Incidence And Prevalence Flashcards
What is prevalence and how can it be expressed?
A measure of how common a disease is
Can be expressed as:
- Percentage
- Number per n people
What are the 3 types of prevalence? Define them.
Point: proportion of individuals with the condition at a specified point in time
Period: During a specialised time interval
Lifetime: At any point in their lives
If you just see the word “prevalence” without qualification, what type of prevalence does it usually refer to?
Point prevalence
How do you calculate prevalence?
No. of people with condition
__________________
No. of people in total
= x (x 100 to get %)
How is prevalence presented when it is low?
Per 1000 or 100,000
What is prevalence used for?
To gauge the burden of disease (but can be affected by disease duration)
What is incidence and how is it expressed?
The rate at which new events occur in a population over a defined period of time
Can be expressed as:
- Per n people per time period
- Per n person-years
What is person-years? How do you calculate it?
A measurement combining the number of people observed and the number of years they were observed for = no. of people x no. of years
How do you calculate incidence?
Number of new cases
___________________
Number of people x years observed
= x (x 1000 for e.g. if for 1000 person-years)
What types of conditions would have high incidence and high prevalence?
Common, not brief conditions e.g. common cold
What types of conditions would have high incidence and low prevalence?
Common, very brief conditions e.g. nose bleeds
What types of conditions would have low incidence and high prevalence?
Uncommon, long-term conditions e.g. type 2 diabetes
What types of conditions would have low incidence and low prevalence?
Uncommon, short term conditions like pancreatic cancer
What are the 4 factors that affect prevalence? Think about the epidemiologists bathtub.
Incidence rate
Recovery (cure) rate
Death/mortality rate
Transfer (migration) rate (bidirectional)
What is statistical inference?
When a conclusion is made about a population after a sample is taken and experiments have been performed on them i.e. given that we cannot know the truth, we make a best guess based on data also describing our level of uncertainty around the best guess
What is point estimation?
Best guess based on sample data (true result is unlikely to be exact is experiment is repeated but probably similar)
What is sampling error?
If the same experiment is repeated multiple times, point estimates from each experiment can be different but will be clustered around true value so the difference between the sample point estimates + truth = sampling error
How can you get rid of/reduce sampling error?
To eliminate it, you would have to test the whole population which is rarely feasible so, you can test a larger sample to make your best guess closer to the truth
What is standard error?
Numerical value that represents the sampling error (can be calculated) -> large SE suggests best guess may be far from truth (& vice versa)
The larger your sample size, the ____ the standard error will be.
Smaller
What are confidence intervals?
The range of plausible values to present level of uncertainty of estimate from a sample i.e. a 95% CI would be you saying you are 95% confident that in reality, the values lie between them 2 values stated - 95% CI includes all values within 1.96 SEs of the point estimate
How do you calculate the lower + upper bound confidence intervals?
Lower = point estimate - (1.96 x S.E.) Upper = point estimate + (1.96 x S.E.)
Why are we given the S.E. equation in the exam?
Because it is variable unlike the CI calculation
Study regarding what side mothers held their babies on preferentially: If a 95% CI for this study was (0.76, 0.84) and the sample proportion was 0.8, how would you interpret this in words?
We found that 80% of our mothers held their babies on the left. However, we can be 95% confident that the true proportion of mothers that hold their babies on the left is somewhere between 76% and 84%.
How can you interpret confidence intervals?
The width of CI gives an indication of how precise our estimate is i.e. a wide CI means you cannot be precise about the truth but larger samples should result in a narrower confidence interval which is more reassuring.
How can confidence intervals be used to compare 2 groups?
If the confidence intervals of 2 different groups overlap, there is no statistically significant difference between the 2 groups e.g. (95% CI [15%, 25%]) + (95% CI [14%, 22%])
The local council wants to determine if there has been a recent reduction in the prevalence of teenage smoking.
In 2010, 20 out of 50 sampled teenagers smoked.
In 2015, 15 out of 60 sampled teenagers smoked.
Work this out.
Standard error: Square root of p (1-p)/n
Prevalence in 2010 = 20/50 = 40%
Prevalence in 2015 = 15/60 = 25%
SE 2010 = SQR (0.4(1-0.4))/50 = 0.07
SE 2015 = SQR (0.25(1-0.25))/60 = 0.06
CI 2010 = 0.4 +/- (1.96 x 0.07) = (0.26, 0.54)
CI 2015 = 0.25 +/- (1.96 x 0.06) = (0.13, 0.37)
In 2010, the prevalence of teenage smoking was 40% although a plausible range for the true prevalence values is 26% to 54%. In 2015, we estimate the prevalence of teenage smoking to be 25% although a plausible range for the truth prevalence values is 13% to 37%. Thus, it is plausible that there has been no change as intervals overlap.
