Small-scale Medical Devices Use IoT to Improve Accuracy

Hospitals are relying on IoT technologies to improve patient care and to help medical staff find the sweet spot for injections. Imaging and artificial intelligence open up deep tissues for more accurate needle placement, resulting in fewer sticks.

Image credit: Unnati Chauhan (Rutgers device)

Hospitals are beginning to introduce more IoT technologies and robotic devices into their daily operations. Robots are used for cleaning and sanitizing instruments and patient rooms, administering medication, and remote patient monitoring. More recently, IoT-based medical devices are helping healthcare workers with injections.

One of the most common procedures, getting blood drawn, can be distressing to patients and practitioners alike. Medical guidance indicates that practitioners make only two attempts before bringing in a more experienced staffer, but as many as a quarter of all attempts fail. That’s a lot of unnecessary pokes.

A Direct Line to a Vein

To alleviate the guesswork, some hospitals are using specialized devices to help nurses and phlebotomists be successful on the first stick. Veebot Systems has developed a robotic device that can locate a vein and draw blood in about a minute. With 83 percent accuracy, the Veebot success rate is better than a human.

The Veebot uses medical imaging, machine learning, and computer vision to identify and track the optimal needle insertion site. Patients place their arm on a narrow platform with a hard arch equipped with an inflatable cuff. The cuff expands, restricting the blood flow, and an infrared light directs the camera. The image is matched against software models of vein anatomy to select the best vein. The system then uses medical imaging to determine the exact depth below the skin, and then aligns the needle for automatic insertion. The Veebot can be used for drawing blood and for inserting IVs.

Researchers at Rutgers University are developing a tabletop device that uses similar technology. Using artificial intelligence, and near-infrared and ultrasound imaging, the device can locate and assess blood vessels for drawing blood or inserting IVs or catheters. The Rutgers device has been successful in trials with humans and with animals, which is key to drug research. The robotic device could support an integrated module to perform blood analysis, which would allow it to be used in emergency rooms, ambulances, and other medical facilities.

White medical machine for imaging

Image credit: Veebot Systems

Injection Success

Artificial intelligence is a key technology in other medical devices that improve patient care as well. Houston-based IntuiTap Medical is testing VerTouch, a handheld device that physicians can use to perform spinal taps and epidurals successfully on the first puncture. Spinal punctures typically are difficult. Doctors must feel the patient’s vertebrae and guess where to insert the needle, resulting in a first-attempt failure rate of as high as 60 percent.

VerTouch improves accuracy by combining 2D imaging with analytics and a needle guide. It provides real-time mapping of the spine, so doctors can see the vertebrae and then it presents the best point of entry for the puncture. IntuiTap’s tests show that the VerTouch increases the first-attempt success rate of spinal punctures to more than 90 percent.

Samsung Medison is enhancing its ultrasound products with NerveTrack, a new technology that improves the delivery of anesthesia. Even using ultrasound-guided regional anesthesia (UGRA), anesthesiologists can find it difficult to pinpoint the surrounding nerves and to see the needle tip when injecting a patient. That poses a risk of nerve damage from improperly placed injections.

To prevent that, Samsung created NerveTrack, an automatic nerve-tracking technology developed using the Intel® Distribution of OpenVINO™ toolkit. NerveTrack is used during the ultrasound scan to identify nerves in real-time using inference modeling technology. It can find nerves and related anatomical structures using proprietary algorithms, even when the nerve isn’t in the ultrasound image. That gives anesthesiologists a better sense of where to safely place the needle.