Prevalence Ratio & Prevalence Difference
Suppose we had data from a cross-sectional survey that had asked whether people currently smoked and whether they currently had problems with wheezing and coughing. You could use a 2x2 table to organize the data from the 100 respondents as in the table below:
|
|
Wheeze/Cough (yes) |
Wheeze/Cough (no) |
Total |
|---|---|---|---|
|
Smokers |
13 |
13 |
26 |
|
Non-smokers |
2 |
72 |
74 |
- Prevalence of wheezing and coughing among smokers = 13/26=0.50.
- Prevalence of wheezing and coughing among non-smokers = 2/74=0.027.
Prevalence Ratio = PR = 0.50/0.027 = 18
Interpretation: Smokers had 18 times the prevalence of wheezing and coughing compared to non-smokers.
Prevalence Difference = PD = 0.5-0.027=0.473 = 47.3 per 100
Interpretation: Among smokers there were 47 excess cases of wheezing and coughing per 100 compared to non-smokers in a given time period.
Based on this cross-sectional data, I might hypothesize an association between smoking and respiratory problems, but the temporal relationship is unclear, and an analytical study would be a good next step.