Medical device worn on body with USB, isolation, 60601-1

Hi,

We are developing a battery powered medical device, an ECG, worn on the body. We'd thinking about using a USB connector for data transfer to PC after the record is done. In such a way, we'd need a proper 5kV isolation...

My question is, if the USB connector is cannot be exposed to the patient while worn on his/her body and recording (the other part that is atteched to prevents this, as it covers it) does we need isolation in this case too? The end user only could plug in a USB cable, while the device is detached (although its back leads could be touched)

If not, a docking station would solve the problem as it would cover the back leads too?

Thanks for all the replies, as it is sometimes hard for me to understand the 60601-1 in such questions, as it is a general standard and the device specific standard doesn't say any more concrete things about this.

If the instructions for use say not to connect to a PC while connected to the patient, then the issue is only a matter of foreseeable misuse, which is dealt with in under risk management.

If you think the risk is significant, you could take different actions ranging from a warning label; software to prevent operation if the system detects the USB is connected while connected to the patient; or if you are really worried, an isolation barrier.

If the PC is IEC 60950 compliant, a genuine risk assessment should find the issue is not that serious (mild shock is possible, but death or serious injury requires about 40mA which is extremely unlikely). It would appear then that an isolation barrier would be overkill.

Good point, peter.

Since the frequency of foreseeable misuse is less than normal condition, but higher than fault condition, it seems warning label could not be sufficient to reduce the risk.

In IEC 60601-1 things are simplified, you have to fit everything into either normal condition or single fault condition.

For foreseeable misuse we can turn to risk management where a more nuanced assessment can be done, looking at the full sequence events, and a realistic overall probability estimate. Significant events in this case are:

- low impedance, double insulation failure in the PC that does not trip any protection (e.g. primary fuse);
- long term connection to the patient;
- low impedance return path from the patient (earthing of the patient);
- user ignores IFU and warning label;
- shock results in death or serious injury (most shocks don't);

It seems likely the probability of all these events is in the negligible range but of course it is up to the manufacturer to make the estimate. Personally I would take the extra step of having the software detect the condition where the PC is used while still connected to the patient (e.g. using electrode impedance), and then prevent any operation; it's not difficult and makes sure that the connection time is short; also the user would only do this once.

Thank you for your answers, they were really helpful! Have a nice day!