IVD Technology Magazine | IVDT Article Index
Originally published January, 1997
GMP inspectional consistency:
Industry comes forward
Leif Olsen and Patricia Shrader
FDA has often dismissed allegations of inspectional inconsistency as more perception than reality. Now, IVD companies are coming forward with examples to prove that such charges are more than just a figment of industry's imagination.
In recent years, members of the medical device industry--including IVD companies--have frequently alleged that FDA is very inconsistent in its inspections of manufacturers. Such allegations are commonly supported by accounts of individual inspections, and encompass both the inspectional practices of the agency's investigators and the outcome of those inspections. Whether or not these allegations are true, the fact that they have been made has chilling consequences for all those involved with the industry and its products.
Allegations of inconsistency are especially problematic for the manufacturers of devices and diagnostics by making it difficult to identify the practices that FDA considers to be industry standards. Most companies know their own practices well, and they may also be aware of the practices of competitors and neighboring companies with respect to manufacturing, labeling, advertising, and so on. Where FDA has established regulations or guidelines to govern such practices--and where industry follows them--individual companies can be comfortable that there is a level playing field. But when existing regulations and guidelines are not uniformly understood, applied, or adhered to, companies often become concerned about whether their practices are in compliance with FDA regulations and about whether overzealous enforcement is putting them at a disadvantage relative to their competitors.
With respect to FDA, allegations of inspectional inconsistency raise several questions, the most obvious of which relate to the objectivity and fairness of the agency and its investigators. Such allegations may also lead to questions about the effectiveness of the agency in educating and controlling its investigators, and in educating regulated industry. Most regulated companies honestly desire to meet regulatory requirements, either because they recognize the business advantages of compliance or because they want to avoid sanctions. But in order for a company to satisfy any set of requirements, they must be clearly conveyed and fairly applied. A regulatory agency that is viewed as inconsistent or unfair will not be respected by industry, and this, in turn, can reduce a company's capacity to learn from the agency as well as its willingness to comply with regulations.
For the purchasers and users of medical devices and diagnostics--and for the patients whose well-being may depend upon them--allegations of inconsistency raise concerns about product quality. Observations made during an FDA inspection may lead end-users to incorrectly conclude that one company lacks the quality systems that its competitors possess. Inspectional inconsistency may also cause end-users to question whether any manufacturer's prod- ucts can be relied upon for quality and consistency. In an era when consumers are extremely sensitive to charges that manufacturers have disregarded public safety, any such suggestion can have a devastating effect on the market for otherwise efficacious medical products.
To FDA's credit, complaints about inspectional inconsistency seem to be less common now than in previous years. But because the results of such occurrences can be so serious, it is still important to identify instances of inconsistency, to evaluate their causes, and to develop corrective actions. To accomplish this, it is essential to look not only at the present but also at the recent past, when complaints of unfairness were at their peak.
Shifting Procedures
In September 1990, FDA issued a statement of enforcement policy indicating its intention to protect the public health through the use of both administrative and judicial enforcement mechanisms. In May 1991, the agency announced two procedural changes intended to strengthen those mechanisms: streamlined procedures that delegated more authority to field offices and reduced requirements for headquarters review; and a new warning letter to replace both the notice of adverse finding and the regulatory letter previously used by the agency. These changes brought with them a host of complications.
Field Authority. The authority that was delegated to FDA district offices permitted them to issue warning letters for a variety of good manufacturing practices (GMP) and other violations without first sending them to be reviewed by the Office of Compliance at the Center for Devices and Radiological Health (CDRH). The intent of this change was to provide companies with quicker notice of violations that district offices considered significant.
Although FDA acknowledged that this change would reduce the ability of the device center to exercise the oversight functions that it had previously carried out, at the time the agency accepted this trade-off in order to speed up its enforcement process. Industry, however, did not necessarily consider the trade-off an even deal. In the view of many device manufacturers, this change in FDA procedure resulted in less consistency, both in the agency's inspectional practices and in other enforcement activities.
Warning Letters.Prior to the 1991 changes, FDA had used two types of letters to communicate warnings to inspected companies. The regulatory letter informed a company of serious violations, and indicated that judicial action would be taken unless it responded satisfactorily within 15 days. The notice of adverse finding was used to inform manufacturers about minor infractions that the agency expected them to correct without threat of further enforcement. Although the regulatory letter was a powerful tool (FDA rarely had to go to court in those days), the agency was less satisfied with the notice of adverse finding. In FDA's view, industry did not regard such notices as having potentially serious consequences, and thus frequently failed to take corrective actions as expected.
By replacing the regulatory letter and the notice of adverse finding with the single warning letter, FDA intentionally blurred the distinction between serious violations and minor infractions. Its intent was to put industry on notice that any such findings were considered very significant, and that it was prepared to take whatever enforcement action might be necessary to bring companies into compliance.
From 1991 through 1994 the number of warning letters issued by FDA increased significantly over the number of notices and letters issued in previous years. For a company in the United States, receipt of a warning letter usually also caused the company to be placed on FDA's so-called reference list. In turn, this resulted in disqualification for (or cancellation of) federal contracts, discontinued review of product applications pending at the device center's Office of Device Evaluation, an increase in inspectional frequency, and other results perceived as punitive. Equally important, because warning letters were issued without regard to the seriousness of the violation, they often caused end-users unjustifiable concern about a company's manufacturing practices and the quality of its products.
During this period, device manufacturers outside the United States were also inspected by FDA, many for the first time. In dealing with these firms, the agency adopted the same strategies it had implemented for U.S. companies. But when a warning letter was issued to a foreign company, it often also resulted in an import ban against its products.
Depending upon the source, a variety of explanations have been forwarded for the increased number of warning letters issued from 1991 through 1994. FDA has said that the increase was a result of more strenuous efforts to identify and take action against companies that were not meeting GMP requirements. Some in industry have suggested that the agency considered the issuance of warning letters to be a measure of its effectiveness, and increased its use of them in order to address congressional concern about weak enforcement. Others have said that the increase merely demonstrated FDA's inability to educate and work with companies. Whatever the reason, warning letters and related inspectional issues rapidly became a significant concern for industry. As a result, companies became more sensitive to real or perceived unfairness or imbalance in the inspectional process.
In 1995 these issues also came to the attention of the newly elected Republican Congress, which quickly made it clear that FDA practices would be high on its list of reform priorities. In response to this pressure, the agency modified slightly its approach to enforcement activities and began looking for ways to work more effectively with companies. Although FDA has not changed its enforcement mechanisms, the number of warning letters issued since the beginning of 1995 is substantially lower than in previous years. It is not clear whether this reduction is a result of better compliance efforts on the part of industry, fallout from FDA's earlier enforcement initiatives, or some other cause.
The Baltimore IVD Roundtable
The Baltimore IVD Roundtable is a group of FDA and IVD industry trade association representatives that meets quarterly to identify and discuss issues of interest to the diagnostics industry and to the medical device industry generally.
The group originated in June 1995, when IVD industry trade associations asked Kenneth Shelin, director of FDA's Baltimore district office, to set up a meeting with FDA representatives. The purpose of the meeting was to explore ways FDA and industry could work better together to achieve common goals.
At that initial meeting, discussion focused on the distrust that had grown between industry and FDA, the sources of that distrust, and possible means of reconciliation. Since then, the roundtable has continued meeting to discuss new issues of interest.
Guideline Inconsistencies
During the period when inspection-related complaints were at their peak, FDA also initiated a number of policy changes specific to the IVD industry. These were embodied primarily in the Guideline for the Manufacture of In Vitro Diagnostic Products, which was published in January 1994. Originally drafted in the mid-1980s, this document had undergone two prior revisions and had long been in the hands of FDA investigators.
Although FDA investigators had access to the guidance before its final publication, they had not been trained to interpret it and were advised not to use it when inspecting companies. Nevertheless, some investigators made use of the guidance prematurely, and as a result the use and interpretation of the document varied widely among investigators and districts.
For example, some investigators assumed that the practices described in the guidance were current industry standards, while others did not. This represented a source of inconsistency with potential to have a significant impact on a company's financial stability. Another source of problems was the fact that compliance with the guideline was considered to be voluntary, which may have led to confusion and misunderstanding by both investigators and industry. A manufacturer that was seeking to match the practices described in the guidance--whether by choice or because an investigator insisted upon it--could find itself investing heavily to develop systems that were not required of its competitors. The following paragraphs offer only a few examples of such potentially bank-breaking practices.
In common with earlier drafts, the final IVD manufacturing guideline classified IVD products as sterile, microbiologically controlled, or microbiologically uncontrolled, depending upon the level of control necessary for the product to meet its stated performance characteristics (section 3.1.1). To achieve an appropriate level of environmental "cleanliness," however, could require a manufacturer to undertake significant changes in its facilities, equipment, processes, and personnel practices (3.4). After publication of the IVD manufacturing guideline, a few companies implemented appropriate environmental controls in accordance with the guideline, while many others did not.
Another area of concern is process validation. Although FDA did not explicitly mention process validation in its 1978 GMP regulation, sections of the regulation had long been interpreted as requiring device manufacturers to conduct this activity whenever appropriate. These requirements were further spelled out by the agency in its 1987 Guideline on General Principles of Process Validation. To these general requirements, the IVD manufacturing guideline added the flat statement that "all new IVDs and/or processes are to be prospectively validated" (3.2). This requirement imposed a new burden on IVD manufacturers--but only if they elected to comply with the guideline or were required to do so as a result of FDA inspections. Other IVD manufacturers were not affected.
Another section of the IVD manufacturing guideline focused on stability studies and expiration dating (3.9). It indicated that FDA would accept only real-time data in support of stability studies, and spelled out a variety of environmental conditions that should be considered for such studies. While some companies undertook to meet these requirements, others did not.
Finally, the IVD manufacturing guideline imposed preproduction controls (3.1.1) and called for the application of statistical techniques to the sampling and testing of IVD products (3.8). For other types of medical devices, however, these two requirements will only become effective when the agency's new quality system regulation goes into effect this June 1.
FDA investigators had access to drafts of the IVD manufacturing guideline even before it was published, and they used them to identify what they believed to be significant deficiencies in company practices relative to the rest of the industry. But FDA investigators did not apply the guideline uniformly, nor did they share with industry a common understanding of its requirements. In conjunction with other inconsistencies in FDA's inspectional practices, these factors eventually had a multitude of adverse effects, including the regulatory side-effects of noncompliance, and, for some companies, financial and competitive disadvantages.
The Baltimore Report on Inspectional Consistency
At the winter 1995 meeting of the Baltimore IVD Roundtable, one of the participants noted that inspectional inconsistency had been identified as a significant issue that merited resolution.
Meanwhile, FDA had already gathered a nationwide task force to investigate such allegations. Under the direction of Ron Johnson, former director of the Office of Compliance at FDA's device center and now director of the agency's Pacific Region, the task force included a number of compliance officers and investigators. The mission of the group was to seek out and identify evidence of inspectional inconsistency.
Despite the efforts of the task force in reviewing establishment inspection reports and FDA-483s for examples of inconsistency, none were identified. In the summer of 1996, the group reported that it was prepared to complete its work and to conclude that inspectional inconsistency in the medical device industry is more perceived than real.
Learning at its summer 1996 meeting of the task force's imminent report, the Baltimore IVD Roundtable formed a working group to address the issue of inspectional consistency with a focus on the IVD industry. Members of the working group include Ken Shelin (Baltimore district office, FDA), Leif Olsen and Ray Watkins (Association of Medical Diagnostics Manufacturers), Paul Touhey (Medical Device Manufacturers Association), Bill Gilbert (Independent Reagent Manufacturers Association), and Pat Shrader (Joint Council of Immunohistochemical Manufacturers).
The working group on inspectional consistency adopted the following objectives for its work:
- To identify and present examples of inspectional inconsistency in the medical device and diagnostics industry, including inconsistencies among districts and investigators, in order to demonstrate that inspectional inconsistency is a real rather than a perceived problem.
- To identify sources of inspectional inconsistency in order to best identify appropriate corrective actions.
- To recommend steps that FDA can take, in conjunction with industry, to help ensure both the reality and the perception of greater inspectional consistency.
Working group members solicited information directly from their association members and from other sources. The information requested included examples of inspectional inconsistency as well as recommendations for FDA and industry efforts to address the issue. The report summarized in the accompanying article is the result of the group's efforts.
Inspectional Inconsistency
The inspectional inconsistency that concerns medical device manufacturers takes many forms, few of which are easy to discover. The following examples were contributed by member companies of IVD industry trade associations and other companies in industry. Some come from the contributors' personal experiences, while others are based on their communications within the industry. At the request of the contributors, the companies are not identified, nor are the specific FDA district offices that were involved in the inspections. All of the following information is a matter of public record. For further information, companies may contact the trade associations that participated in the Baltimore working group (see box below). A citation or explanation of the inconsistency is given in parentheses following each example.
1. A company on the West Coast was cited for failure to have a revision date on a label. The product in question was being manufactured for another company, which had supplied the labeling. All of the company's own products had revision dates on the labels. Because the noncompliant labeling was supplied by the company that contracted for the manufacture of the product, the inspected company felt the citation was inappropriate. (21 CFR 809.10(b)(15).)
2. A company on the West Coast was inspected by FDA and no FDA-483 was issued. The investigator returned to the company after the close-out meeting and asked to inspect the documentation relating to an investigational product. The company informed the investigator that investigational products are exempt from GMP requirements and that therefore the investigator did not have a right to review the documentation for this product. The investigator agreed that the company was correct, but told them his supervisor had made the request. The investigator left without reviewing the documentation. (21 CFR 812.1.)
3. A diagnostics company on the West Coast received a warning letter following an inspection. The FDA-483 issued at the close of the inspection listed only three items, all related to a recall the company had performed more than a year earlier. At the time of the inspection, the company had already implemented short-term corrective actions to address the problem leading to the recall and was able to demonstrate that the corrective action had been effective. The company also had a long-term corrective action in progress. Although the company called these facts to the attention of the investigator, the items were listed on the FDA-483 and a warning letter followed. The result of the warning letter was that review was discontinued for all the company's pending applications. (Compliance Policy Guide, 7382.830, attachment A-1.)
4. An IVD company on the East Coast was inspected by FDA in the spring of 1995. Although the IVD manufacturing guideline was in effect, the company was not asked at that time and had never been asked during previous inspections whether its water system had been validated. Some IVD companies in the same district had been asked about water system validation, while others had not. (Guideline for the Manufacture of In Vitro Diagnostic Products [Guideline], sec. 3.7.)
5. Two companies on the East Coast were inspected by the same FDA district during the same two-year period. One company was asked for its complaint files for the previous two years; the investigator identified complaints for which recalls were not conducted and urged the company to recall certain items, which were indicated on the FDA-483. The other company's complaint files also were reviewed and products were identified for which customer complaints were confirmed as lotwide problems. This investigator did not list the items on the FDA-483, nor did he suggest that the company recall the products. (Compliance Policy Guide, 7382.830, attachment A.)
6. A diagnostics company in the southern United States told the investigator that its products were "microbiologically uncontrolled." The investigator did not ask the company for supporting data to show that control of the products was not necessary to achieve acceptable performance. (Guideline, 3.1.1.)
7. An IVD company with a core facility for the manufacture of "microbiologically controlled" IVD products was audited by another IVD company from a different district within the same region. The auditing company expressed surprise at the sophistication of the auditee's facilities. The auditing company did not have similar facilities and had never been asked by FDA to identify appropriate environmental controls for the manufacture of similar products, while the auditee company had constructed the core facility in direct response to FDA inspectional observations. (Guideline, 3.4.)
8. Following the effective date of the IVD manufacturing guideline, a diagnostics company in the Midwest received an FDA-483 in which the citations included failure to specify a microbiological assurance level for each product, failure to generate data to support weekly environmental monitoring, failure to have statistically significant sampling plans for microbial load analysis, and failure to describe the practice of alternating disinfectants. The company received a warning letter follow-ing the inspection, despite promises of prompt corrective action. The company had a long history of acceptable inspection outcomes and had demonstrated its willingness and ability to implement corrective actions in a timely fashion. A competitor in another district that had recently been inspected by FDA did not have any of the specified controls in place. (Guideline, 3.1.2; 3.4; 3.6; 3.8.)
9. A diagnostics company in the eastern United States labeled IVD products "for laboratory use." An FDA investigator determined, based solely on the company's statements and without verifying the recipients of the company's products, that inspection under the GMP regulation was inappropriate. Other companies in the same area making similar products and selling them to the same facilities were routinely inspected under the GMP regulation. (21 CFR 809.10(d).)
10. Prior to implementation of the IVD manufacturing guideline, a diagnostics company outside the United States had an import ban placed on its products due to failure to adequately validate its water system. There was no indication that this failure had any effect on the performance of the company's products. (Inappropriate use of a draft guideline.)
11. A component supplier for IVD companies was inspected by FDA and received an FDA-483, even though company officials told the agency that they did not believe the company was required to comply with the GMP regulation because it did not manufacture finished products. (21 CFR 820.1.)
12. An IVD company outside the United States received a warning letter for GMP compliance deficiencies, even though the company's products were labeled for research use only. (21 CFR 809.10(d).)
13. A company on the West Coast was inspected in conjunction with a premarket approval (PMA) application. The investigator carried a copy of the Center for Drug Evaluation and Research guidance for WFI water systems. Despite repeated protests by the company that WFI was not required for its products and that the company's water met established specifications, the FDA-483 included observations regarding deficiencies in water system validation. Although there were no performance issues associated with the product, the warning letter that followed the inspection caused a delay in PMA approval. (Inappropriate use of CDER guidance.)
14. An IVD company on the East Coast was issued a warning letter that included an observation that the company lacked a statistical rationale to support the sample size for the serum panel being used in final product testing. The company hired an outside statistician to establish the appropriate sample size and rationale. The statistician was well qualified and had recently retired from FDA after serving as a division director. FDA rejected the consulting statistician's recommendation. After continued negotiations reached no satisfactory conclusion, the company dropped the requirement for use of a serum panel in final product testing, satisfying FDA and resulting in the close out of the warning letter. (21 CFR 820.160.)
15. An East Coast company was cited on an FDA-483 for not manufacturing its IVD products as sterile, even though those products were not labeled as sterile. The investigator later accepted the validation studies demonstrating bioburden control in accordance with the IVD manufacturing guideline. This involved expensive facility renovations to establish a high level of environmental control. The same level of adherence to the guideline had not been required in other districts. (Guideline, 3.1.1.)
Conclusion
The examples cited here demonstrate that inconsistency in FDA inspection and enforcement practices is real, not merely an industry perception. But knowing this is by no means the end of the process. The next steps must be to determine the causes that lead to such inconsistencies, and to develop corrective actions to prevent their recurrence. These will be the subject of another article in the next issue of IVD Technology.
