Nursing homes and coronavirus - why the numbers don't add up

Testing delays and death certificate recording skew Covid-19 mortality rates

The alleged under-reporting of Covid-19 deaths has been raised as an issue in many countries, prompting questions about how accurate Irish mortality figures are.

The issue is of crucial importance because death rates are one of the best ways of measuring how well a country is doing in limiting the damage wrought by coronavirus.

Another way is to measure the number of tests carried out, but in Ireland this process has been subject to long delays and is therefore unreliable as a measure of performance in an Irish context.

Ireland currently ranks 12th in Europe for Covid-19 deaths per million of population, just below Sweden. Much has been made of Ireland's much lower death rate relative to the UK; however, both countries compare poorly to best-in-class performers such as Austria or Iceland.


Different countries collect data in different ways, all of them subject to shortcomings, it has to be noted, so comparisons carry caveats.

There are two main sets of issues relating to the possible under-reporting of Covid-19 deaths. The first relates to the location of death and has grown in importance with the rise in clusters of the disease in nursing homes and other residential centres.

Many of these locations, be they hospitals or other residential locations, have well-worn but differing reporting and surveillance mechanisms for deaths, particularly those associated with a disease outbreak.

While in some other countries deaths outside hospitals were or are not being recorded, it does appear as a rule that deaths here are being logged regardless of whether they happen in a hospital, nursing home or other location.

Residential locations

In residential locations, this happens as a consequence of the investigation by public health officials of local outbreaks of the disease. By Monday, Ireland had recorded 406 deaths, of which 266 had occurred in a hospital environment. Some 222 deaths, including 187 associated with nursing homes, were linked to other residential sites.

The numbers don’t add up because some nursing home patients were transferred to hospital before dying and so may have been counted in both nursing home and hospital categories.

Chief medical officer Dr Tony Holohan cautioned this is the "best understanding" of public health officials "at this moment in time" but carries a "small health warning". The figures refer to registered deaths of patients in community settings, he said.

The second issue relates to how a death is recorded on a death certificate. Many older people dying with Covid-19 also have underlying conditions and it is no easy task to determine what they died of, as opposed to what they died with.

Some years ago, there was controversy when a whistleblower alleged a hospital was failing to record the CPE superbug as a cause of death for elderly patients who had died. Obviously, it would be of concern if something similar were to happen with coronavirus.

‘Index of suspicion’

This seems highly unlikely in hospitals but what about other locations that attracted less attention in the early stages of the pandemic and where a realisation of the dangers posed by the virus was slower to sink in?

“What I couldn’t assure you is that every person who dies has a test for this particular illness or even with the symptoms associated with this illness,” Dr Holohan told The Irish Times on Monday.

Clinically, the “index of suspicion” for the disease would be “a good deal higher” than would normally be the case for flu, he added.

However, in explaining the process followed, Dr Holohan added: “So what we’re [doing] here is identifying the deaths in individuals where there is a laboratory confirmation of Covid-19.”

“We try to give as comprehensive a number as we possibly can,” he later added.

But what about suspected cases? People who had not been tested before they died?

According to HSE chief clinical officer Dr Colm Henry, the doctor filling out a death cert "will give his or her best opinion of the cause of death in the context of a new illness which is lab-confirmed".

From this it is clear our official figures are based entirely on laboratory-tested cases. Eventually, when death certs are filed within the required three-month period, further cases will be logged that add to the figures.

The Coroners Service has instructed doctors to swab test deceased patients with respiratory failure or “circumstantial suspicion” for Covid-19, but this has not been happening all the time, internal correspondence appears to show.

As a result, there is likely to be some level of under-reporting of deaths.

This situation arises with surveillance of many other infectious diseases. Over time, public health officials get around these shortcomings by calculating the excess deaths that have occurred, but this will not be calculable for some time.