Mortality rates—the nuts and bolts

Mortality rates are among the most important indicators in epidemiology. They are used in order to express the risk of dying of a certain disease. In this short 6-minute video from our Epidemiology Essentials course, you'll learn the nuts and bolts of mortality rates:

• What mortality rates are
• How to calculate them
• How to calculate mortality rates for various subgroups
• What proportionate mortality is
• Why proportionate mortality is NOT a measure of risk

Understanding mortality rates is crucial if you want to understand and interpret the medical literature.

Don't forget to sign up for a free trial account!

[00:00:00] Hey, everyone. In this video, we're going to learn something very, very important and that's mortality rates and how they are calculated. Check it out. Let's have a quick look at how epidemiologists measure mortality. Let's take a look at some fictitious numbers again. Here, are the absolute numbers of death for certain disease over time. From looking at this graph, we cannot tell if the risk of dying from that disease increased over time. For that, we would need to know the death rates and we only have the numerator for these death rates, shown here.

[00:00:30] We don't have the numbers of the total population during these time points or in other words, we don't have the denominator. So, if the population would have increased at the same pace, the death rates would have essentially stayed the same. So, if the denominator increases to the same extent as the numerator, the relationship stays the same and in this case, the rates would stay the same. Phrased differently, the absolute number of deaths is the numerator and the population is the denominator. If those two

[00:01:00] increased proportionally, as in this case, the rate stays the same. So you see, absolute numbers are not an indication of mortality risk but rates are. In the following minutes, we're going to cover three measures of mortality. Two of them are rates, one is not. We're going to cover the mortality rate, the case fatality rate, and the proportionate mortality. Let's start out with mortality rates. There are a couple of ways to approach them. The most obvious way would be to calculate the so-called all-cause mortality rate,

[00:01:30] which can be expressed per 1000 population, 10 000 population or 100 000 population. That's totally up to the beholder. The annual all-cause mortality takes the number of deaths from all causes, in a specified time period. Let's say in one year and divides that by the number of all persons in the population at mid-year. Why mid-year? Well, because the population changes over time. Remember, everyone who's in the denominator should be eligible to become part of the numerator.

[00:02:00] In order to obtain the number per 1000 population, we multiply that by 1000. We might not be interested in the entire population but in the mortality rate of a specific subgroup. Let's say we'd like to know the annual all-cause mortality for adults aged 30 to 40 years. Well, then we take the number of deaths from all causes in one year, that occurred in the age group of 30 to 40 and we divide that by the number of persons 30 to 40 years of age, at mid-year. Again, we multiply that by 1000 in order to get

[00:02:30] the number of deaths per 1000 population. Similarly, we might want to get the mortality rate for a specific disease like heart disease. In other words, the annual mortality for heart disease per 1000 population. We take the number of deaths from heart disease in one year, divide that by the number of persons in the population at mid-year and multiply that by 1000. I think you're getting the gist of it. Now, let's take that a step further and combine the two previous mortalities and calculate the annual mortality

[00:03:00] from heart disease, per year, in adults 30 to 40 years of age, per 1000 population. We take the number of deaths from heart disease in one year, that occurred in the age group of 30 to 40-year-olds and divide that by the number of persons 30 to 40 years of age at mid-year and multiply that by 1000. Now, let's turn to case-fatality rate. Case-fatality rates are defined as the proportion of people who have a disease and who are dying from it, in a specified time period. It's actually a measure

[00:03:30] of disease severity. It's calculated as the number of deaths during a specific time period after its onset, let's say heart disease, divided by the number of persons with heart disease in that period. Case-fatality rates are a great way to measure the benefits of a new drug or intervention. Now, let's turn to proportionate mortality. That's defined as the proportion of all deaths that die from a certain disease. Let's take heart disease, again. So, the proportionate mortality from heart disease, in this specific time period, would be calculated as the

[00:04:00] number of deaths from heart disease, divided by the total deaths, multiplied by 100, in order to get a percentage. Let's take a fictitious example again. Proportionate mortality is often displayed in a form of a bar graph, where the entire bar represents 100% of deaths and a certain proportion of that 100% died of a certain disease. Let's pick heart disease, again. So, let's say that the proportionate mortality was 50% in the population in city A. Now, let's take city B and draw the

[00:04:30] proportionate mortality for it. You see, it's also 50%. So, does that mean that the risk of dying from heart disease is the same in those two cities? Think for a moment. Pause the video for 20 to 30 seconds then come back. Okay, the solution is that proportionate mortality in and of itself is not a measure of risk. You would need to know the total mortality rates in both cities in order to say anything about risk. However, if I told you that total mortality per 10 000 population was 10, in

[00:05:00] city A and 20 in city B; knowing that the proportionate mortality was 50% in both cities, you'd be able to calculate that in city A 5 people died of heart disease per 10 000, whereas in city B 10 died of heart disease. So, the risk of dying of heart disease was actually double in city B. You wouldn't have been able to say that just from looking at the proportionate mortalities. Let's pick another example. Let's say that the proportionate mortality from heart disease was 20% in city X and 40% in city Y.

[00:05:30] As we've just learned these numbers are actually not telling you that there is an increased risk of dying from heart disease in city Y. You'd need at least to know the overall mortality rate in both cities. So, let's say that the overall mortality rate was 10 per 10 000 in city X and 5 in 10 000 in city Y, 20% of 10 is 2, and 40% of 5 is also 2. So, there's an equal risk of dying in both cities. So, proportionate mortality is

[00:06:00] not an indicator for risk of dying but what it provides you with is a quick look at the relative contribution of the different causes of death. So, I hope you found this useful. As always, I'd love to hear your thoughts so please leave a comment below.