Medical robots – balancing risks and benefits

“We are going to be more and more dependent on technology for medical use, so safety demands must be high,” says Gurvinder Virk, Professor of Robotics at the University of Gävle.

World-renowned experts

About thirty world-renowned experts in robot safety are meeting to discuss safety standards regarding medical robots and other AI medical equipment.

Work on the safety standardisation of medical robots is a fairly new discipline that started only ten years ago. Previously before medical robots existed, there was what was called medical electrical equipment.

In a hospital today there is so much medical equipment and more than a hundred different standardisations for the different types of such equipment.

An intensive job to create common safety standards within this field is being lead by three groups of experts and Professor Gurvinder Virk is the chairman of one of these groups.

What is a medical robot?

A robot is defined by an ISO standard as an electro-mechanical system that moves and has a high degree of independence, where independence is the keyword.

Until now robots have just been a machine and the regulations for these are one thing.

“Here we are talking about a medical robot and it won’t be just a machine, but a medical device,” says Gurvinder Virk.

If a robot does any of the three things below then it is a medical robot:

• It makes a diagnosis
• It gives treatment
• It rehabilitates

The key is independence

“A medical robot is a robot that is a medical device, that is the definition of a medical robot just now.

The insulin pump that gives the treatment and distributes the insulin could be a medical device. Then there are machines that take blood and are responsible for making a diagnosis.

But all medical devices are not robots, for us a medical robot must have a degree of independence that is the key. If an insulin pump makes its own decisions then it is a robot.”

They must not be able to learn

“Independence can be a good thing but even a risk. You must ask yourself, why am I giving this equipment the power to decide for itself?

The conclusion that the experts have come to is that where there is a big risk, then there should be no independence and when the risk increases by giving the equipment self-determination then it should not be allowed to exist. But when self-determination improves the situation it should be allowed.

If one has a situation where there is medical radiation of cancer where there is a tumour that moves when the patient breathes then one needs an automatic gamma knife, but if the patient should stop breathing then the knife must stop of course.

We look at every possible situation where one wants some degree of independence and determine if it is a good or bad situation and then we advise the manufacturer who wants to develop more effective machines and equipment.

The safety means even that the certificate only applies to robots that do not have the capacity to learn. Medical equipment must not have the ability to change itself because then the equipment loses its certificate.

Essential performance

Another issue concerns medical robots used for surgery, for example insertion and cutting and if a robot is to do this then naturally risk assessment is very important.

Here there are two things, basic safety requirements and something called essential performance.

Basic safety requirements mean that the machine is not able to cause any injury.

“One looks at all the possibilities for faulty decisions and then the consequences of those decisions are assessed and one makes sure that they will not happen. On top of this one defines what the essential performance of the medical apparatus involves. An insulin pump for example must not be able to cause an injury and must under all circumstances deliver the correct amount of insulin.

A manufacturer of medical equipment must be able to state what the essential performance of the equipment is and it must function under all circumstances.”

We have to do it correctly

Medical robots will become necessary due to the lack of qualified staff and at the same time an increase in the population of elderly and sick people. This is already happening, queues for operations must be shortened, and if you are eighty years old you could die before your turn for an operation is reached. The technology is the only way to solve this and medical equipment is now getting more independently intelligent.

“A human being can say OK to a hundred diagnosing robots. This is the way it will work. We must do this and we must do it correctly and we decide how it should be done correctly and this is the purpose of the conference.”

A human being must give their approval

How far away is the hospital where one does not meet another human being?

“We already have a scenario where in certain situations you can get a prescription, but we don’t yet trust machines regarding making diagnoses. It is always a person that makes the final decision about whether the robot has made the right decision. It will be like this for a long time, that the human being is the final judge. We are far from ready to trust machines entirely. I cannot yet see that a person will be removed from that final decision making role. A human being must give their approval….”

For further information, please contact:
Gurvinder Virk Singh, Professor of Robotics at the University of Gävle
Tel: 026-64 87 04
Email: gurvinder.virk@hig.se

Text: Douglas Öhrbom