Thanks so much for this! Very clear explanation!! So my professor decides to include this image in his lecture slides but just insisted on explaining it his way, which took over 3 hours, when it could have been explained in less than 4 minutes.
Hi Rahul, great illustration, thanks a lot! In the pathtub model, there is only one way out (cure or death). I tried to distinguish between people between with and without the disease. Then, two possible exits are necessary. I tried to illustrate this by a "swimming pool model". Best regards, numbergazer
So what is the unit of time for incidence? Is it always agreed to be measured in years? So incidence= the number of new cases of disease in a year? (or is it months or something else?)
0:45 prevalence- how many people have the disease 2:29 increase pravl. 2:43 adds 2:55 p = incidence plus how long water stay in tub(people have disease for how long)
Peter, there will be a video on this soon. Here's how I think of p-value. Think of a court case, innocent until proven guilty. All the evidence makes it seem like the defendant is guilty. You are assuming the defendant is innocent. What is the chance that he is innocent AND all that evidence is true. It's possible but unlikely, maybe a 1% chance. That's what the p-value is - how likely is all this evidence true given that the person is innocent. [see next comment]
Prevalence isn't a rate, it's a single-unit measurement. Incidence is cases per time-unit, prevalence is a snapshot measurement of the number of cases at that time. As a comparison, salary is dollars per hour, savings is a snapshot of what you have in the bank currently.
I have a question here 2:07 . How does here prevalence is increasing if someone cures from the disease? Since he is still in the population so isn't it like cure shouldn't have any effect? However, I'm very grateful for the video. Thanks a lot.
He doesn't say prevalence increase if someone gets cured... he says it decreases (Because if someone is cured, they no longer have the disease so prevalence decreases). He does say that if we prevent death (NOT cure), then the prevalence increases because now those pple are living with the disease
So we arbitrarily picked a number, 5%. If the probability of all that evidence happening if the person is innocent is less than 5%, then he's probably not innocent. He's guilty. So we reject our initial assumption (innocent until proven guilty) if the p-value (probability that all that evidence is true AND he's innocent) is less than our previously set threshold (5%).
Q for epidemiologist: if vaccination reduces death by this bug in the study (1:53)... that would mean less water leave the tub, but would that also mean less water enters the tub (new infections). [if vaccination does not lower incidences or water level, why bother?] So... with less rate (incidence) & water (infected) in the tub, would vaccination DECREASE prevalence?
I'm not an epidemiologist, but vaccines are typically used to prevent disease (decreasing incidence, decreasing the water coming out of the spout). By that way, it would decrease prevalence. But you make a good point. Some vaccines do give partial immunity (like the flu vaccine when we don't predict the strain right) and so can make the disease less severe when contracted.
It is well explained! but I am taking courses to learn how to teach, and according to them, this is too boring to them. So... apparently I have to make this more fun, I am looking for a more entertaining class, and this is not helping because this is basically how I learn and teach.