People in medical device industry keep hiring me to do software testing, but then don’t let me see the code or do any software testing. I think this must be specific to medical device industry.
I feel bad taking their money because I’m not helping them test their software. Maybe someone can clarify. Maybe they don’t know that formal software testing exists. People in other industries do it.
Is this why medical devices suck so bad? I feel like it must be.
In my experience in an adjacent industry:
1: yes there are real software testers
2: the ratio of test writers and test result reviewers and project managers to actual testers, is about 5:1.
@mohe0001, I want your job. Getting paid for not do anything sounds sweet. I write software for a living and there is no doubt that FDA-approved medical software is the worst. I compare DIY Loop and GlucoGram to their commercial counterparts. There is no need to look at the source code. The difference at the user level is mind-boggling. Milk the cow as long as you can. Software testing is not the answer. Eliminating government overreach is needed. I am not against FDA issuing a stamp of approval. There should be no law prohibiting the use of devices or software without FDA stamp. We allow the sale of cigarettes which we know for sure is only bad. I am allowed to buy an assault rifle. It should be legal for me to use a device that keeps my blood sugar in normal range even if some bureaucrat does not understand how it works or has some other reasons to dislike it. I actually prefer not to pay for this bureaucrat who inflictits harm on me. Whoever is in favor of a FDA stamp should pay a little more for a device or software that has the stamp. I am sympathetic to that. Just allow me to have normal blood sugar to avoid more complications. It seems like a reasonable request that unfortunately is being denied. I blame the American justice system for that. The American attitude is that it is always somebody else’s fault. Everybody can file frivolous lawsuits without any repercussion. I know of a county where if someone sues for $1 million and gets awarded $10k, he would get the $10k but pay the prorated legals costs = 99% of the legal costs. To no surprise, frivolous lawsuits are not an issue in that country.
I feel like this stuff ought to be tested, just like stuff is in other industries. We find bugs when we look. Maybe I ought to go to another industry and return with more experience about what is conventional industry practice. That might be valuable.
How is all this stuff getting approved by FDA without any testing. I get that FDA makes people do a lot of paperwork, but it doesn’t seem to require much of anything (other than paperwork) before stuff gets used on humans and is sent to market. Is that true?
Other regulated, critical systems in industry feel different than medical devices, so maybe FDA simultaneously regulates everything and regulates nothing. I’m very confused. What should I do?
I’m doing a TON of work, but I’m not doing the right work in the right way. I can’t find a way to articulate the problem to anyone. If I could, things might be different.
Yesterday, someone in leadership said he went to a conference with FDA speakers where someone said that there is simply no reason to test software. His perspective was that you test it through the adverse events reporting tool. How do I respond to that? Is that perspective common?
I kinda figured I have some civic responsibility to test this ■■■■ and test it well so that we don’t offload all that onto patients who have to do it for free and at peril. Right? I’m so confused about the situation I find myself in.
When it comes to safety, each industry has their own standard for what software defects are safety-critical failures vs non-safety-critical. Or indeed what kind of subsystem is safety-critical vs not.
Train Control Safety, for example, will have the train shut off its engines/motors and apply full brakes in case of an unrecoverable software error and that’s considered the safest state.
In Aerospace, it works very differently, as just shutting down all the airplane engines in event of an in-flight software failure will have very different results.
It’s interesting to see here how, for example, the designers of an insulin pump will take the railroad approach (shut down the pump and deliver no insulin in case of a pump motor vs flow disagreement) when many insulin pump users expected a more graceful aerospace-style failure mode (e.g. don’t shut down the pump).
I’m becoming quickly disenfranchised with white collar work.
My boss told me that the best way to test stuff was to just wait until people started filing Adverse Events on the FDA website. This relationship isn’t gonna work out. Why is the world operating this way? My heart literally broke in that moment.
This tune is dedicated to all the girls at the coffee shop who never did catch the culprit who kept pooping on the bathroom floor while I was there applying to jobs, day after day. The world is a bad place indeed.
@mohe0001 I would keep the job and then work on how you think the job could be improved. You won’t change the bosses or the company, it’s an easier route for them. But if you line up how the problem would be viably solved and the reasons for it and once you have it down, then come up with a plan on how to solve the issue. But I would also verify, because bosses just say things, that it actually isn’t being tested by someone down the line or that he is actually right in what he said.
This pertains to @Tim12 comment about railroads and airplanes and how they treat a problem. My previous Firebird, a wheel speed sensor broke, as soon as it happened it shut down my brakes because the car couldn’t tell how fast I was going. Luckily I was driving in a parking lot when it happened and was easily able to use my emergency brake. I could never figure out how someone could think that shutting down a perfectly good brake system in response to a wheel speed sensor breaking was a good idea.
The market is good now. The market hasn’t been good for twenty years. Gotta dance while the music is playing. It could turn at any time.
The field is too creepy. People hate it. Someone ■■■■ on the floor at my last job. That’s an unconventional professional move from an engineer, but it conveys a difficult truth about the nature of the situation there. You could hardly blame the guy. That was Abbott.