COVID-19: Shielding may not be as effective as expected

Efforts to shield those at highest risk from Covid-19 in the West of Scotland between March and the end of July 2020 may not have been as successful as expected, with those advised to shield experiencing higher rates of infection and death than those not advised to shield. The findings are published in Scientific Reports.

Jill Pell and colleagues analysed data from 1,315,071 patients registered with family practitioners in NHS Greater Glasgow and Clyde between March and May 2020. Of this group, 27,747 individuals had been advised to shield (self-isolate for extended periods of time).A further 353,085 individuals had not been advised to shield but were categorised by the authors as at medium risk of Covid-19 due to health conditions such as diabetes. The remaining 934,239 individuals were considered to be at low risk.

The authors found that, compared to low-risk individuals, people who had been advised to shield were eight times more likely to have confirmed infections, and five times more likely to die following confirmed infection. Moderate-risk individuals were four times more likely to have confirmed infections than the low-risk group, and five times more likely to die following confirmed infection.

In the shielded group, there were 299 (1.1%) confirmed infections and 140 (0.51%) deaths from Covid-19, in the moderate risk group, there were 1,859 (0.53 %) confirmed infections and 803 (0.23%) deaths and in the low-risk group, there were 1,190 (0.13% confirmed infections) and 84 (0.01%) deaths from Covid-19.

The authors conclude that attempts to shield those at highest risk have not been as successful as hoped and that shielding may be best viewed as an intervention to protect individuals that should be implemented alongside other population-wide measures such as physical distancing, face coverings and hand hygiene. The authors caution that while their findings are representative of Glasgow and Greater Clyde, they may be less so for other areas or countries.