The U.S. Food and Drug Administration says it wants to improve transparency around medical device safety, yet the agency is still allowing device companies to file secretive reports of deaths involving novel, permanent implants in the heart.

Earlier this year, the FDA ended a little-noticed but long-running system known as "alternative summary reporting" (ASR), which previously had let device companies avoid publicly filing full narrative reports about specific devices that the agency felt were well-understood.

The ASR program was ended after then-FDA Commissioner Dr. Scott Gottlieb tweeted last March that it was "imperative that all safety information be available to the public." (Gottlieb left the agency the following month.)

Yet last week, the FDA declined to release any information about four deaths summarized in a report about Boston Scientific's Watchman left-atrial appendage closure device -- a novel, permanent heart implant used to cut stroke risk in patients with atrial fibrillation.

And in October, a research letter published in JAMA Internal Medicine said the FDA was still accepting secretive "summary" reports of problems involving Edwards Lifesciences' Sapien 3 aortic heart valve and Abbott Laboratories' MitraClip system for mitral valve repair.

Though the FDA ended alternative summary reporting, it still allowed Boston Scientific to file an adverse event report that summarized four deaths in a few lines of text, and even kept secret the name of the underlying study that uncovered the deaths. FDA declined to release any further detail about the four deaths, including the "spreadsheet" mentioned in the public-facing adverse event report.

An FDA official said in an email that the secret Watchman reports were part of a different summary reporting system that the FDA continues to operate – even though FDA officials are aware that granting such exemptions will lead to further non-public reporting of medical device events.

Last year, the FDA said it was making changes in its device-surveillance division to "free up resources to better focus on addressing the highest risks, such as deaths and serious injuries, associated with medical devices."

Boston Scientific said in a statement Wednesday, "We take all patient safety and any potential product-related complications seriously, and are transparent and comprehensive with our reporting of events per established criteria set by the FDA. Boston Scientific has a robust complaint reporting system and we report all relevant complaint information we receive in compliance with all FDA regulations. We do not withhold or keep in secret any relevant complaint information."

Boston Scientific says the Watchman carries a low complication rate, but the FDA considers it a Class III device, the highest-risk category. There was no device like the Watchman on the U.S. market until it was approved in 2015 – even today, the plug-like system has no direct competitor in the domestic market.

The Watchman faced three different FDA expert-review panels, an unprecedented level of scrutiny for a medical device that required two separate randomized controlled trials. In the end, the device was approved by the FDA with nuanced instructions regarding patient selection.

Like other devices, the Watchman is being studied as part of a large-scale "postmarket" research effort called a device registry. FDA officials say adverse event reports captured in device registries can be exempted from federal laws requiring public disclosure, though it's not clear that the agency has ever published rules on how the system works.

Device registries are expected to become more common in the future. The FDA told the Star Tribune last week that device registries are an important component of the new device-surveillance system the agency is building, called NEST, for National Evaluation System for health Technology.

"Active medical device surveillance will better protect patients by continuously generating, accessing and evaluating large data sets (including registries) on device performance and clinical outcomes associated with device use in routine clinical practice. It also improves the FDA's ability to link adverse events with specific devices, so we can act quickly with manufacturers and health care professionals to make more timely, evidence-based decisions to mitigate device problems and keep patients safe. [The FDA] can encourage registry development by granting limited individual reporting exemptions while maintaining access to the data needed for postmarket surveillance," FDA spokewoman Kristen Pluchino wrote Friday night.

Madris Tomes, a former FDA device-report reviewer who now runs a device-transparency business called Device Events, said the problem with device registries is that they offer no transparency to most doctors or the public.

"Providing these registry exemptions makes it so that deaths can be summarized and very difficult to quantify even to the manufacturers," Tomes wrote in an email. "When a summary report is submitted to the FDA there is no narrative, which is the section that explains the patient outcome. To make matters worse, the patient problem codes (outcomes) are redacted by the FDA. This renders the summary reports almost useless."

The statement from Boston Scientific noted that the Watchman summary reports were filed after receiving information from the FDA-mandated New Enrollment PoST Approval Surveillance Analysis Plan in the National Cardiovascular Data Registry Left Atrial Appendage Occlusion Registry, which independently tracks all patients who have received a Watchman implant.

"The registry was established and is managed by the American College of Cardiology and the data is available to ACC physicians to present or publish on any aspect of the registry, without Boston Scientific involvement. We are given access to only a subset of the data within the registry and made aware of potential post-market complaint information by means of a quarterly data export from the surveillance database," the company said.

Pluchino said it is not true that the FDA is knowingly allowing medical device companies to keep secret the details of fatalities involving novel, high-risk devices.

She said the granting of an exemption for a registry study does not relieve a manufacturer of the requirement to investigate and report problems involving medical devices. However, in the context of a registry run by an outside organization, the FDA understands that manufacturers generally are not provided with all the details they would need, like the name of the patient or the person who reported the event.

Lacking that information "may limit" the manufacturer's ability to investigate event. "Should a manufacturer later obtain any required information that was not available at the time it filed its initial report, then it must submit this information in a supplemental report," Pluchino wrote.

Boston Scientific said the data it receives from the Watchman registry "is limited and de-identified, with no further opportunity to obtain additional detail. When complaint criteria meet established FDA regulations for reporting, we report all relevant provided information to the FDA. All other WATCHMAN events (e.g. those outside of the Registry) are investigated and reported on an individual basis in compliance with FDA regulations."

The Star Tribune wrote about the issue of medical device adverse-event report summaries, at what turned out to be the peak of the secretive program's popularity in 2016, when more than 500,000 adverse events were summarized in a single year.

The program's use quickly fell off, and in March Gottlieb announced the program would end, following critical reports on the ASR program by Kaiser Health News and the International Consortium of Investigative Journalists.

On Sunday, the Star Tribune reported that the Minnesota-invented Watchman device hit a milestone of 100,000 human implants in October. In addition to the summarized deaths, more than 40 death reports were filed for the device in 2019 that contained narratives in the FDA's public MAUDE database.

The Watchman derives much of its safety by enabling patients to stop taking blood-thinning drugs. The Watchman carries a one-time implant complication rate of 1.5%, compared to a 1.3% annual risk for a major bleeding while on warfarin. Boston Scientific said the device reduced disabling and fatal strokes by 55% compared to warfarin patients.

-- Story updated Nov. 13 with comments from Boston Scientific.