. The majority of casualties were civilian[1–5]. Informal observations by physicians in Gaza indicate that many fatalities and casualties could have benefited from bleeding control techniques in the prehospital setting.

Reducing hemorrhagic deaths from limb exsanguination became a cornerstone project of the newly formed Hayat Center for Emergency and Crisis Management, based at the Islamic University of Gaza. The center has been engaged in projects like Advanced Cardiac Life Support (ACLS), Basic Life Support (BLS) and Trauma Life Support (TLS). When the center tried to obtain CAT tourniquets for its new “Stop Bleeding” campaign, they were almost impossible to obtain. In addition to the high cost of USD$50 per unit for 20,000 units paid in advance, approvals to get the supplies through Israel’s blockade could take many months — or never come at all. In contrast, in Canada I could order one from Amazon for CDN$43.30 (USD$33.85) with free shipping.

Even if the CAT could enter Gaza at a reasonable cost, it has fatal flaws when considered for its population. The CAT is targeted at frontline military personnel who are usually men with larger limbs. The Gaza population is mostly children, with 44.78% under the age of 14[6]. CATs do not work on a pediatric population, which would have automatically excluded most of the patient population we expect.

The Glia project developed 3D printing experience working on a high quality 3D printed stethoscope. We use locally manufactured Prusa i3 MK2 printers with locally manufactured filament from recycled and locally available plastic. The whole operation is powered by solar power to bypass the 2–4 hours of electricity our office gets daily. When members of Gaza’s disaster committee and the Hayat Center approached Glia’s Gaza team, it was an easy yes. Using open source techniques, we created the device and designed it for mixed adult and pediatric use.

A terrible decision to deal with a terrible situation: Why we deployed before we were ready

Work started a year ago and reached the field trial stage by early 2018. Glia’s 3D printed tourniquet could be manufactured in Gaza at a cost of 25 NIS (USD$7) and worked well in the lab and in initial field trails. By February 2018, it was obvious that a mix of economic and political factors would make this year’s peaceful protests leading up to May 15th larger than ever before, which raised the potential of violent Israeli suppression of the protests. Our team decided to produce the latest prototype design in quantity and stop work on all other projects. The Hayat Center held weekly training for paramedics, which continues until today.

Tourniquet training at Hayat Center. Gaza, 2018. Source: Glia. License: CC-By-SA

During and after the inaugural March 30th protest, there were many gunshot wounds in upper and lower limbs. Statistics compiled by the Palestinian Ministry of Health in Gaza (MOH) indicate that there were 2,842 injuries to the lower (2,237) and upper (605) limbs between March 30 and May 4, representing 59.6% of all injuries.

Due to the excellent training of first responders, there has not yet been a death associated with limb exsanguination. However, first responders were using makeshift tourniquets and inadequate tools, which led to unnecessary blood loss in clearly identified exsanguination cases.

Image for post
Image for post
Image for post
Image for post
Left: Obvious lower limb injury with marked blood loss in patient Abdulrahman Nofal in Gaza, April 17, 2018. This patient had hemorrhagic shock upon arrival at the ED. He survived with an amputation. Photo from Alray. Right: Distribution of injuries by location. Source: Palestinian Ministry of Health
Improvised tourniquet, Gaza 2018. Source unknown

Deployment on May 11, 2018

We prepared approximately 350 tourniquets. One hundred of these were used in the initial trial runs, trainings and other field testing. 150 were deployed to the field in two stages: 59 on May 11 at the main Gaza protest site, and 90 held in reserve for May 14th and May 15th. 100 more were delivered to the Hayat Center on May 12th for further training and distribution.

There were 25 trained first responders in the field in addition to Dr. Mohammed Al-Attar and myself. Dr. Al-Attar conducted a brief field training for other first responders who were not previously trained. Each first responder received two tourniquets. Ten tourniquets were deployed on patients wounded with live ammunition meeting criteria for tourniquet use. Three tourniquets failed. Seven tourniquets operated as intended. A total of 197 people were injured, though an exact breakdown of where each was injured is unavailable.

Field training on the use of trauma tourniquets. Gaza, 2018. Source: Glia Project. License: CC-By-SA

Failure #1: There are no ideal operating conditions

We dry-tested the tourniquets in laboratory conditions. Even in initial paramedic field trials, tourniquets were generally deployed in single-victim situations with controlled scenes.

In the field, we were trying to put tourniquets on patients while literally running, under live fire, being teargassed, or occasionally all three. This was suboptimal.

Image for post
Image for post
Image for post
Image for post
Left: Israelis launch teargas at protestors using a vehicle-mounted launcher. Right: A gunshot victim with an open tib-fib fracture and partial amputation of his left leg proximal to the ankle is being lifted by his friend. His friend was stopped from lifting him so we could deliver rapid primary treatment before evacuation. Source: AFP via El Comercio[1][2]

Failure #2: The packaging was designed badly

Look at this package. Where should you open it from? For us, it was obvious: Open from the non-label side (bottom), since opening from the label side would be a pain. Instead, first-time users of the package did what we would all expect given our previous packaging experiences as consumers: Open where the label is.

Our attempt to create a faster-opening system actually delayed use of the device. No easy place to grab the seal also lost precious seconds.

We responded to this by unpacking one unit per first responder for placement in an easily-accessible pocket. Since sterility is not an issue, this is not a problem. However, we have no great ideas for how to solve this problem going forward, and will have to trial some options to see what works best.

Failure #3: We deployed an old model we knew had the potential for internal belt failure

When we started training with the tourniquet, we quickly discovered a problem: Twisting the windlass (the stick) too many times resulted in a catastrophic failure of the tourniquet. This was especially true in lower limbs, where higher energy was required. This was largely due to a sharp edge on the windlass, which we discovered quickly.

After making corrections, we tested the new model to failure and found it withstood approximately 5 times more force before non-catastrophic failure resulting in an unstable tourniquet — more than enough for our use-cases.

I made a strategic decision, due to the large number of injuries and the low number of tourniquets for us to deploy the earlier model with the known problem. I thought — incorrectly — that they would survive a single use, which all of our laboratory testing showed that they should.

Approximately one quarter of the deployed units contained this defect. I used one of them on a patient with a partially amputed left leg, I think just below the knee. The tourniquet was applied quickly proximal to the knee while the patient was being carried on a stretcher. The stretcher bearers, paramedics and I were all running. He was not tourniqueted when I arrived. As the tourniquet was not tightly applied before I turned the windlass, I had to turn it more than is ideal. The tourniquet broke, and I used a torn up cloth for a makeshift tourniquet to salvage the situation. I recovered the pieces of the tourniquet after the patient was stabilized. As of today, we have pulled all remaining early models from our inventory.

Image for post
Image for post
Top: Recovered windlass of old model used on patient that failed catastrophically. Note sharp edge in belt slot. Bottom: New model. Note curved edges.

The second failed unit most likely harmed a patient and had a very real risk of killing him. This is the case I will lose sleep over.

The patient had a gunshot in the proximal left thigh that hit a vessel, likely not arterial. We arrived within about 15 seconds of the injury, and the tourniquet was placed by a trained first responder proximal to the injury. As he turned the windlass, it was as though somebody was turning off a tap, and the patient’s bleeding stopped. However, an untrained first responder wanted to turn the windlass a few more times for good measure, unbuckled the windlass from its holder, and turned it several more times until it broke.

The patient’s bleeding was controlled enough by the primary action of the tourniquet (pull and velcro) that we decided to transport rather than reapply another tourniquet. I do not know the outcome of this patient.

First responder applying manual pressure after catastrophic tourniquet failure in patient with left proximal thigh gunshot wound. Source: Glia. License: CC-By-SA

Failure #4: An unexpected buckle failure

A buckle failed on a moderately critical patient while in the first phase of application (tighten belt and velcro) while using the latest model of tourniquet. We have not seen this in any of our testing and did not expect it. Another tourniquet was applied to the patient and behaved normally. I recovered the pieces of the tourniquet for proper dissection in our lab.

Image for post
Image for post


The first large deployment of open source, 3D-printed tourniquets in a battlefield setting has been well-received by patients, bystanders and first responders. We also revealed several moderate and critical errors in the device and our thinking that we will hopefully remedy over the coming days and weeks.

If you would like to participate in this project, please don’t hesitate to contact us at info@glia.org. If you are more technically inclined, check out the tourniquet project’s Github page, where you can file issues or make pull requests.

Edit log

2018–05–12 2304h UTC-04: Added rationale re: pediatrics

Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch

Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore

Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store