Preventive Action

Preventive Action

Many get confused with corrective and preventive actions.  While corrective action deals with existing nonconformances (defects), preventive actions deal with nonconformities that haven’t occurred, but might occur in the future. 

In section 8.5.3 Preventive Action, ISO 9001 states “The organization shall determine action to eliminate the causes of potential nonconformities in order to prevent their occurrence.”  Potential problems are usually identified from the analysis of metrics, or from audits.

For example, an organization has a metric for on time delivery, and will take corrective action when delivery falls below 95% on time.  The metric is above the limit, but examination of the metric chart shows that on time delivery has decreased each month for the past 5 months.  A corrective action is inappropriate because the metric hasn’t fallen below the requirement (95%).  In this case, preventive action is more appropriate.

Whereas corrective action requires that action be taken, preventive action does not.  Consider a control chart for a machine tool.  The chart indicates that the tool is wearing and may need to be replaced soon.  Do we change the tool now or wait until the control limit is exceeded?  An economic analysis might suggest that we wait for the control limit to be exceeded because the cost of the tool is high or the cost of lost production is high.   The customer may be in critical need of the product and we can’t afford to shut down the machine now.  A decision to continue production would close the preventive action.

 A decision to take action would allow the machine tool to continue operation while a plan is developed for implementing a replacement.  In contrast, if this were a corrective action, the tool would be shut down until a replacement is available.

Once the decision to take preventive action has been made, the process becomes very similar to corrective action.  An action is selected, usually by a team.  A plan is developed to implement the preventive action selected, and the implemented action is monitored for effectiveness.

Some organizations use one form for both preventive and corrective actions.  In the case of a preventive action, there is no evidence of nonconformance, and there is no containment plan, since a nonconformance has not occurred.  Root cause analysis still occurs because there may be many choices of actions to take.

When the action(s) have been identified, a plan for implementation is developed.

My steps of preventive action look like this:

1.       Problem statement

a.       State the requirement

b.      State the evidence of potential nonconformance

2.       Root cause analysis

3.       Preventive action plan

a.       What action(s) will be taken?  NOTE: The plan might be to do nothing.

b.      When will they be implemented?

4.       Follow Up

a.       Were the actions taken effective?

In contrast, my 5 steps of corrective (See my August 2012 blog) are

1.       Problem Statement

a.       State the requirement

b.      State the nonconformance

c.       State the objective evidence of the nonconformance

2.       Containment Plan

a.       When did the problem start?

b.      What are we doing to control the bleeding while we are looking for a solution ( short term fix)

3.       Root Cause Analysis

4.       Corrective Action Plan

a.       What will be done to eliminate the root cause(s)?

b.      When will they be implemented?

5.       Follow Up

a.       Was the implemented corrective action effective?

b.      Was the short term fix removed?

 

For more information visit www.rosehillsystems.com

Corrective Action

Corrective action is a critical component of ISO 9000.  The standard specifically requires a documented procedure that describes the corrective action process.

Quality can be free, but it can also be very expensive.  Consider the following example.

A company I was working with was having a problem with a supplier part.  A pneumatic valve appeared to fail at the customer’s assembly plant.  We couldn’t tell if the product was defective from the supplier, we were making it defective, or if the customer was handling the product improperly.  But it was clear that the product coming back from the end customer was defective.  The valve just didn’t work.

Not knowing the source of the problem, we began adding fixes:

•             100% inspection of the incoming valves
•             Added a test fixture to test the valves in a final assembly
•             Reduced the voltage to the valve to weed out any borderline valves
•             A neophyte decided that more is better so the valves were tested 10 times instead of once (as if you can inspect quality into the product)

Still the customer reported defects.  We proposed that the customer test in a piece of test equipment like ours so we could determine if the product was acceptable on arrival, and before it was installed into the customer’s product.  The customer balked.
Finally, a year after the problem initially arose, the customer realized that the wattage of the solenoid should be a quarter watt, but they had specified a 1 watt coil.  Low and behold – problem solved.

If you were to watch the manufacture of the subassembly today you will see an operator testing the valve 10 times.

Depending on who you talk to, there are 4 or 5 or 8 steps to corrective.  Here are my five:

1. Problem Statement
1. State the requirement
2. State the non-conformance
3. State the objective evidence of the non-conformance
2. Containment Plan
1. When did the problem start
2. What are we doing to control the bleeding while we are looking for a solution (short term fix)
3. Root cause analysis
1. There are a myriad of techniques, such as fishbone analysis, control system model, etc.
2. Expect there to be more than one root cause.
4. Corrective Action Plan
1. What will be done to eliminate the root causes?
2. When will it be implemented?
5. Follow Up
1. Was the implemented corrective action effective (did it eliminate the root causes)?
2. Was the short term fix removed?

Corrective action is not rocket science, but root cause analysis can be difficult.  In many cases we see that once the fix has eliminated the pain, it tends to stay in place, and no root cause analysis is performed, or if it is performed, the short term fix is not removed.  Adding inspection and testing which would otherwise not be necessary increases quality cost (and product cost) while not eliminating the source(s) of the problem.

For more information go to www.rosehillsystems.com