CapiaLet’s talk

Analysis Business and society

COVID-19 briefing: coronavirus risk in Northern Norway

COVID-19 briefing: coronavirus risk in Northern Norway

COVID-19 briefing: coronavirus risk in Northern Norway

About the story

Original article

Translated from the original publication. Dates and historical information are retained.

Written for Kunnskapsbanken:

How likely are you really to meet someone infectious with coronavirus in Northern Norway?

Despite strict restrictions on the number of people allowed to gather, Norway’s coronavirus case numbers have increased in recent weeks. Health authorities indicate that stricter measures may be introduced if the situation deteriorates further.

Current recommendations include avoiding public transport. Events may gather no more than 200 people, while private gatherings are limited to 20. Cruise operations have also been paused again following recent transmission incidents.

Many infection-control measures share a common aim: avoiding large numbers of people meeting. Are these measures really necessary? What risks do we actually take in daily life by attending a party with ten or 20 other people, taking a crowded bus to work, meeting colleagues, playing a football match, or attending a concert with 200 people?

What is the probability of meeting an infectious person on the bus, at a party or event, or over an ordinary day or week?

To estimate this, we need figures and assumptions. In particular, we need an estimate of the proportion of infectious people who are neither quarantining nor isolating.

The number infectious today

In its regional report of 13 August, the Norwegian Institute of Public Health’s modelling group estimated how many infectious people there would be in different counties over the next three weeks. The institute stresses that the estimates are highly uncertain.

Dividing infectious people by each county’s population gives the proportion infectious, reported here as infectious people per 100,000 residents.

A newer report, published on 18 August, estimates a somewhat smaller proportion from 23 August and for the following two weeks.

We compare Troms and Finnmark, Nordland, and Oslo.

Oslo has the country’s highest estimated infectious share on 17 August, making it particularly interesting.

For the furthest-ahead estimates, uncertainty is included as 95 per cent confidence intervals in parentheses.

The wide intervals underline the uncertainty.

This article uses the estimates for 17 August. We note that the latest estimates for 23 August are lower for all three counties considered, while uncertainty remains very high.

Quarantine and isolation

It is unlikely that all infectious people are mixing with others. Some have symptoms; others have already tested positive. Hopefully, they remain at home. Others have been infected but have not yet developed symptoms: they are presymptomatic.

The institute’s models also estimate that around 40 per cent of infected people develop no symptoms at all: they are asymptomatic.

Some infectious people without symptoms will nevertheless quarantine because they have infected family members or close contacts, or because contact tracing has informed them that they have been exposed.

What proportion of infectious people might therefore be mixing with others? Half? A quarter? We do not know. Counties also have internal differences: some areas may have much higher infectious shares, while others have very few or no infectious people.

Risk if 75 per cent are quarantining or isolating

Assume that 25 per cent of infectious people are as likely as anyone else to use public transport, shop, work, attend parties and meet others. The remaining 75 per cent quarantine or isolate.

What does this imply for the three counties considered?

If the institute’s estimates and our assumptions hold, after meeting 274 people in Oslo you have roughly a five per cent probability that at least one was infectious.

In Troms and Finnmark you would need to meet 562 people, and in Nordland 970, to reach a five per cent probability that at least one was infectious.

The corresponding numbers for a ten per cent probability are 561, 1,153 and 1,992 for Oslo, Troms and Finnmark, and Nordland respectively.

Risk if 50 per cent are quarantining or isolating

What happens if only half of infectious people quarantine or isolate?

After meeting just 137 people in Oslo, the probability that at least one was infectious is now five per cent. The equivalent numbers are 281 in Troms and Finnmark and 485 in Nordland.

The probability reaches ten per cent after meeting 281 people in Oslo, 577 in Troms and Finnmark, or 996 in Nordland.

How many days does it take you to meet 281, 577 or 996 people on public transport, at events, or elsewhere in daily life?

What do these figures mean on a larger scale?

Let us calculate the expected number of encounters with infectious people across a large group.

If 100,000 people each meet 25 or perhaps 50 other people over a day or week, how many would we expect to meet at least one infectious person? Here we assume that half of all infectious people stay home.

Among 100,000 people who each meet 50 others in Oslo, we would expect 1,861 to encounter at least one infectious person.

In Troms and Finnmark, 910 out of 100,000 would encounter someone infectious; in Nordland, 528 would.

What if those 100,000 people each meet only 25 others? We would then expect 935, 456 and 264 out of 100,000 to meet an infectious person in Oslo, Troms and Finnmark, and Nordland respectively.

Substantial uncertainty

Given the considerable uncertainty in the underlying figures and our assumptions, every figure in this article must be interpreted cautiously.

The probability of meeting an infectious person is also different from the probability of becoming infected.

Infection-control measures help prevent many people who encounter infectious individuals from becoming infected themselves. Nevertheless, these figures illustrate how the probability of encountering someone infectious increases with the number of people we meet and the proportion infectious.

If many people reduce their contacts, the expected number of encounters with infectious people falls substantially.

The report published on 18 August, however, suggests somewhat fewer infectious people in Nordland, Troms and Finnmark, and Oslo in the period ahead. If so, this would reduce the probability of encountering them.

Calculation method

If one per cent of the population is infectious and equally likely to meet others as the remaining 99 per cent, one in every 100 people encountered will be infectious on average. When meeting two people, the probability that neither is infectious is 98.01 per cent (99 per cent × 99 per cent), and the probability that at least one is infectious is 1.99 per cent (100 minus 98.01). Percentages are expressed as decimals in calculations: 99 per cent becomes 0.99. For 100 encounters, (1 − 0.01)^100 = 0.366, meaning a 36.6 per cent probability that none of those 100 people is infectious. This implies a 63.4 per cent probability that at least one is infectious: 0.634 = 1 − 0.366.

Formula: p = proportion infectious; n = number of people encountered. Probability of encountering at least one infectious person = 1 − (1 − p)^n.

Calculating the expected number across a group of 100,000 people who each meet a given number is slightly simpler. With an infectious share of one per cent, the example above gives a 63.4 per cent probability of encountering at least one infectious person among 100 contacts. Across 100,000 people, we would therefore expect 63,400 to encounter at least one: 100,000 × 0.634. Similarly, in 100 coin tosses, we expect 50 heads and 50 tails: 100 × 0.5.

Formula: p = proportion infectious; n = contacts per person; N = number of people each meeting n others. Expected number among N people who encounter at least one infectious person = N × [1 − (1 − p)^n].

Sources and references

Historical context

This story is based on an earlier publication. It describes the work and data at that time and is not an updated market report.

Source

← Back to insights