# Confidence interval

Watch
Announcements
#1
It says: Patients with diabetes had a reduction in risk (95% CI 0.45-0.83) of heart disease.

What does this mean?
0
5 years ago
#2
It means the 95% confidence interval is 0.45-0.83.
confidence interval is
Mean + or - z*standard error

Posted from TSR Mobile
0
#3
(Original post by simonli2575)
It means the 95% confidence interval is 0.45-0.83.
confidence interval is
Mean + or - z*standard error

Posted from TSR Mobile
but what does the 0.45 0.83 stand for ? is it the concentration the therapy would be effective?
0
5 years ago
#4
No, they're the lower bound and the upper bound of this 95% CI.
With 0.45, it means mean - z*s.e. = 0.45

Posted from TSR Mobile
0
#5
(Original post by simonli2575)
No, they're the lower bound and the upper bound of this 95% CI.
With 0.45, it means mean - z*s.e. = 0.45

Posted from TSR Mobile
Ive never done maths as a subject can you dumb it down and tell me what this means in the context of the question?
0
5 years ago
#6
(Original post by asaaal)
Ive never done maths as a subject can you dumb it down and tell me what this means in the context of the question?
I'll try.

It means in 95% of the time, the confidence interval will contain the true mean, so we're 95% confident that the true mean lies between 0.45-0.83.

Posted from TSR Mobile
0
#7
(Original post by simonli2575)
I'll try.

It means in 95% of the time, the confidence interval will contain the true mean, so we're 95% confident that the true mean lies between 0.45-0.83.

Posted from TSR Mobile
so the mean of whatever I'm measuring? so the concentration of my drug in this case would be between 0.45-0.83?

if i have two studies that measured the same thing in different ways and lets say one gave a 95% CI of 0.56 - 0.70 and the other 0.60-0.71 does that increase the validity of the study at the CI's overlapped?
0
5 years ago
#8
Maybe. Though Maybe instead of using their values of CI, I would instead use all of the data to find a new confidence interval.

Posted from TSR Mobile
0
#9
(Original post by simonli2575)
Maybe. Though Maybe instead of using their values of CI, I would instead use all of the data to find a new confidence interval.

Posted from TSR Mobile
How would i do that?

Thank you so much by the way, life saver !
0
#10
does anyone else happen to know as I'm pretty desperate
0
5 years ago
#11
(Original post by asaaal)
does anyone else happen to know as I'm pretty desperate
Hey,
The 0.45-0.83 numbers are are the upper and lower bounds for the risk reduction, not the concentration of the drug. Can you give a link to the paper? You'll need to know if it's relative or absolute risk reduction.
By the sounds of it the paper is saying that the drug has reduced heart disease in diabetics by somewhere in the range of 45%-85% (i.e. 0.45-0.83) compared to something else probably a control?. It's given as a range because of the uncertainty in the study so one cannot say it's reduced heart disease risk by one exact figure, the range (i.e. the confidence interval) is such that you can be 95% sure this is where the actual exact figure lies.
Hope that helped
0
5 years ago
#12
Sorry didn't see your other post,
But because the confidence intervals overlap,
that means the means (technically speaking not the means but the 'true value' for the risk reductions) of the two studies could be the same and hence there is no statistically significant difference between the two.
If the authors concluded there was no significant difference then it would be a valid conclusion but if they said there was then that would reduce the validity of the study.
0
#13
(Original post by ab192)
Hey,
The 0.45-0.83 numbers are are the upper and lower bounds for the risk reduction, not the concentration of the drug. Can you give a link to the paper? You'll need to know if it's relative or absolute risk reduction.
By the sounds of it the paper is saying that the drug has reduced heart disease in diabetics by somewhere in the range of 45%-85% (i.e. 0.45-0.83). It's given as a range because of the uncertainty in the study so one cannot say it's reduced heart disease risk by one exact figure, the range (i.e. the confidence interval) is such that you can be 95% sure this is where the actual exact figure lies.
Hope that helped
Here is one (and i will link the other on a different message as it exceeds the limit of one message).
Ive also attached my own table where i have collated the data from both studies do they are the same units i.e mmol. Now i want to go a step forward and combine data but idk how to !!

http://www.clinchem.org/content/42/12/1938.long
0
#14
(Original post by ab192)
Hey,
The 0.45-0.83 numbers are are the upper and lower bounds for the risk reduction, not the concentration of the drug. Can you give a link to the paper? You'll need to know if it's relative or absolute risk reduction.
By the sounds of it the paper is saying that the drug has reduced heart disease in diabetics by somewhere in the range of 45%-85% (i.e. 0.45-0.83) compared to something else probably a control?. It's given as a range because of the uncertainty in the study so one cannot say it's reduced heart disease risk by one exact figure, the range (i.e. the confidence interval) is such that you can be 95% sure this is where the actual exact figure lies.
Hope that helped
http://www.clinchem.org/content/42/12/1938.long

sorry it wouldn't let me attach it but I've found it online: this is the second paper: http://www.clinchem.org/content/42/12/1938.long
0
#15
(Original post by ab192)
Hey,
The 0.45-0.83 numbers are are the upper and lower bounds for the risk reduction, not the concentration of the drug. Can you give a link to the paper? You'll need to know if it's relative or absolute risk reduction.
By the sounds of it the paper is saying that the drug has reduced heart disease in diabetics by somewhere in the range of 45%-85% (i.e. 0.45-0.83) compared to something else probably a control?. It's given as a range because of the uncertainty in the study so one cannot say it's reduced heart disease risk by one exact figure, the range (i.e. the confidence interval) is such that you can be 95% sure this is where the actual exact figure lies.
Hope that helped
scrap my last messages and studies. So basically...

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2626030/ < This is study A and I would like to find the 95% CI of this

I have this statement from the study B which states: 'A study showed 1,708 patients with diabetes (approximately 33% of the total participants) showed fourty one percent reduction in the risk of cardiovascular events (95% CI: 0.45-0.8) - i am given this CI already in the findings of the study.

I want to compare CI of diabetic patients (as both of these studies used diabetics) who were given EDTA therapy.

I have no idea how to do this....

Q1. I was already given the cI for study B but do have to do anything differently because I am comparing to study A

Q2. how do i find CI of A?

Q3. What does it mean if their CI overlap?

Q4. Can i combine the data to get one CI?
0
#16
Or if anyone else can help id be grateful !!
0
5 years ago
#17
(Original post by asaaal)
so the mean of whatever I'm measuring? so the concentration of my drug in this case would be between 0.45-0.83?

if i have two studies that measured the same thing in different ways and lets say one gave a 95% CI of 0.56 - 0.70 and the other 0.60-0.71 does that increase the validity of the study at the CI's overlapped?
No,

Posted from TSR Mobile
0
X

new posts
Back
to top
Latest
My Feed

### Oops, nobody has postedin the last few hours.

Why not re-start the conversation?

see more

### See more of what you like onThe Student Room

You can personalise what you see on TSR. Tell us a little about yourself to get started.

### Poll

Join the discussion

#### If you don't put your camera on in online lessons, why is that?

My teacher doesn't want us to (8)
19.05%
No one else does (16)
38.1%
I'm embarrassed about my background (4)
9.52%
I feel self-conscious showing my face (12)
28.57%
We don't use a video platform (1)
2.38%
I don't have a camera (0)
0%
Something else (tell us in the thread) (1)
2.38%