The numbers game

When it comes to statistics, how you count matters. Nikki Macdonald and Georgia Forrester explain why our Covid statistics aren’t as clear-cut as they seem, and why hospitalisation numbers for children may be off.

At the 1pm Covid update on Wednesday, November 24, director-general of health Ashley Bloomfield finished up by emphasising that a number of the Kiwis hospitalised during the Delta outbreak have been children.

‘‘Forty-two children under age 12 have been assessed at hospital and, of these, 19 have spent at least 24 hours or longer in hospital being cared for; of those, one was between age 5 and 12, six were age 1 to 4, and 12 were under the age of 1. I just want to make the point that one of the best ways we can protect tamariki of all ages is for adults to be vaccinated.’’

But those statistics are flawed.

Hospitalisation numbers are probably a little inflated

The Health Ministry has confirmed that it counts any person in hospital with Covid as a hospitalisation, whether they’re there because of Covid-19 symptoms or complications, or for something completely unrelated.

This counting method has skewed statistics worldwide, particularly for reported hospitalisations in children. That’s because the majority of children who get Covid have either no symptoms, or mild symptoms. So there’s a higher chance the reason they’re in hospital has nothing to do with the fact that they have the virus.

Peter McIntyre, Otago University professor of child and women’s health, says overseas data clearly show that simply counting positive tests in hospitalised children is a poor indication of the numbers of children needing hospital treatment for Covid.

He points to a Hospital Pediatrics review emphasising the importance of distinguishing between children hospitalised with the Sars-Cov-2 virus, and those hospitalised for it.

‘‘Reported hospitalisation rates greatly overestimate the true burden of Covid-19 disease in children,’’ the authors conclude.

Data from the New South Wales outbreak found 37 per cent of reported hospitalisations in children under 18 were for ‘‘social and vulnerable’’ reasons, rather than medical ones. McIntyre says unpublished New South Wales data found about 80 per cent of children under 10 were in hospital only because their parents couldn’t care for them.

A Californian study found 40 per cent of children in hospital with Covid had no symptoms, and 45 per cent of admissions were unlikely to be because of Covid.

Stuff was blocked from talking to Starship Children’s Hospital paediatricians about what they are actually seeing in children with Covid.

The Health Ministry could not say how many of the cases cited by Bloomfield were hospitalised because of Covid, rather than with Covid. A spokesman says the data covers all children with Covid-19, even when that is not the main reason for hospitalisation, because it affects how they are managed in hospital and might impact on how quickly they recover.

Does this invalidate hospitalisation data?

No. Covid modeller Michael Plank, a Canterbury University professor of mathematics, says it’s important to count people who go to hospital for something other than Covid but test positive on arrival, as that can reveal undetected virus spread in the community.

However, you would ideally separate those cases out from people receiving hospital treatment for Covid, Plank says.

Otago University professor of public health Nick Wilson says it’s difficult to compare hospitalisation statistics from countries where Covid is widespread with those in New Zealand, where Covid is still pretty rare and therefore most hospital admissions of people with Covid are probably Covid-related.

‘‘But if Covid does become much more common in children in New Zealand (eg, if the 5 to 11-year-old vaccinations are delayed for too long) – then it would be good if the Ministry of Health website data included a category of ‘Covid-19 is main reason for hospital admission’ and ‘Covid-19 was not the main reason for admission’.’’

For adults, it might be more difficult to untangle whether hospitalisation is because of Covid, as an apparently unrelated event such as a stroke or heart attack could have been triggered by Covid infection, Wilson says.

‘‘In terms of useful information for the public, hospitalisation data on Covid-19 is still a useful guide to general disease severity.’’

Even the ministry acknowledges the stats are flawed, and is promising improvements.

‘‘In the coming weeks, the Ministry of Health will be making changes to its reporting to more accurately reflect the numbers of people who are in hospital for treatment for Covid-19 related complications, rather than Covid19 cases who have been admitted for unrelated health matters.’’

What about Covid-19 death stats?

If hospitalisation numbers are tricky, death should be simpler, right? But as with most things Covid, it’s complicated.

According to the ministry’s definition, a Covid death includes all cases that died who were classified as an active case of

Covid-19 at the time of death. In some of these cases, the underlying cause of death may have been unrelated to Covid.

Active cases are confirmed (or probable) cases that have not yet met the requirements to be considered recovered.

Deaths where the person had been deemed recovered (ie no longer infectious) but died as a result of complications secondary to Covid-19 are not recorded as

Covid-19 deaths.

This approach, says the ministry, is guided by the World Health Organisation’s definition for deaths due to Covid-19.

By this measure, as at 1pm yesterday, there were 44 deaths recorded in New Zealand since the pandemic began.

What about Covid deaths that are ‘under investigation’?

The ministry says it has recently changed the way it publicly reports deaths associated with the virus.

It still reports on Covid-19 deaths where it has been confirmed that the virus contributed to or caused the death, but it is now also reporting other deaths where the cause of death is not certain, but the person had acute Covid-19. These deaths are reported as ‘‘under investigation’’ until the medical reports accompanying the death have been assessed, and the deceased formally confirmed as having died of, or with,

Covid-19.

In future, Covid death reports from the ministry will be categorised as: those who died where

Covid-19 was the primary cause; those who had Covid-19, but it was not the primary cause of death; and deaths that are under investigation.

It’s also possible reports in future will include a little more detail about a case, such as their other health problems, though this will have to be done in a way that protects the private medical information of those involved. The ministry has recently started publishing more detail on the deaths in the Delta outbreak.

Why some cases are with the coroner?

In 2020, none of the Covid-related deaths were reported to the coroner. But amid the current Delta outbreak in New Zealand, a handful have been.

In a statement, Chief Coroner Judge Deborah Marshall explained that, in New Zealand, if a person dies while in care, such as in hospital or in an MIQ facility, and it is clear to the clinicians that the death is due to Covid-19, the doctor who provided the care will sign a medical certificate of cause of death (MCCD). As a result, the death will not be reported to the coroner or come into coronial jurisdiction.

But, if there is any uncertainty about the cause of death, such that the deceased’s clinician/GP is unwilling to sign the MCCD, then the death will be reported to the duty coroner, who may or may not decide to investigate, Judge Marshall says.

At the moment, there are eight cases referred to the coroner where the deceased had tested positive for

Covid-19, but none of these cases has been finalised. As a result, it is impossible to confirm at this stage whether the deaths were from Covid

19.

It can take a long time for coroners to determine their findings.

How do other countries define Covid deaths?

In the United Kingdom, a Covid death is determined as: the number of people who died within 28 days of their first positive test for Covid-19, or who have Covid-19 mentioned on their death certificate.

The UK Government’s website lists both the daily and total number of deaths within 28 days of positive test, as well as the weekly and total number of deaths with Covid-19 listed on a death certificate.

There can be some differences in reporting data in the UK because of different methodologies used in England, Northern Ireland, Scotland and Wales. Some countries, including England, also calculate the excess death rate, which is another approach to assessing the impact of

Covid-19.

This measures the difference between the reported number of deaths in a given week or month (depending on the country) in 2020-21 and an estimate of the expected number of deaths for that period had the Covid-19 pandemic not occurred.

In the US, Covid-19 deaths are identified through death certificates. A specific code is given when Covid19 is reported as a cause that contributed to death on the death certificate, the Centers for Disease Control and Prevention’s website states. These can include laboratory confirmed cases, as well as cases without laboratory confirmation.

If the certifier suspects Covid-19, or determines it was likely (eg, the circumstances were compelling within a reasonable degree of certainty), they can report Covid-19 as ‘‘probable’’ or ‘‘presumed’’ on the death certificate.

According to the CDC, some states report deaths daily, while other states report deaths weekly or monthly, and the death certificate process itself can take some time.