from the sooooo-sorry-about-the-bullets... dept
Shooting the messenger is the most popular response to reported data breach, making the job of security researcher far more dangerous than it should ever be. The twist in the latest “shoot the messenger” story is the shooter coming back around to ask the shooting victim for help. Bad idea. Even if the body is still warm and breathing, it’s probably not in the best of moods.
Dissent Doe runs databreaches.net, a site that covers all sorts of exposed data stories. Sometimes, Doe is asked by those discovering security holes to disclose the information to the affected parties. (See above paragraph for why.) In early May, Doe tried to alert the Bronx-Lebanon Hospital Center about confidential patient records left exposed by a contractor. The stuff exposed was deeply personal, containing write-ups of patients’ substance abuse problems or mental illnesses.
This didn’t go well. The hospital didn’t want to talk about it or explain why a third-party had so much access to confidential health records, much less why it hadn’t bothered to properly secure the hospital’s database. One day after these mostly futile phone calls, someone (not specified in the post) contacted Dissent Doe to let her know the databases had been secured and thanking her for notifying them.
That should have been the end of the story. But it wasn’t.
It was a brief honeymoon. On May 9, Kromtech published their report and I published my first report on the incident without any statement from the hospital or vendor, neither of whom had provided a promised statement.
Then on May 12, coordinated threat letters arrived via email from external counsel for both iHealth and Bronx-Lebanon Hospital. DataBreaches.net understands that Kromtech Security also received similar letters.
I’ll let that sink in for a minute: they threatened a person who went out of her way to alert them they were leaking protected health information. Instead of saying, “Thank you so much, and can we also ask you to please securely destroy any data you might have in your possession?” they sent me threat letters.
The stupid, angry letters contained stupid, angry threats. First, the letters accused Doe of improper access. Then they went on to demand she and everyone else in possession of this data delete it and send a certified letter (or something) back to the hospital and vendor confirming the destruction of the data. They also demanded she reveal her sources and not post anything further about the breach.
Doe didn’t think much of the demands, but she did retain counsel just in case. An angry, non-stupid response letter from her legal rep changed the tone of the demands into more polite requests. Not that the change in tone won Doe over. A bridge only needs to be burnt once to render it useless. And, in one sense, the angry, stupid threat letter did work: while Doe didn’t cave, it appeared that Kromtech did delete the data it had discovered. That resulted in a problem.
Apparently, the hospital and vendor forgot about their earlier bridge-torching efforts. They approached Doe again, this time asking for help identifying which patients had had their personal info exposed in order to notify them.
Now the entities could just notify everyone who had PHI/PII on the server, of course, but it seemed like they were trying to narrow the universe to only those whose data wound up in Kromtech’s hands – or this site’s – or NBC News’ hands. And now Kromtech could not tell them which patients had data in the 500 mb of data they had downloaded and then destroyed.
But Kromtech had sent a subset of that data to DataBreaches.net, who had not destroyed the data it possessed. If DataBreaches.net wanted to be helpful, it could go through all the data and let the entities know which patients had data in there, right?
But why should Doe do this? The two affected entities had already expressed their gratitude using legal threats, not exactly the best foundation for future collaborative efforts.
I might have been able to spare the vendor and hospital some notifications if I was willing to donate my time to going through files to compile information for them, but I’m not willing.
I’m not willing, in part, because I do not want to be going through PHI if it’s not for my reporting purposes. And I’m not willing because why should I have to spend my valuable time compiling information for entities that tried to bully me and who now need my help to help them clean up their mess??
Shooting the messenger kills potential allies. But far too many entities think it’s better to shoot first and live with their regrets later. Security researchers aren’t the enemy of privacy, but they’re often treated as criminals and malcontents by entities who have screwed up their own security efforts.