A THIRD-PARTY TRAINING INSPECTION at a Philips Medical Systems plant in Germany resulted in a warning letter, to the firm's surprise. Philips volunteered for the audit in order to support FDA's Accredited Persons inspection program, which is languishing due to difficulties in scheduling training sessions for the participating third-party auditors. A Philips exec describes what happened at this inspection and a separate, more successful AP training session in Andover, Mass., while a Gambro manager details a training audit at her firm that went well. Only nine out of 55 third-party auditors have completed the training necessary to perform independent FDA inspectionsWARNING LETTERS: Baxter, GE and St. Jude cited for CAPA and other QS violations; Cochlear implant maker Advanced Bionics faulted for handling of moisture problems in devicesIN BRIEF: Diagnostic Products Corp. receives FDA nod for corrective plan to ensure reliable clinical data in submissions; agency issues guidance on device software

Firms Report Mixed Experiences In Hosting AP Training Inspections

Two device companies that have allowed Accredited Persons (AP) auditors to be trained during inspections at their facilities emerged with vastly different stories: one received only minor observations from inspectors, while the other incurred an unexpected warning letter.

The warning letter went to Philips Medical Systems on May 6. It followed a Nov. 15-18 quality systems inspection at the firm's Boeblingen, Germany facility, during which an FDA investigator trained an employee from TUV America, one of the 15 auditing firms participating in FDA's third-party inspection program.

Prior to receiving the warning letter, Elisabeth George, VP-Quality, Regulatory, Sustainability & Product Security for Philips' cardiac and monitoring systems business, had told "The Silver Sheet" that the Boeblingen inspection went less smoothly than usual.

She described apparent tensions between the FDA and TUV America auditors, and expressed concern that the poor dynamic may have contributed to an unusually high number of observations on the FDA-483 issued at the end of the inspection.

Philips' experience could make it harder for FDA to recruit additional volunteers to host AP training inspections. With the AP inspection program taking longer than expected to become operational, the agency has been actively encouraging device firms to volunteer for the sessions.

AP auditors are required to complete three of the joint inspections with agency trainers before they can perform FDA inspections on their own ("The Silver Sheet" July 2004).

Philips' George said that despite the Boeblingen outcome, she would be willing to go through the process again. An additional AP training inspection Philips hosted at its Andover, Mass. facility last summer went very well, she pointed out.

The quality and regulatory affairs exec is confident that Philips will be able to resolve FDA's warning letter concerns. The agency cited the firm for several complaint handling and corrective and preventive action lapses, including failure to generate complaint files in response to numerous customer calls concerning product performance issues, and failure to thoroughly investigate the cause of alarm malfunctions in bedside monitors.

FDA further faulted the company for not providing evidence in its post-inspection communications that the violations have been resolved.

According to George, the firm has now sent the agency "all of the procedures and all of the records" on its corrective actions, and has offered to meet with FDA compliance officials.

Meanwhile, a representative of Gambro BCT remains pleased with her company's experience hosting an AP training inspection in March 2004. Lynette Makowski, regulatory affairs and quality assurance manager, said Gambro would volunteer to host AP auditors moving through the next of their required three training audits whether her firm was due for an FDA inspection or not.

Gambro's eagerness is motivated mainly by a desire to accelerate the AP inspection program, which "sounds like it is dragging pretty badly right now," Makowski said.

The training sessions, which count as official FDA quality systems inspections, are the final step in implementing FDA's Accredited Persons inspection program, established more than two years ago under the Medical Device User Fee & Modernization Act.

Once it is up and running, the program will allow device companies to hire selected third parties to conduct their FDA quality systems inspections. It is expected to benefit manufacturers that are subject to frequent FDA and European/Canadian inspections by allowing them to undergo both types of inspections during a single visit from an "accredited person" auditor.

In theory, manufacturers that take advantage of the program will have fewer inspection-related disruptions to their operations. In many cases, they will be able to use a third-party firm they already know and feel comfortable with to perform their FDA inspections.

While many device firms have indicated they would like the option of using a third-party auditor for their FDA inspections, relatively few have been willing to host the AP training inspections required by FDA before the program can begin - in part due to concern that the training audits will be more rigorous than a normal inspection.

As recently as May 24, CDRH Director Dan Schultz implored device companies to help get the AP program off the ground. "We are going to ask for, and expect, your cooperation in making this happen," he told industry stakeholders at AdvaMed's annual device submissions workshop in Arlington, Va.

Schultz said there is "no reason, especially for so-called low-risk, standard inspections, that [firms] need to be inspected 10 times a year by 10 different jurisdictions, and why those jurisdictions need to spend their resources doing an inspection that can easily be done by one well-trained qualified party. It is time to get this program moving."

FDA "Very Systematic" At Philips' Andover Facility

Philips Medical Systems signed on for two AP training sessions last year in the spirit of supporting an FDA program the company believes offers several potential benefits.

With manufacturing operations in Germany, the Netherlands, Israel and the Asian Pacific, the firm looks forward to having AP auditors conduct inspections of its foreign facilities, rather than waiting for FDA to travel to the distant sites.

"Philips in general feels that partnership with the FDA is a much better way to go," said George. "It allows us to help them to learn more about our industry, because we've got such complex products. And it also allows us to better understand the focused areas FDA has."

The company also likes the idea of having an AP auditor check on compliance with both ISO 13485 and the FDA Quality System Regulation during a single visit - the AP program's main appeal for most device firms. "There is a clear perceived value in only having one inspection rather than two," said George.

Participation in the AP program also may provide marketing benefits, allowing Philips to claim that facilities have had annual inspections under both FDA and ISO 13485 standards, George noted. In the case of some units that make Class II devices and have not had an FDA inspection for several years, "one of the questions that has come to light is, 'Are they still compliant to today's interpretation of the FDA laws?' The only way you can know is to be inspected," she commented.

With these future benefits in mind, Philips offered to host AP trainees at two facilities with previously strong compliance histories. The first inspection took place last July at the company's Andover, Mass. electrocardiograph manufacturing site. The second was in November at the Boeblingen plant, which makes patient monitoring devices.

Both facilities are subject to FDA inspections about every 18 months, and both had received "no action indicated" (NAI) outcomes - FDA's most favorable classification - for their most recent previous inspections.

In Andover, FDA investigator George Allen, from the agency's New England District Office, trained two AP auditors. The trainees were from the Mississauga, Ontario-based Quality Management Institute and the Hsinchu, Taiwan-based Center for Measurement Standards/Industrial Technology Research Institute (CMS/ITRI); the Philips inspection was the first of three required qualifying audits for both individuals.

Like all the AP qualifying audits, the Andover inspection was a comprehensive Quality System Inspection Technique inspection, meaning it covered all aspects of the firm's quality system, as spelled out in FDA's QSIT manual. The facility's previous four FDA inspections had been CAPA-plus-one audits, in which the FDA investigator reviews the corrective and preventive action subsystem, plus one subsystem of his or her choosing.

The FDA investigator "was very, very systematic," said Philips' George, noting that the investigator's strict adherence to the official QSIT approach made for an orderly, predictable inspection process. "He took the QSIT manual and walked through the manual, so there were no surprises there," she said.

The Andover inspection "was extremely well done," she maintained. FDA investigator Allen "took the time as he walked through each of the questions to explain to [the trainees] what records they were looking at, and how they were looking at it. He was very slow and methodical in his approach."

Although the investigator's approach made for long days, George appreciated the constructive interaction between the FDAer and the two AP trainees.

The trainees "made sure they understood the records they were looking at. And when an issue would be identified by the FDA, the FDA took the time to pull out the very specific paragraph in the regulation that was at issue," recalled George.

FDA investigator Allen said he approached the AP training session in the same way he would any inspection. He provided the trainees a copy of the inspection flowchart in the QSIT manual "and we just went step by step by step," he explained. "I said, 'You should approach it this way - just go through step by step.' There was no anxiety that it had to be tougher or anything else."

According to Allen, the two AP trainees were well equipped for the job. "I found these people to be very knowledgeable. It wasn't any more difficult than training some of our own people," he said.

"They had experience with the products that were being investigated. They understood the inspection technique." Although the APs' frame of reference was ISO 13485, "it was easy for them to make the transition" to the quality system reg, Allen added.

Philips' George agreed. "Both of the APs were very experienced people. They would frequently ask me for confirmation as to how [an inspection item] correlated to ISO 13485. They would do a little bit of sanity checking."

Albert Li, a manager at Taiwan-based CMS/ITRI, which participated in the Andover inspection, said his firm's AP auditor found it invaluable to get a close-up look at the manufacturing operations of a major device firm.

"Philips is a big...global company. It has much more experience [with] quality system improvement than many Taiwanese device companies. To observe what a good quality medical device manufacturer does was the biggest thing [he] learned during the training inspection at Philips," Li said.

The CMS/ITRI auditor also was struck by the strength of the Quality System Regulation as a compliance tool, according to Li. "The biggest difference between conducting a notified body ISO 13485 audit and conducting an FDA inspection is that FDA investigators can handle a whole inspection powerfully and flexibly because of clear and detailed law and regulation," he remarked.

Li's firm hopes its two AP auditors will be able to complete the qualifying audit process in 2006. Once the third-party inspection program is operational, CMS/ITRI expects to find potential clients for its AP services from among the 400 device manufacturers in Taiwan, a third of whom currently export to the U.S., he said.

According to George, each day of the five-day Andover inspection ended with a discussion between the auditing group and company reps. "They would close the day saying, 'At this point, evidence would show that this may end up as a nonconformance,' and…then they would ask us, 'Do you think that you'll have anything to refute that?'" On the following morning, she said, the auditors allowed Philips to present evidence showing why the company felt the particular matter did not represent a nonconformance.

At the close of the Andover inspection, Philips received an FDA-483 with two observations. FDA ultimately classified the inspection as "voluntary action indicated," the agency's designation in cases where nonconformances noted by the investigator are relatively minor and no regulatory action is anticipated.

A Rougher Inspection At Philips Plant In Germany

Things did not go as smoothly at Philips' Boeblingen facility, where FDA Florida District-based investigator Victor Spanioli trained an auditor from Danvers, Mass.-based TUV America.

The inspection began on an unsettling note, with Philips' George learning only as the inspection team walked in the door that the AP auditor was undergoing the second of his three qualifying audits. At this stage, the AP trainee leads the inspection, with the FDA investigator offering input as needed.

"I was under the impression that it was going to be the same as the Andover inspection - that the FDA would be leading the inspection," said George.

Although the team followed the QSIT manual, the rapport between the AP auditor and the FDA investigator was not as good as at Andover, according to George. "We did feel that there were a number of times where the FDA and the AP were a little bit confrontational with each other on an issue," she said.

"When the AP identified what he felt was a nonconformance during the inspection, I almost felt that the FDA…said, 'Well, that's a nonconformance, but you missed this as well - almost like a little bit of one-upsmanship. And we did feel that the results of the inspection were a little bit more negative than we would have expected."

Furthermore, end-of-the day discussions were less open than in typical FDA inspections, George said. For example, the Boeblingen auditing team did not always reveal all of their concerns from the day's investigations, "which left us a little bit in the lurch," said George. Moreover, in some instances "they would emphatically state, 'This is a nonconformance,'" leaving Philips no opportunity to discuss the matter.

The Boeblingen inspection resulted in an FDA-483 with seven observations. In accordance with Philips' internal policy, George sent a formal response to FDA within two weeks. "We actually committed to a follow-up response because we knew that in the 14 days I wasn't going to be able to commit the complete response," George explained.

The follow-up, sent later, provided an update on corrective actions planned or taken in response to the FDA/AP observations. "In some cases we actually attached samples of the records so they could see that we had in fact done what we had committed to," she said.

George noted that a few of the quality system issues identified during the Boeblingen inspection were somewhat difficult to respond to. "We have a single quality system for all seven of the facilities that I am responsible for. So in some cases if we deemed that it was potentially systemic…because it was a process issue, we had to take the time to roll out" the corrections. The quality system changes potentially affected up to 2,500 employees.

George can't help suspecting that the Boeblingen training inspection turned out worse for Philips than if it had it been a regular FDA inspection - even with the same FDA investigator. The fact that the quality system at the Boeblingen plant is nearly identical to the one in Andover only added to her discomfort.

George has managed more than 50 FDA inspections for her company, including six since last August. The May 6 warning letter was the first Philips has received at a facility under her watch.

Despite the less-than-ideal experience at the Boeblingen facility, George reiterated that she would host another AP training audit. "I would, but I would be a lot more diligent in making sure I understood the roles and responsibilities [of the parties involved] - who is the follower, who is the decision-maker, and what methods will be used."

As someone who is "very comfortable interacting with" agency investigators, George admits she preferred the stage-one AP training audit, in Andover, where the FDA investigator clearly drove the process.

Gambro BCT: An Enthusiastic Volunteer

A firm's willingness to host a joint FDA-AP inspection probably depends in large part on the extent of cooperation between the company and the agency in general.

Gambro BCT, for instance, has worked with the Denver FDA field office on various training-related programs and a level of trust has been established over the years, according to RA/QC manager Lynette Makowski.

The Gambro AB subsidiary, based in Lakewood, Colo., manufactures blood collection and processing devices such as the Trima collection system. It has operations in North America, Latin America, Asia and Europe.

"We are pretty much familiar with the FDA personnel from our local office," Makowski told "The Silver Sheet." And the firm is comfortable working with each of them, she explained.

Volunteering for the joint AP/FDA visit, which took place during a regularly scheduled quality systems inspection in March 2004, just seemed like a "good idea," Makowski said. "We want to work with our FDA [branch office]. This is a cooperative effort," she explained.

Gambro BCT had independently sought to host BSI, its ISO notified body, in an AP training inspection, but when those arrangements were delayed, FDA contacted the firm about Orion Registrar, another approved AP organization with whom Gambro BCT had not previously worked.

The 1,550-employee firm gladly accepted the offer to host an Orion auditor at its Lakewood facility, Makowski said. "The person that was coming in was on the first phase [of training]," she explained. "I guess we had some assumptions that they wouldn't actively be making observations...that this person was truly interested in learning."

Joint Audit Was "Organized," "Cordial," Gambro Says

Gambro BCT sees the potential to obtain more timely inspections with auditors that have a sophisticated understanding of the company's technology as one of the main advantages of the AP program.

"Not that we are dying to have FDA audits out here every two years," Makowski noted. "But it is a lot easier if somebody has been here within some reasonable time period...because otherwise there are too many changes."

"If we can get more people through the [training] process...I would think that there would be a benefit to everybody."

But she also reported that the training inspection itself offered some specific advantages, most prominently in the area of efficiency.

"Historically in audits, schedules are never really kept well," she pointed out. "Things kind of drag. Even though you may know you are going through a QSIT inspection, things are totally undefined. In this particular case, it was much more organized."

The efficiency of the audit was partly due to the fact that the Orion trainee was only available for one week. This put a closure date on a type of inspection that is typically open-ended and could sometimes take up to three weeks, according to Makowski.

She suggested that the training aspect of the audit made for a more orderly and quicker process. Most notably, the FDA inspector defined, much more clearly than usual, how each day would proceed to both the company and his trainee.

The shorter time period did not translate into a less in-depth inspection, Makowski noted. "I wouldn't say that...the FDA negotiated on how much they would review," she noted. "They were thorough and certainly had the challenging questions. From that perspective, it felt the same [as past inspections]."

The FDA trainer established an effective working relationship with the Orion employee from the start, Makowski observed. "The investigator was very diligent and tried to make sure that everything was in place for this person so that he could truly go through a good training process."

When the pair arrived - after Gambro BCT had gotten more than the typical five-day advance notice - it was clear to Makowski that the Orion auditor had been given a good introduction to the process. And during the inspection, she noticed, the FDAer was thorough in explaining his questions to the trainee.

"It was a very cordial inspection. I won't go to the degree of saying [it was a] friendly inspection, but it was very 'working together,'" she noted.

Collaboration between the FDA trainer and the AP trainee seemed to grow throughout the inspection, Makowski said. "I really think they walked out with kind of a bonding," she observed.

"At the beginning of the audit, the guy from Orion was pretty quiet, and then as it started moving along...you could see a relationship building." By the latter part of the process, he asked questions more openly, of both the FDAer and Gambro reps.

The Orion auditor's technical competency and ISO auditing experience allowed him to make "some good points" that improved the quality of the inspection, Makowski said.

Several FDA-483 observations were identified, but those were dealt with soon after the inspection, according to Makowski. The end of the process went smoothly and Gambro received the establishment inspection report quickly, she noted. FDA classified the inspection as "no action indicated" (NAI).

Overall, Makowski had no complaints. She acknowledged, however, that her firm might have been somewhat lucky. The two inspectors "worked it through, which was nice, because they could have had a real conflict," she said. "Or the FDA investigator could have had an attitude. But I didn't feel that any of that was present."

Makowski feels there is no reason for a medtech firm in good standing with FDA not to volunteer for an AP training audit. If a company "already has a reasonable relationship with FDA, and has a reasonable history of regulatory compliance, there really shouldn't be anything for them to be afraid of," she said.

The Gambro exec believes that training inspections can offer something to all of the parties involved, including FDA. "The notified bodies have a lot of education and knowledge and background, [and] sometimes they have more hands-on" industry experience than FDA investigators, she claimed. The training sessions offer a chance for an AP auditor's know-how to rub off on the agency, she suggested.

Questions About How APs Will Handle Internal Audits

Two years into the roll-out of the AP program, a number of stakeholders remain anxious for guidance from FDA on how to handle one tricky aspect of third-party inspections.

The issue relates to the fact that FDA investigators generally are not authorized to examine a firm's internal audit reports or management review meeting minutes, while ISO 13485 auditors are. Indeed, notified bodies often begin their ISO 13485 inspections with an examination of these documents.

Device firms are wondering how an AP hired to perform both an ISO 13485 audit and an FDA inspection will manage the conflicting authorities.

According to Philips' Elisabeth George and others, the general understanding is that an AP auditor would complete the FDA portion of an inspection first, issuing an FDA-483 if necessary. The AP would then proceed to the ISO 13485 review, examining the internal audit and management review records and any other required items not covered during the FDA inspection.

This approach would prevent the AP from being influenced by internal audit and management review information during the FDA inspection. However, it leaves open the question of what the AP auditor would do if he or she discovered a potential quality system violation not previously detected during the FDA inspection.

The possibility that FDA would want its inspection reopened in such a scenario worries some firms. "There is significant concern in that area," said George, predicting that the issue could cause many companies to avoid using the AP program.

"I don't think there is sufficient guidance" on the matter, she added.

According to FDA investigator George Allen, the agency's lack of authority to view a firm's internal audit and management review reports was the most important point he emphasized to the AP trainees at Andover. The AP auditors were already aware of the limitation, he noted. But "there is still a lot of confusion about how [APs] are going to separate the two programs," he said.

Could AP Program Harm NB-Client Relationships?

Although Gambro BCT's Makowski is a strong supporter of the AP program, she expressed some concern that adding FDA inspections to the duties of her company's notified body (which also is accredited under the AP program) could disturb the relationship between the two firms.

"[The] relationship is broader than just audits," she explained. "You actually have a partnership with this notified body and they help you do other things with your quality system and certifications and all that....You start to develop a relationship with your client manager, who comes on a routine basis [and] gets pretty familiar with your products and your processes."

"When you are certified to the ISOs...there is always a feeling that the notified bodies are actually there to help you market your product, where FDA might be there, you know, to enforce the law," she added.

Orion Registrar President Paul Burck, however, asserts that this is not a significant issue, at least at his firm. Due to policies Orion has set up for "education and disclosure" to manufacturers, he is confident that clients have an understanding about how its auditors would juggle new responsibilities under the AP program.

"We reinforce that [when we do an FDA inspection], it is different than what we do as a notified body...and they realize that," he said.

"We set up protocols for that because our clients are very concerned," Burck said. The protocols, he explained, "put up proper walls and proper controls" between diverging aspects of FDA, European and Canadian standards.

"We are very concerned from our end that people... remember what hat [we] have on," Burck said.

FDA Scheduling Difficulties Delaying Program

FDA acknowledges that the AP training process is progressing slower than initially expected. According to the agency's current count, only five of the 15 AP firms have auditors who have completed all the required training and are ready to conduct independent FDA inspections.

Moreover, the auditors completed their three required training inspections not under the AP program, but under the U.S./European Union Mutual Recognition Agreement, which carries similar training requirements.

Spread among the remaining 10 AP firms are 46 individual auditors still awaiting their full training. Of these individuals, two have completed two training inspections, 21 have completed only one training inspection, and 23 have had no training audits.

A letter sent last November by FDA's Office of Regulatory Affairs to 5,000 U.S. companies urging their participation in the training process produced some volunteers, according to Bill Sutton, deputy director of CDRH's Division of Small Manufacturers, International and Consumer Assistance. However, some of the respondents were ineligible to host training audits because they have never had an FDA inspection, Sutton said.

At the same time, it has proven complicated to schedule training audits for the firms that volunteered and are eligible, with coordination required between the CDRH Office of Compliance, the FDA Office of Regulatory Affairs, and field offices, Sutton explained.

The device center staffer stressed that a paucity of firms offering to host joint audits remains a major obstacle to program implementation. "The more volunteers we get, the more chance we will have to find firms that are eligible" and can fit within the Office of Compliance, ORA, and field agendas. "It is a big coordination activity," he said.

Meanwhile, frustration is mounting among some stakeholders. Orion President Burck says he had hoped to have some auditors ready to perform independent FDA inspections last year. As of mid-May, however, Orion's only AP training inspection was the one with Gambro BCT, over a year ago. The same auditor is scheduled for a second-phase joint inspection this summer at another manufacturer's facility.

"A year ago, everything was starting to click," he mused. "We had our first [training inspection] done, we had feedback, we had further developed...a lot of protocols...then all of a sudden we walked off the cliff into the abyss."

ORA has not provided ample opportunities for his firm to attend joint inspections, Burck lamented. And when inspection dates are offered, it is with very short notice - sometimes only a month in advance, when his employees are already booked solid.

This means that other appointments must be broken or the training doesn't happen. Burck said that two Orion auditors have yet to work out a time for a training inspection with ORA.

Despite the frustrations, the Orion exec remains highly interested in getting his auditors out on independent AP inspections from a business standpoint. "Investment in the AP program is substantial for us."

FDA Warning Letters
The following chart lists medical device-related warning letters released by FDA from April 19 through May 10. If an inspection led to the warning letter, the location of the inspected facility and dates of inspection are noted; otherwise, the location of the warning letter recipient is listed.
Description	Summary
Advanced Bionics Corporation Warning letter date: Feb. 1, 2005 Location: Sylmar, Calif. Inspection dates: Aug. 25-Sept. 15, 2004	Manufacturer of cochlear implants cited for QS reg deviations. FDA described numerous QS deficiencies that played a role in moisture being hermetically sealed in *HiRes 90K* and other cochlear implants. The firm did not address the moisture issue in management reviews, and the internal audit procedure lacked provisions for auditing manufacturing processes and process validation; additionally, there was no procedure detailing how audit checklists should be prepared, reviewed and approved in advance of audits. Root cause of the moisture in hermetically sealed devices was not identified, and products continued to be manufactured and distributed, exposing patients to risks of device failure and associated risks of surgical intervention and permanent hearing loss. Vacuum bake ovens used for moisture removal were not validated and process controls were not effective in detecting gross leaks. There was also a lack of in-process or finished device sampling and testing to determine if devices conformed to moisture limits. Employees lacked training in basic principles of process validation. There were no design controls for the modification of the *PowerPak* and no written determination that such controls were not necessary. The water content limit for the HiRes 90K device was not specified in a design input document, and risk analysis was inadequate. Manufacturing process validation did not ensure that non-hermetically sealed devices were identified as unacceptable for distribution and implantation. The firm failed to establish and maintain procedures for monitoring and control of vacuum bake process parameters. Process revalidation was not performed to ensure that moisture was being removed from the cochlear implant prior to welding of the vent hole, which hermetically seals the device. Products were not sampled and tested to ensure hermetically sealed products were compliant with moisture specs; finished devices were not screened for gross leakage; finished device acceptance did not include verification that vacuum was maintained throughout the vacuum bake oven process; and there were no incoming acceptance procedures to ensure that the helium/argon gas mixture met specs. Nonconformity investigations were incomplete, and the firm did not identify all of the actions needed to correct and prevent the recurrence of nonconforming product. In addition, the company failed to determine the extent of electronic circuit board damage that may have resulted from moisture. Available information about device failures was not always forwarded to FDA as updates to MDR reports or to comply with PMA conditions of approval. Finally, the firm did not furnish information concerning residual moisture in explanted devices in MDR reports to FDA. The warning letter notes that Advanced Bionics agreed to cease shipment of devices until compliance issues were resolved. [Issued by Los Angeles District Office]
Baxter Healthcare Warning letter date: March 23, 2005 Location: Largo, Fla. Inspection dates: Oct. 12-22, 2004	QS deviations cited at facility that manufactures hemodialysis devices. Baxter failed to verify or validate corrective and preventive actions to ensure their effectiveness. Specifically, the firm did not verify that a letter to users effectively explained the potential dangers of using unqualified bloodlines with *Meridian* hemodialysis machines, although the firm had received reports of deaths and serious injuries possibly related to use of unqualified bloodlines supplied by Baxter. Additionally, the company did not perform a risk analysis when it substituted qualified bloodlines with ones that did not meet the specifications required in the Meridian operator's manual. Problems with unqualified bloodlines were observed at both the McGaw Park, Ill. and Largo facilities, FDA noted, leading to agency concern that Baxter's corrective and preventive actions "may be inadequate and appear to be a systemic problem" throughout the company. The warning letter describes six additional serious product problems for which Baxter's corrective and preventive actions were inadequate or unverified/unvalidated, including a software anomaly that could cause patient overfill in *Homechoice* peritoneal dialysis machines, air detection alarm problems in Meridian hemodialysis machines andSystem 1000 hemodialysis machines, and alarm system inactivation in hemodialysis machines. In addition, Baxter did not use appropriate statistical methodology to detect recurring quality problems; for example, there was no documentation showing that the firm analyzed nine incidents of needle dislodgement as multiple incidents rather than as one complaint, complaint information received by the firm was not analyzed by failure modes, and MDRs were not analyzed for problems occurring at the Largo facility, FDA said. Baxter also failed to adequately investigate the cause of product nonconformities, waiting nine months to place System 1000 machines on hold after noticing that the machines were not meeting air detection specs. Baxter also failed during design validation to identify a software anomaly in the *Renalsoft* Software Suite which resulted in the delivery of the wrong medication to one or more patients. In at least 15 complaint investigations, Baxter failed to identify which dialysis device was used with which type of bloodline; in three other incidents, machines and bloodlines were identified, but Baxter failed to determine whether the bloodline was qualified for use with the subject dialysis machine. Finally, the company failed to submit MDRs for serious injuries related to *Medisystems* blood tubing, and failed to report several corrections and removals to FDA. [Issued by Florida District Office]
Catch Incorporated Warning letter date: April 6, 2005 Location: Bothell, Wash. Inspection dates: Feb. 9-11, 16, 18 & 24, 2005	Specification developer of Homogeneous Enzymic Homocysteine Reagent test system cited for QS reg violations. The firm had no approved written procedure for handling a complaint involving the homocysteine assay, and failed to maintain a device master record. Purchasing controls were deficient, with the firm failing to establish and maintain requirements to be met by a supplier of enzymes used in the manufacture of the test system. The company also had no approved written CAPA procedures for verifying or validating a corrective and preventive action involving the homocysteine assay and ensuring the effectiveness of a subsequent change in the package insert. Finally, design control procedures were not established and maintained to include all necessary elements. [Issued by Seattle District Office]
ConvaTec Warning letter date: April 18, 2005 Location: Skillman, N.J. Inspection dates: Jan. 27-Feb. 8, 2005	Bristol-Myers Squibb subsidiary cited for QS reg deviations associated with the manufacture of various ostomy care products. ConvaTec failed to investigate devices associated with serious injuries and deaths, including the death of a child; specifically, the firm did not investigate whether the devices failed to meet specifications. In addition, the firm did not identify actions needed to correct and prevent recurrence of product defects and other quality problems involving a pouch device. ConvaTec failed to submit an MDR report to FDA concerning a device that may have caused or contributed to a patient's artery being cut, an injury that required medical intervention at a hospital. [Issued by New Jersey District Office]
Dotolo Research Corporation Warning letter date: April 5, 2005 Location: Pinellas Park, Fla. Inspection dates: Jan. 26-28, 2005	Manufacturer of colonic irrigation system cited for QS and MDR reporting violations. Employees responsible for quality control said they did not have time to obtain training and knowledge of the QS and MDR regs, and consultants hired to help write procedures and conduct audits focused on ISO standards rather than FDA regs. Internal audit procedures failed to address the QS regs. CAPA procedures did not require verification/validation of the effectiveness of corrective and preventive actions prior to implementation. The firm did not adequately document an investigation and corrective actions related to complaints about carbon filter failures, and failed to warn users that the filter could be adversely affected when unused for long periods. Dotolo failed to establish and maintain quality requirements to be met by suppliers and contractors; among other supplier control problems, the firm did not verify that the carbon filter included with the *Toxygen BSC-UV* device removes chlorine and toxins from water, and vendor-supplied UV lights were not verified to meet light intensity requirements or the claim that they render 99.99% of microorganisms harmless. The firm did not assure that chlorine residue on the speculum meets acceptable levels after rinsing and that bacteria on rectal hoses and specula are removed by exposure to UV light. Design control procedures did not include requirements related to labeling or packaging design changes, and changes to labeling in the owner's manual and website did not include design control to verify, validate or address risk analysis. There was no documentation of complaints received from 1999 to 2001. Finally, MDR procedures lacked a full definition of an MDR event, and employees did not understand the definition of an MDR. [Issued by Florida District Office]
PIE Medical International Warning letter date: April 27, 2005 Location: Duluth, Ga. Inspection dates: March 1-3, 2005	Specification developer for Pulsed Irrigation Evacuation (*PIE*) Bowel Prosthetic System cited for QS reg violations. There was no management oversight of employees responsible for manufacturing, finished device release, distribution or quality system record maintenance. The firm did not conduct or document investigations into 12 complaints involving possible device failures. In addition, 18 complaints had no documentation in the complaint files to determine whether an investigation was necessary. The firm had no documentation showing that a new software version for the controller unit was validated. In device acceptance and purchasing control lapses, the firm had no procedures to assure that finished devices met specs, did not perform finished device testing, had no documentation of contract manufacturer audits, and did not define the controls to be exercised over suppliers and contractors. Finally, the company failed to file an MDR report concerning peritonitis and a rectal perforation possibly caused by the PIE speculum. [Issued by Atlanta District Office]
General Electric Warning letter date: March 31, 2005 Location: Salt Lake City, Utah Inspection dates: Nov. 15-Dec. 1, 2004	QS violations cited at facility that manufactures x-ray and fluoroscopic systems and accessories. GE executive management failed to conduct periodic checks or audits of testing records, which would have revealed that test technicians were not completing original test procedure documentation, but instead copying test results from one device to another for at least a year. There was no assurance that testing was actually performed by the employees who falsified the records. Further, executive managers did not investigate the breadth of test result copying or perform risk analysis to determine how the copying may have affected finished devices; the firm initiated corrective and preventive actions only after FDA investigators noted the issue. The firm performed risk assessments for product failures without documenting how the severity of likelihood of occurrence used in the assessment was determined. The firm's CAPA procedure did not refer to electronic database systems used by GE to document, track and control CAPAs, and did not describe steps necessary to close CAPAs. The validation of equipment used in final device testing did not include a process capability challenge, did not ensure that the equipment was capable of capturing results at both the high and low ends of test specifications, and did not include challenges with known failures to ensure the equipment detected fault conditions. In addition, the validation was accepted despite the fact that computational errors were noted in the test procedure report. Personnel training was inadequate, as evidenced by the fact that supervisors and managers who signed off on test records did not detect that test technicians were using photocopied test results; additionally, errors noted in device history records were not detected by quality assurance personnel, and MDRs filed by the complaint coordinator were incomplete and not submitted on time. There was no evidence that numerous complaint records were evaluated for patient involvement or MDR reporting. Complaint investigations were not completed documented; for example, there were no records showing that device history records were reviewed, and some complaint records were missing related email correspondence, field reports and complaint investigators' notes. Device history records lacked signatures of employees who completed the work, and numerous DHRs did not include the required labeling. A review of DHRs indicated that the quality assurance unit did not consistently review and verify that rework was performed, and internal procedures did not require documentation of the re-release of systems for distribution. In addition, GE failed to submit an MDR report to FDA within the required 30-day timeframe, and failed to include all required patient information on two MDR reports. Finally, GE failed to report recalls related to *Axcess* and *Insta Trak 3500 Plus* systems. FDA judged some of the company's responses to the FDA-483 adequate and others inadequate, and ordered GE to promptly correct the quality system at the Salt Lake City facility and all other locations. [Issued by Denver District Office]
Gerald Grubbs, MD Warning letter date: April 11, 2005 Location: Venice, Fla. Inspection dates: Dec. 29, 2004-Jan. 21, 2005	Clinical investigator cited for violations of IDE and human subject protection regs. Grubbs transferred study records - including case report forms, informed consent forms and source documentation concerning residual stenosis and color Doppler ultrasounds - to another party for storage without obtaining the sponsor's approval, and could not produce the records during the inspection, apparently because they had been discarded, according to FDA. Grubbs also failed to notify the agency within 10 days after transferring custody of the records. In addition, there was no clinical investigator financial disclosure information available for review during the inspection. [Issued by CDRH]
St. Jude Medical Warning letter date: March 17, 2005 Location: Caguas, Puerto Rico Inspection dates: Sept. 27-Nov. 9, 2004	QS deficiencies noted at facility that manufactures Class III *Angio Seal* vascular closure devices. St. Jude did not adequately evaluate, investigate and address the cause of nonconformances in Angio Seal devices, and released batches of the device without adequate justification. In addition, numerous reject material reviews did not identify the cause of nonconformity results and failed to record the actions required for disposition. Procedures failed to clearly describe the necessary steps to evaluate and conduct nonconformance investigations. St. Jude also failed to monitor the polyfoil pouch sealing process, despite nonconformances. Further, the firm released batches of product that had out-of-specification results during acceptance tests; FDA asked St. Jude to furnish studies supporting the firm's conclusions regarding other batches that were tested and found within specification. There were no written procedures describing the methods and criteria for retests of nonconforming product and no procedures describing criteria used to sort units into an acceptable or reject group. Lots that failed the polyfoil peel force test were retested using different sample sizes and test methods, and disposition of similar nonconforming lots were based on different justifications and rationales. CAPA procedures did not clearly specify the criteria or circumstances necessary to trigger CAPA investigation and related corrective and preventive actions, and lacked provisions for performing a risk analysis during investigations of quality problems. Additionally, CAPA investigations lacked due dates for completion, and were not opened in a timely manner. FDA asked the firm to ensure that any CAPA opened at St. Jude's Minnetonka, Minn. facility affecting products made at St. Jude Medical Puerto Rico are promptly referred to the Puerto Rico plant for evaluation. Finally, complaint data was not routinely analyzed for trends of potential quality problems to determine if a formal investigation and/or CAPA is necessary; 15 complaints related to Angio Seal-STS platforms were not thoroughly investigated to identify the probable root cause; and an investigation or formal study addressing problems in the Angio Seal STS platform was not available. St. Jude's response to the FDA-483 was found deficient. [Issued by San Juan District Office]
Inoveon Corporation Warning letter date: April 27, 2005 Location: Oklahoma City, Okla. Inspection dates: Dec. 7-21, 2004	Manufacturer of the Diabetic Retinopathy *3DT* System cited for QS reg violations, many relating to the firm's handling of software problems. In the design control area, Inoveon did not document a formal risk analysis of the original system design and software changes to correct software bugs; did not document how design reviews were linked to software release notes and did not indicate who attended design reviews or the dates of review; did not explain the impact of unresolved or known software bugs; and did not document design changes to explain why certain changes were not implemented to correct software bugs. In addition, the firm was not able to explain what changes had been made to the device design validation document when the document version was changed. Design change testing to release a new software version did not clearly explain the exact functionality or performance problems detected during testing and the impact of these problems, and a workaround to resolve a bug in the new software version was not clearly explained. The company failed to establish and maintain procedures for analyzing all sources of quality data to identify causes of nonconforming product; for example, tracking of product issues did not clearly identify whether the issues were related to customer complaints or internal nonconformances detected by software engineers; there were no procedures clearly explaining how nonconforming items are resolved; and the specific nature of trended product defects and nonconformities, and their impact on product functionality or performance, were not clearly explained. In addition, Inoveon failed to document whether an alternative algorithm, intended to correct a software bug in the 3DT system, was implemented; the firm also did not justify reasons for not implementing the algorithm. Signatures needed to release several 3DT software versions for distribution were not documented, records on device installation at clinical sites were not maintained, there was no device master record index identifying all the necessary specs and procedures for the three subsystems of the 3DT system, and there were no procedures defining the information and acceptance records to be included in device history records. Inoveon did not evaluate suppliers for their ability to meet the firm's requirements, did not define the quality requirements each supplier must meet, and did not define the frequency of supplier evaluations or document such evaluations. The firm had conducted no internal quality audits in the previous five years. Finally, the firm failed to develop, maintain and implement written MDR procedures. FDA acknowledged receipt of a letter from the firm indicating that the firm will correct the FDA-483 observations, outline a general corrective action plan to be completed by May 2005, and complete a comprehensive, baseline third-party audit of the quality system by June 2005. The agency noted that Inoveon's response was incomplete until the firm completes all corrective actions and provides FDA with the auditor certification report upon completion of the baseline quality system audit. [Issued by Dallas District Office]
Kenneth Rosenthal, MD Warning letter date: April 11, 2005 Location: Great Neck, NY Inspection dates: Dec. 29, 2004-Jan. 19, 2005	Clinical investigator associated with Rosenthal Eye and Facial Plastic Surgery cited for IDE reg violations. During the inspection, Rosenthal did not provide records documenting the disposition of study devices; FDA noted that this was a repeat violation from a 2001 inspection for which Rosenthal received an FDA untitled letter. In another repeat violation, study subjects' records contained numerous inaccuracies and inconsistencies. For example, a monitoring report noted that the case report forms for nine study subjects had inconsistent and discrepant data as compared to the source record for medications, complications and other information; the monitoring report also noted that clinical records were not available to verify specific study visit information reported on case report forms for four subjects. Rosenthal did not perform several study procedures required by the protocol, and did not send the reviewing IRB timely progress reports, despite several reminders. FDA requested a complete list of all clinical trials in which Rosenthal has participated in the last five years. The agency also requested information on the electronic medical record system used by Rosenthal to maintain clinical data for study subjects; specifically, FDA requested documentation related to the system's compliance with the Part 11 electronic records and signatures regulation. [Issued by CDRH]
Charles Winters, MD Warning letter date: May 1, 2005 Location: Baltimore, Md. Inspection dates: Dec. 6-8 & 14, 2004 and Jan. 11, 2005	Clinical investigator cited for violation of IDE and human subject protection regs. Winters did not obtain informed consent or screening consent from some subjects before procedures were performed, and failed to conduct a procedure within the timeframes required in the study protocol. Case history documents for all subjects were incomplete, not maintained, or non-existent; there were no progress notes or previous medical histories of the subjects, and diagnostic test results were not available for review. Finally, records related to the receipt, use and disposition of study devices were not present at the clinical site, and it appeared that the sponsor's reps completed and maintained these records. FDA could not determine when the product disposition log was completed, and could not judge the log's accuracy or whether maintenance of the log was delegated by Winters to the sponsor's reps. The clinical investigator's responses to the FDA-483 were deemed inadequate. FDA asked Winters to provide a list of his current investigational studies. [Issued by CDRH]
Sanjay Yadav, MD Warning letter date: April 11, 2005 Location: Cleveland, Ohio Inspection dates: Jan. 6-18, 2005	Clinical investigator cited for violations of IDE and human subject protection regs. Yadav failed to report to the IRB the use of non-investigational stents in several subjects; use of the stents was a protocol deviation. Yadav also deviated from the protocol by treating a subject who met exclusion criteria, and failed to report the deviation to the IRB and sponsor. The investigator failed to ensure that subjects met inclusion/exclusion criteria prior to study enrollment, and failed to conduct diagnostic evaluations, including pre-procedure coagulation tests and neurological examination stroke scales, in accordance with the investigational plan. In addition, Yadav failed to administer protocol-required heparin during the carotid stent implantation procedure in one subject. He also failed to report to the sponsor major adverse device events, including a seizure, a myocardial infarction, an embolic stroke and a neurological event, within 24 hours of learning of the events. [Issued by CDRH]

In Brief

New president for F-D-C Reports: Former Medscape pharmaceutical sales & client relations VP Mike Squires took the helm at the Chevy Chase, Md.-based publisher of "The Silver Sheet" May 2. Squires departed Medscape (now owned by WebMD) in 2002, following five years of service, to serve as a pharmaceutical marketing consultant to companies including Elsevier/International Medical News Group. His news career began with the advertising sales department for the Chicago Sun-Times and Chicago Daily News, where he rose to VP & East Coast manager. He also directed North American sales operations for the Times of London and Sunday Times. "I look forward to listening to our readers and customers, learning from our expert reporting and publishing teams, and leading our news gathering and distribution operations forward," he stated May 6....

DPC foresees near-term AIP removal: FDA accepted Diagnostic Products Corp.'s corrective action plan submitted in response to an application integrity policy issued in February 2004, CEO Michael Ziering reported during a May 10 earnings call. Ziering expects FDA to conduct its verification inspection, required before an AIP can be lifted, by mid-July. The policy, which halted scientific review of all FDA submissions from the company, was issued based on discrepancies between the findings of a Nov. 10-Dec. 30, 2003 inspection and information presented two months later in DPC's premarket submission for a Chagas disease diagnostics product ("The Silver Sheet" May 2004)....

Vanguard can resume reprocessing trocar devices: Lakeland, Fla.-based firm receives 510(k) clearance for reprocessed dilating tip and blunt trocar April 8. The product was removed from the market last fall following MDUFMA-mandated requests from FDA for additional data on cleaning, sterility and functionality validation to determine substantial equivalence to predicate devices ("The Silver Sheet" December 2004). The trocar is intended to provide a pathway for entry of instruments during minimally invasive surgery. OEM versions of the device include Ethicon's Tristar blunt tip trocar, Endopath dilating tip trocar and Endopath Optiview optical trocar....

Medical device software guidance: FDA's "Guidance for the Content of Premarket Submissions for Software Contained in Medical Devices," issued May 11, updates a May 1998 document on device software and January 1997 guidelines for blood establishment computer software. It recommends that firms describe in their submissions the role of software in "causing, controlling, and/or mitigating hazards that could result in injury to the patient or operator." The agency outlines criteria for major, moderate and minor levels of concern and explains risk-based regulatory recommendations for each level. Device components of combination products are within the guidance's scope....

Class I recalls: BioMérieux's Durham, N.C. facility revised its characterization of a March 2005 recall of tissue thromboplastin reagents, acknowledging in a May 14 release that the problem rises to the level of clinical significance. The amended alert, affecting 236 U.S. customers and more worldwide, was released "because mislabeling problems with certain lots of this product could result in inaccurate test results, which in turn could lead to improper patient treatment - in some cases resulting in serious or life-threatening injury," the firm said....Glucose meters: J&J/LifeScan's worldwide notification regarding unit of measure-based misinterpretation of glucose meter results was categorized as Class I, according to a May 17 FDA release ("The Silver Sheet" April 2005)....AED short circuit:Welch Allyn/MRL's AED20 automatic external defibrillator action was initiated April 28 following a complaint of malfunction during clinical use that resulted in a death in the U.S. The malfunction reportedly occurs when impact to the exterior of the AED20 causes an internal electric short, resulting in failure of the device to analyze the patient's heart function. MRL, Inc., a Welch Allyn division, recalled 597 products....AED adaptor cables: Laerdal Medical Corp. initiated a recall April 29 of more than 3,000 products after several user deaths. The cables, which allow the use of Laerdal's HeartStart multifunctional defibrillator pads with various Laerdal and Philips defibrillators, are susceptible to breakage, which in some cases might prevent the AEDs from delivering a shock....

Firms Report Mixed Experiences In Hosting AP Training Inspections

More from Archive

More from Medtech Insight