A question on statistics

[Agian]

I don't know how to word this question in a scientific way. So excuse my ignorance.

When they say someone has a 1% chance per year of having an 'event', does that mean that this risk accumulates? I mean 2% after two years etc. I'll give an example: Let's say someone has a 1% chance per year of having a stroke. Does that mean that after 50 years the chances of having had a stroke are 50%? I don't think this is how it works; but I'm struggling to get my head around it.

 

[Ola Thoresen]

In theory, this is how it works. However, this is statistics, so it doesn't work on an individual level. IE. it is not 100% chance to have a stroke after 100 years.

It is probably more correct to think about it on a group level. So if you have 100 people with 1% chance, statistically 1 of them will have a stroke every year.
But then again, 100 people is a too small group, but if you find 1000 people, then the probability of 10 of them having a stroke each year is quite high.

 

[Ola Thoresen]

Oh, and also. If you do think of it as a group, another interesting point is that if you have a group of 1000 people, and on average 1% (10) of these have a stroke on year one. You can not deduct that ALL of them will have a stroke in 100 years. You would have to add another 10 people to the group the next year, and again, 10 of the 1000 would have a stroke (possibly one of the newly added). And this should be done every year. So in reality, you could end up with a few of the people from the original 1000 still not having had a stroke after 100 years.
Statistics is fascinating - but it is still just statistics...

 

[pellicle]

When they say someone has a 1% chance per year of having an 'event', does that mean that this risk accumulates?

its rather like flipping a coin ... you have a 50% chance of getting a head. So if you flip it again its not dependent on the last flip (unless you're playing 2 up at the Crown Casino)

 

[Agian]

Thanks guys. That's what I thought; but the cardiologist I saw today freaked me out. You'd understand why it's relevant to most if not all of us. I was told that if he patched up the hole in my heart, the recurrence rate would approach the 1% per year rate of someone who has had a 'cryptogenic stroke' without a pfo (d'oh).

Then I thought about some of the other estimates I've read about here. I've realised a lot of this heart stuff is a bit of a percentage game.

 

[pellicle]

-I've realised a lot of this heart stuff is a bit of a percentage game.

without a word of a lie all of life is a percentage game. Know a fella who's wife died the day after giving birth, from a brain aneurysm ... while in hospital. Some of us take a number and wait, others get called unexpectedly.

In reality, ya just never know anything

 

[newmitral]

When they say someone has a 1% chance per year of having an 'event', does that mean that this risk accumulates? I mean 2% after two years etc. I'll give an example: Let's say someone has a 1% chance per year of having a stroke. Does that mean that after 50 years the chances of having had a stroke are 50%? I don't think this is how it works; but I'm struggling to get my head around it.

Sorry to those answering above, but that is NOT how the math works, and those answers are only partially correct.

You must multiply the probabilities of individual independent events to get the total probability, you do not add them.
So, using the numbers in your example, if there is a 0.01 probability of a stroke per year, then there is a 0.99 probability of NOT having a stroke in a given year. To go 50 years without a stroke, you have a probability of 0.99 to the power of 50, which equals 0.605. So if the probability of NOT having a stroke in 50 years is 0.605, then the probability of HAVING a stroke in 50 years would be 1.0 - 0.605 which equals 0.395 or 39.5% - NOT 50%.

I gave a more in-depth explanation of how this works about a year ago in an answer to another fellow in this thread:
http://www.valvereplacement.org/forums/showthread.php?41083-understanding-stats-and-probability-discussion-and-tutorials

 

[pellicle]

Sorry to those answering above, but that is NOT how the math works, and those answers are only partially correct.
...
I gave a more in-depth explanation of how this works about a year ago in an answer to another fellow in this thread:
http://www.valvereplacement.org/forums/showthread.php?41083-understanding-stats-and-probability-discussion-and-tutorials

and look, here I am (that fellow) again making that same mistake!

thanks for pointing out my error (again). I will try to remember it this time

To go 50 years without a stroke, you have a probability of 0.99 to the power of 50, which equals 0.605. So if the probability of NOT having a stroke in 50 years is 0.605, then the probability of HAVING a stroke in 50 years would be 1.0 - 0.605 which equals 0.395 or 39.5% - NOT 50%.

I was thinking in strict probabilities where there is no relationship between events. Ignoring that it was for the probablity over a 50 year period.

 

[newmitral]

Hi Pellicle,

Although you participated in the earlier discussion, you were not the one that made the mistake before. That discussion was to correct the statistical errors made by a different user (Fundy).

 

[dick0236]

.........then the probability of HAVING a stroke in 50 years would be 1.0 - 0.605 which equals 0.395 or 39.5% - NOT 50%.

]

These discussions regarding stroke probabilities, while mathematically interesting, have questionable relevance in the "real world". After 46+ years post surgery, I have a probability of around 1% of having a stroke THIS year. While it may be statistically accurate to say that my probability of having HAD a stroke sometime within my post-op years might be around 46%, it is around 1% THIS year and will be 1% next year.......actually, my personal probability of having HAD a stroke has been 100% since seven years post-op (stroke in 1974, due 100% to ignorance)....or, my probability of NOT HAVING HAD a stroke in the past 39 years has been 100%. You know the old saying.....figures don't lie, but they can often be massaged to prove any point.

 

[pellicle]

You know the old saying.....figures don't lie, but they can often be massaged to prove any point.

Virst, wee shall torture zee data ... und zen we will extract what we need from it.

Igor! Pass me zee elektrodes und zee Excel sheet

 

[neil]

sometimes you gotta look beyond statistics and live you life,

 

[Rachel]

Virst, wee shall torture zee data ... und zen we will extract what we need from it.

Igor! Pass me zee elektrodes und zee Excel sheet

LOL Pretty funny!!!

 

[Agian]

Sorry to those answering above, but that is NOT how the math works, and those answers are only partially correct.

You must multiply the probabilities of individual independent events to get the total probability, you do not add them.
So, using the numbers in your example, if there is a 0.01 probability of a stroke per year, then there is a 0.99 probability of NOT having a stroke in a given year. To go 50 years without a stroke, you have a probability of 0.99 to the power of 50, which equals 0.605. So if the probability of NOT having a stroke in 50 years is 0.605, then the probability of HAVING a stroke in 50 years would be 1.0 - 0.605 which equals 0.395 or 39.5% - NOT 50%.

That's scarily almost 40%.

 

[newmitral]

That's scarily almost 40%.

Hi Agian,

It is only scary if you believe that the annual probability number you started with applies to you.

Remember, I showed you how to do the math properly, but I made the major caveat:
using the numbers in your example

Personally, I don't believe that 1% per year number is valid as a starting point, at least for us typical valvers. I agree with Dick0236, that these numbers have to be treated skeptically. That 1%/year number gets thrown around a lot for valvers, but as far as I can tell it seems to have its roots in older studies where the population was generally elderly and had other health issues as well, did not necessarily stay within their INR ranges (if applicable) and had older valve designs. I can't speak to "cryptogenic stroke" or "pfo" issues and how that effects risk, but I'm sure you can research that.


More modern heart valve studies (I referenced a few on the other thread shown in my prior post) indicate that a person with a mechanical valve who maintains proper INR has about the same stroke risk as a patient with a tissue valve. It's important to note that a patient with a tissue valve does not have a 0% stroke risk either. Furthermore, someone who has never had any heart surgery and still has their "factory original" native valves and other parts still does not have a 0% stroke risk.

From the "National Stroke Association" fact sheet:
"Stroke can happen to anyone at any time, regardless of race, sex or age."

All you can do, whether you have a heart problem or not, whether you have a tissue valve or a mechanical valve, or your native valve, is to take whatever steps are reasonable and appropriate to minimize your risks. This is no different from driving a car, crossing a street, or approaching any health issue or life issue in general.

I don't want to get hung up in philosophy, but I mainly wanted to show you how to do the math properly. The numbers you start with have to be thought through very carefully to determine whether they apply in your specific case or not. That is much harder than getting the math right.

[Edit-update]
I just did a quick bit of googling and found:
Other terms used in the literature to describe cryptogenic stroke include cryptogenous stroke and infarcts of unknown, uncertain, or undetermined cause.
So what your doctor appears to be saying, at least to put a positive spin on it, is that your risk of stroke after he does his repair would be about the same as that of anyone else - those without any known high risk factor. I would take that as a great answer and not worry about the numbers.

 

[Agian]

I'm starting to think that Warfarin is a good deal overall. If it wasn't for the 'risk of hemorrhage', I think it would be more commonly suggested. However, this too can be minimised through self-monitoring. I actually asked to go on Warfarin after my minor eye TIA.

 

[AZ Don]

Newmitral has the math right. I can't really comment on the accuracy (or inaccuracy) of the annual risks assessments, but risks certainly do accumulate over time. I think the statistics are more relevant for those in the waiting room, where there is a decision to be made about when to intervene surgically. Some risks can be mitigated, at least somewhat. Take stroke risk for example, this is a very real risk for people without valve issues and there are a number of accompanying risk factors within our control.