>>Return to Tell Me About Statistics!
The “risk ratio” or “relative risk (RR)” is often found in epidemiological survey results by the Ministry of Health, Labor and Welfare, academic societies’ published guidelines, and medical papers.
The “odds ratio (OR),” is very similar to the risk ratio and the difference between these two metrics can often be confusing.
Here, we will explain the difference between risk ratio and OR.
What is risk ratio (RR)?
The risk ratio is “a measure of the degree of risk of contracting a disease between a person exposed to a certain condition and a person not exposed to it.” Here, the risk is the rate (probability) of contracting a disease.
For example, Table 1 looks at how much more likely smokers are to die from cardiovascular disease (stroke, myocardial infarction, etc.) than non-smokers.
[Table 1] Percentage of people who died of cardiovascular disease in the first 10 years of starting the survey
From these results, we calculate the proportion of deaths due to cardiovascular disease among smokers and non-smokers.
[Table 2] Contingency table for the survey in Table 1
As shown in Table 2, the percentage of deaths due to cardiovascular disease was 7% among smokers and 3% among non-smokers.
To determine how much more likely smokers are to die from cardiovascular disease than non-smokers, we divided the percentage of smokers by the percentage of non-smokers.
7 (%) ÷ 3 (%) = 2.3.
This value is the “risk ratio.”
In this way, the risk ratio can be obtained very easily, but what is important is not how to obtain the risk ratio, but how to interpret it.
In this example, the risk ratio is 2.3, which can be interpreted as “the risk (proportion) of smokers dying from cardiovascular disease is 2.3 times higher than that of non-smokers.” This is the RR, sometimes described as “RR = 2.3” in the medical literature.
Thus, the risk ratio (RR) is easy to interpret. The higher the value, the higher the risk of contracting a disease or dying for a person under a given condition compared to a person without the condition.
What is the odds ratio?
Odds is a familiar word because it is often used in gambling, such as horse racing.
Odds are defined as the proportion (probability) of a situation that is more likely to occur than another. We will explain the OR using the previous example.
The number of deaths from cardiovascular disease among smokers divided by the number of deaths among non-smokers are the “odds.” Similarly, odds are the number of smokers who survive divided by the number of non-smokers who survive.
[Table 3] Contingency table for calculating case OR
When focusing on the number of deaths due to cardiovascular disease, the number of deaths among smokers (700) is 2.3 times than among non-smokers (300), that is, the mortality odds are 2.3.
When focusing on the number of people who did not die, smokers(9,300) are 0.96 times than non-smokers(9,700), that is, the non-fatal odds were 0.96.
The important point here is that the ratio of the odds of fatalities to the odds of non-fatalities is called the “odds ratio.” It is calculated as 2.3 ÷ 0.96, which is 2.4. The OR is 2.4, suggesting that smoking status is an influencing factor in cardiovascular disease mortality.
The key here is: do not say smokers are 2.4 times more likely to die from cardiovascular disease than non-smokers! You cannot go wrong with this.
So, while we know that smoking is bad for health, we do not know how much more likely smokers are to die from cardiovascular disease than non-smokers. This is because the odds of smoking among those who die from cardiovascular disease and the odds of smoking among those who do not die does not tell us how much smoking increases the risk of dying from cardiovascular disease. It cannot be derived and can only be interpreted as a risk ratio (RR).
Thus, the OR may seem less useful than the risk ratio in clinical practice. However, in actual clinical research, the OR is often used. Some of you may wonder why ORs that are so difficult to understand are used in clinical research. This is because there is a specific way to use it.
Next time, we will explain how to use the OR.
>>Return to Tell Me About Statistics!
Comments are closed