Imagine you have two balloons. One is filled with water, and the other with air. They appear the same, but if you push on them, each will feel very different. That’s how organs feel to a doctor. When a patient needs a needle biopsy, or a gall bladder drain, or a cortisone injection to the spine, or a venous catheter, the doctor inserting a needle must be able to feel the build up and release of pressure as that needle pushes into, and eventually punctures each successive tissue.
“The character of the tissue gives you force feedback, and your brain figures that out, and can use that to interpret lots of different things,” says David Han, professor of surgery and radiology at Penn State. “If you’ve touched a lot of livers and you’ve touched a lot of spleens, sometimes with your eyes closed you can tell which is which.”
But it’s really not easy. Research over the last 30 or more years has shown complication rates that range from 5 to 21 percent in central vein catheterization, and the fallout is infection or increased hospital time and cost, or even death. Experienced doctors are much better at it, partly because it takes a lot of practice. (In many cases, ultrasound guidance helps, but even with a visual cue it’s easy to go just a bit too far, and into the wrong tissue.)
How do med students learn this technique? In some cases, a mannequin built to resemble particular tissues provides feedback, but more often the students watch an experienced doctor, and then they try it. “I’m really good at it,” says Han. “So I have somebody standing next to me who wants to learn how to do it, and I sort of lean over their shoulder and say, try this, or that.”
A team of researchers at Penn State University had a different idea. Led by Han, in 2017, they published research describing a robot that would hold the end of a needle and provide mechanical feedback — as the student pushes the needle into a lump of silicon, the robot arm pushes back. Unlike a mannequin, it can be programmed to follow different force curves, made to match the pressure profile of a needle sliding into different tissues, and even representing different body types. “What you want to be able to do is to have people prove their competency in a simulated environment before you hand them the controls,” says Han.
But some of the other researchers Han was working with had a further insight: They could make a tool that would do the same thing, sans robot, for far cheaper. Instead of a robot arm, the force feedback would be provided by a mechanism housed within a simulated syringe. The researchers filed a provisional patent application this year and received a grant from the Penn State College of Engineering to develop the device as a business.
“We could create those forces a bit more simplistically by having this, essentially, material fracturing inside these cartridges create our haptic force,” says Jason Moore, an associate professor of mechanical engineering who led the team. “And then we could still provide the user with a lot of feedback about how they performed the needle insertion.”
Though the provisional patent application describes several means of simulating pressure (including electromagnetic, magnets, friction, hydraulics, and others), the group has chosen to focus on a version actuated by a series of membranes housed within the body of the syringe. Upon pushing against a surface, the needle retracts into the body of the syringe. As it does, it abuts the membranes in sequence. Each one deforms and eventually breaks, just like human tissue. By varying the configuration, thickness and material of the membranes, the device simulates different force profiles without the need for an expensive robot arm.
Han, Moore and Moore’s collaborators, associate professor of engineering design Scarlett Miller and associate professor of anesthesiology Sanjib Adhikary, aren’t the only ones working on devices for training students in ultrasound-guided injections. “Everybody’s trying to come up with different ways and means to make it look better, or make it more user friendly,” says Adhikary. “But no one has got the Holy Grail.”
In 2015, a company called Blue Phantom released a sophisticated training model for knee joint injections, complete with simulated femur, tibia, patella and bursa — but it costs $3,800, and is only useful for practicing injections into the knee. There are even DIY solutions featuring gelatin-filled balloons, with rubber tube vessels. David Gaba, a professor of anesthesiology at Stanford, has been building needle injection simulators for more than 30 years, including plastic trainers for lumbar injections. He even uses pork shoulder tissue as a substitute for human.
“Just because something can be simulated by a computer/hardware combo to portray the haptics doesn’t necessarily mean that it will achieve miracles of learning or skill,” says Gaba. “Unless there is clear-cut evidence that a particular device makes a big difference, ultimately it will be the marketplace that determines whether any particular engineering advance has legs as compared to other approaches.”
There still must be a balance, points out Han. Remove too much of the realism and students won’t properly connect the practice tool to the reality. But any computerized apparatus can provide valuable and quantitative feedback — a report card of sorts — into the performance of the students learning the technique.
As they work toward a marketable device, Moore, Miller and Adhikary are building an accelerometer into the cartridge, which will pair with custom software to give similar feedback on insertion angle and force profile. Their prototype, including sensor and replaceable cartridge, cost them around $100.
“The idea is worth pursuing, especially if it can be sold at $100,” says Paul Bigeleisen, a professor of anesthesiology at the University of Maryland. But injection molding and wide distribution, possibly through schools and training hospitals, could drive the cost per unit even lower.
“If we can make these new medical students or very early future doctors be very good at their hand motions, be very steady, could that have a positive impact on their skill much farther down the road?” says Moore.
That’s the hope, he adds.