ChatGPT actually gives a very good explanation of this:

Person-years is a measurement used in epidemiology and demographic studies to estimate the number of people at risk for a certain event (like death or disease) over a specific period. It helps to account for both the number of people in a population and the length of time they are at risk.

To calculate deaths per 100,000 person-years, follow these steps:

1. Determine the study population: Identify the group of people you are studying, such as a specific age group or people with a particular health condition.

2. Calculate the total person-time at risk: This is the sum of the time that each person in the study population is at risk for the event of interest (in this case, death). Person-time at risk is usually measured in person-years.

To calculate person-time at risk, multiply the number of people in the population by the number of years they are observed. If individuals have different follow-up times, you would sum the individual follow-up times for each person in the population.

For example, suppose you are studying a group of 1,000 people over a 5-year period. The total person-time at risk would be:

1,000 people x 5 years = 5,000 person-years

3. Count the number of deaths: Determine the total number of deaths in the study population during the observation period.

4. Calculate the death rate: Divide the number of deaths by the total person-time at risk.

For example, suppose there were 50 deaths during the 5-year period. The death rate would be:

50 deaths / 5,000 person-years = 0.01

5. Calculate the death rate per 100,000 person-years: Multiply the death rate by 100,000.

0.01 x 100,000 = 1,000 deaths per 100,000 person-years

So, in this example, the death rate would be 1,000 deaths per 100,000 person-years.