FMEA scoring - Reliance on customer judgement and follow through

yessir

Registered
Hi All - we are developing a new product. In our fmea there are a couple of control measures that rely on the customer reading and adhering to a warning and instructions. How should we think about the probability of occurrence, and efficacy of detection when we are relying on a human?

One hypothetical example to illustrate my question: Imagine we sell a heavy light fixture. One failure mode is that the fixture could fall after installation. A possible cause is if the mounting surface is not strong enough, the fixture could fall and hurt a building occupant. The envisioned process control is to explain the necessary strength of the mounting surface along with a warning in the instructions.

Severity would be a 9 or 10

How do we think about scoring the Occurrence and Detectability? Are their semi standard rule of thumb when humans are involved?

Occurrence? 1-10 - What is the possibility that a customer would choose a weak mounting surface?

Detectability? 1-10 - Will the customer read the instructions and follow them?

Thank you for your thoughts!!
 

Ed Panek

QA RA Small Med Dev Company
Leader
Super Moderator
If you suspect the risks are significant and the need for human reading and understanding is important I suggest a human factors study. We have done this internally in the past. Determine the types of users (Layperson, clinical, expert, technician, etc), create the label, and test how humans interpret the label during real or simulated use.

This reduces the "Guessing about human behavior factor" somewhat.
 

Tidge

Trusted Information Resource
Can you tell us more about the specific industry? There are some industries that have either/both specific regulatory stakeholders and consensus standards that would have to be considered. For example: In the Medical Device industry direct use of IFU as a risk control to reduce risk isn't generally allowed... although well-understood and limited User Classes can offer some indirect leverage. As @Ed Panek mentions: the human factors study is generating Objective Evidence of some sort... but it would have to be a study of the IFU.

I think I can see what the example (commercially available light fixture) is trying to do... but I can't imagine light fixture manufacturers actually constructing a (Design) Failure Modes Effects Analysis without some external mandate... for example: home use electrical fixtures do have meet national/state/local consensus standards for the wiring. I'm aware that for commercially available products many manufacturers make an effort to make products dummy-tolerant and try to make instructions clear (with appropriate warnings) but I don't think these will inoculate a manufacturer against responsibility for injuries or damage.... especially if the device is available on the market for anyone to purchase.
 

yessir

Registered
If you suspect the risks are significant and the need for human reading and understanding is important I suggest a human factors study. We have done this internally in the past. Determine the types of users (Layperson, clinical, expert, technician, etc), create the label, and test how humans interpret the label during real or simulated use.

This reduces the "Guessing about human behavior factor" somewhat.
Hi Ed - good idea and doable over time for sure. Any intuition for rules of thumb from what you've leaned as a starting point for us?
 

yessir

Registered
Can you tell us more about the specific industry? There are some industries that have either/both specific regulatory stakeholders and consensus standards that would have to be considered. For example: In the Medical Device industry direct use of IFU as a risk control to reduce risk isn't generally allowed... although well-understood and limited User Classes can offer some indirect leverage. As @Ed Panek mentions: the human factors study is generating Objective Evidence of some sort... but it would have to be a study of the IFU.

I think I can see what the example (commercially available light fixture) is trying to do... but I can't imagine light fixture manufacturers actually constructing a (Design) Failure Modes Effects Analysis without some external mandate... for example: home use electrical fixtures do have meet national/state/local consensus standards for the wiring. I'm aware that for commercially available products many manufacturers make an effort to make products dummy-tolerant and try to make instructions clear (with appropriate warnings) but I don't think these will inoculate a manufacturer against responsibility for injuries or damage.... especially if the device is available on the market for anyone to purchase.
Hi Tidge - thanks for the help. The details of the product itself is still under wraps. We can say that It is purely mechanical and does not have any moving parts or electricity or other sources of potential / kinetic energy besides movement from the item it is mounted to. It could be used across a few different industries (boat, aftermarket auto, camping, home) so we've looked into a few different bodies (e.g. cpsc, nhtsa) and have been advised by the bodies that we should look into self certification like ANSI Z34.2.
 

Ed Panek

QA RA Small Med Dev Company
Leader
Super Moderator
MS Word has the ability to determine the education reading level of your label. 8th grade is what we use for all nonclinical communication. Experts are generally more educated and require less comms. For example, the instructions for a patient vs a doctor on a device might be different due to experience and education differences.

Alternatively, if a public database of complaints/injuries etc exists for your market you can use this data analysis to create an informed plan of action.

For your questions of "Will the customer read the instructions and follow them?" it depends on the experience of the user. My car has a 200-page owner manual but I may have read 5 pages out of it ever. I know I need to adjust my mirrors for safety without being told. Does this device you are creating have a similar experience for users? Is expertise required to use this or for laypeople?

You can also be creative. Maybe to access the mounting bracket holes a piece of bright red or yellow tape with your one-sentence warning has to be removed first. They have to be intentionally not seeing it to miss it.
 

AndrewK

Involved In Discussions
Привет всем - мы разрабатываем новый продукт. В нашем FMEA есть несколько мер контроля, которые основаны на чтении и соблюдении клиентом предупреждений и инструкций. Как нам следует думать о вероятности возникновения и эффективности обнаружения, когда мы полагаемся на человека?

Один гипотетический пример, иллюстрирующий мой вопрос: представьте, что мы продаем тяжелый светильник. Одним из видов неисправности является то, что приспособление может упасть после установки. Возможная причина: если монтажная поверхность недостаточно прочная, приспособление может упасть и причинить вред жильцу здания. Предполагаемый контроль процесса заключается в разъяснении необходимой прочности монтажной поверхности с предупреждением в инструкции.

Серьезность будет 9 или 10.

Как мы думаем об оценке возникновения и обнаруживаемости? Являются ли их полустандартные практические правила, когда дело касается людей?

Вхождение? 1-10 - Какова вероятность того, что клиент выберет ненадежную монтажную поверхность?

Обнаруживаемость? 1-10 – Будет ли клиент читать инструкции и следовать им?

Спасибо за ваши мысли!!
тake a look on new handbook VDA-AIAG- there is severity and several variants occurence and detection tables -base on them you might estimate potential score.
 
Top Bottom