Hacker News Reader

You reminded me of an interaction I had recently while getting my blood taken at a Quest office. To make small talk while I was sitting there in a chair waiting for the needle, I asked the tech if he had heard about Theranos or Elizabeth Holmes. He had not.

I explained that she ran a Silicon Valley startup that was trying shrink this little cubicle area we're in here down to the size of an ipad.

I thought he might identify with the difficulty if not absurdity of that mission. His response surprised me:

"Good idea."

"Well, actually, it turned out to be a giant fraud. She's been indicted and is awaiting trial."

"Not surprised. You know how much money this industry is worth? Big Blood isn't going to let that happen."

I let him focus on his work after that.

There was one time I went I for a physical from a "doc in a box". I was new in town and needed a physical to get some money for my HSA before a deadline.

I drank water before going in for my urine test. I never did a urine test... And remember vividly peeing in the waiting room as I left, thinking it's weird I didn't do a urine test. But I didn't care as it was through away results, I just needed a formed sign for my money.

Well, somehow I got results back for a urine test. When I questioned it... They said, nah no way its wrong... I did a test.

that could easily be fraud. you should report things like this to your insurance company.

Were you surprised? The job is not that different in what is asked of you and what is given to you than any other unskilled work. We just have a tendency to think of doctors as smarter than auto mechanics when they both are pattern-matching repair jobs, and lab techs as more capable than CVS cashiers when they are both just routine operator jobs.

Ya know - I agree with the Tech's first statement - it's a FANTASTIC idea. Too bad it's technically infeasible. really too bad that she lied about it over and over.

I'm not really surprised that he'd never heard, though. The story was much more popular in the startup/tech space than the medical space. I know a bunch of people working in molecular diagnostics who hadn't heard about it while it was happening - though they've all been incredibly fascinated and engaged by the recent book.

> field tests are highly suspect, and even lab tests need to be carefully considered - all of which requires a solid grasp on scientific/critical reasoning (which is in short supply.)

Similarly, I've heard stories from MDs about techs in pharmacies at hospitals not knowing enough to raise an eyebrow at an order for something off by 1000-fold (mg vs. µg). Patient almost gets a lethal dose of something because sleep deprived residents are running the show and incompetent / underpaid people are filling orders.

YES.

Dimensional analysis and a basic understanding of metric prefixes. Again, this is just ass-u-me'd to be well understood by all.

Pharmacies are increasingly reliant on software to catch interactions. I wonder if the same is true for dosage

Yes, yes it is. I've worked on such software, in fact. Dosage calculators can, obviously, only work with the inputs they're given. Even when prompted to verify and re-verify data entry for critical things, most of the actual users are blanked out on auto pilot.

I've been involved in countless conversations with program/hospital system administration types (at _the_ largest hospital systems in the US) about all of the ways we might try to mitigate against these types of issues. These admin folks are clever, too, they know how to work _our_ reporting systems through sales and product people to CYA, as it were. We typically walk into these meetings under the guise of "training issues," but it's anything but... This software we're talking about is form fields, if we're being honest about this stuff, and the real issue is overworked and underpaid staff. If you say that, though, you're on the hook for potentially losing these massive contracts.

It's all very corporate, on every side, is the problem, as others are alluding.

I almost died this way. When I was about 5 years old, I had pneumonia. I went to a doctor, they misplaced a decimal point in my weight, and prescribed me a 10x over dose. The nurse didn't bother to double check it, and just blindly gave me the shot. Then the second nurse, who somehow didn't know the first nurse had given me the shot, gave me a second shot, so I was 20x over dosed.

They called my parents a couple of hours later when they figured it out and told them to take me to a hospital immediately. I ended up in the hospital for a week.

> why would a lab tech lie?'

"'Stunning' Drug Lab Scandal Could Overturn 23,000 Convictions"

"She worked at the William A. Hinton State Laboratory Institute in Boston for nearly a decade before her misconduct was exposed in 2012. She admitted to tampering with evidence, forging test results and lying about it. "

https://www.nbcnews.com/news/us-news/stunning-drug-lab-scand...

Very true.

I was thinking more of third party labs. A lot of these services are outsourced due to costs / need.

For labs that are a part of the system - yeah, I definitely can see how a person working would feel the need to be a 'team' member and support their team. Sad but happens all the time.

In my city, where I was a criminal trial prosecutor for a few years, all lab work in criminal cases is done by the medical examiner's office, run by the county. The conflict of interest should be obvious but I don't even know whether the general scheme has ever been challenged.

If you think that the Dookhan case is unique, here is another one: https://www.dothaneagle.com/news/crime_court/lab-owner-accus...

This time it's made-up paternity tests and made-up drug tests in custody cases. The lack of cross-checks in the judicial system is concerning, it absolutely invites corruption.

The worst part about this case was in the custody situations. Even if as a parent you prove the test was false, if they take your kid early enough and for long enough they still won't let you have your kid back. Even if you are totally innocent. And that time span is relatively short.

This woman must be firmly entrenched in her community in rural Alabama if she could get a contract with Dale County Human Services not long after she pled guilty to credit card fraud in 2013.

Just look at her photo! She is righteous, and her victims were all drug-taking adulterous noogoodniks. If the case goes to trial the jury will give her a hug, a bible and a year on probation.

Well, there's this.

https://www.nbcnews.com/news/us-news/epic-drug-lab-scandal-r...

"A mass dismissal of wrongful convictions ─ perhaps the biggest ever ─ unfolded in a Boston courthouse on Tuesday, as prosecutors in seven Massachusetts counties dropped more than 21,000 low-level drug cases tainted by the work of a rogue lab chemist."

I was given a breathalyzer test at a road block after admiting I had two bottles of beer with dinner about an hour before. The test came back with a blood alcohol level of 0.02 which is well below the legal impairment level. The officer asked me to sign a statement confirming the reading and the amount I had consumed, so that in the event that the calibration was disputed by someone they could call on me to testify in court. I agreed and was the 7th or 8th person on their list that had signed with results.

That's a funny way to do a validation study...

Sadly, that 'data' will be used in a manner (most likely) inconsistent with any rigorous scientific underpinnings - but passed to policy makers as such.

Did the officer ask you, or instruct you to sign the document? Such a situation lends itself to feel coercive, as the officer has already detained you for a search.

Legally, it's consensual. Morally, it's an order.

If anyone except a police officer made a similar request, they'd be guilty of brandishing a firearm and any contract (or agreement to have sex) would be invalid. But if you accede to a police officer's "request", when they have a legal right to kill you for any reason during the encounter, it's considered consent.

You can be critical of the abuses of power that exist in the police force without resorting to hyperbole like "they have the legal right to kill you for any reason during the encounter." They most certainly do not legally have that right, and that's one reason it's reported on so heavily when people are shot by cops.

> and that's one reason it's reported on so heavily when people are shot by cops.

Is it though? I would say that the reason is that it is reported because the police are not expected to shoot you. As for legally being allowed to shoot you, there have been a lot of cases in the recent decades proving that point. For example https://en.wikipedia.org/wiki/Ruby_Ridge and https://en.wikipedia.org/wiki/2017_Wichita_swatting

You need to get out of the city from time to time. I've done plenty of business where one or both parties were packing, yet no threat of violence, no sense of coercion.

I for one, am quite happy to live in places where it is not considered cool, necessary or culturally appropriate for anyone to carry around lethal weapons, on any kind of ordinary basis.

I for one am quite happy to live and work in places where exercising the constitutional right to bear arms is generally not infringed and where the culture around me generally accepts and sometimes encourages the exercise of that right.

You, and few more billions of people in the worlds decided that is "not considered cool, necessary or culturally appropriate for anyone to carry around lethal weapons".

Let's not forget it.

I'm happy to live in a rural place with virtually no crime, and a high enough trust between citizens that guns are really no big deal.

That's not the point. Those people were "mere peasants" like you or I so while they possesses the tools to use deadly force they couldn't use them on flimsy pretexts and then have the system back them up. In those transactions you were on more or less equal footing. In any transaction with a cop there is an unspoken threat of state violence being used against you if you are not complaint enough.

The situations are not comparable. If you and I meet, each with our guns, and you shoot me for no reason, you will be arrested for murder and go to jail. If a police officer does the same, the worst consequence he is likely to suffer is an administrative suspension with reinstated back pay.

The officer did ask me to sign not order me, but it still left me with an un easy feeling. Once I got home I thought about it, and was dumb to comply because if I had been in an accident that document could have been used to either refuse or reduce my insurance, and possibly be used in civil action by the other party.

Yeah, I don't know if I would have done that myself. They were collecting evidence from you. It's like talking to a police officer: it's best to not - they can and will use what you say against you.

You're kind of leading up to the topic, so thought I would comment. The US concept of jury nullification is fascinating to me. That such a powerful tool exists, and that the prosecutor has so many aveneues to forbid it being discussed or used.

One particular example that bothers me a lot:

"One of the other jurors wrote a note to the judge informing on Kriho, and the latter was subsequently charged with and convicted of contempt of court for her actions, for intentionally serving on a jury with the intent to obstruct justice. " https://verdict.justia.com/2017/05/18/fresh-look-jury-nullif...

> so this doesn't necessary attract the best of the best

Reminds me of fingerprint analysis where the standards and interpretation is all over the map and it could be just done by an officer who says "yeah that matches".

> Eg: One team I was an auditor for, didn't realize that the instrument they were using had effectively two (2x) ON/OFF switches - and the detector was turned off. 2 years of data had to be rejected.

How is this kind of negligence legal?

One code base I inherited at a job had a unit test suit with no assert statements. No one verified that the tests would fail with a faulty implementation. It's the same kind of error. In the lab no one verified that detector would be able to detect.

As to forgetting switches, on lab equipment, especially custom built equipment it's sometimes not obvious, which switch is which.

Also as a student I worked in a lab where many experiments where built by one or two generations of students and then continued by others so those producing data. Working with an existing setup is like inheriting a code base. You never know what surprises await you.

The reason why the code failed to realize there was even an issue?...(in the case I mentioned earlier)

The software looked at the calibration data using linear best fit (standard stuff) and provided the r^2 was within the specified range - that instrument and the data was good to go.

Well, if you run your Standards{1,2,3,4,5} and all output "0.0". That is a straight (perfectly so) line.

They had no idea how to read the computer monitor and the real time data coming off the instrument. They had zero real training / understanding of even how the instrument worked.

I see this a lot still - people using an instrument without a great understanding of how the thing even works. Why? There's just no money/resources being spent to ensure correctness - self regulation is (and has been) all the rage for decades here in the US.

Honestly, I believe in open software - and this type of software NEEDS to be opened up. I can only imagine the horrors that will be found.

PS - regarding the actual switch. Agreed.

This instrument was of a poor UI. One switch was in the front, the other in the back.

Still, this was a regulated contract laboratory with cert's.

Upon discovery - they did perform corrective actions and such.

As for "in school" - oh yeah. It's a wild wild west sort of situation. Custom hardware (with who knows what thought being given) and software that someone wrote years ago (and having zero documentation and zero standard coding practices observed).

Many years ago I was briefly kicked out of school over a false positive on a small-town police department's lab test. I was able to prove that it was a false positive and get myself back into school through a complicated and ridiculous sequence of events, but I'll always wonder how many other people probably suffered much worse negative consequences due to the lab's negligence.

It sounds like you continued to be rarely selected :-)

So I take it the defense lawyer dismissed the MD offhand, but on your turn, the DA broke out into a sweat and dismissed you?

I really really really really wanted to be on that last one too. Especially when they asked, "does anyone have a problem with the police." I think I began to salivate. ;-)