“Ninety percent of doctors I know are fed up with medicine.” • Hacker News Board

Re: “Ninety percent of doctors I know are fed up with medicine.”

#31

Post by nradov » Thu, Jun 19, 2008, 11:37 PM UTC

dividendium wrote: post #28

Maybe we need to redefine the problem. Some of the suggestions below were about automating the paperwork. Why not automate the treatment? I mean consider how unscalable it is to have a doctor, an actual person, see each patient. At most they can get through one patient every 15 minutes. A doctor normally looks at a patient, asks them a few questions to rule out certain conditions and then prescribes a pill or request…

Sounds good. This will be practical just as soon as you invent an AI smart enough to make it through medical school and pass the boards.

Re: “Ninety percent of doctors I know are fed up with medicine.”

#32

Post by kingkongrevenge » Thu, Jun 19, 2008, 11:51 PM UTC

arn wrote: post #29

kingkongrevenge wrote: post #24

Earlier quoted context omitted.

> Modern medicine is based on the scientific method and millions on dollars are spent on studies to prove that therapies work. Sure. Take Ambien, for example. The studies show it works! People who take it fall asleep a whole 15 minutes sooner than placebo! For thousands of dollars and unquantified long term consequences you can have something pathetically less effective than lifestyle changes or some cognitive therap…

"[statins] have absolutely no affect on mortality rates in the typically prescribed case" http://www.ncbi.nlm.nih.gov/pubmed/12114036 The Heart Protection Study (HPS) randomly assigned 20,536 subjects to simvastatin (40 mg/day) or placebo After an average follow-up of 5.5 years, the following significant benefits with simvastatin were observed: - A 13 percent reduction in all-cause mortality (12.9 versus 14.7 percent…

This study supports my position. 12.9 vs 14.7. From taking my statins I get a measly 1.8% lower chance of dying over a five year period, and that's a shaky assertion. It costs a fortune, I'm impotent, and my short term memory is screwed up, but hey, the doc says to take it for that 1.8%. And I suspect the 1.8% is misleading. The profile of the people who died in both groups was probably distinctly identifiable, so in effect the vast majority probably got no benefit.

And I think I've seen this methodologically refuted elsewhere. Some studies have definitely got the opposite result.

Re: “Ninety percent of doctors I know are fed up with medicine.”

#33

Post by dazzawazza » Fri, Jun 20, 2008, 12:05 AM UTC

streety wrote: post #7

dazzawazza wrote: post #3

If they want they can come and work in the British NHS. A good GP here can get £100,000 a year. Of course you'll have to put up with the MRI taking weeks unless it's an emergency but it seems better then what that doctor suffered. I think the biggest difference here is that although people in the UK moan a lot about the National Health Service they do respect the doctors and staff and value their input to society. Wh…

Probably not such a good idea. We can barely find enough positions for the doctors trained in this country.

ahh, but there are a lot of openings for GP's. Our medical students all want to be brain surgeons and don't seem to realise that without GP's most people would be dead before needing a brain surgeon.

Re: “Ninety percent of doctors I know are fed up with medicine.”

#34

Post by yummyfajitas » Fri, Jun 20, 2008, 1:28 AM UTC

nradov wrote: post #31

dividendium wrote: post #28

Maybe we need to redefine the problem. Some of the suggestions below were about automating the paperwork. Why not automate the treatment? I mean consider how unscalable it is to have a doctor, an actual person, see each patient. At most they can get through one patient every 15 minutes. A doctor normally looks at a patient, asks them a few questions to rule out certain conditions and then prescribes a pill or request…

Sounds good. This will be practical just as soon as you invent an AI smart enough to make it through medical school and pass the boards.

All the AI needs to do to be medically useful is obey the samurai principle: meet/exceed the effectiveness of a doctor on some subset of patients, and `raise GoSeeARealDoctorException` on the rest.

I'm sure regulators/politicians/medical establishment won't like it. But the technical problem isn't all that bad (and has actually been solved by machine learning for a few use cases).

Re: “Ninety percent of doctors I know are fed up with medicine.”

#35

Post by copenja » Fri, Jun 20, 2008, 1:59 AM UTC

My uncle is a surgeon and he complains about the same things:

Overheads are getting huge, profits are shrinking, etc.

Meanwhile, however, he lives in a friggin mansion. Seriously,a mansion.

Not saying this is the common case, just one observation.

Re: “Ninety percent of doctors I know are fed up with medicine.”

#36

Post by iamelgringo » Fri, Jun 20, 2008, 2:30 AM UTC

kingkongrevenge wrote: post #32

arn wrote: post #29

Earlier quoted context omitted.

"[statins] have absolutely no affect on mortality rates in the typically prescribed case" http://www.ncbi.nlm.nih.gov/pubmed/12114036 The Heart Protection Study (HPS) randomly assigned 20,536 subjects to simvastatin (40 mg/day) or placebo After an average follow-up of 5.5 years, the following significant benefits with simvastatin were observed: - A 13 percent reduction in all-cause mortality (12.9 versus 14.7 percent…

This study supports my position. 12.9 vs 14.7. From taking my statins I get a measly 1.8% lower chance of dying over a five year period, and that's a shaky assertion. It costs a fortune, I'm impotent, and my short term memory is screwed up, but hey, the doc says to take it for that 1.8%. And I suspect the 1.8% is misleading. The profile of the people who died in both groups was probably distinctly identifiable, so in…

Dude, I've worked as an ER/Critical care nurse for years. I push a drug, a person's heart rate changes. I tweak a drug on an IV pump, a person's blood pressure comes out of shock and normalizes. The idea that pharmacology doesn't help people is quite simply wrong.

Sure, the protocol for Advanced Cardiac Life support changes every 2 years, because some therapies that we use to thing were very effective, are proven to be less effective than new ones. But, there is a strong self-corrective factor in Medicine. The ideal isn't to do what the pharma companies push the doctors to do. The ideal is to do whatever the best science says increases lifespan, quality of life and decreases morbidity and mortality.

At times, studies are funded by pharma companies, and the research is tainted. Current best practices are a moving target, because the science changes. But, over all, good medical practice is a big plus.

Sure there are drugs that have great promise, are pushed through FDA approval and then are found to have bad side effects and/or not as effective as once thought... (e.g. Vioxx, Celebrex... both recalled several years after approval).

But, for every drug that isn't effective, or has bad complications, there are dozens that do work. And, many more are found to have other beneficial uses. For example, mucomyst is a drug that was developed to help break up mucous plugs in people's lungs. Then, is was found to work as the antidote for one of the most common and toxic poisonings: tylenol. It's also used to protect people's kidney's during IV contrast injections.

I'm sorry but the anti-pharma, anti-vaccination stuff that flies around the internet is one of the clearest examples of FUD I've seen, and it's very harmful. I've seen a number of people die because they didn't trust medicine and they didn't trust doctors. Please, check what you are saying before someone gets permanently hurt by your ideas.

Re: “Ninety percent of doctors I know are fed up with medicine.”

#37

Post by arn » Fri, Jun 20, 2008, 2:34 AM UTC

kingkongrevenge wrote: post #32

arn wrote: post #29

Earlier quoted context omitted.

"[statins] have absolutely no affect on mortality rates in the typically prescribed case" http://www.ncbi.nlm.nih.gov/pubmed/12114036 The Heart Protection Study (HPS) randomly assigned 20,536 subjects to simvastatin (40 mg/day) or placebo After an average follow-up of 5.5 years, the following significant benefits with simvastatin were observed: - A 13 percent reduction in all-cause mortality (12.9 versus 14.7 percent…

This study supports my position. 12.9 vs 14.7. From taking my statins I get a measly 1.8% lower chance of dying over a five year period, and that's a shaky assertion. It costs a fortune, I'm impotent, and my short term memory is screwed up, but hey, the doc says to take it for that 1.8%. And I suspect the 1.8% is misleading. The profile of the people who died in both groups was probably distinctly identifiable, so in…

it's 1.8% percentage points which is not the same as 1.8% lower chance. It's a 13% lower chance. And that's just from dying. It prevented many others from having a heart attack or stroke which can be pretty bad to.

You can dismiss it if you like, but what these numbers mean in more practical terms is this:

Over 5 years on treatment, it will prevent up to 1 person out of 10 from dying, having a stroke, or heart attack.

And there may be a negative study here or there, but there are many more positive studies. It's like flipping a coin 100 times in a row. There will be times when you get 70 heads, 30 tails, and times you will get 30 heads, and 70 tails. If you do it enough, it will average around 50% heads and tails. This is the "true" result or chance over time of getting heads or tails.

If you do a single study, you might get a negative result (no improvement), if you do another study you might get a really positive result (huge improvement). If you do it enough, the results with trend towards the "true" result... which for statins is that it saves a significant number of lives and prevents heart attacks and strokes in certain patients.

Re: “Ninety percent of doctors I know are fed up with medicine.”

#38

Post by kingkongrevenge » Fri, Jun 20, 2008, 2:48 AM UTC

arn wrote: post #37

kingkongrevenge wrote: post #32

Earlier quoted context omitted.

This study supports my position. 12.9 vs 14.7. From taking my statins I get a measly 1.8% lower chance of dying over a five year period, and that's a shaky assertion. It costs a fortune, I'm impotent, and my short term memory is screwed up, but hey, the doc says to take it for that 1.8%. And I suspect the 1.8% is misleading. The profile of the people who died in both groups was probably distinctly identifiable, so in…

it's 1.8% percentage points which is not the same as 1.8% lower chance. It's a 13% lower chance. And that's just from dying. It prevented many others from having a heart attack or stroke which can be pretty bad to. You can dismiss it if you like, but what these numbers mean in more practical terms is this: Over 5 years on treatment, it will prevent up to 1 person out of 10 from dying, having a stroke, or heart attack…

Semantics. I change my odds by 1.8% in the naive assumption that these numbers are applicable to myself at face value. If you like these odds I would like to sell you some lottery tickets.

Your thinking is fundamentally flawed. 1.8 people out of 100 are prevented from dying, and not for free: there are side effects and financial costs. How many died in year six as opposed to year five is also unclear. We're talking about drugs they want to keep people on for the rest of their lives.

The cumulative arithmetic of how many escaped cardiovascular "events" and death cannot be computed clearly from the numbers presented, as you seem to allude.

Re: “Ninety percent of doctors I know are fed up with medicine.”

#39

Post by kingkongrevenge » Fri, Jun 20, 2008, 2:54 AM UTC

iamelgringo wrote: post #36

kingkongrevenge wrote: post #32

Earlier quoted context omitted.

This study supports my position. 12.9 vs 14.7. From taking my statins I get a measly 1.8% lower chance of dying over a five year period, and that's a shaky assertion. It costs a fortune, I'm impotent, and my short term memory is screwed up, but hey, the doc says to take it for that 1.8%. And I suspect the 1.8% is misleading. The profile of the people who died in both groups was probably distinctly identifiable, so in…

Dude, I've worked as an ER/Critical care nurse for years. I push a drug, a person's heart rate changes. I tweak a drug on an IV pump, a person's blood pressure comes out of shock and normalizes. The idea that pharmacology doesn't help people is quite simply wrong. Sure, the protocol for Advanced Cardiac Life support changes every 2 years, because some therapies that we use to thing were very effective, are proven to…

Key words are ER and Critical Care. That stuff works great and makes progress all the time. It's also somewhat irrelevant to the bulk of medical practice and health care costs, which pertain to chronic conditions and lifestyle disease.

Re: “Ninety percent of doctors I know are fed up with medicine.”

#40

Post by arn » Fri, Jun 20, 2008, 2:55 AM UTC

kingkongrevenge wrote: post #38

arn wrote: post #37

Earlier quoted context omitted.

it's 1.8% percentage points which is not the same as 1.8% lower chance. It's a 13% lower chance. And that's just from dying. It prevented many others from having a heart attack or stroke which can be pretty bad to. You can dismiss it if you like, but what these numbers mean in more practical terms is this: Over 5 years on treatment, it will prevent up to 1 person out of 10 from dying, having a stroke, or heart attack…

Semantics. I change my odds by 1.8% in the naive assumption that these numbers are applicable to myself at face value. If you like these odds I would like to sell you some lottery tickets. Your thinking is fundamentally flawed. 1.8 people out of 100 are prevented from dying, and not for free: there are side effects and financial costs. How many died in year six as opposed to year five is also unclear. We're talking a…

I agree that the cost is not free. The point being is that these medications do do something. They aren't sugar pills.

You're first assertion was that these medications don't actually do anything. I simply am pointing out they do do something.

It comes at a cost of $$, possible side effects, sure. But if you are arguing that point, then you have to concede that if the price is low enough or the side effects are low enough, then it's worth it.

If it's not personally worth it to you to take the statin because you have side effects from it... that's fine. don't take it. But you can't use it as a blanket to say all medications don't do anything, just because you had a bad experience with one or don't agree with the cost/benefit analysis.