Failure Investigation: Treating the Root Cause, Not the Symptoms

For Medical Device

The formation of a problem is the most essential part of problem solving.

-Albert Einstein

Introduction

Why should medical device manufacturers perform failure investigations? One reason is that it is a regulatory requirement. The Code of Federal Regulations (21 CFR 820) in areas such as Non-Conforming Product, Corrective and Preventive Action and Complaint Files describes the requirements for failure investigations. Performing effective failure investigations is also good business practice which can lead to reduction in scrap and rework, process improvements and the elimination of recurring issues. However, the most significant reason is that customers (i.e, the patients), expect safe, reliable and effective products which will perform as intended.

From a regulatory perspective, are medical device companies adequately performing failure investigations? The United States Food and Drug Administration (FDA) inspectional observations indicate persistent issues with inadequate investigations or lack of investigations, failure to determine root cause, inadequate corrective and preventive action and lack of documentation to support investigations performed. Approximately 70 percent of FDA-issued Warning Letters to device manufacturers in 2007 and 2008 cited Failure Investigation related issues!

On that note, it's important to recognize that inadequate failure investigations can be due to a number of factors, including:

1. An inaccurate problem statement - the most fundamental part of conducting effective failure investigations is defining the actual issue in clear terms. Failure to do so will result in implementing a corrective action which may not address the reason for the failure;
2. The investigation reveals the symptoms, but not the true root cause, i.e., the quick fix or correction;
3. Inadequate corrective actions are implemented - many times an issue has been identified, failure investigation has been conducted, corrective actions have been put in place and the same issue arises again;
4. Incomplete or missing documentation;
5. Inadequate and/or incomplete effectiveness checks.

Failure Investigation Steps

So what steps can medical device manufacturers take to help ensure the adequacy of a failure investigation? The components of any failure investigation should include the following:

• defining the problem statement
• identifying the true root cause
• assessing risk
• evaluating the products, processes and system impact
• identifying and implementing appropriate corrective and preventive actions
• implementation of appropriate effectiveness checks

In defining the problem statement, questions such as: "what do you know," "what are the facts," and "what data do you have" are all critical and will help define the scope and complexity of the investigation. For example, compare the following problem statements:

a) Complaints have been received indicating an issue with Device ABC

b) The complaint data from Jan 2005 to Jan 2006 demonstrates a trend in audio failures related to the main PCB board, PN 1234, of Device ABC.

It is quite evident which problem statement is better! In summary, the problem statement should be clear, concise and measurable leaving no room for ambiguity.

The Quality System Regulation Preamble states, "At times a very in-depth investigation will be necessary, while at other times a simple investigation followed by trend analysis or other appropriate tools will be acceptable," (reference comment 161, 52634). The scope and level of the investigation performed should be commensurate with the complexity of the issue and the degree of risk. The use of failure investigation tools varies, but typically the selection will depend on the type and complexity of the issue being investigated. There are many investigation tools available, such as Failure Mode and Effects Analysis (FMEA), Fault Tree Analysis (FTA), Ishikawa or Fishbone diagram and the 8D approach. Probably the most simple and effective tool is called the 5 Whys. Let's look at the following example:

Problem Statement: You are on your way home from work and your car stops in the middle of the road.

1. Why did your car stop?
   • —Because it ran out of gas.
2. Why did it run out of gas?
   • —Because I didn't buy any gas on my way to work.
3. Why didn't you buy any gas this morning?
   • —Because I didn't have any money.
4. Why didn't you have any money?
   • —Because I lost it all last night in a poker game.
5. Why did you lose your money in last night's poker game?
   • —Because I'm not very good at "bluffing" when I don't have a good hand.

As you can see from the answers to the 5 Whys, the true root cause in this example might be very different from what it may have been if we had stopped earlier in the process. Very often there is a tendency to stop asking critical questions too early and this may have ramifications later as the corrective and preventive actions identified do not address the true root cause.

In establishing true root cause in an investigation, some traps include conclusions such as, "Equipment failure - it just broke" or "Human error - the procedure is inaccurate or the employee made a mistake, the customer did not use the device as intended..." To drill down to the true root cause in cases where user error may be a factor, one has to ask if a comprehensive investigation and analysis has been performed to validate that the cause of the issue is user error. Similarly, in the case that the device might have been incorrectly used, has the design validation study and results been reviewed for robustness and adequacy? Another potential pitfall could be stopping too early in the investigation process, only addressing symptoms and not identifying the true root cause.

Risk assessment is also an integral part of the investigation process. The level of risk to the patient or user needs to be determined and appropriate action taken. It is important to note that containment action is a fluid process as the level of information will increase over time. The initial decision may be to place a product on hold and then as more data is received in the interim period, the product hold may be extended or reduced.

With regard to Corrective and Preventive Action, it is vital to avoid the quick fix (i.e, correcting the immediate issue and not going beyond the specific issue to review the processes and systems) which may be the ultimate cause of failure. Good investigations need to look at the broader picture, answering questions such as: Have there been similar issues reported via complaints, non-conforming product reports, repair reports, trend reports? Has horizontal deployment been considered? Are there similar processes in the facility or in sister plants which could be impacted? Can preventive action be instigated in similar processes and technologies before it becomes an issue there?

One of the most critical steps in the failure investigation process is the implementation of appropriate effectiveness checks. This step can determine if the corrective and preventive actions which have been put in place actually work. Effectiveness checks should be clear, concise and measurable, with clearly defined acceptance criteria.

Failure Investigation Lessons Learned

There are some common themes emanating from poor failure investigation processes, including:

• Stopping too soon - not asking "why" enough times and not determining the true root cause
• Performing an incomplete investigation
  • All factors are not considered
  • All associated lots are not identified and/or evaluated
  • Conclusions are not supported by facts and data
• Jumping to conclusions - "this happened before and we fixed it by..."
• Including unsubstantiated opinions, assumptions or guesses in place of facts and analysis
• Not having the correct individuals involved in the investigation process.

In conclusion, to ensure the robustness and adequacy of failure investigations and to determine true root cause, the problem statement needs to be clearly defined and an investigation conducted. The investigation conducted is commensurate to the issue and needs to be completed by a multi-functional team and well documented, with adequate effectiveness checks implemented to validate the proposed actions.