Medical Device Link.

In an IVD manufacturing environment, the quality department has to evaluate nonconformities identified by various sources and investigate them if necessary. Per FDA regulations, companies are responsible for conducting thorough and timely investigations to determine the causes of nonconformities. Although the causes of some nonconformities are obvious and can be identified immediately, IVD companies must investigate many of them first before they can identify the root causes or the most probable causes.

The investigation system supports and interacts with many quality subsystems (e.g., discrepant materials, customer complaints) that are important to an IVD company’s compliance structure. A particularly important subsystem for the investigation system is its relationship with the corrective and preventive action (CAPA) system. Corrective and preventive actions enable a company to restore controls back to the quality subsystems that prompted the nonconformities. This article discusses some essential elements for a well-established investigation management system.

Nonconformities can emerge in any IVD manufacturing process. The sources of such nonconformities include discrepant materials, manufacturing deviations, out-of-specification results, customer complaints, and returned products. In fact, FDA’s quality system regulation (QSR) requires IVD companies to investigate “the cause of nonconformities relating to product, process, and the quality system.”1 A company therefore must establish, implement, and maintain a system that can manage investigations.

Figure 1. (click to enlarge) The investigation management system.

If IVD companies do not manage their investigations properly, more-serious problems could emerge, especially if they do not identify the root causes of the nonconformities correctly or in a timely manner. Since the investigation system is linked to many quality subsystems, poorly managed and conducted investigations could cause such subsystems to be ineffective or dysfunctional (see Figure 1). Many FDA inspectors and quality auditors prefer to review the investigation system for its adequacy, effectiveness, and relationship with other quality subsystems.

Resources

A key element for any successful investigation process or system is the availability of resources in terms of people, equipment, and facilities. A successful investigation management system requires sufficiently trained personnel, or investigators, to conduct the investigations properly. The number of investigators employed should be based on the average number of investigations performed per year and the average time needed to complete an investigation. IVD companies can conduct a cycle-time study to establish the baseline (i.e., the average number of days) for completing an investigation.

Training and Qualification. In order for investigations to be conducted in a thorough and timely manner, IVD companies must select personnel that possess the experience and skill sets necessary for the investigative tasks. If a company selects individuals with limited or no experience and inappropriate skill sets, the investigations could lead to the root causes not being identified or incorrectly identified, and the implementation of inappropriate or ineffective corrective actions.

Since conducting investigations requires special skill sets, IVD companies should establish a plan to train the investigators in the following areas:

• QSR and other relevant requirements (e.g., customer).

• Manufacturing operations and processes.

• Materials, components, products, and equipment.

• Risk management principles and risk analysis.

• Root cause analysis (RCA).

• Report writing skills (i.e., technical writing).

• Computer skills (e.g., Microsoft Word or the software program the company uses to generate investigation reports).

Infrastructure. The infrastructure (equipment, facilities, etc.) of the investigation system should appropriately support the investigative workload as determined by the quality department. Based on the number of investigations conducted, an IVD company should determine whether to implement an automated system (i.e., a software application capable of recording and processing investigations) to manage all investigations.

An automated system is especially helpful if an IVD company conducts a large number of investigations and/or has multiple divisions and locations. The automated system should be customized to meet the company’s needs and should be able to do the following: log new investigations; track all open investigations; approve investigations by electronic signatures; provide notification when an investigation is overdue; search the databases with different sets of criteria (i.e., queries); and generate metrics and reports from the databases.

Risk Management

All nonconformities are not the same and do not require the same level of investigation. Some nonconformities do not pose any material or product risks, and in-depth investigations are not necessary. However, IVD companies should document such nonconformities on their associated quality records for future reference.

On the other hand, IVD companies should conduct in-depth investigations including using appropriate RCA methods for those nonconformities that pose a quality risk on the products. The depth of such investigations should depend on the severity of the nonconformity, which is based on the following factors: the impact of the nonconformity to the safety, efficacy, and quality of the material or product; and the risk as related to statutory, regulatory, and customer requirements.

IVD companies should use a risk-based approach when evaluating nonconformities and determining the level of investigation required. They should defer to ISO 14971 regarding the application of risk management to IVD devices.

Conducting Investigations

Technically speaking, the investigation process begins when the quality department decides to investigate an observed nonconformity. For most FDA-regulated IVD companies, an investigation should be completed in 30 calendar days, which is especially true when investigations are linked to events that fall under the Medical Device Reporting requirements.2 The following workflow shows that actually conducting the investigation is the major part of the process:

Nonconformity identified ? Risk analysis to determine the level of investigation warranted ? Conducting the investigation (including developing the investigation plan, collecting evidence and information, identifying the root cause and CAPAs, summarizing the results, writing the investigation report, reviewing and approving the report) ? Investigation closed.

The 30-day time frame is a tight window for a lot of activities to occur and requires a strategic approach to ensure that an IVD company can meet the established time requirements. To achieve this goal, the company should develop a written investigation plan. The investigation plan should include the elements (e.g., processes, documents, areas, etc.) that are identified for review by the investigation team and the associated timeline for completing the review. The investigation team should be a cross-functional team with participants from manufacturing, engineering, quality, regulatory, etc. The investigation plan is similar to an audit checklist that could provide structure to the investigation.

Figure 2. (click to enlarge) Investigation template (simplified version).

Investigations should be conducted in a timely and thorough manner, and should be based on factual evidence. However, evidence could get lost and distorted over time. An IVD company should develop a template to assist the investigators in capturing such documentation when nonconformities are first identified. Timely gathering of such information and data can greatly affect the success of the investigation. Figure 2 is a simplified version of an example of an investigation summary template. Each company should develop its own template that best suits its operational needs.

Root Cause Analysis

The main purpose of conducting an investigation is to identify the root causes so that an IVD company can implement CAPAs to eliminate the nonconformities from occurring again. In order to reach this objective, the selected root cause analysis (RCA) method must provide an answer to the question, “Why did the nonconformity occur?” If the RCA method does not answer the question, then the cause of the nonconformity has not been identified. The two RCA methods widely used for achieving this objective are the Five Why’s and the Cause and Effect (Fishbone) methods.3,4

The Five Why’s method requires investigators to ask why multiple times until they identify the underlying cause or the contributing factor of the nonconformity. The method does not insist on asking why five times, and the investigator can ask more than five times or less. One frequently asked question relates to the end of the Five Why’s methodology. The key to the Five Why’s technique is to pare down to such a level that an action can be determined and taken to eliminate the cause. This RCA method is best suited for those nonconformities related to operator errors and equipment malfunctions.

The Cause and Effect (Fishbone) method requires investigators to review, at a minimum, the following items as they relate to the nonconformities:

• Materials, components, and products.

• Methods, documents, and procedures, including measurements and records.

• Machines and equipment that were caused or affected by the nonconformities.

• People, including the work performances and training of the operators.

• Environment, facilities, and utilities.

Figure 3. (click to enlarge) The fishbone diagram.

The Cause and Effect method was developed by Kaoru Ishikawa in the 1960s, and is often depicted by and referred to as a fishbone diagram because of its shape (see Figure 3). The fishbone method is best suited for nonconformities that involve foreign matters, contaminations, out-of-specification results, and packaging defects such as broken units and low counts or fills.

These two RCA methods are powerful quality tools and could identify the root causes of most nonconformities. Nevertheless, IVD companies should avoid the common pitfall of identifying operator error as the root cause for all nonconformities.

Figure 4. (click to enlarge) RCA classification/coding scheme (example).

To facilitate the collection of data for reporting metrics, IVD companies should consider using a classification scheme for the different root causes identified during the investigation process. A classification scheme can be a good starting point for those companies that are interested in creating one (see Figure 4).

Corrective and Preventive Action (CAPA)

IVD companies implement CAPAs to correct the nonconformities and eliminate their root causes from recurring in the future. To facilitate the correction, elimination, and prevention of nonconformities, investigators should at least state the CAPA objective and what each action will correct or prevent. In addition, companies should develop a check plan to determine the effectiveness of the CAPA being implemented.

IVD companies should identify the method of performing the CAPA effectiveness check in the original CAPA plan. Besides, the effectiveness check should take place one to three months after the completion of the CAPA to ensure sustainability of the corrective measures being taken. If the effectiveness check indicates that the corrective measures were not effective, companies should implement a new CAPA. In some instances, IVD companies should consider using another RCA method to reconfirm the original root cause identified for the nonconformities.

One of the common pitfalls for IVD companies in implementing a CAPA is the frequent use of training and retraining as a convenient fix for operator error. As Kristina R. Spitler stated in her article, “Ensuring Training Is an Effective Part of CAPA,” that “If we really want training to have a corrective and preventive effect, we need to carefully examine retraining events to ensure that they rectify the original problem, if it truly is a training problem at all.”

Metrics

IVD companies should develop a mechanism to monitor the continual performance of the investigation system in terms of process, system, and people. Companies should establish metrics or key performance indicators so that trend analysis (e.g., Pareto charts) can be performed to ensure that few vital problems exist. Since not all nonconformities require in-depth investigations, the quality department should review all quality records at defined intervals to ensure proper decisions have been made on evaluating nonconformities that require such investigations.

Based on the metrics and related trending, the IVD company’s management could quickly identify if an adverse trend exists and take appropriate actions to correct the trend. The company can also use the metrics and related trending as inputs for continuous improvement of the manufacturing processes and quality systems. Some common metrics that most IVD companies use to determine the performance of the investigation system are described as follows:

• Process: Number of closed investigations per defined interval (week, month).

• Process: Number of overdue investigations greater than 30/60/90 calendar days.

• System: Number of new investigations generated per defined interval (week, month).

• System: Number of open investigations per functional area or product.

• People: Cycle time in completing an investigation.

• People: Number of investigations completed per investigator per defined interval such as week or month (measuring productivity).

Conclusion

A well-established investigation management system will not only contribute to restoring controls back to the processes that prompted the nonconformities but will also provide the necessary inputs for continuous improvement activities. The pathway to a successful investigation system is to provide adequate resources, utilize a risk-based approach to evaluate nonconformities, conduct thorough and timely investigations, identify appropriate root causes and CAPAs, and develop metrics as management inputs for continuous improvements. The elements discussed above have provided some insights into setting up an investigation system that will meet the needs of an IVD company’s operations.

Stephen Sheng is director of operations at Quality Environmental Safety Inc. (San Gabriel, CA). He is also associated with The Quantic Group and Intertek Systems Certification. He can be reached at auditing@aol.com.