Assessing risk where harm is indirect - Generic devices / accessories / intermediates

ThatSinc

Quite Involved in Discussions
Hi All,

Looking for some thoughts on assessing risks where harms are indirect, or a fault with the device in question causes another device to cause harm, with the specific example being low-pressure hose for medical gases.

There are numerous risks where the harm is direct with these types of devices - if there are contaminants in the hose it can cause harm, if the hose bursts it can cause harm, it can be a trip hazard if too long and cause harm, etc. etc. these are relatively straight forward.

However, the hoses are generic in that they connect to a terminal unit or regulated outlet and then the other end connects to a device that performs a medical function. There are no limits on what devices can be connected - if the device requiring gas is designed to take air from a terminal outlet (in accordance ISO 9170-1), then the hose can do it, based on its compliance with ISO 5359 which specifies performance requirements for low-pressure hoses.

One of the risks is that if the hose is too long the pressure drop will be too much and the device will not function.
As a control there are limits on the length of hose that the customer can order as a part of the quality system documentation for the hoses, verified through evidence that hoses at the max length meet the pressure drop requirements of ISO 5359.

Where the connected device could be an oxygen therapy system, an anaesthesia machine, a bone saw, or any other number of non-specified devices - how would you document a harm for this within the risk file?

Cheers,

TS.
 

ThatSinc

Quite Involved in Discussions
Is this hose even a medical device?

By virtue of being an accessory, yes.

See excerpt from the standard - also suggesting these hoses to undergo clinical investigation...
And the risk management section explicitly references 14971.

1645652390469.png
 
Last edited:
Oh okay. This is a tough one. I'm interested to hear other's thoughts as well. Seems to me that you will need to know all of the devices and gasses that may be used with the accessory. You can't really know exactly how the attached medical devices function, so I would try to keep the harms as simple as possible.
 

ThatSinc

Quite Involved in Discussions
Oh okay. This is a tough one.

You wait until you try writing a clinical evaluation for them! That's something I'm really looking forward to negotiating with a TF auditor!

So all of the gases that may be used is fine - as all connections are gas specific and colour coded, we have a range of gases and specific hoses for them (it's wonderful when risk control measures are given for you in a standard)
Knowing all of the devices is broadly possible, in terms of types of devices - but at the end of the day, you've got a lack of delivery to the device, that is going to provide it to the patient (typically).
So for Oxygen you'd be looking at Hypoxia, and for Nitrous Oxide you'd be looking at just "pain" as it's typically used as an analgesic, but it's also used in anaesthesia systems where a lack of it would cause more issues with how the device functions.
For Medical Air for surgical tools - that's where it gets tricky as it could literally be any tool that is driven by air that wouldn't function.
 

Tidge

Trusted Information Resource
A couple of quick thoughts. I may circle back with more specific talk.

The harms exist independent of any combination of devices. You probably want to focus on different hazardous situations.

P2 (probability of hazardous situation leading to a harm) is something the mechanical system relying on supplied gas can affect by design choices, like:
  • filters for particulate
  • feedback mechanism to check for input pressure
P1 (probability of being exposed to a hazardous situation) is something I'm not inclined (at immediate blush) to think about trying to reduce through design choices... but I may not be thinking entirely clearly or with consistency on this point.

I want to spend a little bit of time thinking how the circumstance described in the OP could be analogous to electrical safety. In electrical safety we design for safety in the "single fault" condition, and don't particularly worry about what gave rise to that circumstance ("NRTLs will NRTL"). For non-electrical safety issues, there MAY be some guidance in 60601-2-xx particular standards... if not, my gut instinct is to consider the ancillary issues with a feeder device (that is, the hose) as part of the use cases for the hazard analysis.
 

Hi_Its_Matt

Involved In Discussions
This may be one of the situations where you consider the often-neglected second half of the definition of harm..."injury or damage to the health of people, or damage to property or the environment."

Instead of thinking about the health-related harms to patients/users, what about considering the operational-related harms to the connected instruments.

Essentially, a failure in your hose could result in a direct harm to a connected device, in that the connected device could have reduced or completely impaired performance. Going beyond that would require you to have knowledge of the risks of the connected device, which (1) isn't your area of expertise, and (2) as you said, could be boundless given the variety of devices.

Philosophically speaking, if a completely new air-driven device were released tomorrow, should you be obliged to update your risk management file to reflect the new indirect harms your hose could contribute to? Personally I think not.
 

Tidge

Trusted Information Resource
This may be one of the situations where you consider the often-neglected second half of the definition of harm..."injury or damage to the health of people, or damage to property or the environment."

Instead of thinking about the health-related harms to patients/users, what about considering the operational-related harms to the connected instruments.

I wouldn't go too far in this way of thinking. Classically (in the 14971 sense) the environmental concerns are restricted to considering things like:
  • emissions (including fumes, heat, electromagnetic, fluids) from the device
  • "leftovers" (disposal of environmentally dangerous materials used in construction, e.g. batteries, plastics)
  • recently (per 24971, annex F) security concerns can also factor into the question of environmental risks for interconnected information systems
The physical reliance on other elements should be (IMO, based on 24971 and my own experience) be factored into the use cases/circumstances of use. If the reliance is subtle (e.g. the diesel-powered anesthetic-delivery system requires sufficient air-flow to maintain combustion and vent exhaust fumes) that sort of stuff ought to be explicitly considered in the circumstances of use (could be a use specification, could be in the use cases) and ought to inherently be part of both the risk management and validation testing.
 

Hi_Its_Matt

Involved In Discussions
I wouldn't go too far in this way of thinking. Classically (in the 14971 sense) the environmental concerns are restricted to considering things like:
  • emissions (including fumes, heat, electromagnetic, fluids) from the device
  • "leftovers" (disposal of environmentally dangerous materials used in construction, e.g. batteries, plastics)
  • recently (per 24971, annex F) security concerns can also factor into the question of environmental risks for interconnected information systems
You're right, I probably should have stopped my underline at property. Your bullets do align with my experience for environment.

It is admittedly a tricky situation. My experience is more with finished devices, not so much generic accessories. A very similar question was raised here [Risk Analysis of a Medical Device Accessory], but it didn't get much traction or definitive resolution.

The one thing to keep in mind is the overall purpose of risk management (which I'm admittedly simplifying): Identify ways in which your device can create/contribute to harm, and create controls to minimize those risks. If, in order to do this, you require a detailed analysis of all the different individual devices, or types of devices the hoses may support, then that's what is required. But if you can accomplish the same goal in another more efficient manner, then do it.
 
Top Bottom