Labeling colors

Hi

Our home use electrical products(currently classified as cosmetics) are about to be reclassified as medical devices and are will require conformity with IEC 60601.
One (of many) issues is labeling.
Our rating label on the device is currently B&W laser etching directly on the device.
This may be problematic with color coded symbol like:
Will it be sufficient to have the blue symbol on the product case and user manual while having it in B&W on the device?
Are there reference in the standard itself for this?

Thanks

Not clear ...

The informative annex D states: "The colours of symbols are not specified except for the background of the AP and APG symbols".

There are also special mention of colour requirements for safety symbols (yellow on black).

But otherwise, the statement appears to say any colour is OK for general symbols.

However, this looks like a design creep issue. The intention of the above text may have been to say the symbols presented in black and white can be any colour e.g. black on white, white on black, brown on yellow etc etc. The addition of the blue symbols came later (A1?) and the text might not have been updated to identify this exception.

Also there is a general requirement that symbols have to conform to the ISO/IEC standards. If those standards specify blue, then that would take precedent over Annex D which is informative only.

Upshot, you could try to point out the text in Annex D to convince a test lab it's OK. But it might not work

Another possibility could be to eliminate the symbol altogether. The standard says Symbol 10 may be required if RA and UA indicate it is needed. Otherwise Symbol 11 "may" be used. "May" has a special meaning in the standard, and it is not mandatory. But some test labs are pushing it as mandatory (they say it has to be either 10 or 11).

I agree that a NRTL is likely to insist on the blue color.

My understanding of the "blue man" symbol is that it is NOT supposed to be used as a stand-in for "RTFM", but in circumstances where there is a specific (potentially unacceptable) hazard that can only be explained in the accompanying documents.

Never heard of RTFM before but it was a literal LOL (or at least a big smile) to see my guess was right after looking it up!

The confusing point is that most warnings in a manual are of a kind that a user can reasonably be expected to know without reading the manual. The purpose of these warnings is not for general users, but to cover cases where a irresponsible user ignores well established convention, causes an incident, and then the claims that there was not a warning in the manual.

These are not risk controls, they are liability controls.

To make it clear, manufacturers should not refer to liability type warnings as risk control measures in the risk analysis. That not only makes it clear with respect to the symbols on the label, but also reduces the workload for usability testing for the effectiveness of warnings and cautions.

I am in complete agreement with the above, and so I will quote it!

Peter Selvey said:

The confusing point is that most warnings in a manual are of a kind that a user can reasonably be expected to know without reading the manual. The purpose of these warnings is not for general users, but to cover cases where a irresponsible user ignores well established convention, causes an incident, and then the claims that there was not a warning in the manual.

These are not risk controls, they are liability controls.

To make it clear, manufacturers should not refer to liability type warnings as risk control measures in the risk analysis. That not only makes it clear with respect to the symbols on the label, but also reduces the workload for usability testing for the effectiveness of warnings and cautions.

Click to expand...

I really like the distinction between risk controls and liability controls. No amount of certification to a standard is going to inoculate a manufacturer against a lawsuit.

In the brief but glorious era when it wasn't considered heretical to identify IFU as risk controls that had the potential to reduce risk (for specific line items in a Hazard Analysis), I used to advise that the "blue man reading a book" symbol ought to be used for those very few cases where it was not legitimately possible to reduce the risk of specific hazards in specific circumstances of use, with a few qualifiers:

• The IFU had to have a separate, dedicated section for quick reference to those risks, so that the user could find and understand them.

My reasoning was motivated by usability principles: If there is a serious risk, I don't want users to have to flip through a manual to find the explanation of the risk. While I accept that general use sections of an IFU can contain this sort of information, when that sort of information appears in multiple sections (i.e. because there are multiple risks which prompt the use of the "blue man") I really don't want the user flipping between multiple sections of the IFU to find the specific explanation.

• There shouldn't be too many of these types of risks (that indicate use of the "blue man") for any device.

Too many cases of the "blue man" implies to me that the device doesn't have an otherwise "safe" risk profile. There is also the possibility that the risk-benefit analysis for the device is lop-sided... and my instinct would be that the device is being marketed to the "wrong" user class (i.e. too general, such as offering a clinical device for home use) or that the use cases are wrong (i.e. the device is being promoted for use in a circumstance that isn't medically well-motivated). The "blue man" isn't going to turn an otherwise unqualified septuagenarian into a qualified x-ray technician.

• By limiting the use of the "blue man" to specifically identified risks, it is possible to make an assessment (during user validation) of the effectiveness (of communicating specific risks) for this sort of IFU control.

I'm not claiming that the specific risks are necessarily reduced via IFU, my suggestion is that it is possible to better guarantee that (by providing focused information) that the appropriate user class has the information necessary to defend the manufacturer's assessment of user classes and possibly P1, P2 for those risks.

This is an example of the broader issue that arises when the effort required to justify no action exceeds the effort to take action. What's the harm in adding a warning? It's just a couple of lines. Who cares.

The problem here is cumulative.

Yes, two lines is not significant. But if we apply it to all similar situations, we end up accumulating 100 pages of warnings that no one ever reads, obscures the few warnings that are really important, uses up trees for the paper, extra fossil fuels for shipping and raises the blood pressure of test engineers stuck wading through the manuals trying to decide if they meet applicable standards (which also contain a lot of requests for useless warnings).

In safety we are taught to consider each case on its own merits; the need to be aware or cognisant of the broader, cumulative impact of a particular interpretation is not something we are trained to keep an eye out for. It's not mentioned in ISO 14971.

This is not just warnings and cautions, but applies to production tests, record keeping, supplier control and many other aspects of modern medical device regulations, where many small actions are justified on the basis of "better to be safe than sorry", which then accumulate to become counterproductive and even dangerous by obscuring the cases where actions are really important. Such as a blue man to indicate RTFM.

Peter Selvey said:

In safety we are taught to consider each case on its own merits; the need to be aware or cognisant of the broader, cumulative impact of a particular interpretation is not something we are trained to keep an eye out for. It's not mentioned in ISO 14971.

Click to expand...

In another life, I got significant mileage from the 14971-2007 Informative Annex. This text I include below is (now) included in TR 24971-2020, with some modifications made (the main one being that we now call out benefit-risk analysis, as if there was some sort of "divide-by-zero" problem). All emphasis mine. I have copied the 14971-2007 text.

D.6.3 Criteria for risk/benefit judgments

Those involved in making risk/benefit judgments have a responsibility to understand and take into account the technical, clinical, regulatory, economic, sociological and political context of their risk management decisions. This can involve an interpretation of fundamental requirements set out in applicable regulations or standards, as they apply to the product in question under the anticipated conditions of use. Since this type of analysis is highly product-specific, further guidance of a general nature is not possible. Instead, the safety requirements specified by standards addressing specific products or risks can be presumed to be consistent with an acceptable level of risk, especially where the use of those standards is sanctioned by the prevailing regulatory system. Note that a clinical investigation, in accordance with a legally recognised procedure, might be required to verify that the balance between medical benefit and residual risk is acceptable.

As a practical matter: I found myself often trapped in a low level document (e.g. a Failure Modes Effects Analysis) where the "action" (e.g. a specific design decision) would be chosen and analyzed/tested/evaluated and within that same document I wanted to demonstrate ownership of the responsibilities (see above)... but at such a low level it is hard to demonstrate that the responsible team is (properly) considering the cumulative impact of the individual decisions. I don't think it is impossible, but it definitely takes discipline and determination to do this. I have a deep sympathy for the test engineers who end up having to do miniscule-value added work simply because folks who should know better won't apply some filtering.

This was personally very frustrating for me because the Risk Management system in use required my teams to do risk acceptability analysis for every line of the low level documents (reminder: such docs don't study risks, just failure modes) and every once in a while I'd have a team member (or boss!) that would "stand his ground" and say (with a straight face) that "we can't apply risk-benefit criteria in the low level document because the low level documents are ignorant of the benefits".

Oh the time i've spent on this symbol. I'm not going to spend too much more extremely long on this. Authorities will force you so if you're looking for the certification advice then just do it™, but sanely speaking BW should be fine if the system works; for context, this:
The blue colour with the circular shape are intended to communicate a mandatory nature, versus
the red colour with the circular+diagonal bar shape which indicates a prohibition, versus
the yellow colour with the triangular shape, indicates the need to (be aware in certain conditions) to perform or abstain from an action. It is accented with a black border surrounding it because yellow is able to fade/bleed over into other colours so easily the shape is lost.

The colours emphasize the shapes, but are redundant to the core message of the shapes, which are maintained even when fading or printed in BW.
A symbol that can be contained in any of the above has a contrasting colour so it is highly identifiable and visible.
The colours are specified to a very high degree of accuracy in the colour spaces in those standards, but generally no-one cares as long as humanly seen they don't deviate too much from the primary colours which I've written above, because we're only human after all™.

Why does all this matter, especially with regards to the 60601-1?
The old booklet (Symbol 1641) was used as a general information sign in black and white at the should level, and as the symbol within the blue circle old (Safety sign 01 of 60878). Without the colour these could not necessarily be distinguished well from each other.
The new symbol, man holding booklet (M002) is very different from just the booklet (Symbol 1641). The blue colour thus is no longer essential to communicate the difference in intent. That intent however, is hardly used appropriately.
It mandates you to consult the manual. Risk management wise, it should be used on the location and preferably only during the time it is relevant, and not when it detracts from more important information. However, most people get through certification by sticking it somewhere on the device, once, without reference to the relevant part of the manual you would need at that point (which is logical, because section numbering can change). Long live human flexibility in interpretation.

Source: I've bought all the relevant standards, read the manual on "Writing and Designing Manuals and Warnings, fifth edition", and had the approval to go into lengthy discussion on this with varied people because of similar reasons. Advice: save the money, go the colour route. Some people will use it as a pet peeve find, because it is easy to identify and hard to argue against and ultimately cheaper to give in.
Fun note: practically speaking the presumption of compliance for these symbols is not reflected by reference to study on their effectiveness, because (from OBP) "Test data obtained according to ISO 9186-1:2007 are not available. Consequently, a supplementary text sign shall be used to increase comprehension except when the safety sign is supplemented by manuals, instructions or training.". Or in other words: mass-delusion.

Also @Tidge ++ man. Every time. Also @Peter Selvey ++.

We have a similar issue. We have a small product that is classified as a Class I medical device but the markings are etched (white) on the product (black), so blue color is impractical. Initial test house safety review has pointed out that it needs to be blue which cannot be done by etching. Looks like we will therefore remove the symbol from the device itself and leave the warning in the IFU but what about our initial production run that was made with the samples sent for testing? Are we OK to leave etched white on black or will we need to apply a label over the top ?