The successful implementation of medical devices hinges on multiple factors, including the patient, surgeon, and technical details related to how the devices are placed. When devices fail, the suspected cause of failure may ultimately be found in the granular detail of those factors. Yet, too often we see examples of either a lack of data or significant delay in data acquisition to evaluate potentially harmful devices. As a result, patients may be subjected to those devices for far too long. One notable example is that of transvaginal mesh (TVM), utilized for pelvic organ prolapse and stress urinary incontinence. In 2008, the FDA issued a Public Health Notification (PHN) to inform clinicians of potential adverse events associated with the performance of these procedures and guide discussions with patients. The PHN was based upon the review of over 1,000 Medical Device Reports (MDRs) gathered from the FDA's Manufacturer and User Facility Device Experience (MAUDE) database. Over 4 years later, the FDA issued an update to its PHN, stating that adverse events indeed were not rare. Based on an additional subsequent 2874 MDRs, the FDA convened an advisory panel in 2011 to further discuss TVM and consider regulatory changes surrounding the device. Initially, TVMs were classified as Class II, or medium-risk, on a scale of I–III where Class I is considered low risk and Class III is considered high risk. Ultimately, TVMs were reclassified as Class III devices. Importantly, the lack of clinically meaningful data regarding the performance of TVM hindered the ability of the FDA to identify the issue sooner. In absence of a registry to provide us with insight to the performance of these devices, were we not poised for such a catastrophe to occur? Could these adverse events have been different if our regulatory process had been different? Moreover, how many more women were potentially harmed because our system failed us?
An examination of the current means of reporting device failures and the limitations of the MAUDE database serves to illuminate some of the gaps in our current system.1 The MAUDE database has been criticized for its passive collection of data as well as underreporting of events, quoted as potentially only proportionally 0.5% of all device failures.2 The MAUDE database is largely reliant upon manufacturers, who are required to report adverse events related to devices and device failures. This ultimately depends on the end user, typically a physician, reporting the event to a Field Safety Representative (FSR). The frequency with which this happens is presumably fairly low, as is direct reporting from physicians in the MAUDE database. Once a complaint reaches the FSR, the process for evaluation and reporting of the complaint to the FDA is proprietary per manufacturer and is not publicly described. Once an internal review is completed, it may be that no MDR reaches the MAUDE database if the manufacturer determines that the device is not responsible for the adverse event. The inherent conflict of interest within the current reporting system remains an additional barrier to device surveillance.
Although these stories may be viewed as a bleak and bitter indictment of the status of medical device regulation in the United States, the FDA should be lauded for its subsequent attempts to transform post-marketing medical device surveillance from a passive endeavor to an active one. In fact, the active surveillance of disease-specific clinical registries housing device-specific data has become a priority of the FDA,3 in addition to several other ongoing initiatives such as Unique Device Identifiers for hernia mesh. Still, change is slow, particularly as it relates to complex systems where innovation must be allowed to flourish and safety must also take priority.
There are, however, shining examples of success which involved the leveraging of clinical registries to either recall or change the Indication For Use (IFU) of certain devices. Perhaps the best-known is the recall of the Articular Surface Replacement (ASR) hip system manufactured by DePuy Orthopedics, based on concerns regarding an unexpectedly high rate of revision using the prosthesis, reported by the Australian Orthopedic Association National Joint Replacement Registry (AOA NJRR). Owing to this, a recall was issued in Australia in 2009. Further data from the National Joint Registry of England and Wales led DePuy to issue a worldwide recall of the devices in 2010. Although unfortunate for the estimated 93,000 patients who did receive the device, this event also clearly demonstrates the ability of registries to contribute to the effective post-market surveillance of devices.4
Recently, the Americas Hernia Society Quality Collaborative (AHSQC) was able to affect a change in the IFU for a broadly marketed hernia mesh device after several surgeons raised concerns of central mesh fractures, resulting in unexpected hernia recurrences. The AHSQC data registry is a prospectively maintained, surgeon-entered, point-of-care continuous quality improvement registry, which includes data entered from academic, community, and academic-affiliated surgeons in a variety of clinical settings. The AHSQC is a multistakeholder collaborative that involves surgeons, industry, patients, and the FDA to improve the outcomes of hernia patients. In this particular instance, the process was initiated when several surgeons who participate in the Quality Collaborative raised concerns with their industry FSRs. The concerns were then presented to the manufacturer of the device. Distinct from industry-standard internal reviews, the manufacturer of the device in question was able to engage in a dialogue with leadership from the AHSQC, in which granular, deidentified, disease-specific and device-specific data were provided to the manufacturer. Subsequently, the manufacturer issued a medical device correction, citing hernia-specific details: information available only as a result of surgeon participation in a disease-specific registry. The medical device correction advised against the usage of the device in the context of a specific type of hernia repair. The number of patient adverse events that occurred before the issuing of the medical device correction was 27. The intervening time frame from societal collaboration with the manufacturer was approximately 9 months.
Although those familiar with the aforementioned device and its use may question the ultimate impact of the arguably small-scale change afforded to the IFU of this device, it is our belief that this event signifies the success of a long-vaunted initiative of the FDA to leverage Coordinated Registry Networks to surveil medical devices once they reach the market.5 Notably, this includes devices which are not considered to be or classified as high-risk. This level of collaboration highlights the importance of all stakeholders being engaged in the process of post-market surveillance to allow rapid detection and correction of potential safety issues.
There is a continuum in the life cycle of a device, from its introduction to the market until it either stands the test of time or is recalled because of device-related adverse events or failure. We believe that registries are indeed the missing link that can be used to actively monitor new devices within the early phase of their life cycles. Herein, we have presented the golden moment in which a clinical registry was able to identify an early signal, and corroborate the signal using device- and disease-specific data to ultimately affect a change. Considering this in the greater context of the ecosystem of medical devices, it is hard to deny the distinct possibility that a disease-specific registry similarly could have presented the golden moment in which far <1000 MDRs would have led to the surveillance and ultimate reclassification of TVM. Although registries may still be somewhat hampered by voluntary reporting and loss to follow up of patients who are doing well, we believe that there is a role for leveraging the wealth of information that is being collected for quality improvement purposes rather than ignoring this data because of its inherent limitations. Still, the creation and maintenance of a sustainable model for device surveillance through Coordinated Registry Networks will require significant engagement from all stakeholders, including the FDA, device manufacturers, physicians, and patients to foster an environment where innovation and safety are not mutually exclusive. Nevertheless, we are buoyed by the concept that this type of collaboration may provide the safe space in which innovation may continue to occur, and harm may finally be minimized.
1. Sandberg JM, Gray I, Pearlman A, et al. An evaluation of the manufacturer and user facility device experience database that inspired the United States food and drug administration's reclassification of transvaginal mesh. Investig Clin Urol
2. Resnic FS, Matheny ME. Medical devices
in the real world. N Engl J Med
3. Sedrakyan A, Campbell B, Graves S, et al. Surgical registries for advancing quality and device surveillance. Lancet
4. De Steiger RN, Miller LN, Davidson DC, et al. Joint registry approach for identification of outlier prostheses. Acta Orthop
5. Coordinated Registry Networks. Available at: http://mdepinet.org/crn-community-practice/
. Accessed February 3, 2019.