Researchers find ways to automate the insertion of needles
AROUND one in ten Americans is a trypanophobe, stricken with a dread of needles. Some may have inherited a tendency to faint when jabbed; others develop the phobia after a painful injection or a botched blood sample. Sadly, their fears are occasionally well founded.
Central venous catheters, or central lines, are commonly used with critically ill patients to administer drugs, fluids, food or blood products close to the heart. However, placing needles inside veins deep in the body is notoriously difficult. Some 15-30% of attempts suffer complications, mainly punctured arteries that can lead to infection (around 250,000 cases in America annually), but also bleeding, collapsed lungs and even cardiac arrest. Failure rates in children can be higher still. A study in 2013 by Stanford University School of Medicine and Lucile Packard Children’s Hospital, found that over half the attempts to place a central venous catheter in children failed on the first go.
Portable ultrasound can be used to produce an image of the position of blood vessels to help insert needles. This has reduced errors, but the technique still requires a highly trained physician. Now a collaboration between Cincinnati Children’s Hospital and Ben Gurion University of the Negev in Israel hopes to automate the entire process. A team led by Hugo Guterman, a robotics expert, has built a prototype device that uses ultrasound, machine vision and a robotic needle-dispenser to make placing a central venous catheter a push-button affair.
The operator lays the wireless device on a patient’s arm, leg or neck and views an ultrasound image on a nearby computer screen. The system then identifies the centre and edges of each blood vessel, as deep as 15cm inside the body and as narrow as 0.5mm in width, making it particularly useful for treating children. Using a joystick, the operator aligns a target icon over a vein. The system uses a tracking algorithm to keep the blood vessel aligned. When ready, the operator simply presses a button to insert the needle.
The operator does not have to be a doctor. Dr Guterman envisages his device being used by most clinical staff, including paramedics. In a demonstration, your medically untrained correspondent used the prototype to carry out a simulated needle placement in under two minutes—less than a quarter of the time that a doctor typically needs for a manual procedure. The system will not deploy the needle if it cannot sense a suitable vein at the target location, minimising complications.
The prototype is designed to place the special needles required for paediatric and adult central lines, but Dr Guterman expects the technology will also work well for other needle-based procedures. These could include breast, prostate and liver biopsies, some cardiac procedures, and the precise delivery of radiotherapeutics or anti-cancer drugs to tumours.
The system has already undergone successful testing on live, anaesthetised pigs and is now heading for pre-clinical testing and regulatory approval. But America’s Food and Drug Administration is understandably wary of any fully autonomous medical devices. “Because it is an invasive procedure, you have to let a human make the final decision,” says Dr Guterman.
The ultimate barrier to adoption will be persuading doctors that a robot can do a good job. Dr Guterman is not alone in thinking that it can. A number of startups are also developing robots to insert intravenous needles. Some use ultrasound, others rely on infra-red vision. As such robots can fire needles through skin and flesh much faster than a human hand can, it causes less damage to tissues and should mean less pain, too. Trypanophobes aren’t the only ones who would welcome that.