;, . w United States General Accounting Office Report to the Chairman, Subcommittee ‘*-r on Health and the Environment, . Committee on Energy and Comrkerce, House of Representatives May 1990 MEDICAL DEVICES Underreporting of Serious Problems With a Home Apnea Monitor United States GAO General Accounting Office Washington, D.C. 20548 Program Evaluation and Methodology Division B-237533 May 31,199O The Honorable Henry Waxman Chairman, Subcommittee on Health and the Environment Committee on Energy and Commerce House of Representatives Dear Mr. Chairman: In February 1989, we reported to the Subcommittee that our review of the implementation of the medical device reporting regulation had found evidence that some medical device manufacturers may have been over- reporting problems with devices, while others either were not reporting at all or were underreporting.’ In its comments on the report, the Food and Drug Administration (FDA) said that our conclusion that the indus- try was underreporting was “questionable” and that FDA’s medical device reporting regulation compliance inspection strategy was suffi- cient to identify compliance problems. Since the release of our report, we have received additional information from several sources that sug- gests that problems are underreported and that underreporting is not always identified through FDA'S inspection program.’ On September 18, 1989, you asked us to investigate a citizen’s report to the General Accounting Office (GAO) of numerous unreported deaths of patients associated with the Aequitron Medical, Inc., Model 8200 home apnea monitor and to include our investigation in our ongoing review of FDA'S postmarketing surveillance of medical devices for the Subcommit- tee. With the concurrence of the Subcommittee staff, we undertook a case study based on three specific questions: (1) How many complaints involving the death of patients have been associated with the Model 8200 apnea monitor? (2) Did the device manufacturer comply with FDA'S existing problem-reporting regulations and procedures? (3) When FDA received information from the device manufacturer or other sources that Model 8200 had been associated with numerous deaths, what actions did FDA take in response to that information? This letter presents our findings, conclusions, and recommendations. ‘See IJ.S. General Accountmg Office, Medical Devices: FDA’s Implementation of the Medical Device Reporting Regulation, GAO/PEMD-8%10 (Washington, D.C.: February 1989), p. 3. ‘The sources included citizen reports, device industry publications, consultation with members of our expert review panel, and rewew of indwidual recalls in connection with our earlier report entitled Medud Dewce Recalls: Exammation of Selected Cases, GAO/PEMD-90-6 (Washington, D.C.: October 1989). Page 1 GAO/PEMD-90.17 Underreporting Home Apnea Monitor Problems 5237533 overall efficacy of the FDA medical device reporting regulation compli- ance program and, in particular, the finding that a number of FDA inspections had discovered instances in which reportable serious inju- ries and deaths had been recorded in a manufact,urer’s files but not reported to the agency. This study of an apnea monitor is consistent with our earlier finding of differences in the interpretation of medical device reporting require- ments.’ In this case, a manufacturer’s interpretation of the requirements has resulted in the underreporting of serious problems associated with its device. It also illustrates a weakness in the compliance inspection process. Although the manufacturer had been the subject of several inspections, it was nearly 4 years after the medical device reporting reg- ulation went into effect before FDA'S inspection program identified and attempted to resolve the underreporting. ___. __- Medical devices include almost everything, other than drugs, that Background health-care professionals use to diagnose, treat, or prevent illness, improve human functioning, and support and sustain life.” More than 1,700 different types of medical devices are available in the United States today. They represent an industry of more than $14 billion a year. FDA is authorized to regulate medical devices during all phases of their development, testing, production, distribution, and use. FDA has identified the apnea monitor as a “critical device.” Critical devices are intended for surgical implant into the body or to support or sustain life. Their failure to perform when used properly in accordance with instructions provided in the labeling can be reasonably expected to result in a significant injury to the user. “Medical Devices, p. 4. “Section 201(h) of the Federal Food. Drug, and Cosmetic ACT of 1938, w amended by the Medical Device Amendments of 1976, defines “drvice” as an instrument, apparatus, implement, machine, con- trivance, implant, in vittw rragvnt, or other similar or related art~rlc, inchrding any compment, part, or accessory, that is (1) rwogmzed m t.hr official National Formulxy or the 1J.S. Pharmacopeia or any supplement to them, (2) mntendtsd for USCin the diagnosis of disease or other conditions or in the cure, mitigation, treatment, or prevention of disease, in humans or other animals, or (3) intended to affect the structure or any function of thr human body or bodies of other animals, and that dws not achieve any of its principal intended pwposcs through chrmical action within or on the body and does not depend upon being metabolized in order to achieve any of its principal intended purposes The effect of the amendments was to enlarge the 1938 definitmn of “device” to include (1) devices intended for use in the diagnosis of concbtions other than diwasc. wrh as pregnancy, (2) m vitro diagnostic prod- ucts, and (3) specific productr prcv~wsly regulated as new drugs, including soft contxt 1~~s. bone cement, and wltures Page 3 GAO/PEMLNO-17 Underreporting Home Apnea Monitor Problems FL237533 Aequitron Medical, Inc., Model 8200 home apnea monitor. The table con- tained 82 complaints with four categories of information for each com- plaint.!’ (The table is reproduced in appendix I.) Our investigation of the origin and contents of the HID file determined it to be a list of complaints derived from the device manufacturer’s “haz- ard, injury, or death” GMP record. The list contained complaints that the device manufacturer had received about its Model 8200 apnea monitor dated between January 1983 and January 1989.1” Our analysis of the HID file showed that it listed 68 complaints in which the word “death” was included in the category of reasons for the device’s return to the manufacturer. Further research found that a simi- lar list drawn from the manufacturer’s general complaint record con- tained 2 additional complaints in which the word “death” was included. There were therefore a total of 70 complaints in which the allegation of a patient’s death was included in the complaint description11 The remaining 14 complaints in the HID file were complaints that included allegations of either hazards to safety or injuries. (See table 1. The remainder of the table is discussed below.) Table 1: Complaints and Medical Device Reporting Regulation Reports on Aequitron Medical, Inc., Model 6200 Apnea Monitor Medical device reporting regulation reports to FDA Complaints in Aequitron’s files Before FDA’s May After FDA’s May In “H/l/Death File” In “general” file Total 1968 inspection 1966 inspection Total Deaths 68 2 70 5 2 -- 7 Nondeaths 14 66 60 78 193 271 Total 62 66 150 63 195’ 276 a150 of these reports were submltled in response to FDA compliance actlons “The four data categories for each complaint were serial number, date, reason for return, and analy- sis. It is important to note that with regard to the “analysis” category, the general limitations of the technology employed in apnea monitors or the specific limitations of the design of a particular model may cause a monitor to fail to detect apnea events in some circumstances. Such an occurrence is known as a “false negative.” If this happens, later testing of the monitor would not necessarily reveal that a component had malfunctioned, and the device could be found to be “within specifications.” “‘Model 8200 was introduced Into the market m June 1982. According to the manufacturer, approxi- mately 30,000 of the monitors were distributed between 1982 and 1987, but it is not possible to cstrnate the number in actual II.% or their frequency of use. ’ ‘The manufacturer confimwd I hat these were complaints alleging a death had been associated with the use of the device. Page 6 GAO/PEMD-90-17 Underrepmthg Home Apnea Monitor Problems B-237533 record, FDA found that 10 unreported complaints should have been reported, including 4 that alleged the death of patientsl:l It is important to note that not all incidents in which a device is associ- ated with the death of a patient must necessarily be reported to FDA under the medical device reporting regulation. The regulation requires reporting of incidents to FDA only if the information in the possession of the manufacturer “reasonably suggests” that a device may have caused or contributed to a death or a serious injury. If a health care profes- sional states that this has happened, then the manufacturer is required to file a medical device reportt.14But in the case of a report from a layperson, if an immediate investigation by the manufacturer reveals that a patient was not connected to an apnea monitor at the time of death, or that the monitor’s alarm sounded and the caregiver was alerted even though the patient could not be revived, then a report might not be required. FDA has characterized the circumstances in which home apnea monitors are used and the limitations of the technology they employ as sometimes making it difficult to determine whether a problem is reportable under the medical device reporting regulation. - Question 3: FDA’s Actions When FDA received information from the device manufacturer or other sources that Model 8200 had been associated with numerous deaths, what actions did FDA t,ake in response to that information? One of the principal tools of FDA'S postmarketing surveillance of medical devices is biennial inspections for compliance with the GMP regulation. In addition, FDA conducts “for cause” inspections when they are warranted by com- plaints or other evidence of problems with devices.‘” FDA assesses device manufacturers’ compliance with the medical device reporting regulation by executing a special medical device reporting inspection program as “‘A more da&d discusswn of Ihls rewew IS contained in the following section of this report on FDA’s acrions. “The “per se” reporting rule states that whenever a health care professional advises a manufacturer that one of its devices may have caused or contributed to a serious iqjjury or death, the manufacturer is “per se” m receipt of information that “reasonably suggests” that a device may have caused or contributed to a serious injury or death, and it therefore must report the event. It does not, however, imply that reports from perx~ns other than health care professionals are not reportable. “A pt-inc~pal rationale for “for GWX” inspections is information developed by FDA analysts who monitor and compare reports submitted through FDA’s voluntary problem-reportii program, device recalls, and the medical deww reporting system. Page 7 GAO/PEMDCW-17 Undermpmting Home Apnea Monitor Problem 5237633 “per se” provision of the medical device reporting regulation and com- pany policy, Aequitron did not submit these types of reports. In our earlier study of the implementation of the medical device report- ing regulation, we reported that the evidence suggested an undeter- mined amount of overreporting by some device manufacturers and that others were either not reporting or underreporting.lH The most fre- quently identified dimension of noncompliance noted by FDA inspectors was failure to establish adequate procedures for handling complaints to determine their reportability (20 percent of all such citations for the first series of medical device reporting regulation compliance inspec- tions and 53 percent for the second series).“’ We also encountered varia- tions in the interpretation of reporting requirements among the FDA officials and staff we interviewed. As a result of the October 1988 notice-of-adverse-findings letter and negotiations between FDA officials and the apnea monitor manufacturer, the manufacturer agreed to review its complaint records and revise its medical device reporting policy. Subsequently, Aequitron submit,ted medical device reports on 6 of the 10 incidents listed in the notice-of- adverse-findings letter. including 2 of the complaints in the HID file involving deaths. These 2 reports of death were submitted 1 year and 9 months, respectively, after the events they described. The manufacturer also submitted 144 reports of malfunctions involving confirmed alarm failuresYil Aequitron thus submitted at least 150 medical device reports in response to FDA compliance actions. (See table 1.) Aequitron also submitted a revised medical device reporting policy to FDA, and FDA notified the manufacturer that its revised policy was ade- quate. We found certain aspects of the revised reporting policy to be inconsistent with the medical device reporting regulation. Specifically, we believe that it is improper to condition the reporting criteria on the confirmation or observation of a malfunction. (Our complete analysis is contained in appendix III.) %e Medical Devices, p.61. “‘See Medical Devices, p. 59. ‘“Apnea monitor alarms meet FDA‘s definition of a “critical device component”-that is, any rompo- nent of a critical device whose failure to perform can be reasonably expected to czuw the fadurr of a critical device or to affed its safety or effectiveness Page 9 GAO/PEMD-99-17 Underreporting Home Apnea Monitor Problems B-237533 FDA did not provide evidence that its inspection procedures included any systematic evaluation of trends in the frequency, type, or severity of complaints involving the failure of alarms that were made to the manu- facturer or reported under the medical device reporting regulation. There were also no comparisons of overall complaint rates of alarm fail- ures of Aequitron’s Model 8200 to those of other monitors. The inclusion of these procedures in the inspection program could have served to more quickly identify both problems of underreporting and potentially seri- ous device problems. Conclusions, Recommendations, and Agency Comments Conclusions We conclude that the evidence from the current case of an apnea moni- tor is consistent with our earlier finding that there are differences in the interpretation of medical device reporting requirements among device manufacturers and between manufacturers and FDA.” In this case, the manufacturer’s interpretation resulted in an undetermined amount of underreporting of serious problems with a device. We found that FDA'S review of a sample of Aequitron’s records deter- mined that some complaints alleging that the device was associated with hazards to safety, injuries, or death had not, been reported to the agency because of the manufacturer’s interpretation of the medical device reporting requirements. As a result of FDA'S intervention in this case and the device manufac- turer’s review of its o\vn records, 144 additional medical device reports of malfunctions were submitted. We believe that malfunction reports should be considered as seriously as reports of serious injury or the death of a patient, especially as a preventive measure. Malfunction reports describe problrm occurrences that were not associated with the injury or death of a pat,ient. Ilowcvcr. if thr problem should recur, it may result in injury or even death. Page 11 GAO’PEMLWO-17 Underreporting Home Apnea Monitor Problems B-237633 and consider developing specific guidance for the manufacturers of any devices, such as apnea monitors, for which determining reportability is thought to present special difficulties. Agency Comments and Our The Department of Health and Human Services (HHS) agreed with our recommendations and reported that FDA had taken actions consistent Response with them. FDA conducted a comprehensive good manufacturing prac- tices and medical device reporting regulation inspection of Aequitron and is currently reviewing the results to determine what regulatory action is warranted. The agency is also considering the development of device-specific guidance on medical device reporting and has established a study group to analyze the medical device reporting regulation and program, identify problems, and recommend solutions. HHS'S general comments indicated a concern that this case study is not an accurate representation of the way the medical device reporting reg- ulation is implemented by most device manufacturers and that its find- ings cannot be generalized to the program as a whole. We agree that the case does not by itself support generalizations about the implementation of the medical device reporting regulation. The general statements in the report about the implementation of the medical device reporting regula- tion and its associated compliance inspection program are based on our earlier, general study of these issues. The case does, however, illustrate the principal findings of that report regarding compliance with the med- ical device reporting regulation. HHS'S comments, along with our detailed responses, are reproduced in appendix IV. HHS also provided technical comments on the report. We reviewed them and made changes in the report as appropriate. In one technical com- ment worth noting, HHS agreed with our opinion in appendix III, acknowledging that FDA'S Center for Devices and Radiological Health made an error in allowing Aequitron to condition its reporting of com- plaints on the confirmation or observation of a malfunction. HHS stated that the firm was notified of the correct interpretation at the time of the inspection referred to above. This review was a follow-up to our earlier general overview of FDA'S Objective, Scope, and implementation of the medical devices reporting regulation. The objec- Methodology tive of this study was to conduct a case study of medical device problem reporting and FDA act,ions related to the Aequitron Medical, Inc., Model Page 13 GAO/PEMKNC-17 Underreporting Home Apnea Monitor Problems Page 15 Contents Abbreviations FDA Food and Drug Administration GAO General Accounting Office GMP Good Manufacturing Practices 1111s Department of Health and Human Services HID Hazard, Injury, Death Page 17 GAO/PEMD-SO.17 Underrepmting Home Apnea Monitor Problems Appendix I H/I/Death Pile Number Serial number Date Reason/return Analysis 40 110139 12/86 Checkout only-death Dewce rn-spec 41 204165 12/86 Checkout death-no-alarm Devrce In-spec .__~_ 42 108272 l/87 Checkout only-death Devrce rn-spec 43 204768 l/87 Checkout only~death Devrce rn-spec 44 54319 1187 Checkout o&death Devrce In-soec - 45 101216 4/87 Checkout only~death Devrce In-spec 46 110870 4187 Checkout death-no alarm Devrce in-soec 47 108197 4/87 Checkout only-death Devrce In-spec 48 105515 6/87 Checkout only~death Devrce in-spec 49 63510 6/87 Checkout only-death Device In-spec 50 54121 8/87 Checkout only-death Devrce rn-spec 51 106034 6/87 No 10 set apnea alarm Dewce rn-spec 52 202394 lo/87 Checkout only-death Devrce In-spec 53 60184 lo/87 Checkout death-no apnea alarm Devrce In-spec 54 66740 12/87h~o~nly-death- Devrce In-spec 55 110447 12/87 Checkout only-death Dewce In-spec ____ 56 63934 l/88 Checkout-death-no alarm Devrce fully functconal 57 111579 7/88 Checkout death No other details avail __- 58 104348 S/88 Checkout only-death Devrce rn-spec 59 103552 l/88 Checkout only~death Devrce In-spec 60 50657 1i88 Checkout onlv-death Device In-scec 61 204894 l/88 Checkout only~death .__-_-~___ Device rn spec 62 104635 l/88 Checkout only-death Devrce in-spec 63 52877 2/88 Checkout only~death- Dewce In-spec 64 201228 3/88 Checkout only-death Dewce fully functronal -.__ 65 66086 3188 Checkout death-no audible alarm Devrce In-sbec 66 62873 3/88 Checkout only-death Devrce In-spec 67 50903 3/88 Checkout onlvdeath Device rn-scec 68 54569 3/88 Checkout death-no alarm Devrce fully functronal_- 69 250197 4/88 Checkout only-death Device rn-spec 70 205002 l/88 Checkout only-no apnea alarm Devrce rn-spec 71 108359 l/88 Checkout only~no brady alarm Devtce fully functronal 72 250215 -L 6/88.-~_. Checkout onlv-death Device In-saec 73 202134 6/88 Checkout death-no alarm Devrce rn-spec 74 205881 8/87 Checkout only-death Devrce crave constant_- al 75 110094 6/88 Checkout only-death Device not recerved 76 201785 5/88 Checkout only-no alarm Device not received 77 66250 11/88 Checkout onlv-death - ~___ Devrce rn-soec 78 201785 1 l/88 No apnea alarm-pt Ok Devrce In-spec 79 51130 l/69 Checkout only-death not rn use Dewce In-spec 80 109126 1 l/f38 Checkout no audible alarm -~.__ ~ -~ -~ In-spec Devrce (contrnued) Page 19 Appendix 11 Chronology of Contacts Between FDA and Aequitron This chronology has been adapted from one supplied to us by FDA'S Center for Devices and Radiological Health during our data collection. We did not use every entry we were given but selected items from the original chronology that were germane to the subject of this report. August 18,1984 $71~conducted a GMPinspection of the manufacturer, covering the manu- facturing of Model 8200. No significant deviations were noted. December 19,1984 FDA conducted a limited inspection of the manufacturer in response to a complaint of the failure of Model 8200 to sound its alarm and a device distributor’s complaint of overheating wires in the monitor. After 10,000 units were distributed, the manufacturer received 91 units returned with burned wires as well as 17 returned with problems relat- ing to the alarm. FDA inspected the device distributor to follow up on a complaint about a Model 8200 monitor that delayed in sounding its alarm, although the light came on. Follow-up revealed that Model 8200 was designed with a 6-second delay. The distributor had approval from the manufacturer to eliminate the delay by replacing the switch if the customer so requested. During this inspection, the distributor identified a problem with “burn out” or “melting” wires with 6 of approximately 300-350 units. The manufacturer had repeatedly attributed the problem with the burned wires to user intervention and not design. However, an indepen- dent engineering firm hired by the distributor believed the problem was design-related. The distributor was not aware of any serious injuries or deaths relating to “burn out,” and the manufacturer disputed the find- ings of the engineering firm. - December 28, 1984 FDA conducted a limited inspection of the device distributor to collect recall data on 30 Model 8200 monitors that the distributor had replaced in home visits and to collect a failed unit for laboratory analysis. January 4,1985 An FDAdistrict office submitted a recall recommendation to FDA’SCenter for Devices and Radiological Health for the 30 Model 8200 monitors that the device distributor had replaced. The recommendation noted that there was much confusion regarding the “burn out” problem and specifi- cally requested a technical evaluation concerning the validity of the Page 21 GAO ‘PEMD-90-17 Underreporting Home Apnea Monitor Problems Appendix n Chronology of Contacts Between FDA and Aequitron that the action constituted a “Safety Alert” and that FDA testing con- firmed the information presented by the manufacturer. The manufac- turer agreed to submit a 510(k) application to add a circuit breaker to the device to resolve the “burn out” problem. June 25,1985 The manufacturer sent. a letter to the distributors informing them about the “burn out” problem. July 25, 1985 FDA classified the manufacturer’s .June 25 letter as a “Safety Alert.” August 7,1985 In its Weekly Enforcement Report, FDA published information on the manufacturer’s “Safety Alert.” December 26,1985 The manufacturer sent a letter to FDA that requested a “Certificate of Export” for five models of monitors of various kinds, including Aequi- tron’s Model 8200. February 3,1986 FDAsent a letter to the manufacturer approving the export of four of the monitors but denying the manufacturer’s December 26 request to export Model 8200. February 11,1986 Responding to FDA'S denial of a request for export, the manufacturer claimed it was in compliance and requested that the export of Model 8200 be approved. March 4, 1986 FDA approved the export of Model 8200. January 27,1987 FDA conducted a GM? inspection and reviewed all injury and death com- plaints received by t,hc manufacturer since 1985. FDA'S inspector ques- tioned whether two of the complaints on the Model 8200 monitor should have been reported under the medical device reporting regulation.’ ‘This inspection was not m~-lr~Ivd m t hc chronology provide by FDA but was documented m a sepa- rat? mspctction report Page 23 GAO/PEMD-SO-17 Underreporting Home Apnea Monitor Problems Appendix II Chronology of Contacts Between FDA and Aequitron adverse-findings letter be issued for the manufacturer’s failure to sub- mit 10 medical device reporting regulation reports for Model 8200. May 20,1988 FDA conducted an inspection in response to an anonymous letter alleging that six Model 9216 apnea monitors, intended for hospital use, had caught fire. The inspection confirmed that the failure resulted from a transistor that had been used in a circuit application outside its electri- cal specifications. June 6,1988 The manufacturer wrote to inform FDA that it had instituted a field replacement program and a labeling change for all the Model 9216 hos- pital apnea monitors. June 15,1988 FDA visited the manufacturer and collected recall data. July 18,1988 FDA classified the manufacturer’s field replacement and labeling change for Model 9216 as a Class II recall. October 18,1988 In response to the deficiencies in medical device reporting found in the May 1988 inspection, FDA issued a notice-of-adverse-findings letter to the manufacturer. FDA also conducted an inspection of the manufacturer in response to five complaints received by FDA on Model 8200.2 December 28, 1988 The manufacturer responded to the October 1988 notice-of-adverse- findings letter, stating that it would change its reporting policies and practices. June 22,1989 FDA notified the manufacturer by letter that it appeared to have ade- quately addressed the concerns listed in October 1988 notice-of-adverse- findings letter and that a review of the manufacturer’s revised medical device reporting regulation policy found that the reporting guidelines appeared to be adequate. ‘This inspection was not included in the chronology provided by FDA but was documented in a sepa- rate inspection report. Page 25 GAO/PEMB9&17 Underreporting Home Apnea Monitor Problems Appendix III Our Analysis of Aequitron’s Medical Device Reporting Criteria a manufacturer receives or becomes aware of information from a layperson that its device may have caused or contributed to serious injury or death, the medical device reporting regulation requires the manufacturer to determine whether the information “reasonably sug- gests” a link between the operation of the device and the injury or death. Information from a layperson reasonably suggests the link if a reasonable person would reach that conclusion. Thus, if a reasonable person would conclude that the device may have caused or contributed to serious injury or death, a report must be made. The regulation does not require, expressly or by implication, that a mal- function need occur or be suspected in order for a reasonable person to conclude that the device is implicated in serious injury or death. A rea- sonable person may conclude that the device may in some way have caused or contributed to serious injury or death, thus necessitating a report by the manufacturer, even when serious injury or death is attrib- utable to user error, poor maintenance, or the use of the device beyond its useful life. FDA'Srationale for this approach is explained in guidance accompanying the publication of the final regulation in the Federal Register: “One comment asserts that a death or serious injury should not be required to be reported unless it is associated with or related to a device malfunction. “FDA disagrees with the comment. A device that performs to its specifications or otherwise performs as intended does not ‘malfunction’ as defined in the final rule. However, because of flaws m its labeling or because of user error , such a device could cause or contribute to a serious injury or death. “In this context, the phrase ‘contribute to’ means to play a part in the serious injury or death.” (49 Fed. Reg. 36.330 (1984); see also 49 Fed. Reg. 36,338 (1984)) FDA believes that it needs to be notified of deaths resulting from user error and problem labeling because these could indicate the need for educational programs, user notification, voluntary recalls, or corrective labeling to prevent future deaths or serious injuries. (49 Fed. Reg. 36,331.332, 36,339 (1984)) Page 27 GAO/PEMD-90-17 Underreporting Home Apnea Monitor Problems Appendix JII Our Analysis of Aequitron’s Medical Device Reporting Criteria a report to FLIAwould be required if the information given to the com- pany suggests to a reasonable person that a recurrence of a malfunction would likely cause or contribute to a future serious injury or death. In the second, the per se rule would require a report. As explained above, neither confirmation nor observation of a malfunction is required before a company may be required to report to FDA. Page 29 GAO/PEMD-9017 Underreporting Home Apnea Monitor Pmblems Appendix IV Comments From the Department of Health and Human Services - COMMENTS OF THE DEPARTMENT OF HEALTH AND HUMAN SERVICES ON THE GENERAL ACCOUNTING OFFICE (GAO) DRAFT REPORT ENTITLED "MEDICAL DEVICES: UNDERREPORTING OF SERIOUSl9PgROOBLBNSWITH A HOME APNEA MONITOR," MARCH 9, Although we agree that implementation of the Medical Devices Reporting (MDR) Regulations has been slower than we would like and the regulations are subject to interpretation, we disagree with the following conclusions of the draft report. 1. Although the report states that the study design precludes generalization to the medical devices industry, it proceeds to make statements such as "...there are persistent differences in the interpretation of the medical device reporting requirements,..." and "...we have received additional information from several sources which suggested that problem underreporting does occur and that it is not adequately identified through FDA's inspection program." We believe these and similar generalizations without apparent foundation should be deleted or sufficient data should be presented to substantiate the allegations and allow the Food and Drug Administration (FDA) to evaluate their validity in the context of the overall MDR program and competing priorities. We believe that the case study of apnea monitors is not an accurate representation of how MDR is implemented by most firms and the report is skewed, setting forth a worst- case scenario rather than a balanced presentation of NDR implementation. The report characterizes implementation of MDR as "flawed" based upon this one case. Yet, FDA has received thousands of reports from other firms that have led to regulatory actions such as recalls, seizures, and injunctions, as well as safety alerts, bulletins and even withdrawals of premarket applications (PNA) and premarket notifications (510(k)). MDR is primarily intended to elicit reporting from health care professionals who generally are in a position to observe product performance first hand and be able to appropriately assess the reportability of an event. Unlike most other devices, however, apnea monitors are prescription devices used in home care settings where health care professionals are unlikely to observe use failures and problems first-hand. Many reports are, therefore, made by nonprofessionals. Evaluation of data on apnea monitors both by the manufacturer and FDA is, therefore, more complicated than for most other devices. 2. The report faults FDA inspection strategy for failing to identify serious problems and for failing to recognize that the firm was misinterpreting the regulations, thus resulting in underreporting. Page 31 (;AO/PEMD-90.17UnderreportingHomeApnea MonitorProblems AppendixIV CommentsFromtheDepartmentofHealth andHumanServices 3 incorrectly created a presumption against reporting unless an event was confirmed. However, the preamble to the NDR regulation indicates that while mere allegation is insufficient, confirmation is unnecessary. Aequitron's criteria were also worded so as to make the time frames work against reporting. In fact, the presumption should be to report unless testing shows, within the time frames, that there is no reasonable suggestion there has been a malfunction. GAO also correctly notes that the Aequitron criteria Suggest that a health care professional must observe a malfunction before he/she reports it (unless tests confirm the malfunction). However, the rule that an event is per se reportable is invoked when the health care professional reoorts the malfunction even if he/she has not observed it. It should also be noted that in its response to FDA's Inspectional Observations (Form FDA-483) given to the firm at the conclusion of a full GNP inspection from September 1989 through November 1989, Aequitron changed its policy. Aequitron stated: "At this point, Aequitron wishes to advise FDA that the company has elected voluntarily to report a incidents of death or serious injury that the company becomes aware of as NDR events to FDA whether justified or not. . . . This action was taken by Aequitron voluntarily to avoid any debate pertaining to an appropriate construction and interpretation of the NDR regulation." This action by Aequitron was precipitated by FDA's Minneapolis District’s discovery of Aequitron's error in interpretation of the NDR as stated in their reporting criteria. As a result of Aequitron's change in policy, the firm has been submitting significantly more NDR reports than had previously been the case. FDA will continue its normal postmarket surveillance efforts of reviewing every NDR report that is submitted and determining reporting rates. If a significant fluctuation from the normal reporting occurs, FDA will follow up to determine the cause. By following these procedures FDA will be able to immediately determine if the firm deviates from the course of action it had committed to follow. Finally, it should be noted that in the Fiscal Year 1991 budget proposal the Administration has requested an increase of eight full time equivalents and $600,000 for the NDR program. Our comments on the recommendations are as follows. page33 GAO/PEMD-L)O-l7Underreportin~HomeApneaMonitorProblems Appendix IV Comments From the Department of Health and Human Services The following are GAO'S comments on the HHS April 27, 1990, letter IIHSprovided three types of comments on our report-general com- ments, comments on our recommendations, and technical comments. Our responses to the first two are contained in this appendix. We have responded to the technical comments by revising the body of the report as appropriate. agreed with our recommendation that a comprehensive good manu- Comments on HHS facturing practices and medical device reporting regulation compliance Recommendations inspection of Aequitron Medical, Inc., giving special attention to all com- plaints listed in the HID file should be conducted. A full GMP inspection, including a determination of medical device reporting regulation compli- ance of Aequitron was conducted in December 1989, after we completed our fieldwork. The agency is currently evaluating the results of this inspection to determine what, if any, regulatory action is warranted. HHS also concurred with our recommendation to review its medical device reporting regulation and its guidance to device manufacturers on problem reporting for clarity and effectiveness and to consider develop- ing specific guidance for the manufacturers of devices such as apnea monitors for which determining reportability is thought to present spe- cial difficulties. HHS stated that FDA has established a medical device reporting regulation study group to analyze the medical device reporting program and regulation, identify problems, and recommend solutions. ISIS expects that this study group will complete its work later this fiscal year. We agree that our case study does not by itself indicate how the medical General Comments device reporting regulation is implemented by most firms. We have stated in the report that this case illustrates some of the concerns that we identified in our earlier study of the implementation of the medical device reporting regulation.’ This earlier report was based on the more than 50,000 medical device problem reports received during the first 3 years of the regulation’s implementation, as well as the results of com- pliance inspections on 575 manufacturing establishments. It also reviewed FDA'S report-processing and data-handling procedures and included analyses of regulatory actions and other initiatives FDA took on ‘Se? 11,s. General Accounting Officr, Medical Devices: FDA’s Implementation of the Medical Dcwre litwtting Regulation. GAOPEMD-89-10 (Washington, D.C : February 1989). p. 4. Page 35 GAO/PEMD-90-17 Underreporting Home Apnea Monitor Problems Appendix IV Comments From the Department of Health and Human SeWiCeS first years of the program’s operations. The point here is that manufac- turers are obliged by regulation to review problem reports and make judgments about whether the events described warrant a report to FDA. Page 37 GAO,‘PEMD-90-17 Underreporting Home Apnea Monitor Problems Page 39 GAO/PEtdD99.17 Undemporting Home A9nea Monitor Problent, .. Requests for copies of GAO reports should be sent to: U.S. General Accounting Office Post Office Box 6015 Gaithersburg, Maryland 20877 Telephone 202-275-6241 The fiit five copies of each report are free. Additional copies are $2.00 each. There is a 25% discount on orders for 100 or more copies mailed to a single address. Related GAO Products Medical Devices: Early Warning of Problems Is Hampered by Severe Underreporting, GAOIPEMD-87-l. Washington, D.C.: December 1986. “Medical Devices: Early Warning of Problems Is Hampered by Severe Underreporting,” statement of Eleanor Chelimsky, GAO/T-PEMD-87-4. Washington, D.C.: May 1987. Medical Devices: FDA'S Forecasts of Problem Reports and FTEs Under H.R. 4640, GAO/PEMD-88-30. Washington, D.C.: July 1988. Medical Devices: FDA'S Implementation of the Medical Device Reporting Regulation, GAO/PEMDRR-10. Washington, DC.: February 1989. Medical Device Recalls: An Overview and Analysis 1983-88, GAO/ PEMD-89.15BR. Washington, DC.: August 1989. Medical Device Recalls: Examination of Selected Cases, GAO/PEMD-90-6. Washington, DC.: October 1989. “Medical Devices: The Public Health at Risk,” statement of the Comp- troller General, GAO:/T-PEMD-90-z. Washington, D.C.: November 1989. “Medical Devices: IJnderreporting of Problems, Backlogged Systems, and Weak Statutory Support,” statement of Eleanor Chelimsky, GAO/T- PEMD-90-3. Washington, D.C.: November 1989. (973293) page40 GAO/PEMDSO-l7UnderreportingHomeApneaMonitorProblems Major Contributors to This Report Gerald L. Dillingham, Assistant Director Program Evaluation L. Joseph Sonnefeld, Evaluator and Methodology Elaine L. Vaurio, Evaluator Division Julian P. Klazkin, Attorney Office of the General Counsel Page 38 GAO/PEMD-90.17Underreporting Home Apnea Monitor Pmblems Appendix IV Comments From the Department of Health and Human Services the basis of medical device problem reports. It is our earlier report that forms the basis for general statements about the implementation of the regulation and its associated compliance inspection program. The pre- sent report’s conclusion is only that the evidence in the Aequitron case was consistent with the findings of our earlier work. In addition, we did not find support in the text or preamble of the medi- cal device reporting regulation or in associated guidance for HHS'S state- ment that the medical device reporting regulation is primarily intended to elicit reporting from health care professionals. The regulation is focused on the reporting responsibilities of device manufacturers who learn of potentially serious problems associated with the use of their product, regardless of the source of information. In the report, we have acknowledged that the determination that a complaint is reportable may be more difficult for home apnea monitors than for some other devices, in part because many reports of problems with home monitors are received from laypersons as opposed to health care professionals. The report does not fault FDA'S inspection strategy, although we agree that it would have been desirable to conduct the first Aequitron compli- ance inspection earlier than 2 years after the promulgation of the regu- lation. We agree that the inspection scheduling strategy appears reasonable. It is only the effectiveness of the inspection program in establishing and resolving the problem with the firm’s reporting criteria, despite the more-frequent-than-normal inspections, that the report calls into question. The remainder of nns’s general comments review the manufacturer’s compliance history and FDA'S actions in more detail but are not inconsis- tent with our report. ~11sdoes give a quotation from Aequitron’s response to the “Inspectional Observations” given to the firm at the con- clusion of the GMP and medical device reporting regulation compliance inspection conducted between September and November 1989. In this response, Aequitron stated that it planned to submit reports on all inci- dents involving serious injury or deaths that the company becomes aware of. This response was not available to us during the period of our data collection, but we are concerned that this reporting policy may not represent a satisfactory general solution to the interpretation of medical device reporting regulation requirements in the case of home apnea monitors. Rather, it could tend to contribute to the problem of “overre- porting” identified during our earlier review. This problem appeared to affect FDA'S ability to handle the volume of reports submitted during the Page 36 GAO/PEMD-90-17 Underreporting Home Apnea Monitor Problems AppendkIV Comments From the Department of Health and HnmanSen-ifes 4 GAO Recommendation We reconrmend that FDA take the following actions to improve the effectiveness of medical device problem reporting and to ensure the overall safety and effectiveness of medical devices. 1. Conduct a comprehensive good manufacturing practices (GWP) and medical device reporting regulation compliance inspection of Aequitron Medical Incorporated, giving special attention to e complaints listed in the HID file. HIi. Comment In December 1989 FDA completed a full GWP inspection that included the WDR compliance status of Aequitron Medical Incorporated. The results of the inspection are currently under review in FDA. We, therefore, do not believe another such inspection is warranted at this time. GAO Recommendation 2. Review its medical device reporting regulation and its guidance to device manufacturers on problem reporting for clarity and effectiveness; and consider developing specific guidance for the manufacturers of any devices, such as apnea monitors, for which determining reportability is thought to present special difficulties. fitiS Comments We concur. FDA has established an WDR study group to analyze the WDR program and regulation, identify problems and recommend solutions. This group has been working on the WDR and plans to have its study completed later this fiscal year. The concept of developing specific guidance for certain industries, devices, etc., has been discussed by the CDRH and is part of this study group's objectives. In addition, a new guidance document for device firms has been drafted and is currently under review in FDA. It should be noted, however, that the MDR regulation covers some 1,700 categories of devices and 30,000 - 40,000 individual devices. It would be virtually impossible for FDA to craft language that would inform a firm about every event that the Agency would want reported for each device and/or every generic category. (Additional technical comments were provided and have been incorporated, where anpropciats, throughout the reQ0Kt.j Page34 GAO/PEMD-SO-17Underreporting HomeApneaMonitar Problems Appendix IV Comments From the Department of Health and Human Services 2 PDA agrees with GAO that it took longer than we would have liked to reach full implementation of the MDR regulation. The first of these inspections of Aequitron did not occur until 1987, at which time all complaints in the "H/I/Death File" (HID) file were evaluated for compliance. The inspectional program appears to have been successful in uncovering violations. Moreover, under normal circumstances, FDA would have inspected Aequitron biennially. However, because of problems with the firm, FDA has visited the firm at least ten times in the last 6 years. A great deal of regulatory activity occurred with this firm as a result of those inspections including recalls, safety alerts, etc. It was during one of these inspections that the MDR problems were discovered and appropriate recall action initiated. It should also be noted that the Aequitron WDR policy in effect at the time of the 1987 inspection appears to more closely follow the MDR regulation as written than does the firm's later policy found during the May 1988 inspection. The firm was issued a Notice of Adverse Findings (NAF) as a result of the May 1988 inspection. Subsequent inspections have shown that the firm's HID complaint file was not always complete during previous inspections. We believe that in this case FDA's inspectional program was reasonable. We did adopt an enhanced program on October 1, 1988. Districts are now directed to conduct an MDR inspection during the course of every other "good manufacturing practices" (GMP) inspection unless certain situations exist, in which case, "for cause" inspections are to be made as needed. Also, the time required for FDA to identify the underreporting problem is not, in our opinion, a reflection of failure of the inspection strategy. As previously noted, this firm was inspected at least twice as often as would be expected during the time period of GAO’s study. As problems were identified, regulatory actions were initiated that resulted in the submission of additional reports to FDA and the modification of the firm's MDR reporting criteria. See above. The GAO report correctly observes that Aequitron's stated reporting criteria were overly liberal. As noted in FDA's Center for Devices and Radiological Health (CDRR) guidance document "Medical Device Reporting Questions and Answers," a report from a layperson, in contrast to a report from a health care professional, can be analyzed to decide whether it reasonably suggests that the event occurred (p. 19). Aequitron Page 32 GAO/PEMD-90-17 Underreporting Home Apnea Monitor Problems Appendix Comments From the Department of Health and Human Services Note: GAO comments -1 supplementing those in the report text appear at the end of thts appendix. DEPARTMENTOF HEALTH & HUMAN SERVICES O,flce 01 lnSw?ctOlGeneral APR 27lm MS. Eleanor Chelimsky Assistant Comptroller General United States General Accounting Office Washington, D.C. 20548 Dear Ms. Chelimsky: Enclosed are the Department's comments on your draft report, "Medical Devices: Underreporting of Serious Problems with a Home Apnea Monitor." The comments represent the tentative position of the Department and are subject to reevaluation when the final version of this report is received. The Department appreciates the opportunity to comment on this draft report before its publication. Sincerely yours, ector General Enclosure Page 30 GAO/PEMDWl7 Undrrreporting Home Apnea Monitor Problems Appendix III Our Analysis of Aequitron’s Medical Device Reporting criteria Second, when a report of a malfunction is received, confirmation of its existence by the manufacturer is not a precondition for the manufac- turer’s obligation to report. FDA explains in its published guidance why the regulations do not permit the manufacturer to refrain from report- ing merely because it is unable to confirm a malfunction: “FDA believes that it IS not speculative, inefficwnt, meffective, or uneconomical for the agency to require the submissmn of information about reportable events before such events have been confirmed by the manufac%urer or importer. It would bc irresponsible for FDA to wait for confirmation of a reportable event by a manufac- turer or importer bcfow requiring reporting of any information about such went.” (49 Fed. Reg. 36.330 (September 14, 1984) Thus, although confirmation of a device malfunction may help a manu- facturer determine whether its device actually caused or contributed to serious injury or death. lack of confirmation does not mean that a report is not required. Aequitron’s reporting criterion requiring that a health care professional observe a malfunction before it will report (unless a test by Aequitron confirms the malfunction) is also inconsistent with the medical device reporting regulation. When a health care professional notifies a com- pany that one of its devices may have caused or contributed to serious injury or death, the company must file a report with FDA. No “reasona- ble person” analysis need be performed because, under the regulation, information from a health care professional reasonably suggests “per se” the required link. thereby triggering the reporting requirement. (49 Fed. Reg. 36,336 (1984)) The company has no discretion in this circum- stance, even if it cannot confirm a device malfunction or can attribute the serious injury or death to user error. poor maintenance, or use of the device beyond its useful life. Nothing in the regulation requires that a health care professional must have observed a malfunction before the professional’s report will be accepted as per SCreasonably suggesting that the device may have caused or contributed to a serious injury or death. Finally, Aequitron’s reporting criteria are also inconsistent with the reg- ulation in situations in which the company is made aware of a malfunc- tion that did not result, in serious injury or death but that may cause or contribute to a future one. Aequitron’s criteria suggest that a report to FDA is not required if ( 1) the company cannot confirm a malfunction reported to it by a layperson or (2) the company is notified by a health care professional who did not observe the malfunction. In the first case, Page 28 GAO/PEMD-W-17 Underreporting Home Apnea Monitor Problems Appendix III Our Analysis of Aequitron’s Medical Device Reporting Criteria The Aequitron reporting criteria set out the conditions under which it must report to FDA when it receives information that one of its devices has malfunctioned. It appears that Aequitron’s policy is not to report to FDA when someone other than a health care professional notifies the company that one of its devices has malfunctioned and may have caused or contributed to serious injury or death, unless the company is able to confirm a malfunction within a specified time. Additionally, if an initial report of a malfunction associated with serious injury or death comes from a health care professional, Aequitron’s criteria appear to require either that the professional must have observed the malfunction or that the malfunction be confirmed before the company reports to FDA. Reports of malfunctions not resulting in serious injuries or death, but with the possibility of causing either in the future, are handled in simi- lar fashion. FDA indicated in a letter to Aequitron that its reporting crite- ria appeared to be adequate. We believe these policies are inconsistent with FDA'S medical device reporting regulation (21 C.F.R. 803) in several respects. Under the medical device reporting regulation, a manufacturer must file a report with FDA whenever it receives information from a person other than a health care professional from which a reasonable person would conclude (that is, information that reasonably suggests) or a statement from a health care professional concluding that a device (1) may have caused or contributed to a serious injury or death or (2) has malfunc- tioned and a recurrence of the malfunction would be likely to cause or contribute to serious injury or death. (21 C.F.R. 803.3(f), 803.24(a)) According to Aequitron’s reporting criteria, Aequitron will not report under the following circumstances (among others): “The reported malfurr~~tion was not observed by a health care professional and the reported malfunction vannot be confirmed by testing of the device by adequately trained persomwl w11hm reporting time frames as established by the regulation. ” Thus, Aequitron’s criteria seem to contemplate that, in order for a report to be required, there must have been a report of a malfunction, and the malfunction must either have been confirmed by Aequitron or have been observed by a health care professional. These criteria appear to be inconsistent with the regulation in several respects. First, the requirement to report is not limited by the regulation to only situations in which a malfunction is reported to the manufacturer. When Page 26 GAO/PEMD9017 Underreporting Home Apnea Monitor Problems Appendix II Chronology of Contacts Lktween FDA and Aequitron February 12,1988 The manufacturer requested an export certificate for a second home apnea monitor, Model 9200, plus three other devices. March 9, 1988 The manufacturer submitted a medical device reporting regulation report indicating that Model 9200 failed to sound its alarm. According to the manufacturer, this was the first failure confirmed by the manufac- turer’s testing. Testing confirmed a lo-percent failure rate in the alarm component. The manufacturer had 24 reported failures of 4,800 units in distribution. The failures were related to a variation in voltage that would be likely to occur only when the monitor is running on its battery. March 14,1988 The manufacturer notified FDA of a recall of Model 9200, having issued a letter to dealers recommending the following: specified monitors should be used with only a particular model of battery charger and with AC power, the specified units should not be used on battery power only, and the monitor’s audible alarm should be verified daily. The manufacturer advised that an interim solution would be made pending a permanent one. FDA visited the manufacturer to collect data on the recall. April 4, 1988 FDA classified manufacturer’s action as a Class I recall, based on a health hazard evaluation determining that the alarm problem presented a high risk. FDA sent a letter to the manufacturer advising that FDA was classi- fying the manufacturer’s action as a Class I recall. April 6, 1988 FDA disapproved the February 12 export request for Model 9200 but approved export of three other devices noted in its February 12, 1988, request. April 15, 1988 Following FDA'S March 14 inspection of the manufacturer, FDA sent a notice-of-adverse-findings letter to the manufacturer, noting deficiencies in inventory control. May 11,1988 FDA conducted a GMI' inspection of the manufacturer, including a medical device reporting regulation compliance inspection. Six deficiencies were noted on the observation form; the manufacturer promised to correct all the deficiencies. The FLhi district, office recommended that a notice-of- Page 24 GAO/PEMD90-17 Underreporting Home Apnea Monitor Problems - Appendix II Chronology of contact.3 Between FDA and Aequltmn recall, with an engineer to review the electronics test conclusion and a physician to evaluate the delayed-alarm feature. January 8,1985 FDA inspected the manufacturer to follow up on the recall recommenda- tion for Model 8200. The inspection determined that the manufacturer believed that the device distributor had modified the 30 units without approval. January 29,1985 FDA conducted a limited inspection of the manufacturer in response to an assignment from the FDA Center for Devices and Radiological Health to collect mechanical and electrical drawings to assist in an analysis of the units and as follow-up to a medical device reporting regulation report on Model 8200. The manufacturer’s analysis revealed the monitor to be performing within specifications. FDA'S inspector recommended routine follow-up. May 28,1985 The manufacturer submitted 30 medical device reporting regulation reports on Model 8200. All were unrelated to the problem with burned wires. The manufacturer considered all 30 to be random failures. Most monitors were repaired and returned to the customers. May 31,1985 FDA hand delivered a letter to the manufacturer advising of the results of FDA'S evaluation of the health hazard involved in the “burn out” prob- lem with Model 8200. The letter indicated that the risk to users was high and that there was not enough evidence to show that the problem was limited to the 30 units modified by the distributor. The manufacturer indicated that it would prepare a list of options for dealing with the “burn out” problem and present them to FDA. June 18,1985 At a meeting between FDA and the manufacturer regarding the May 30 health hazard evaluation letter, the manufacturer reasserted that the wire “burn out” problem with Model 8200 occurs not during use but as the device is being set up. The manufacturer indicated that it was going to issue a letter to distributors and that it believed that the letter would constitute a “Safety Alert.” FDA staff concurred with the manufacturer Page 22 GAO/PEMD9I%17 Underreporting Home Apnea Monitor Problema Appendix I H/I/Death File Number Serial number Date Reason/return Analysis 81 206310 IljE8 No alarm/no Ilght-In hospital pt.Ok Ic u4/cap c5 62 106327 l/69 Checkout only-death/alarmed Device in-spec Page 20 GAO/PEMLWO-17Underreporting HomeApnea Monitor Problems Appendix I H/I/Death F’ile Number Serial number Date Reason/return Analvsis 50461 l/83 Checkout only~death Device in-spec 6/83 Checkout-death-no alarm noted Device In-sDec 3 51048 7183 Checkout o&death Device In-soec 4 52362 8/83 Checkout-death brady alarm Device In-spec 5 52435 9183 Checkout-death-bradv alarm Device In-soec 6 53512 9/83 No apnea alarm-brady alarm (seizure) Device In-spec 7 54078 1l/83 ElectrIcal burn (unfounded) Device In-spec a 53360 12/83 Checkout onlv~death rhme I”-S”PC 9 54441 l/84 No audible alarm (unfounded) Device in-spec 10 51224 2/84 Checkout only death (unit not In use) Not received 11 51994 7184 Checkout only death Device In-spec 12 -- 52944 8/84 Checkout-no apnea alarm Device In-spec 13 52306 8/84 Checkout only-death Device In-spec 14 60034 12/84 No apnea alarm (perlodlc breathing) Device in-spec 15 53228 1 l/84 Checkout only~death Device In-spec 16 xxxxx 12/84 Death Device status unknown 17 IO/84 Death-no audible alarm Device In-snec 18 64183 1 l/84 Death checkout only Device In-spec 19 54711 12/84 Death-no alarm Device In-spec 20 64512 12/84 Checkout only-death Device In-spec 21 .- -101467 l/85 Checkout only-death Device In-spec 22 102224 5/85 Checkout only-death Device In-spec 23 102842 8/85 Checkout only~death Device In-spec 24 101457 8/85 Checkout only-death Device In-spec 25 102748 8/85 No apnea alarm Device In-spec 26 53881 lo/85 Checkout-death~no alarm Device In-spec 27 54775 lo/85 Checkout only~death Device In-spec 28 ..- 60496 12/85 Checkout-death-no alarm Device In-spec 29 .- 106149 12/85 Electrical burn~(unfounded) Device In-spec 30 52701 l/86 Checkout only-deatti Device In-spec 31 65221 l/86 Checkout only~death Device i In-spec 32 65942 2/86 Checkout only~death Device In-spec 33 52950 6/86 Checkout-death-no alarm Device In-soec 34 62944 6/86 Checkout-death (shock/burni Device status unknown 35 201411 7186 Death Device status unknown 36 101472 7/86 Checkout only-death Device fullv functlonal 37 54267 8/86 Checkout only-death Device in-spec 38 -- 66150 1i/86 Check&t only-death Device In-spec 39 109286 12/86 Checkout only-death Device In-spec (continued) Page 18 GAO/PEMD-W-17Underreporting Home Apnea Monitor Problems Contents Letter 1 Appendix I 18 H/I/Death File Appendix II 21 Chronology of Contacts Between FDA and Aequitron Appendix III 26 Our Analysis of Aequitron’s Medical Device Reporting Criteria Appendix IV 30 Comments From the Comments on Recommendations 36 36 Department of Health General Comments and Human Services Appendix V 38 Major Contributors to Program Evaluation and Methodology Division 38 Office of the General Counsel 38 This Report Related GAO Products 40 Table Table 1: Complaints and Medical Device Reporting Regulation Reports on Aequitron Medical, Inc., Model 8200 Apnea Monitor Page 16 GAO/PUKB9@17 Underreporting Home Apnea Monitor Problems B-237533 8200 home apnea monitor. Our fieldwork was conducted from Septem- ber 1989 through December 1989. The information on which this report is based was obtained from multi- ple sources and required both qualitative and quantitative analysis. We conducted a selective review of the available technical literature, includ- ing apnea monitor engineering studies and the National Institutes of Health consensus report on the status of apnea monitors, in order to provide ourselves with background for identifying and understanding the relevant issues.‘? We systematically reviewed FDA documents, including records of inspec- tions and regulatory actions related to the device manufacturer, and cor- respondence between FDA and the device manufacturer. We also analyzed reports submitted to FDA under the medical device reporting regulation. To clarify, supplement, and confirm the documentary evi- dence, we conducted structured interviews with representatives of the device manufacturer, FDA officials in the Center for Device Radiological Health, district and regional offices, and other knowledgeable persons. Our review was conducted in accordance with generally accepted gov- ernment auditing standards. We will send copies to the Secretary of Health and Human Services, to the Director of the Center for Devices and Radiological Health, and upon request to others who are interested. If you have any questions or would like additional information, please call me at (202) 275-1864 or Dr. Michael J. Wargo, Director of Program Evaluation in Physical Systems Areas, at (202) 275-3092. Other major contributors to this report are listed in appendix V. Sincerely yours, Eleanor Chelimsky Assistant Comptroller General “Satwnal Institutes of Health Infantile Apnea and Home Momtoring (Bethesda, Md.. October 19%). Page 14 GAO/PEMD-90-17 Underrepurting Home Apnea Monitor Problems B-237533 Although FDA has provided all registered device manufacturers with some guidance on reporting requirements in the form of a “questions and answers” document, the example we have presented suggests a potentially serious problem in the manufacturer’s interpretation of the requirements and formulation of a reporting policy based on that inter- pretation, rather than random or isolated failures to report. Our study illustrates the serious consequences that shortcomings in the implementation of the medical devices reporting regulation and subse- quent inspection program can have. In this case, FDA did not have infor- mation on a number of adverse experiences with the device in question when the agency made critical decisions with respect to recalls and other regulatory actions. FDA found that this manufacturer’s interpreta- tion of the reporting requirements differed from the agency’s in ways that put it in noncompliance with the regulation. The compliance inspec- tion program did not identify and resolve these differences for a sub- stantial period of time after the regulation went into effect. During that time, FDA was unable to make valid comparisons of the monitor’s prob- lem rates or trends with those of other, similar monitors made by other manufacturers, thus compromising one of the most important uses for data from the medical devices regulation. It was beyond the scope of our study to conduct an in-depth review of the device manufacturer’s prob- lem reporting policies, but we did analyze certain aspects of the manu- facturer’s revised reporting policy. The design of our study precludes generalizing to other devices or manu- facturers. However, the findings, taken with the findings of our earlier report, raise a concern that the problem-reporting and inspection issues may pertain to a much broader segment of the device manufacturing industry and to the safety and effectiveness of medical devices in gen- eral. Therefore, we believe they are worth further attention by FDA. Recommendations We recommend that FDA take the following actions to improve the effec- tiveness of medical device problem reporting and to ensure the overall safety and effectiveness of medical devices. 1. Conduct a comprehensive good manufacturing practices and medical device reporting regulation compliance inspection of Aequitron Medical, Inc., giving special attention to all complaints listed in the HID file. 2. Review its medical device reporting regulation and its guidance to device manufacturers on problem reporting for clarity and effectiveness Page12 (;AO/PEMD-9017UnderreportingHomeApneaMonitorProblems 5237533 The time required in this case for FDA to identify the underreporting problem raises questions about the effectiveness of the agency’s inspec- tion program and its capacity to identify potentially serious device problems through monitoring complaints. The inspection strategy is designed to include special emphasis on firms manufacturing the types of devices that have demonstrated reportable problems and scheduling more frequent medical device reporting regulation inspections for those manufacturers. Nine of the complaints in the HID file involving alarm problems and the death of patients are dated before the manufacturer underwent its first GMP inspection. By the end of the manufacturer’s third year of opera- tions and before a second FDA inspection, it had received an additional 7 complaints of alarm problems and deaths. According to the device man- ufacturer, all the complaints that had been received with allegations of hazards, injuries, or deaths were placed in the appropriate GMP record. The first GMP inspection, in 1984, found “no significant deviations.” Dur- ing the second inspection in 1984, in response to a complaint of the mon- itor’s failure to sound its alarm, FDA noted that the manufacturer had 17 monitors returned for problems with the alarm. During this time, there were several inspections and other interactions between the manufacturer and FDA about a variety of problems associ- ated with various models of apnea monitors. FDA reported that during an inspection in January 1987, it reviewed all complaints of hazards, inju- ries, and deaths received since 1985. It was not until a May 1988 inspec- tion, 4 years after the first GMP inspection, that FDA dealt with the reportability of complaints about alarms. By that time, FDA had received a total of 83 medical device reporting regulation reports on the moni- tor.2L (See table 1.) More than 88 percent of these involved allegations of an alarm failure, including the death of two patientsY2 “Many of these reports did not originate m complaints to the manufacturer. Reports must be made under the medical device reporting regulation, not only in response to complaints but also whenever a manufacturer acquires informatmn from any source that reasonably suggests that one of its devices may have caused or contributed to serious 1q~uz-yor death or has malfunctioned in such a way that, if the malfunction were to recur. it would be likely to cause or contribute to serious i&y or death. Some of the other sources of such mfonnation include the manufacturers’ own research, testing, or servicing of their devices as ~41 as the mcdlcal and scientific literature. “‘Of 278 reports submitted by September 1989, consisting of 271 malfunction reports and 7 reports of deaths, the manufx~urer’s tests confirmed that Model 8200 had malfunctioned in 271 of the com- plaints. However, Aequitron’s tests did not confirm that the device had malfunctioned in any of the 7 complaints in which the death of a patient occurred Page 10 GAO/PEMD.90.17 Underrepmting Home Apnea Monitor Problems B237633 part of its GMP inspections. The results of these inspections can lead to additional actions by the agency.l’l We found that between August 1984 and June 1989, FDA had various contacts with Aequitron, including at least eight formal inspections. Three were GMP inspections, and two of these included the medical device reporting component. Five were “for cause,” including three that were initiated in response to complaints FDA had received. One was a follow-up to a medical device report on Model 8200, and one was a fol- low-up to an FDA district office’s recommendation to recall Model 8200. (A selective chronology of contacts between FDA and Aequitron Medical, Inc., is given in appendix II.) During a May 1988 GMP inspection, FDA examined the manufacturer’s complaint records and identified 10 unreported complaints that FDA inspectors believed met the medical device reporting regulation’s defini- tions of reportable events. Four of these complaints contained allega- tions of the death of patients. As a result of this May 1988 inspection, a notice-of-adverse-findings letter indicating “noncompliance” with the medical device reporting regulation was issued to the manufacturer in October 1988.‘; According to FDA'S inspection report, Aequitron’s complaint records were reviewed for the previous 15 months (January 1987 through April 1988). However, the earliest of the unreported complaints in the HID file that FDA cited in its notice-of-adverse-findings letter was dated only 6 months prior to the inspection, in October 1987. Representatives of the manufacturer stated that FDA'S finding of “non- compliance” resulted from a difference in the interpretation of the medi- cal device reporting rcquirrments. According to the manufacturer, many of the unreported complaints had not been made by health care profes- sionals and could not be confirmed by the company within the required reporting time. Therefore. in accordance with its interpretation of the “‘According to FDA, the mspv&on strategy adopted by the agency will result in a medical device reporting regulation compliance inswction for every firm manufacturing medium-risk (class II) and high-risk (clazs III) deviws HTleast NMX ewry 4 years an d incorporating manufacturers of low-risk (class I) devices less freqwntl!, “A notice-of-adverse-findln#s letter may be sent to a manufacturer when an inspection reveals that a manufacturer or indwidual is m vwlation of the laws and regulations or when there LSinformation that an existing condition or prwtw may lead to a violation if left uncorrected (although the agency has concluded that the narurc of thv wolation does not require immediate action against the manufac- turer or indiwdual). Page 8 GAO/PEMD-90-17 Underrrporting Home Apnea Monitor Problems R237533 Question 2: Compliance Did the device manufacturer comply with FDA'S existing problem-report- ing regulations and procedures? FDA'S primary source of information about problems associated with the use of medical devices consists of the manufacturer’s reports generated by the requirements of the medi- cal device reporting regulation (21 C.F.R. 803). This regulation requires that device manufacturers telephone an initial report to FDA on serious injuries and deaths within 5 calendar days, and it requires that this be followed by a more complete written report within 15 working days. Reportable malfunctions that do not involve serious injury or death must be reported within 15 working days of the manufacturer’s receiv- ing the device-problem information. One important source of informa- tion that leads to medical device reports is complaints to the manufacturer, which can be made by health care professionals or other users of devices.LY We compared the 70 complaints that contained the word “death” with FDA'S record of medical device reports and found that 14 complaints were dated before the medical device reporting regulation was promul- gated. The significance of this is that before the regulation was promul- gated, the device manufacturers’ obligations were fulfilled by maintaining general complaint records and making them available to FDA during GMP inspections. We found 56 complaints associated with deaths whose listed dates fell after the medical device reporting regulation was implemented. Only 4 reports in FDA's medical device reporting data base as of September 1989 could be matched with the 56 complaints as hav- ing been reported to FDA in accordance with the provisions of the regula- tion. One additional medical device report on a death associated with Model 8200 was reported to FDA during this time but could not be identi- fied with a specific complaint on the HID file, because the manufacturer had not submitted a serial number with the medical device report. Thus, at, least 5 1 of the 56 complaints of death had not been reported before F~)Abegan to take compliance actions. The information in the HID file alone was not sufficient for us to make a definitive judgment about the reportability of the complaints or to establish causal connections between the device and the safety hazard, injury, or death. The examination of the manufacturer’s complete record for each complaint, which would be necessary for such assessment, was beyond the scope of our review. In a partial review of the complaint - ’ ‘Accordmg to the GMI’ regulation. a complaint is a written or oral expression of dissatisfaction wgarding the identRy, quahty. duratnlity, reliability, safety, effectiveness, or performance of a device Page6 GAO/PE~-SB17UnderreportingHomeApneaMonitorProblems 5237633 Apnea is a prolonged lack of respiration that can result in low blood oxygen levels, which can lead in turn to brain damage and death. The condition can be induced by a variety of underlying medical disorders. However, premature and low birthweight infants are particularly prone to apnea. Apnea monitors are electronic devices intended to detect episodes of apnea. In a typical device, when either breathing or heart rate falls below set levels or when the device’s electrical leads are improperly attached to a patient, both audible alarms and flashing lights are trlg- gered. Specialized models of apnea monitors are designed for hospital and home use. To avoid lengthy hospital stays, home apnea monitors have been increasingly used in recent years. Findings Question 1: Complaints of How many complaints of the deaths of patients have been associated Death with the Aequitron Medical, Inc., Model 8200 home apnea monitor? We received information through our fraud hotline that there was evidence of serious nonreporting or underreporting of problems associated with Model 8200.’ We were also told that many of the unreported complaints involved the deaths of the patients. FDA'S good manufacturing practices regulation requires that manufac- turers maintain two types of records regarding the complaints they receive from users about their products.* The first is a general record of all users’ complaints. The second is a record exclusively devoted to com- plaints alleging that hazards to safety, injuries, or deaths are associated with a medical device. Evidence subsequently provided to us included a table labeled “H/I/ Death File” (HID file) and identified as users’ complaints related to the ‘This information was simultaneously provided to the staff of the Subcommittee on Health and the Environment of the House Committee on Energy and Commerce. ‘Section 520(f) of the 1938 Federal Food, Drug, and Cosmetic Act, added by the Medical Device Amendments of 1976, authorizes FDA to promulgate regulations that specify practices in the manu- factwe, packaging, storage, and installation of devices. The good manufacturing practices established by the GMP regulation include controls over manufacturing, specifications, processing procedures, device components, packaging, labeling, manufacturing equipment, and records. Page 4 GAO/PEMINO.17 UnderreportIng Home Apnea Monitor Problem E-237533 With regard to the first question in the paragraph above, we found that Results in Brief Aequitron had received at least 70 complaints that the deaths of patients were associated with the use of the Model 8200 monitor.” With regard to the second question, we found that the manufacturer had maintained the required record of complaints but had not fully complied with the reporting requirements of the medical device reporting regula- tion.’ A partial review of the manufacturer’s complaint record by FDA found that 10 unreported complaints should have been reported, includ- ing 4 that involved the death of patients. We could verify that only 6 of the complaints of deaths dated after the implementation of the medical device reporting regulation were reported to FDA. Two of these com- plaints of death were reported only after FDA compliance actions, nearly 1 year after the events. With regard to the third question, we found that when FDA received information about the association of the monitor with deaths, it investi- gated whether a sample of complaints should have been reported to FDA. Following the investigation, FDA cited the device manufacturer for non- compliance with the medical device reporting regulation and, in concert with the manufacturer, reviewed the manufacturer’s problem-reporting policy. FDA then reviewed and approved a revised problem-reporting pol- icy submitted by the manufacturer. These actions resulted in the sub- mission of more than 150 additional reports to FDA. Our case study methodology precludes generalizing from these findings and conclusions to other devices and manufacturers. Instead, the study’s function is to illustrate some of the critical concerns that we identified in our earlier generalized work on the implementation of the medical device reporting regulation. Included there were concerns about the ‘We found that an Acquitron dorument entitled “H/I/Death File” (or “Hazard, Irljury, Death (HID) File”) contained information abstracted from the special section of the record of complaints reserved for hazards to safety. injuries, and deaths that a dewe manufacturer is required to maintain under the good manufacturing practxes (GMP) regulation (21 C.F.K. 820.198). The HID file listed 82 com- plamts. 68 of which referred to deaths. Two additional complaints of death of were contained in a similar list drawn from the “gt~~eral” portion of Aequitron‘s GMP record of complaints. (See appendix 1.) According to the manufacturer, the HID file contains complaints alleging that serious ir\juries or deaths were associated wth rh(%dcmrr. hut some of the complaints do not allege that a malfunction of the monitor occurred WV did not independently investigate each complaint on the list to determine the circumstances of tht, evrnts: causal connerrwns between the device and the safety hazard, idjury, or death; or the actual occwwnw of thr wents listed. ‘The medical deuce reportmy rcgulatlon, cffectwe December 13, 1984, requires that device manufac- turers report to FDA whenrwr they hwomr aware of information that reasonably suggests that one of their devices may hwc causrd or rontnbuted to a serious injury or death or has malfunctioned in such a way that, if the malfunctmn were LU recur, the device would be likely to cause or mntribute to a death or serious in,jwy or drath. Sw Mrdxal Devices for a detaded discussion of the medical device reporting regulation Page 2 GAO/P&MD-90.17 Underreporting Home Apnea Monitor Problems
Medical Devices: Underreporting of Serious Problems With a Home Apnea Monitor
Published by the Government Accountability Office on 1990-05-31.
Below is a raw (and likely hideous) rendition of the original report. (PDF)