Cristian Fracassi and Alessandro Romaioli are two engineers from Brescia, a city within Lombardy, the Northern Italian region hardest hit by the ongoing covid-19 pandemic. The duo owns a five-year-old engineering firm – Isinnova together. Recently, they came to know the supplier is facing difficulties in producing sufficient ventilator valves as a result of overwhelming demands. These valves are an important part within the ventilator which brings a mixed of pure oxygen and air to patients and need to be replaced constantly.
From 3D printed ventilator valves to new open dataset
Fracassi and Romaioli thought they could do something to help after finding out more from a local hospital and decided to replicate the valves using 3D printing. The duo joined forces with Lonati, a local 3D printer company but the initial attempt ended in failure. The less than a millimeter in diameter hole and tube within the valve are challenging the making process. Nevertheless, Fracassi and Romaioli managed to generate four prototypes in the end; each comes at a cost of less than a Euro and roughly takes about three hours to design.
They brought the valves to the hospital for testing and they work perfectly. As Fracassi and Romaioli are printing more valves to help several hospitals in Italy, they have expressed their willingness to share the 3D model, so others can do the same. At the same time, they noted the model they have recreated was Venturi valve and different ventilator may rely on valves with different designs. Fracassi and Romaioli also proposed to turn a snorkeling mask which can be easily obtained from the market into a ventilation-assisted mask for hospitals in need.
Fracassi and Romaioli are not alone, the pandemic may have slowed down economy but surely, it’s hastening technology adoption. On 16 March, researchers and leaders from the Allen Institute for AI (Artificial Intelligence); Chan Zuckerberg Initiative; Georgetown University’s Center for Security and Emerging Technology; Microsoft and the National Library of Medicine at the National Institutes of Health launched the COVID-19 Open Research Dataset (CORD-19).
The dataset comes with more than 29,000 published or on preprint server scientific articles. It’s believed to be the most comprehensive literature collection around the topic of coronavirus. Most notably, all articles come in a format that is easily digest by computers. This is to facilitate the work of fellow AI experts, in hope that they will develop new solutions or techniques to battle against the ongoing pandemic or other related diseases.
Why weren’t we doing this before?
Some large healthcare institutions like Tampa General Hospital in Florida is now relying on a new AI system to distinguish visitors who may be having a fever via a simple face scan. Created by Care.ai, the software has successfully cut down the hospital’s foot traffic by 75%. Others such as Sheba Medical Center in Ramat Gan, Israel has set up two remote hospital units to especially treat covid-19 patients. The units come with AI-powered monitoring set up to predict whether the patients will be at risk of developing sepsis or other respiratory complications. They have also employed virtual clinical assistants to screen patients before they enter the hospital premise.
In the UK, healthcare practitioners had replaced face-to-face visits at the clinics with either phone or online consultations as the government imposed movement restrictions. Some patients questioned “why weren’t we doing this before” after they experienced the convenience of exchanging conversations with their physicians via social media. While experts see this as a “kick forward” to ensure staff are no longer “technophobes” but readily prepared for the digital age.
As AIMed reported earlier, although AI shows the potential to stop an epidemic, it remains unknown whether it can truly stop an epidemic, let alone the ongoing pandemic. What is known to us is, it’s making people accept and adopt technology that’s already available which is more crucial than having stunning AI-driven tools in place.