4 ms·

Engineers and doctors and nurses... Eh... Maybe not so fast...

by mantas 3y ago

inglor_cz 3y ago
Possibly yes, but on emulators/training data/virtual patients. One of the advantages of today's digital world is that you don't have to jump to the real thing immediately. Come to think of it, maybe even laws should be tested in a laboratory first. We cannot emulate the human society fully, but some unwanted effects could be detected this way anyway.
- mantas 3y ago
  It's easy to find issues in laws with some brainstorming over a pizza. The problem is people pushing certain laws frequently don't give a damn about side effects. Why would they care about results coming out from a simulation?
  - inglor_cz 3y ago
    It is easy to find theoretical issues through brainstorming, but weeding out the false positives is much less easy. As an example, religious conservatives will brainstorm a lot of issues that are bound to rise (according to them) from gay emancipation. In practice, these failed to manifest.
    - mantas 3y ago
      You can easily manipulate results in a lab test too by messing with the environment and cherry picking results. And focusing on different outcomes. I'm pretty sure same religious conservatives will claim that issues they saw did come up. While the other extreme would just shrug off the same issues as non-issue. Different people value different things... More at 11.
wongarsu 3y ago
Maybe the statement could be refined to "must be given opportunity for failure". For some professions that might include running some projects at a smaller scale, as a simulation or even as role play exercise
evandijk70 3y ago
For doctors looking to innovate, there are actually a lot of ways to fail fast and safely if you introduce a new drug. First, establish that a drug binds to the target in laboratory conditions (relatively quick). After that, test the drug in an animal model. After that, proceed to clinical trials, which are phased (phase I, II, III) to first establish whether a dosage exists that is safe (Phase I), what an effective dosage is (Phase II), and only after than testing if it improves upon the current standard of care (Phase III). Each phase is more costly (both in time and money) than the previous phase and will eliminate a significant portion of drug candidates. "The success rate of each drug discovery stage in academia was 31.8% for preclinical, 75.1% for phase I, 50.0% for phase II, 58.6% for phase III." [1] https://pubmed.ncbi.nlm.nih.gov/24406927/ https://pubmed.ncbi.nlm.nih.gov/24406927/
- dahart 3y ago
  Doctors? This sounds more like the activity of a Pharmaceutical company?
mordae 3y ago
Even those. When the cost of failure is too high, we need to invest in practices that allow them fail safely. And get as much from failures as possible. It means giving doctors and nurses simulators, someone to look over their shoulder when learning new techniques, giving them high-quality followup data, finding top performers and make them explain what they do and so on. Which is currently very hard to do, because teaching others means you lose your advantage. It also means learning from failures of others. That means writing about the fuckups and telling them. Which is currently almost impossible to do, because admitting failure can often result in an end to your career. And all of this is exacerbated with consolidated, privatized healthcare where the shots are called not by outcome-oriented caregivers, but by profit-oriented MBAs.
taopai 3y ago
My grandfather, a photographer, told me this story: 20 years ago in a doctor's graduation ceremony held somewhere in Spain the old doctor who gave the speech told the graduates: "Now your job is to kill people. Along the years you'll learn, and if you are good, you'll kill less people and save some."

Engineers and doctors and nurses... Eh... Maybe not so fast...

by mantas 3y ago

Engineers and doctors and nurses... Eh... Maybe not so fast...

Possibly yes, but on emulators/training data/virtual patients. One of the advantages of today's digital world is that you don't have to jump to the real thing immediately. Come to think of it, maybe even laws should be tested in a laboratory first. We cannot emulate the human society fully, but some unwanted effects could be detected this way anyway.

mantas 3y ago
It's easy to find issues in laws with some brainstorming over a pizza. The problem is people pushing certain laws frequently don't give a damn about side effects. Why would they care about results coming out from a simulation?
- inglor_cz 3y ago
  It is easy to find theoretical issues through brainstorming, but weeding out the false positives is much less easy. As an example, religious conservatives will brainstorm a lot of issues that are bound to rise (according to them) from gay emancipation. In practice, these failed to manifest.
  - mantas 3y ago
    You can easily manipulate results in a lab test too by messing with the environment and cherry picking results. And focusing on different outcomes. I'm pretty sure same religious conservatives will claim that issues they saw did come up. While the other extreme would just shrug off the same issues as non-issue. Different people value different things... More at 11.

wongarsu 3y ago

Maybe the statement could be refined to "must be given opportunity for failure". For some professions that might include running some projects at a smaller scale, as a simulation or even as role play exercise

evandijk70 3y ago

For doctors looking to innovate, there are actually a lot of ways to fail fast and safely if you introduce a new drug. First, establish that a drug binds to the target in laboratory conditions (relatively quick). After that, test the drug in an animal model. After that, proceed to clinical trials, which are phased (phase I, II, III) to first establish whether a dosage exists that is safe (Phase I), what an effective dosage is (Phase II), and only after than testing if it improves upon the current standard of care (Phase III). Each phase is more costly (both in time and money) than the previous phase and will eliminate a significant portion of drug candidates. "The success rate of each drug discovery stage in academia was 31.8% for preclinical, 75.1% for phase I, 50.0% for phase II, 58.6% for phase III." [1] https://pubmed.ncbi.nlm.nih.gov/24406927/ https://pubmed.ncbi.nlm.nih.gov/24406927/

dahart 3y ago
Doctors? This sounds more like the activity of a Pharmaceutical company?

mordae 3y ago

Even those. When the cost of failure is too high, we need to invest in practices that allow them fail safely. And get as much from failures as possible. It means giving doctors and nurses simulators, someone to look over their shoulder when learning new techniques, giving them high-quality followup data, finding top performers and make them explain what they do and so on. Which is currently very hard to do, because teaching others means you lose your advantage. It also means learning from failures of others. That means writing about the fuckups and telling them. Which is currently almost impossible to do, because admitting failure can often result in an end to your career. And all of this is exacerbated with consolidated, privatized healthcare where the shots are called not by outcome-oriented caregivers, but by profit-oriented MBAs.

taopai 3y ago

My grandfather, a photographer, told me this story: 20 years ago in a doctor's graduation ceremony held somewhere in Spain the old doctor who gave the speech told the graduates: "Now your job is to kill people. Along the years you'll learn, and if you are good, you'll kill less people and save some."