Comment by munificent | Hacker News Reader

Comment by munificent | Hacker News Reader

Parent Full thread

> According to the literature 33 out of 100 patients who underwent this operation in the US within the past 10 years died. 90% of those had complicating factors. You [ do / do not ] have such a factor.

Everyone has complicating factors. Age, gender, ethnicity, obesity, comorbidities, activity level, current infection status, health history, etc. Then you have to factor in the doctor's own previous performance statistics, plus the statistics of the anaesthesiologist, nursing staff, the hospital itself (how often do patients get MRSA, candidiasis, etc.?).

And, of course, the more factors you take into account, the fewer relevant cases you have in the literature to rely on. If the patient is a woman, how do you correctly weight data from male patients that had the surgery? What are the error bars on your weighting process?

It would take an actuary to chew through all the literature and get a maximally accurate estimate based on the specific data that is known for that patient at that point in time.

No one said anything about a maximally accurate estimate. This is exactly the sort of obtuse attitude I'm objecting to.

By complicating factors I was referring to things that are known to have a notable impact on the outcome of this specific procedure. This is just summarizing what's known. It explicitly does not take into account the performance of any particular professional, team, or site.

Something like MRSA is entirely separate. "The survival rate is 98 out of 100, but in this region of the country people recovering from this sort of thing have been exhibiting a 10% risk of MRSA. Unfortunately our facility is no exception to that."

If the recipients of a procedure are predominately female and the patient is a male then you simply indicate that to them. "The historical rate is X out of Y, but you're a bit unusual in that only 10% of past recipients are men. I'm afraid I don't know what the implications of that fact might be."

You provide the known facts and make clear what you don't know. No weasel words - if you don't know something then admit that you don't know it but don't use that as an excuse to hide what you do know. It's utterly unhelpful.

So, while you are correct, you are missing an important piece:

most people cannot think like this

I'm not talking about patients, I'm talking about everyone, including doctors. They just can't think in a probabilistic sense. And you'll counter that it's just reporting facts, but they don't even know which ones to report to you, how to report them, none of it. It just doesn't seem to fit in many peoples heads.

Fair enough. It's a depressing thought but you're probably right.

this is part of the mindset had by doctors that makes some people want to “do their own research” rather than trust their physician. A medical intervention has to have positive expected value for it to be a good idea, and figuring out the expected value has to involve some quantification of risks. If doctors don’t want to do that because they could get sued if they don’t give a maximally accurate estimate and producing a maximally correct estimate would be too much work, then fine, it’s a free country and I don’t want to make doctors do anything they don’t feel like doing, but they are creating a situation where parents who want to figure out if something is a good idea have no choice but to start googling things themselves.

I’ve undergone some surgeries that were not without risks and every time, i’ve been stonewalled by doctors when asking for basic information like “in your personal practice, what is the success rate for this surgery?”. Always something like “Oh, everyone is different, so there’s no way to give any estimates.” The only options are, either they have some estimate they think is accurate enough that they’re comfortable recommending the surgery but they won’t tell me (in which case they’re denying me useful information for their own benefit), or they have no idea and are recommending the surgery for some other reason (a very concerning possibility lol). Either way, it instantly makes our relationship adversarial to some extent, and means I need to do my own research if I want to be able to make an informed decision.

Don't they use quantifications of risk to determine treatment plans to at least some extent?

I doubt doctors do: my guess would be most doctors follow a list of best practices devised by people like malpractice actuaries and by their sense of the outcomes from experience.

Reply on news.ycombinator.com