Human error corrective action

Biju29j

Registered
Hi
During our periodic revalidation process, one of the dimensional measure using manual drop indicator found oos for CPK. The Cpk for that particular dimension check is only part of validation, since the value is within specification requirement the technician who performed the test didn't pay attention on his test accuracy. The same tested samples were retested by a second technician and cpk found conforming.
The root cause identified as "Human Error" but my validation department is not in agreement with this comment on our deviation report, Since the device is very straight forward and simple method no additional modification can be made to improve to reduce the probability of error.

looking for your expert comments /opinion , thank you in advance.
 

Jen Kirley

Quality and Auditing Expert
Leader
Admin
Welcome Biju29j!

It seems to easy to say "Human error!" when this type of issue occurs. We are instead expected to ask:
  • Why did the human perform the error? (escape)
  • Where else might such an error occur? (containment)
  • Why was the error not noticed? (detection)
  • How can we ensure the human does not perform a similar error in future? (process)
There are, in fact cases of simple human error, but we do need to investigate enough to feel sure this was, in fact an isolated case of simple human error before we call it such.

I hope this helps.
 

Steve Prevette

Deming Disciple
Leader
Super Moderator
"the technician who performed the test didn't pay attention on his test accuracy."

This could mean several things. Did we have roundoff error? Did they misread the measurement device? How were they trained? How were they qualified to do the task? Did the person actually say "I didn't pay attention" - or if not how was it determined they were not paying attention? Were there distractions? Has the person involved provide any statement or explanation? What has happened to the technician involved?
 

Bev D

Heretical Statistician
Leader
Super Moderator
Could you describe the failure better?
As I interpret your post an inspector measured a sample of parts and all of hte parts were within specification. However, either they failed to calculate the Cpk value or they did calculate it and it was smaller than required and they didn’t fail the lot. But when the second inspector measured the parts the Cpk value was larger than required. How close were the two Cpk values?
 

Biju29j

Registered
"the technician who performed the test didn't pay attention on his test accuracy."

This could mean several things. Did we have roundoff error? Did they misread the measurement device? How were they trained? How were they qualified to do the task? Did the person actually say "I didn't pay attention" - or if not how was it determined they were not paying attention? Were there distractions? Has the person involved provide any statement or explanation? What has happened to the technician involved?


Thank you Steve,
Did we have roundoff error : No
Did they misread the measurement device : No
How were they trained : Trained with WI , according to procedure and hands on training on actual measurement with sample part and verified by supervisor , documented the training and qualified for testing.
Did the person say " I didn't Pay attention" : Since the tested value is within the release spec limit the technician just completed the test on all sample parts without getting the exact value.
Has the person involved provide any statement: Yes, small difference in value on device preset is noticed , for this drop indicator which measure the height of the part the preset value is set to "Zero" from the base of the fixture. If a difference is noticed in preset value the tech has to preset the device to "zero" and do a verification check using standard. Second point is the centering of part to the tip of the drop indicator. if the part is not placed centered a small difference in value is reported as the reference point of measurement is different. This small difference usually get ignored by techs since the value still falls within spec limit.
The result for this particular dimension is not verified for CPK on regular production batches but is a requirement for Validation.
What happened to the technician involved : Got retrained. The same parts were re-inspected by a second tech and values found centered and conforming for cpk
 

Steve Prevette

Deming Disciple
Leader
Super Moderator
"Since the tested value is within the release spec limit the technician just completed the test on all sample parts without getting the exact value. "

Gee, that same thing happened to me at a facility I supported. I was informally reviewing the calibration notebooks (passing the time waiting for a meeting) and noted every recorded value for any given piece of equipment was exactly the same, down to several decimal places. Whole pages of calibrations filled in with the same number. Come to find out, the tech said the training they received was if the item within limits, record the "expected value". Even though the column was labeled "As read value". What a mess, and they claimed several folks had inspected the calibration records over the years.

Perhaps techs need to be more involved in the trending calculations (Gage R&R or CpK) for the actual as-found values, so they become more aware of why it is not just a "go no-go" test.
 

Biju29j

Registered
Could you describe the failure better?
As I interpret your post an inspector measured a sample of parts and all of hte parts were within specification. However, either they failed to calculate the Cpk value or they did calculate it and it was smaller than required and they didn’t fail the lot. But when the second inspector measured the parts the Cpk value was larger than required. How close were the two Cpk values?
Could you describe the failure better?
As I interpret your post an inspector measured a sample of parts and all of hte parts were within specification. However, either they failed to calculate the Cpk value or they did calculate it and it was smaller than required and they didn’t fail the lot. But when the second inspector measured the parts the Cpk value was larger than required. How close were the two Cpk values?


Thank you Bev
First test Cpk Value is 1.18 and retest value with second tech is 1.77
This measurement using simple drop gauge , to verify the height of a plastic molded component . Particular dimension is part of QC regular release process with Cpk check. Only for validation ( in this case its a periodic revalidation process) we have to calculate Cpk for this dimension.
The tech who performed the test reported value within specification range ( 26.20mm to 27.20mm) but values were identified as not centered, on retest of same parts by a second tech identified that the actual dimn for all parts tested is more centered ( less variability).
Our validation team is not satisfied with the justification of retraining.
 

Steve Prevette

Deming Disciple
Leader
Super Moderator
Our validation team is not satisfied with the justification of retraining.

Some suggestions - Schedule management observations of future such tests. Would help managers understand quality principles. Of course, give them a good guide of what to look for. Schedule reviews / monitor future CpK results and not just for this specific test. For that matter, do you have a management observation schedule. Have the inspector involved have to perform x number of observed tests, and to participate in y number of CpK calculations.

A good question to answer is - how long had this been going on, and what should have caught it earlier.
 

Bev D

Heretical Statistician
Leader
Super Moderator
What sample size is used? Also is Cpk calculated on only that sample size or is it calculated on all measurements since last validation?

aside from why the first operator didn’t reject the sample I question the use of Cpk as a validation criteria if the sample size small and/or a single point in time sample. I also question why the second operator got ‘better’ results than the first operator if the same parts were measured.
 
Top Bottom