Regulatory Open Forum

This message was posted by a user wishing to remain anonymous

I work for a tiny start up and am implementing a QMS that we hope to get certified in the future. One area I am not sure about are the procedural requirements for trending quality metrics/data. Does anyone have any guidance on how to do this for a company that has no products on the market yet but yet needs to have a process in place? If anyone has a template for an SOP that they are willing to share, I would be very grateful.

You didn't say, so I infer you are asking about 820.100(a)(1).

First, trending is not a requirement. There is a good discussion of this in the QSR preamble which resulted in using the phrase "appropriate statistical techniques".

There is a difference between having the method in place and having data to analyze. For example, you may have data on audit nonconformances, but no data on complaints.

At an early stage, make your approach very simple. You should plan to collect data in each of the areas listed in 820.100(a)(1). I recommend using two methods at the beginning.

For each area plot a time series, i.e., the number of non-conformances over time. A common approach uses calendar months for the time. Plot the data as a line graph using Excel. On the line graph, ask Excel to include the linear regression line with its equation. Because you are plotting nonconformances the improvement direction should be down. Also plot a target line on the same graph. If an actual data point goes above the target, this is a signal. Investigate and take action.

For each area, develop a set of classifications of the nonconformances. Code them in some simple way. I recommend avoiding codes with "intelligence" since, over time, the system will not work. For audit nonconformances, for example, code them as A01, A02, etc. Create a Pareto chart of the codes. This will give you a rank order for the nonconformance frequency. If a code moves up two or more places in the rank order, this is a signal. Investigate and take action.

These are statistical techniques, so don't forget to include them in your 820.250 documentation.

There is a good GHTF guidance document in this area. GHTF/SG3/N18:2010 Quality management system – Medical Devices – Guidance on corrective action and preventive action and related QMS processes. It has more detail than you need, but has a nice framework for a system.

------------------------------
Dan O'Leary CQA, CQE
Swanzey NH
United States
------------------------------

Original Message:
Sent: 21-Aug-2019 11:26
From: Anonymous Member
Subject: Tending/Data Analysis SOP

This message was posted by a user wishing to remain anonymous

I work for a tiny start up and am implementing a QMS that we hope to get certified in the future. One area I am not sure about are the procedural requirements for trending quality metrics/data. Does anyone have any guidance on how to do this for a company that has no products on the market yet but yet needs to have a process in place? If anyone has a template for an SOP that they are willing to share, I would be very grateful.

This message was posted by a user wishing to remain anonymous

Sorry for the delay in acknowledging your reply. I was out of the office. Thanks so much for this advice, I will follow up with the GHTF guidance documents.
Original Message:
Sent: 22-Aug-2019 10:08
From: Dan O'Leary
Subject: Tending/Data Analysis SOP

You didn't say, so I infer you are asking about 820.100(a)(1).

First, trending is not a requirement. There is a good discussion of this in the QSR preamble which resulted in using the phrase "appropriate statistical techniques".

There is a difference between having the method in place and having data to analyze. For example, you may have data on audit nonconformances, but no data on complaints.

At an early stage, make your approach very simple. You should plan to collect data in each of the areas listed in 820.100(a)(1). I recommend using two methods at the beginning.

For each area plot a time series, i.e., the number of non-conformances over time. A common approach uses calendar months for the time. Plot the data as a line graph using Excel. On the line graph, ask Excel to include the linear regression line with its equation. Because you are plotting nonconformances the improvement direction should be down. Also plot a target line on the same graph. If an actual data point goes above the target, this is a signal. Investigate and take action.

For each area, develop a set of classifications of the nonconformances. Code them in some simple way. I recommend avoiding codes with "intelligence" since, over time, the system will not work. For audit nonconformances, for example, code them as A01, A02, etc. Create a Pareto chart of the codes. This will give you a rank order for the nonconformance frequency. If a code moves up two or more places in the rank order, this is a signal. Investigate and take action.

These are statistical techniques, so don't forget to include them in your 820.250 documentation.

There is a good GHTF guidance document in this area. GHTF/SG3/N18:2010 Quality management system – Medical Devices – Guidance on corrective action and preventive action and related QMS processes. It has more detail than you need, but has a nice framework for a system.

------------------------------
Dan O'Leary CQA, CQE
Swanzey NH
United States

Original Message:
Sent: 21-Aug-2019 11:26
From: Anonymous Member
Subject: Tending/Data Analysis SOP

This message was posted by a user wishing to remain anonymous

I work for a tiny start up and am implementing a QMS that we hope to get certified in the future. One area I am not sure about are the procedural requirements for trending quality metrics/data. Does anyone have any guidance on how to do this for a company that has no products on the market yet but yet needs to have a process in place? If anyone has a template for an SOP that they are willing to share, I would be very grateful.