Expert Commentary

How she did it: Reporter finds millions of hidden reports of medical device injuries

Christina Jewett revealed a hidden FDA database that concealed millions of medical device malfunction and injury reports from public view.

Lorraine Bonner, a retired doctor and sculptor from Oakland, Calif., says she spent a year recovering after surgical staples were used to seal her colon. (Heidi de Marco/KHN)
Lorraine Bonner, a retired doctor and sculptor from Oakland, Calif., says she spent a year recovering after surgical staples were used to seal her colon. (Heidi de Marco/KHN)

Annually, the Shorenstein Center on Media, Politics and Public Policy awards the Goldsmith Prize for Investigative Reporting to a stellar investigative report that has had a direct impact on government, politics and policy at the national, state or local levels. Six reporting teams were chosen as finalists for the 2020 prize, which carries a $10,000 award for finalists and $25,000 for the winner. This year, as we did last year, Journalist’s Resource is publishing a series of interviews with the finalists, in the interest of giving a behind-the-scenes explanation of the process, tools, and legwork it takes to create an important piece of investigative journalism. Journalist’s Resource is a project of the Shorenstein Center, but was not involved in the judging process for the Goldsmith Prize finalists or winner.  “Copy, Paste, Legislate” — a collaboration among The Arizona Republic, USA TODAY and the Center for Public Integrity — was named the winner on March 23.


Christina Jewett’s “Hidden Harm” series for Kaiser Health News revealed a hidden U.S. Food and Drug Administration database that concealed millions of medical device malfunction and injury reports from public view.

Malfunction and injury reports associated with medical devices are typically reported to the FDA and made public in a database known as MAUDE, short for Manufacturer and User Facility Device Experience.

But starting in 2000, medical device manufacturers began taking advantage of reporting exemptions, which were created to cut down on extraneous paperwork, Jewett reports. Companies with exemptions could batch hundreds of injury reports into a single “alternative summary” filed in an internal repository.

“Many reports were so similar that reviewing them individually was ‘mind-numbing,’” Jewett writes, quoting an FDA official. “The manufacturers would instead send the FDA a spreadsheet of injury or malfunctions each quarter, half-year or year.”

At first, in 2000, the list of exempted devices was public, though at an unspecified later date and for unknown reasons that subsequently changed.

Jewett’s reporting uncovered half a million reports of injuries or malfunctions tied to breast implants; some 66,000 surgical stapler malfunctions and more than 50,000 deadly incidents tied to an implantable heart defibrillator, which corrects abnormal heart rhythms.

Her big story was published in March, 2019. Two months later, the FDA announced it was ending the “alternative summary” program.

And it all started with one tip.

“I had it on my story list for months, maybe more than a year, to look into medical device oversight,” Jewett explains. “And so I called a woman who had previously worked at the FDA, and she’d been quoted in the media, various stories, and I just kind of was fishing — you know, what’s interesting, what’s going on, what’s going wrong, who’s doing what, trying to figure out where I could plant my shovel and find a good story.”

The source mentioned the existence of “alternative summary reports,” in which medical device manufacturers batched injury reports associated with a particular device into one report that was not made accessible to the public. Jewett found this interesting, knowing that the FDA plays a key role in compelling medical device manufacturers to publicly report injuries, deaths and malfunctions associated with their projects. “I realized that there was a pretty big blind spot in the data,” she said.

The story took about six months to report.

“It was a lot of just, working the fundamentals,” Jewett says. “Just kind of the basic things you do as an investigative journalist. And persistence, and also just narrowing the scope and being pretty laser-like about what do I need to get the story out there in a way that’s going to really have an impact?”

We asked Jewett – a reporting team of one – how she did it. Her responses, which provide great tips for journalists, were edited for length and clarity.

  1. Get the minimum story.

In investigative journalism lingo, the so-called “minimum story” is basically the core story you can promise to deliver to your editor – even if some of the reporting fails to yield certain juicy details and other elements you’re seeking.

Initially, “I had no minimum story,” Jewett says. “I just knew it was happening. I didn’t know how much, I didn’t know to what degree and for which devices.

“So I went to the FDA really early and started asking questions. I would send a link [to an FDA document referencing an ‘alternate summary report’] and say, ‘Hey, it looks like the makers of this device have some kind of exemption to report. Can I get the records or get some kind of explanation of why they got an exemption and how many reports they filed that way? And the answer was always, ‘File a FOIA [a Freedom of Information Act request].’ And that was turning out to be kind of a dead end, because I did file FOIAs, and then I would seek to get expedited processing. They refused to grant me expedited processing. In one of the letters I got back, they said, ‘You’re here at the end of the regular line, you can expect your records in 22 months.’ To which I said, ‘Oh, boy, this is not good.’

“I did a lot of Googling, terms and quotes, and I found some PowerPoint presentation from some industry conference where an expert was presenting to device companies. A lot of people who are former device executives or FDA employees go into consulting, and I worked the daisy chain of consultants who knew about this, and I found the one who said, ‘Yeah, I don’t know how many [alternative summary reports] there are; I think I read about it a couple of years ago,’ and she noodled around and sent me the link [to an analysis in Medtech Insight] and I was like, “that’s significant.” That really kept me going because I’m starting to see the shape of a minimum story. I’m like, okay, this is a thing, and here’s how much this happens.

“And so I sent a link to the FDA press office and said, ‘Hey, this [story] shows the number of these reports… do you mind just updating these numbers for me?’ And so that was kind of an easy ask, the kind of thing this press office probably handles all the time. And that got me an early break in terms of figuring out there were more than a million of these reports, and just a couple years show that.”

  1. Focus on humanizing the story.

“The next step was, what device am I going to focus on to really turn the story from an abstract data story to a more human story? I was starting to look up litigation related to some of the devices I was finding, to see the types of things that can go wrong when there are malfunctions or injuries. And I ran across the surgical stapler.

“I’d never even heard of it, never thought about, I’d never even had any idea what it was. But after reading some of the lawsuits, I realized that when they malfunction, it can be devastating, even deadly for patients. And so, I started digging into those cases where the stapler would sever a vessel, but then failed to actually fire and close it, and people were just bleeding out. They’re horrible, upsetting cases to read. I found a case in Toledo where the attorney gave me access to the family, and flew out there and spent some time with the family of Mark Lovering. He had a liver surgery and had such a massive bleed that he had extensive brain damage, and could no longer walk, came out of the coma just barely able to function at all.”

  1. Persistence pays.

“In 2016, publicly, I think there were 84 stapler injuries reported, whereas in the alternative summary data that year, there were nearly 10,000. That’s a powerful enough stat. But that came after asking [the FDA] for all the data for 20 years, and the justification for the exemption. ‘No, go away.’ Asking for three years [of data]. ‘No, sorry, go away.’ And then finally saying, ‘Please just give me one year, one year.’ And then I got one year. So, it was kind of enough that it gives you a sense that there’s something up there.

“This story kind of runs counter to my typical advice, which is to sometimes try digging into several different stories and then just go down the path where the doors are starting to open. I mean, this one, the doors were not opening but I still continued down the path. Just because I felt like it was such an interesting, important topic.”

  1. Review past notes as the story develops.

“I talked to a lot of experts in the field. And I was really frustrated that I was just getting so many people saying, ‘I’ve never heard of that. I don’t know what that is, I have no idea.’ I reflexively take notes, and it felt like just like throwaway material, useless, but then when I did find out it was a thing, really a big, big thing, suddenly that material seems relevant, like, this could actually be something I use in the story to show that the people who dedicate their career to studying medical device malfunctions don’t even know this exists.”

  1. Industry consultants can be valuable sources, actually.

“An investigative reporter might see a consultant for industry who’s always been part of that corporate club as kind of a boogeyman, but I approached the consultants for the medical device companies as humans. The ones I talked to, I asked all of them, put on your hat as a mother or a sister or a father or an uncle. And, when you’re approaching the medical system and trying to find out about a device, just the fact that so many of these reports are somewhere else, does that bother you? That really brought down some walls where they were like, ‘Yeah, I mean, you can’t trust this data.’”

  1. Think about how framing a question to sources can compel answers.

“Medtronic, which owns Covidien [a surgical stapler manufacturer], I’m sure didn’t want to talk to me. But I framed my question in a way that would to compel them to answer.

“Their exemption ended in 2017. And so in 2018, it looked like there was this huge spike in surgical staple injuries. So I just asked them, I already knew the answer, but I just put it to them in a way that would compel them to answer, which is basically, ‘Did your surgical staplers suddenly become incredibly dangerous in 2018? Is this a public health concern?’

“It put them in a position to say, ‘No, our reporting exemption ended,’ when their first instinct would be to just blow me off or not really answer directly.”

  1. You don’t always have to hold your cards close.

“There’s always a debate over, when do you go to the ‘target,’ you know, like the government agency that’s the architect of something you’re investigating. And there’s often an impulse to hide what you’re doing until the very end, and then you have this list of 15 penetrating questions and you give them one week to answer and then your investigation comes out. That’s just not always the right approach.

“With this story, I didn’t know [at first] if I had a minimum story, if I should just move on to something else. There was almost no choice. They knew everything about this, and I knew nothing about this. I really had to go early and just start asking questions about what these exemptions were.

“This was a really good case study for when it’s helpful to go to the agency early to just help them unpack: What is this? And those are often the kinds of questions that press offices are good at answering. What’s this program? How does it work?

“Sometimes it’s helpful to forward a message from a government agency to a doctor, or a lawyer or someone to you want to talk to about this, when no one’s heard of it.

“The impulse is to be really protective of that kind of material, but when you’re doing a story where nobody knows what you’re talking about, at some point, you have to cut some things loose and say, ‘This is what I’m talking about.’ It’s kind of counterintuitive and hard to do. But when you really need to get people to comment on something they’ve never heard of, it can be helpful.”

About The Author