# COVID-19 & Age in Ireland

## An Analysis of COVID-19 Cases and Hospital Admissions in Ireland based on Age Group

At the time of writing (1/4/2020) Ireland’s COVID-19 numbers have been moving in the right direction, with daily increase in cases, hospitalisations, and ICU bed all trending downwards since mid-March, when the current restrictions began. The restrictions are working but we need to keep going to remain on the glide path for a safe landing for the healthcare system.

The daily Irish COVID-19 datasets include information abou the age profiles of confirmed cases and hospitalisations so I thought it would be useful to have a look at what they tells us about the relationship between infection risk and age. We have been told that the seriousness of COVID-19 tends to increase with age, because in general those that require hospitalisation tend to be older, but how much additional risk do older citizens have to bear?

As we attempt to answer this question it is useful to have baseline age profile for the Irish population. The most recent official data on this comes from 2016 census and is presented below as a graph of the total number of people (plus, males and females) of a given age. While obviously these people are now ~4 years older we can expect the relative proportions to be similar today, and so we can use these data as a baseline for what follows. It is worth noting that the COVID-19 dataset groups cases and hospitalisations into age ranges, rather than yearly ages, andso we will be grouping the census data in a similar fashion.

In theory, the number of confirmed cases by age, compared with the population age data, should tell how likely someone with a given age (range) is to become infected. The census data tells us that approximately 13% of the population are between 25–34 years old, and so, all other things being equal, we might expect 13% of confirmed cases to be in this age range. However, as the graph below shows, about 17% of the cases are aged between 15 and 34; 30% higher than the percentage of 25–34 year-olds in the general population. In general, the graph shows a greater percentage of COVID cases among people >25 years old (the blue bars ) than we might otherwise expect, given the share of these age groups in the census data (the orange bars).

We might be tempted to conclude that this means there is an increased risk of people older than 25 becoming infected. However, this is not a safe conclusion to make: if younger people suffer from milder symptoms then they will be less likely to be tested, and therefore less likely to show up in the official case statistics.

Using hospitalisations instead of confirmed cases is one way to get a sense of this and the graph below shows the percentage of COVID-19 hospitalisations for each age-group, compared to the population baseline. This time it is people who are 45 years and older who make up a greater share of COVID-19 hospitalisations, than their share in the population as a whole. The difference is especially striking for those who are 65 and older. Over 40% of COVID hospitalisations are in this age group, even though the group makes up just over13% of the general population.

Finally, we can compare age groups in terms of their share of cases and their share of hospitalisations. This is a useful comparison to answet the following question:

If I have tested positively for COVID-19, then what is the likelihood that I will need to be hospitalised?

To answer this we need to calculate the number of hospitalisations as a proportion of the number of cases, per age group; this estimates the *percentage hospitalisation risk*. However, this is not quite correct because, presumably, there is usually a gap between a case being confirmed and a patient needing hospitalisation. If so, then we need to compare the number of hospitalisations today to the number of confirmed cases a few days ago.

But how many days ago? It’s difficult to know, and adjustment needed probably varies from country to country, depending on the status of testing, and perhaps even from age group to age group. For example, testing delays will truncate the time between a confirmation of diagnosis and a hospital admission; indeed some admissions might occur without a positive test. For this analysis we use a short 2-day delay so that the number of hospitalisations are divided by the number of cases two days ago when calculating the percentage hospitalisation risk.

As an aside, if we use a longer delay then the hospitalisation risk estimate increases for all of the age groups, because there are fewer confirmed cases. In fact, if we use 4 days or more for the adjustment, then we start to see risk estimates that are greater than 100%, for 65+ year olds, which should not be possible. If and when testing improves, then we should find more COVID-19 confirmations earlier in the disease progression and it will be possible to use longer adjustments for this calculation.

The results, using a the 2 day adjustment, are shown in the above graph. At the time of writing, there are 388 people in hospital in the 65+ age group, compared with 511 confirmed cases for this age group (2 days previously), leading to a hospitalisation risk of just over 75%. In contrast, the hospitalisation risk for the 45–54 and 55–64 age groups is closer to 30%, and it continues to decrease for younger age groups.

These risk values are undoubtedly over-estimating the true risk because of the lack of testing and the delays associated with the testing that is taking place. There is little doubt that the number of positive cases is considerably higher than as presented, which will lower the hospitalisation risk across all age groups. Nevertheless the estimates as presented serve to highlight the relative difference in hospitalisation risk across different age groups.