Wolfgang Schmitt is Director of Operations at Pharmaceutical Consulting Alliance (PCA) a business unit of Concept Heidelberg GmbH, Heidelberg, Germany.

Wolfgang Schmitt

The handling and examination of nonconformances and deviations is becoming more and more important to the pharmaceutical industry, primarily for two reasons. Firstly, the ongoing modernisation of pharmaceutical quality management systems has made the industry recognise the value and benefit of "failure" detection and CAPA (Corrective and Preventive Action). Continuous process improvement tools have a long and successful history, especially in the automotive industry, and these are now being adopted and implemented in the pharmaceutical industry. Applied properly, nonconformances can be prevented and processes can be continuously improved. 

Secondly, the regulatory requirements and expectations of government agencies worldwide have continued to increase in recent years. In November 2009, for example, the UK Medicines and Healthcare Products Regulatory Agency (MHRA) published a "Deficiency Data Review"

  for the period April 2008 to March 2009, in which it cited "anomalies" as their most frequent observation of regulatory deviations, with nonconformances and deviations in conjunction with CAPA being the third most common observation. This is similar to the findings of the recently published FDA Warning Letters Report from the European Compliance Academy (ECA).

  Here, the deficiencies in the "Production Record Review" (21 CFR 211.192) are among the most frequently reported GMP deviations. According to the report, more production record review deviations were observed in 2009 than in any other years and, on closer examination, most of the objections applied to the poor handling of deviations and incorrect CAPAs. In most cases, the companies did actually detect the deviations themselves, but did not properly examine them and did not define appropriate corrective and preventive actions. So what went wrong?  

It seems that one of the biggest challenges for companies is to complete investigations and actions in a timely manner. In many cases, incorrect assumptions are made that everything is an isolated incident. In other instances, problems are not corrected and everything is blamed on a single employee or a simple laboratory error, or the system fails to ensure that a problem does not extend to other lots, and the incident recurs. The ultimate criterion for adequate correction is to ensure that it doesn't happen again! 

CAPA was adopted as a new quality management tool following the introduction of the ICH Q10 guideline. According to the ICH Q10 document, which was adopted by the FDA in April 2009 as an industry guideline, a pharmaceutical Quality Management System (QMS) consists of four central elements: 												

process performance and product quality monitoring

management review of process performance and product quality. 

The guideline states that a pharmaceutical company should have a system in place to detect and evaluate nonconformances to take respective corrective and preventive actions. Among other things, the information regarding nonconformances can result from complaints, deviations, recalls, observations at audits and inspections, or from monitoring findings. The examinations within the system must have the objective of determining the actual root cause. As a result, the process and product should be better understood so that improvements can be derived from it.  

The EU Commission has now published a suggestion for the revision of chapter 1 of the EU GMP Guide to incorporate the recommendations of ICH Q10. Now, specific requirements for a CAPA system shall be included. A comprehensive description can be found in section 1.8 items 7, 8 and 9, in conjunction with the requirements for quality risk management systems according to ICH Q9 (identical to annex 20 of the EU GMP Guide). Accordingly, the extent of the actions, technical complexity and documentation of the necessary CAPA actions have to be managed according to a risk assessment. In doing so, it will become more and more important to fulfil these requirements and expectations quickly and smoothly, whilst keeping an eye on the economical and operational situation.  

However, neither of the above documents provides precise instructions on how to operate an efficient CAPA system. In this scenario, a 2009 proposed guidance on the Global Harmonisation Task Force (GHTF), titled "Quality management system — Medical Devices — Guidance on corrective action and preventive action and related QMS processes", might offer assistance. Within this guidance, the CAPA system is described as an integral part of a QMS and is divided into four phases (

). Although suggested for medical device manufacturers, this document also provides comprehensive guidance on how a pharmaceutical company could implement a functioning CAPA system. 

Identifying and documenting nonconformances

The first study of non-conformance was recorded at the beginning of the 20th century by the German educator Hermann Weimer (1872–1942), who established a psychologically oriented scientific study of nonconformance. Weimer was a pioneer in this field; he differentiated between a mistake and an error — an error being based on the lack of knowledge of a fact, while a mistake was defined as a deviation from what is correct, for which there is an individual personally responsible for the mistake. His theories, though groundbreaking, have naturally since been superseded. In modern nonconformance management, a nonconformance is understood as an actual situation or process, which deviates from a standard and includes all types of procedural mistakes. By no means should a nonconformance lead to the respective employee being immediately made accountable. This approach is, in fact, counterproductive to an efficient GMP failure detection strategy. 

The establishment of positive failure awareness is essential for a successful CAPA system. Simply training the assumed responsible person on a continuous basis will not lead to the desired prospect of success, but will lead to negative failure awareness. Employees as well as superiors must accept that people make mistakes. This should not be regarded as a flaw, but rather as an opportunity because otherwise, there is a risk that employees may try to cover up or play down mistakes. It is crucial to understand that all mistakes must be addressed and completely documented if a CAPA system is to be successful.  

In order to derive the correct — and therefore the effective — actions from nonconformances, the root cause must be determined. The following simple example highlights some of the factors to consider when identifying a root cause. 

An in-process control shows that spots, derived from machine oil, consistently appear on tablets. The reason for this is apparent — the tabletting machine is showing signs of oil contamination. In this case, it would be very easy to simply clean the machine and put it back into operation; however, the precise cause has not been established, which means the problem could arise again in the future. One must question why the tabletting machine leaked in the first place. If, for example, the gaskets are found to be the cause, these should be replaced, but not before finding out why the gaskets leaked. They may not have been correctly assembled or could be of poor quality. In the latter case, it must be examined why the gaskets were in-built. The real reason for the contamination of the tablets could be due to the installation of cheap replacement gaskets and a technically immature change control system. As such, the gaskets should be thoroughly inspected. 

Many options are available to perform failure analysis, for example: the Pareto chart, Ishikawa and Scatter diagrams, and the classical fault tree analysis. These tools provide a clear and simple visual representation of the steps that are involved in failure analysis and can facilitate the understanding, explanation and systematic analysis of complex processes and associated risks. 

Other techniques compare deviations with existing facts to get the results using systematic problem isolation (where/when/how did the failure occur in the process chain and where/when/how did it not and why?). Derived possible causes are continuously examined and adjusted.  

Overall, the following steps must be taken to identify the root cause of a failure: 								

possible associated or related facts must be considered

annual reports and trends should be evaluated

all equipment involved in the process should be part of the investigation (equipment maintenance records, qualification of equipment and the validation of process if applicable)

knowledge gained during process development should be documented

precision and accuracy of analysis must be verified. 

Definition, implementation and inspection of actions

Once the root cause of nonconformance has been established, the necessary actions must then be precisely defined. A risk analysis may help to assign action priorities. Responsible persons for the defined activities must then be named and realistic timelines specified. The processing should be tracked and evaluated to assess efficiency and, once all defined actions and action plans are formally concluded, the results must be conveyed to all parties involved. 

Often a CAPA system should be integrated into a QMS alongside existing systems, such as deviation management, complaint management, OOS handling, change control and observation processing systems. Before implementing a CAPA system, however, pharmaceutical companies should consider the following two options: 

In this case, the CAPA system replaces and completely integrates all other management and control systems. The advantages of such a comprehensive system are obvious: all non-conformances are recorded and processed in one system, making it easier to detect correlations, to cross-reference, plan and assign resources. 

In general, there is less risk of getting bogged down in the detail. The effectiveness of the actions is easier to recognise because of the increased transparency of the system and the correction of all activities is simplified. However, implementation of a new system is resourceintensive; many processes must be adapted and often it is necessary to implement new IT systems, resulting in increased time and effort for training. Initial difficulties will also ensue. 

In this scenario, all of the quality management systems continue to exist more or less independently, and only larger activities or projects are transferred to the CAPA system. In this case, it is of course more difficult to recognise correlations and an overview of the overall activities and processes is often missing. Even though such a multiple system represents the supposed more costefficient alternative, it should be taken into account that higher costs could develop in the longer term because of possible inefficiencies. Several parallel systems also mean more validation activities, maintenance work and training. 

Before implementing or improving a CAPA system, it is very important to understand where the bottlenecks are in existing processes and why they occur to avoid future inconsistencies and problems. Examples include: 								

interrelationships between different SOP processes

little resources for analysis and management

prioritisation between deviations and daily work

difficulty checking past data to establish recurring problems. 

Each company must also evaluate whether it requires a fully electronic system to display the CAPA process or whether a system based on paper is sufficient. The advantages and disadvantages are listed in 

Table 1: The advantages and disadvantages of a fully electronic CAPA system versus a paper system.			

The size of the company has a significant influence on the decision. A paper or suitable hybrid system can be sufficient for a smaller firm; however, for the larger firm that produces more batches and evaluates more processes, the advantages of an ITbased system far outweigh those of a paper system. It is, however, essential not to implement a computer system to automate paperbased processes; eworkflows must be designed before any computer system to support the implementation of a QMS.

No matter how the CAPA system is ultimately drawn up, it is important that all bottlenecks are identified and the required information is accessible. All deviations should be detected, the root cause determined and respective actions defined and implemented. This is the only way to learn from the non-conformances and to enable a continuous improvement process. 

Articles

An effective quality management system will identify the root cause of non-conformances and put measures in place to ensure that they do not recur, but is the pharmaceutical industry choosing the right methods to make sure this happens?