Step I odds ratio??? help
Started by sunflower2020
You look back at some company records for a manufacturing company and see that of the 100 people that worked in the chemical factory for 10 years, 10 developed liver cancer. Of the 100 people that worked in the office for 10 years and were not exposed to chemicals, only 1 developed liver cancer. What is the chance that a newly hired person will develop liver cancer after 10 years depending on whether or not they work in the factory or the office?
...............cancer....no cancer
factory........10..........90.......
office............1..........99......
The odds of developing cancer in the factory is 10:90 and in the office is 1:99 so the odds ratio is 10/90 divided by 1/99 or 10*99/1*90 or 11:1
11 people with cancer worked in the factory for every 1 that worked in the office.
In other words if you had a person with cancer in front of you, you would break even in the long run if you bet 11:1 that they worked in the factory.
...............cancer....no cancer
factory........10..........90.......
office............1..........99......
The odds of developing cancer in the factory is 10:90 and in the office is 1:99 so the odds ratio is 10/90 divided by 1/99 or 10*99/1*90 or 11:1
11 people with cancer worked in the factory for every 1 that worked in the office.
In other words if you had a person with cancer in front of you, you would break even in the long run if you bet 11:1 that they worked in the factory.
GOT IT!! For some reason I made it into a bigger dilema than it really is. I hope it's stress, but everything just looks foreign to me.
Thank you!!!
Thank you!!!
Advertisement - Members don't see this ad
There is a mistake above which I corrected. It was just my last statement. I used the calculated odds ratio to make a statment about relative risk and it should have been about odds.
Don't confuse odds ratio with relative risk. Relative risk is for a cohort study and is sometimes called relative ratio (I know, way too confusing). In that case you would start with the two groups of people and follow them for 10 years. Then take the ratio of the incidence rates (not the ratio of odds). So in this case it would by (10/100)/(1/100) = 10
Your risk of developing cancer by working in the factory is 10 times greater than your risk of developing cancer when working in the office.
In the case of a disease that is not too prevalent, the odds ratio is an approximation of the relative risk.
Don't confuse odds ratio with relative risk. Relative risk is for a cohort study and is sometimes called relative ratio (I know, way too confusing). In that case you would start with the two groups of people and follow them for 10 years. Then take the ratio of the incidence rates (not the ratio of odds). So in this case it would by (10/100)/(1/100) = 10
Your risk of developing cancer by working in the factory is 10 times greater than your risk of developing cancer when working in the office.
In the case of a disease that is not too prevalent, the odds ratio is an approximation of the relative risk.
Im gonna try to rephrase this to see if i get it, please correct me if i dont alright?
Relative Risk : if one out of ten in group A and 2 out of ten in group B get something, the relative risk is 1:2 because basically youre comparing the likelyhood of getting something if youre in one group versus if youre in another
Odds ratio... is a ratio of the odds in each group, as above would be [(1/10) / (2/10) ]
ok if this is right then can someone explain in english what odds ratio tells you...aka why do we care?
thanks
~Brooklyn
Relative Risk : if one out of ten in group A and 2 out of ten in group B get something, the relative risk is 1:2 because basically youre comparing the likelyhood of getting something if youre in one group versus if youre in another
Odds ratio... is a ratio of the odds in each group, as above would be [(1/10) / (2/10) ]
ok if this is right then can someone explain in english what odds ratio tells you...aka why do we care?
thanks
~Brooklyn
The odds ratio is the only way you can report your results if you aren't doing a cohort study. Retrospectively you are looking at people with a disease and saying how many were exposed to something and how many weren't. So that when someone comes in with say, mesothelioma, you can say there is a 20:1 shot they were a shipbuilder or some such nonsense. It really tells you jacksquat except that many studies are done retrospectively and hence this data must be reported in these terms...and statistically it is a good approximation of the relative risk.
BrooklynDO said:
Close, except that you would say that you have twice the risk of getting a disease being in Group B than Group A. You would not report it as ratio since it is telling you the risk, not the odds.
soo,
if a question is talking about a population of some cities and the incidence rates in them then it is safe to assumue that you are asked about relative risk.
if a question is talking about a small study, smaller sample, where a doctor or you are performing at a hospital, then most likely it is going to be odds ratio.
practically, it all depends on the study that you are using.
Name the study then calculate according to it.
You think it's ok to think like that to answer these questions?
if a question is talking about a population of some cities and the incidence rates in them then it is safe to assumue that you are asked about relative risk.
if a question is talking about a small study, smaller sample, where a doctor or you are performing at a hospital, then most likely it is going to be odds ratio.
practically, it all depends on the study that you are using.
Name the study then calculate according to it.
You think it's ok to think like that to answer these questions?
If it's restrospective, meaning looking back in time or in other words a case-control study, then you would use the odds ratio. If it is a cohort study...looking forward, meaning you select a group of people and say half of these will get a medication and half won't and so forth then you use relative risk.
