FOI 3545
Document 1
OFFICIAL
Preliminary Estimates of Long-COVID
Industry Analysis Unit, MECD,s 22
Key Points
• In the absence of survey data, Treasury has undertaken preliminary analysis to estimate the number
of long-COVID cases in Australia and its potential impact on workforce absenteeism.
• The analysis suggests that long-COVID is likely to be a modest but growing component of overall
COVID-related worker absenteeism. The analysis attributes around 12 per cent of COVID-related
absenteeism or 31,000 workers per day to long-COVID in June, based on an estimated 100,000
long-COVID cases in June (see
Chart 1, Table 1). Note that long-COVID is calculated as a subset of
our existing COVID-19 absenteeism estimates based on the ABS Labour Force Survey (LFS).
• Given limited survey data, the analysis is based on key assumptions informed by domestic and
international studies on long-COVID.
– Long-COVID is defined as ongoing symptoms for 4 or more weeks following initial infection.
– 4.5 per cent of people who test positive for COVID-19 will have ongoing symptoms for 4 weeks,
decreasing to 2.3 per cent at 8 weeks, 1.2 per cent at 12 weeks, and 0.2 per cent at 6 months.
– 10 per cent of long-COVID cases work zero per cent of their normal hours prior to infection,
30 per cent work half their normal hours, and 60 per cent work full normal hours.
• The estimates of long-COVID and related absenteeism are preliminary and rely on imperfect data
and the results presented here should be considered in the context of these constraints.
Chart 1: Long-COVID no-long-COVID absenteeism, average per day in month, as a % of workforce
Source: ABS Labour Force Survey, CovidBaseAU, Treasury Analysis. Note: Modelled long-COVID cases count as long-COVID from day 1.
Table 1: Estimated absenteeism (long-COVID and no-long-COVID) per day, average in month
Results 2022
January
February
March
April
May
June
Average absentee rate per day as a share of
3.4
1.2
1.0
2.6
2.2
1.9
workforce (LFS method) (%)*
Average no. of absentees per day
453,000
155,000
129,000
346,000
294,000
259,000
Share of existing absenteeism due to long
2.3
8.0
5.6
5.3
8.8
12.0
COVID** (%)
No. of absentees per day due to long-COVID
10,000
12,000
7,000
18,000
26,000
31,000
*The absentee rate is calculated using IAU’s Labour Force Survey methodology. Refer to IAU’s monthly COVID-19 absenteeism updates for detail.
**Source: Refer to detailed assumptions and methodology sections below for method to calculate proportion of absenteeism due to long-COVID.
Insight note | Macroeconomic Conditions Division | Preliminary Estimates of Long-COVID 11 August 2022 | 1
Official
Detailed assumptions
•
We have followed the UK’s National Institute for Health and Care Excellence (NICE) definition of
long-COVID
(NICE, 2021):
–
“Long-COVID is commonly used to describe signs and symptoms that continue or develop after
acute COVID-19. It includes both ongoing symptomatic COVID-19 (from 4 to 12 weeks) and
post-COVID-19 syndrome (12 weeks or more).”
•
Using the findings of the Kings College London Study
(June 2022), we have assumed that 4.5 per cent
of people who test positive for COVID-19 will have ongoing symptoms for 4 weeks or more (see
Additional details from supporting research for more information). We have assumed this decreases to
2.3 per cent after 8 weeks, 1.2 per cent after 12 weeks and down to 0.2 per cent after 6 months.
–
This step-down function follows a similar trajectory as studies by the UK Office for National
Statistics (ONS)
(2021) and Liu et al.
(2021).
–
We have chosen 4.5 per cent at 4 weeks rather than a higher level identified in other studies as
most of Australia’s cases were of the Omicron variant and occurred when the population was
mostly vaccinated, both factors which greatly decrease the chance of long-COVID.
•
We assumed the impact of long-COVID on workforce absenteeism will be a constant share of
long-COVID cases. Ten per cent of long-COVID cases are assumed to work zero per cent of their normal
hours prior to infection, 30 per cent work half of their normal hours, and 60 per cent work 100 per
cent of normal hours. The evidence underpinning these assumptions is mixed, and research into the
direct effect on work is limited. Three studies which provide some insights are listed below:
–
Walters and Wernham
(2022) found that around 10 per cent of people who suffer long-COVID
stop working, with sufferers generally going on sick leave. They also found no evidence that
getting long-COVID makes workers more likely to reduce their hours to some amount above
zero.
–
A study in the UK found that for long-COVID patients, who tested positive both 4 weeks and
12 weeks prior, around 60 per cent reported some limitation on day-to-day activity, 18 per cent
reported their activity had been limited a lot and the rest had no activity limits
(ONS, April 2021).
–
In a survey with 3,762 participants spanning 56 countries, Davis et al.
(2021) found that around
45 per cent of long-COVID cases reported requiring a reduced work schedule and 22 per cent
were not working at all at the time of the survey.
Methodology
•
We apply the assumptions above to our existing case-based absenteeism model to calculate how much
absenteeism can be attributed to long-COVID. This proportion is then applied to our existing LFS
absenteeism model to calculate absenteeism due to long-COVID.
–
It is more appropriate to use absenteeism estimates from the LFS than a case-based approach, as
health advice suggests that a large portion of COVID-19 infections are not being picked up
through reporting mechanisms. Nevertheless, the case-based model assists in identifying the
share of absenteeism that can be attributed to long-COVID.
•
First, we take total reported COVID-19 cases each day and apply the assumed proportions of cases that
will have long-COVID and the duration. For example, if we had 10,000 reported cases on one day, we
would expect 450 (4.5 per cent) of those to experience long-COVID for at least the next month.
•
We then apply the assumed impact of long-COVID on hours worked to get the effect on absenteeism
on any given day. For example, of the 450 people with long-COVID, 10 per cent work no hours, 30 per
cent work half of their hours, and the rest work their usual/full hours – resulting in 79 people being
Insight note | Macroeconomic Conditions Division | Preliminary Estimates of Long-COVID
11 August 2022 | 2
Official
absent from work on any given day during the month due to long-COVID (Note: This is also multipled
by the share of people assumed to be employed: 70 per cent).
•
We compare absenteeism due to long-COVID to the total amount of absenteeism due to COVID-19 as
calculated through our existing cased-based model, to get the share of absenteeism due to
long-COVID. For example, these 10,000 cases would lead to around 6,000 absentees per day over 7
days (this is inclusive of long-COVID). We then take absenteeism due to long-COVID identified above
and divide it by this number to obtain the long-COVID share (which is less than 1 per cent). Note, this
produces a lower share of long-COVID absenteeism than our estimate for 2022 as it only considers
cases from one day. As cases cumulate over time the share of long-COVID absenteeism increases.
•
Finally, we apply this share of long-COVID-related abseteeism to our LFS estimates of daily
absenteeism. We assume that the LFS category 'Own illness, injury and sick leave' captures
absenteeism due to both normal COVID-19 and long-COVID. We use the shares of COVID-absenteeism
versus long-COVID absenteeism derived from our case-based model assumptions above, to calculate
the split.
–
Note: we may miss some of the impact of long-COVID as we are not including persons who leave
the labour force due to long-COVID. While this has been noteable in other countries, for example
the UK has had around a 12 per cent rise in workers who have dropped out of the workforce due
to long-term illness since December 2019
(UK ONS), we are yet to see a material impact of this in
the Australian LFS data. This is possibly due to a higher proportion of Australia's infections
coming at a time of high vaccine protection, and with the relatively less severe Omicron variant.
Additional details from supporting research
•
The Kings College London
(June 2022) found that among Omicron cases, 4.5 per cent experience
ongoing symptoms for 4 weeks or more, compared to 10.8 per cent for Delta cases.
–
The odds of experiencing long-COVID were between 20 and 50 per cent less during Omicron
versus the Delta period. Further, Omicron cases were less likely to experience long-COVID for all
vaccine timings. They also found that the chance of getting long-COVID increases to around 5.2
per cent for those aged 60 or over who test positive for COVID-19.
•
According to the ONS, 3 per cent of the UK population had long-COVID by the end of May 2022, of
whom 72 per cent reported that their day-to-day activities were limited by the condition and 21 per
cent were limited ‘a lot’
(ONS, 2022).
–
Of the triple-vaccinated adults, 4.5 per cent (Omicron BA.1), 4.2 per cent (Omicron BA.2) and 5.0
per cent (Delta Variants) self-reported having long-COVID 12 to 16 weeks after a first
lab-confirmed COVID-19 infection. Of double-vaccinated adults, 4.0 per cent self-reported having
long-COVID 12 to 16 weeks after a first infection with Omicron, compared with 9.2 per cent for
those with the Delta variant
•
Long-COVID research and data in Australia has been limited, especially regarding Omicron.
–
The largest study in Australia is a whole population cohort study in New South Wales which
followed 94 per cent (2,904) of COVID-19 cases diagnosed between January and May 2020. They
found that 80 per cent of those with COVID-19 recover within a month, but almost 5 per cent
continued to experience symptoms 3 months later
(Liu et al., 2021). Note, this study was based
on the initial COVID-19 variant from 2020 on cases not protected by widespread vaccination.
–
Hensher and Angeles
(December 2021) used ONS data and Doherty Institute forecasts to
estimate that Australia was likely to see between 80,000 and 325,000 people with long-COVID
for at least 12 weeks. However, these estimates were made prior Australia’s Omicron wave.
Insight note | Macroeconomic Conditions Division | Preliminary Estimates of Long-COVID
11 August 2022 | 3
FOI 3545
Document 2
