Minor Biostat Question.

[Old Style Nanny]

---

Members don't see this ad.

Why does FA say Prevalence is cases/population while Incidence is cases/population at risk? I was under the impression both had the same at risk population as the denominator?

Thanks.

 

[Phloston]

Why does FA say Prevalence is cases/population while Incidence is cases/population at risk? I was under the impression both had the same at risk population as the denominator?

Thanks.

"Population at risk" means those with the disease are no longer at risk to get it because they already have it. For example, in a population of 1000, if 20 get the disease during 2010, the incidence is 20/1000 for 2010. However, in 2011, the at risk population is now only 980, so if 20 more get the disease, the incidence is 20/980.

 

[Old Style Nanny]

"Population at risk" means those with the disease are no longer at risk to get it because they already have it. For example, in a population of 1000, if 20 get the disease during 2010, the incidence is 20/1000 for 2010. However, in 2011, the at risk population is now only 980, so if 20 more get the disease, the incidence is 20/980.

Let us forget about the actual terms, and get a clearer understanding. Hypothetically, let us say there is a disease that only affect the males in the population. Then, is prevalence = cases/males in population or cases/population. That is my question. Thanks.

 

[Phloston]

Let us forget about the actual terms, and get a clearer understanding. Hypothetically, let us say there is a disease that only affect the males in the population. Then, is prevalence = cases/males in population or cases/population. That is my question. Thanks.

My guess would be that the denominator for prevalence refers to the entire population unless otherwise specified.

If it's 100% known that the disease only affects males, then the prevalence of the disease in the population still includes females in the calculation.

I'd be happy to hear someone else's thoughts on this too.

 

[withrye]

They're both population at risk. Prevalence for prostate cancer is does not include women in its calculation. The only difference between prevalence and incidence is that prevalence is everyone who is currently in disease state, while incidence is new acquired disease during a time period. You basically just need to know that for chronic diseases that don't kill you, prevalence tends to be much higher than incidence. For acute diseases that resolve quickly that you can catch again and again, they tend to be equal.

Lastly, if you have a chronic disease where prevalence > incidence, and you start treating end-stage mortality outcomes of the disease, prevalence will go up because those people are living longer while incidence stays the same (you haven't changed the epidemiology of the original disease).

 

[Old Style Nanny]

They're both population at risk. Prevalence for prostate cancer is does not include women in its calculation.

If you are sure about this, it answers the question completely.

 

[Phloston]

I'm not so sure it's that simple.

With XR diseases, women are not excluded due to the possibility of skewed lyonization.

In the case of Rett's, for instance, where only females are affected, if you asked, "what is the prevalence of Rett's in the United States?" It would be the # of disease cases / total population. I believe you'd actually have to say "what is the prevalence of Rett's in females?" for the at-risk population to be assessed only.

 

[withrye]

I'm not so sure it's that simple.

With XR diseases, women are not excluded due to the possibility of skewed lyonization.

In the case of Rett's, for instance, where only females are affected, if you asked, "what is the prevalence of Rett's in the United States?" It would be the # of disease cases / total population. I believe you'd actually have to say "what is the prevalence of Rett's in females?" for the at-risk population to be assessed only.

Of course not. The prevalence of Rett syndrome includes everyone because everyone's theoretically at risk. When you consider the prevalence of breast cancer, both genders are included in the at risk group. The whole "population at risk" thing isn't meant to be tricky. It's meant to take out people once they've gotten something they can only get once (mumps) or to remove a gender if it's impossible to get the disease. Otherwise, practically everyone is counted as "at risk."

The prevalence of ovarian and prostate cancers only includes women and men, respectively. It simply doesn't make sense as a statistic otherwise.

 

[Phloston]

Of course not. The prevalence of Rett syndrome includes everyone because everyone's theoretically at risk. When you consider the prevalence of breast cancer, both genders are included in the at risk group. The whole "population at risk" thing isn't meant to be tricky. It's meant to take out people once they've gotten something they can only get once (mumps) or to remove a gender if it's impossible to get the disease. Otherwise, practically everyone is counted as "at risk."

The prevalence of ovarian and prostate cancers only includes women and men, respectively. It simply doesn't make sense as a statistic otherwise.

Everyone's not theoretically at risk for Rett's. It only affects females because males die in utero.

 

[withrye]

Everyone's not theoretically at risk for Rett's. It only affects females because males die in utero.

http://www.ncbi.nlm.nih.gov/pubmed/11392517

 

[Phloston]

http://www.ncbi.nlm.nih.gov/pubmed/11392517

Interesting. Thanks for posting that.

 

[withrye]

Interesting. Thanks for posting that.

It's trivia, and I was just being difficult. You usually see prevalence reported with the relevant population. "Prevalence of Rett Syndrome is X in XXX in females." It's almost nonexistent in males, so it's not really useful to dilute the prevalence in half just for those few cases.