key: cord-0766963-s2e56ety
authors: Jani, B. D.; Ho, F. K.; Lowe, D. J.; MacBride-Stewart, S.; Mair, F. S.; Pell, J. P.
title: Comparison of COVID-19 outcomes among shielded and non-shielded populations: A general population cohort study of 1.3 million
date: 2020-09-21
journal: nan
DOI: 10.1101/2020.09.17.20196436
sha: 4378867a7ad37f43315382f8f874c83d5818f7ec
doc_id: 766963
cord_uid: s2e56ety

Background Shielding (extended self-isolation) of people judged, a priori, to be at high-risk from COVID-19 has been used by some countries to protect the individuals and reduce demand on health services. It is unclear how well this strategy works in either regard. Methods A general population study was conducted using linked primary care, prescribing, laboratory, hospital and death records up to end of May 2020. Poisson regression models and population attributable fractions were used to compare COVID-19 outcomes by overall risk category, and individual risk criteria: confirmed infection, hospitalisation, intensive care unit (ICU) admission, population mortality and case-fatality. Results Of the 1.3 million population, 32,533 (2.47%) had been advised to shield, a further 347,374 (26.41%) were classified as moderate risk. Testing for COVID-19 was more common in the shielded (6.75%) and moderate (1.99%) than low (0.72%) risk categories. Referent to low-risk, the shielded group had higher risk of confirmed infection (RR 7.91, 95% 7.01-8.92), case-fatality (RR 5.19, 95% CI 4.12-6.53) and population mortality (RR 48.64, 95% 37.23-63.56). The moderate risk had intermediate risk of confirmed infection (RR 4.11, 95% CI 3.82-4.42) and population mortality (RR 26.10, 95% CI 20.89-32.60), but had comparable case-fatality (RR 5.13, 95% CI 4.24-6.21) to the shielded, and accounted for a higher proportion of deaths (PAF 75.27% vs 13.38%). Age [≥]70 years made the largest contribution to deaths (49.53%) and was associated with an 8-fold risk of infection, 7-fold case-fatality and 74-fold mortality. Conclusions Shielding has not been effective at preventing deaths in those with highest risk. To be effective as a population strategy, shielding criteria would need to be widely expanded to include other criteria, such as the elderly.

Background 1 2 Early in the COVID-19 pandemic, one of the greatest concerns was that the demand on 3 health services would exceed capacity in terms of hospitalisations, admissions to intensive 4 care units (ICU) and requirement for ventilation 1 . Based on past experience, it was assumed 5 that sub-groups of the population would have worse prognosis and, therefore, contribute 6 disproportionately to adverse outcomes and healthcare demands. 7 8 Asian countries have generally relied solely on population-wide strategies 2 . Early, 9

widespread 'test, trace, isolate' strategies were made possible by higher testing capacity and a 10 greater willingness to monitor and enforce compliance. In contrast, European countries have 11 generally adopted a two-pronged approach 2 , whereby general population interventions, such 12 as physical distancing and hand hygiene, designed to reduce transmission in the population as 13 a whole, have been supplemented by shielding of those assumed to be at higher risk. Notably, 14

Sweden, an outlier in having not applied lock-down, nonetheless mandated shielding 3 . 15 Studies suggest that shielding can have negative physical and psychological impact on those 16 required to undertake strict isolation over a protracted period. 4,5 17 18 In the United Kingdom, a Shielded Patient List (initially referred to as the Vulnerable Patient 19

List) was produced comprising two categories(Supplementary Table 1) 6 . The two groups 20 have been labelled high risk, highest risk or clinically extremely vulnerable and moderate 21 risk, at risk or clinically vulnerable by various UK organisations. For the purposes of this 22 study they are referred to as shielded and moderate risk, with the remaining population 23 labelled low risk. In the United Kingdom, the shielded group received individual letters 24 strongly recommending that they self-isolate over a protracted period, by not leaving their 25 homes and avoiding non-essential contact with their household members, and were provided 1 with support at home such as the delivery of food packages. The advice to the moderate risk 2 category was simply to be vigilant in adhering to the general population measures. 3 4 The definitions of high and moderate risk were based largely on expert opinion informed by 5 our understanding of previous viruses and the critical need for better definition of high risk 6 has been highlighted 7 . Studies are emerging of the individual risk factors associated with 7 COVID-19 outcomes. For example, among two million UK community-based app users self-8 reported heart disease, kidney disease, lung disease, diabetes and obesity were associated 9 with self-reported hospital admission and respiratory support for COVID-19 8 . Similarly, 10 linkage of primary care records of 17 million people in England demonstrated that a wide 11 range of long-term conditions were associated with in-hospital death from COVID-19 12 including: respiratory, heart, liver and kidney disease, diabetes, cancers, stroke and organ 13 transplantation 9 . Unfortunately, the investigators did not have access to deaths in the 14 community. COVID-19 risk scores are being developed in an attempt to improve 15 identification of high risk individuals who would be advised to shield 10 but attempts to 16 investigate the potential contribution of a shielding strategy to population-level outcomes and 17 healthcare demands have so far been limited to abstract mathematical modelling 11-19 . 18 19

The aims of this study were to compare those classified, a priori, as high risk (and therefore 20 advised to shield) and those classified as moderate and low risk, in terms of their individual 21 risk of COVID-19 infection and outcomes and the extent to which they accounted for 22 COVID-19 related outcomes at a population level. The CHI register provided sociodemographic information (age, sex, socioeconomic 11 deprivation). Socioeconomic deprivation was derived from the Scottish Index of Multiple 12 Deprivation (SIMD), an area-based measure derived from seven domains -income, 13 education, health, employment, crime, housing, and access to services -which 14 was categorised into general population quintiles. The Electronic Communication of 15 Surveillance in Scotland (ECOSS) database collects laboratory data on infectious diseases, 16 including test date and result. EMIS and Vision are the primary care electronic health record 17 systems used in NHS GGC. Data are extracted using Albasoft software. The PIS collects data 18 on all medications prescribed by community-based healthcare workers, including general 19 practitioners. The RAPID database collects real-time data on every patient admitted to a 20 general (i.e. non-psychiatric) hospital in Scotland including dates of admission and discharge, 21 and type of ward (e.g. ICU) but does not record disease codes. Subsequently, the Scottish 22

Morbidity Record 01 (SMR01) records the same information, in addition to disease codes. 23

Death certificates provide the date and cause of death for all deaths, whether they occur in-24 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint

The copyright holder for this this version posted September 21, 2020. . hospital or in the community. Follow-up data were available until the end of May 2020, 1 before the recommendation to shield was lifted. 2 3 Supplementary Table 1 lists the criteria for the shielded and medium risk categories being 4 applied in the United Kingdom at the time of data extraction. All remaining patients are 5 categorised as low risk. The Scottish list of high-risk individuals is compiled, and regularly 6 updated, centrally using data obtained from a number of sources including Albasoft 7 extraction of general practice data, SMR01 (hospital admissions), SMR06 (cancer registry), 8 national disease registries, and the PIS. The list is sent to individual general practitioners to 9 check for completeness and accuracy, before letters are sent to patients with advice to shield. 10

The NHS GGC Shielding List contains the validated data including the criterion satisfied. We 11 ascertained people who met the criteria for moderate risk via Albasoft extraction of EMIS 12 Laboratory-confirmed cases were defined as patients who had a positive PCR test for 23 COVID-19. Clinically-confirmed cases were defined as patients who either had a positive 24 PCR test or who died from COVID-19 without testing. COVID-related deaths were defined 25 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint

The copyright holder for this this version posted September 21, 2020. . as deaths with International Classification of Diseases 10 th revision (ICD-10) code U07.1 or 1 U07.2 recorded on the death certificate. COVID-related hospitalisation was defined as an 2 SMR01 record of hospitalisation with an ICD code U07.1 or U07.2 or, for more recent 3 admissions, a RAPID record of hospitalisation with a positive COVID-19 test taken between 4 two weeks before and two days after admission to hospital. ICU admission during such 5 hospitalisations was assumed to be COVID-related. 6 7

The sociodemographic characteristics of people were compared by risk category using chi-8 square tests. Poisson regression models with robust standard errors were used to compare the 9 risk ratios (RR) for the high and moderate risk categories referent to the low risk category for 10 each outcome in turn. The models were run univariately and then adjusted for sex and SIMD 11 quintiles as potential confounders. Age was not included as a covariate because it was a 12 moderate risk criterion. The models were re-run using the individual criteria for the high and 13 moderate risk categories as the exposure variables, referent to the low risk category. 14 15 Population attributable fractions (PAFs) were calculated, from prevalence and adjusted RR, 16 to determine the proportion of each outcome that could be attributed to being high and 17 moderate risk, as well as the proportion due to each of the individual criteria. The PAFs of 18 individual criteria were proportionally calibrated so that the sum of PAFs of individual 19 criteria equated to the overall PAF of the relevant risk category. PAF confidence intervals 20 were estimated using bootstrapping (x 1000). 21

Ethical approvals 23 24 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint

The copyright holder for this this version posted September 21, 2020. . followed similar patterns as testing. It increased with age, was higher in women, and was 25 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint

The copyright holder for this this version posted September 21, 2020. . highest in the shielded group and lowest in the low-risk category (Table 2) . After adjustment 1 for sex and deprivation quintile, the risk of laboratory-confirmed infection remained higher in 2 the moderate-risk category and highest in the shielded group (Table 3).  3   4 Overall, 1,661 people were admitted to hospital for COVID-19. Within the general 5 population, hospitalisations increased with age but were comparable between men and 6 women ( Table 2 ). Hospitalisations were more common in the moderate-risk category and 7 most common in the shielded group (Table 2 ) and remained so after adjustment for sex and 8 deprivation quintile (Table 3) . Overall, 122 people were admitted to ICU wards for COVID-9

19. ICU admissions were significantly more common among people aged 45-64 years of age 10 than among older people (Table 2) . Compared with the low risk category, the shielded group 11

were 18 times more likely to be hospitalised but only 4 times more likely to be admitted to 12 ICU (Table 3) . Overall, 1,027 (0.08%) people died from COVID-19. Within the general 13 population, mortality increased with age but was similar in men and women (Table 2) . 14 Population mortality was higher in the moderate-risk category and highest in the shielded 15 group (Table 2 ) and remained so after adjustment for sex and deprivation quintile (Table 3) . 16 17 Among the sub-group with laboratory-confirmed (test-positive) COVID-19 infection, 1,661 18 (49.6%) were hospitalised. Hospitalisations increased with age but were comparable between 19 men and women (Table 4 ). The moderate-risk category was more likely to be hospitalised 20 and the shielded group most likely (Table 4 ) and remained so after adjustment for age and 21 deprivation quintile (Table 5) . Among the sub-group with laboratory-confirmed COVID-19 22 infection, ICU admissions were more common in men and more common in people aged 45-23 64 years than those older (Table 4 ). Low risk cases were more likely to be admitted to ICU 24 than those in the moderate-risk category and shielded groups (Tables 4 & 5) . Among the sub-25 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint

The copyright holder for this this version posted September 21, 2020. . group with clinically-confirmed (test-positive or COVID-19 related death) COVID-19 1 infection, 1,027 (26.70%) died (Table 4) . Case-fatality increased by age and was higher in 2 men than women. It was lowest in the low-risk category but not significantly different in the 3 moderate-risk category and shielded group (p=0.64) ( 28.8% of the population would have had to receive the current level of shielding including 10 those with five criteria currently classified as moderate risk (Supplementary Figure 1) . 11 12 Individual risk criteria 13 14 Due to insufficient numbers, the models for individual risk criteria could not be run for 15 pregnant women with severe heart disease or for COVID-19 related ICU admission in the 16 high-risk category. All of the remaining individual risk criteria were associated with higher 17 likelihood of being tested for COVID-19 (Table 1 ) and higher likelihood of having 18 laboratory-confirmed infection (Table 2) . They were all associated with higher risk of 19 hospitalisation, population mortality (Table 3) and case-fatality (Table 5 ) independent of sex 20 and deprivation quintile. Among the moderate-risk category criteria, age ≥70 years and 21 weakened immune system had risks of population mortality (Table 3) and case-fatality (Table  22 5) that were at least as high as the overall shielded group. Apart from the 0.13% of people 23 with relevant rare diseases and inborn errors of metabolism, the strongest associations were 24 observed for those aged ≥70 years who were eight times more likely to have confirmed 25 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint

The copyright holder for this this version posted September 21, 2020. . infection (Table 3) ; seven times more likely to die following confirmation of infection (Table  1 5); and seventy four times as likely to die overall (Table 3) Table 2 ). Among those admitted to hospital for COVID-19, the 5 likelihood of being admitted to ICU was significantly lower for all of the individual risk 6 criteria included in the moderate risk category, other than diabetes (Table 5 ). In particular, 7 hospitalised patients over 70 years of age were 14 times less likely to be admitted to ICU than 8 low risk hospitalised patients (Table 5) . 9

10 Discussion 11 12

The 2.47% of people who had been advised to shield were, nonetheless, eight times more 13 likely to have confirmed infections than the low risk category, five times more likely to die 14 following confirmed infection and 49 times more likely to die from covid-19 overall. Whilst 15 selective testing of potential cases might explain the first outcome, it does not explain higher 16 overall mortality which suggests that the shielding strategy has not been as effective as was 17 hoped. 18

One quarter of the population were classified as moderate risk and not advised to shield. 20 Nonetheless, they were four times more likely to have confirmed infections than the low risk 21 category, five times more likely to die following confirmed infection and 26 times more 22 likely to die overall, suggesting that consideration should be given to expanding the shielding 23 criteria to include many currently classified as moderate risk. In particular, older age needs to 24 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint

The copyright holder for this this version posted September 21, 2020. . be considered since the elderly are both at high individual risk and contribute significantly to 1 population burden due to their relatively high numbers. 2 3 Paradoxically, people in the shielded and moderate risk categories were less likely to be 4 admitted to ICU following hospitalised for COVID-19 in spite of their poor prognosis. 5

Patients over 70 years were particularly unlikely to be admitted to ICU. This is likely to be 6 due to selective admission policies applied in order to avoid demand exceeding supply, as 7 experienced by some other countries. This finding reinforces the importance of prevention in 8 those with the worst prognosis. 9

10 Comparison with existing literature 11 12 Our finding of 26.41% percent of people satisfying the moderate risk criteria is consistent 13 with the limited existing evidence. A previous study linking English primary and secondary 14 care records on 3.9 million people reported that 20% of the population satisfied similar 15 criteria 20 . Similarly, a study using data from the Global Burden of Diseases Study estimated 16 that 22% of the global population are at increased risk of severe COVID-19 disease 21 . A USA 17 study using data from the Behavioral Risk Factor Surveillance System reported that 45.4% of 18 444,649 adults had one or more of a longer list of morbidities that may be associated with 19 higher risk from COVID-19 22 . Another USA study estimated that 14.2% of participants in the 20 National Health Interview Survey had more than two-fold risk and 1.6% had more than 10-21 fold risk 23 . 22

The evidence on COVID-19 related complications among those people classified as high risk, 24 and therefore advised to shield, has mainly come from case series, expert opinions, 25 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint

The copyright holder for this this version posted September 21, 2020. . unpublished reports. Evidence from case reports found higher COVID-19 related 1 complications among organ transplant recipients 24,25 , patients receiving active cancer 2 treatment (chemotherapy, radiotherapy or immunotherapy) 26,27 , and patients with 3 haematological cancers 28 . Systematic review of case reports found higher COVID-19 4 complication risk among COPD patients, but the effect of COPD severity was not 5 investigated 29 . Patients with cystic fibrosis and sickle cell disease were regarded as high risk 6 for COVID-19 complications based on expert opinions 30,31 . While pregnant women with 7 COVID-19 were found to have higher risk of poor maternal and perinatal outcomes 32,33 , 8 outcomes were not investigated specifically for pregnant women with heart disease. There 9 was no evidence found for higher COVID-19 related complications for patients on various 10 immunosuppressants 34 . A large community study in England found strong association 11 between severe asthma (hazard ratio 1.25) and COVID-19 related mortality but they did not 12 investigate the risk of COVID-19 infection or hospitalisation 9 . 13

In common with previous studies, we demonstrated that age was associated with one of the 15 highest relative risks of death. Additionally, we showed the large extent to which age 16 contributed to adverse outcomes at a population level with 79.14% of deaths attributable to 17 age ≥70 years. The higher mortality in the elderly was mediated in part by their higher case-18 fatality but they also had a higher incidence of infection, possibly due to transmission within 19 care homes. Their lower likelihood of being admitted to ICU once hospitalised for COVID-20 19 may have contributed to their higher case-fatality. Previous studies have reported that men 21 are at highest risk of from Our study demonstrated that men are less likely than 22 women to be tested for COVID-19, less likely have confirmed infection and slightly less 23 likely to be hospitalised. They have comparable overall mortality from COVID-19, due to 24 their lower incidence, but their case-fatality is significantly higher. 25 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint

The copyright holder for this this version posted September 21, 2020. This study adds to the existing evidence of the possible effectiveness of a shielding strategy 4 which is currently limited to mathematical modelling of population effects based on 5 assumptions [11] [12] [13] [14] [15] [16] [17] [18] [19] . Ours was a large-scale, unselected general population study. The data 6 cover a period when shielding was in place. Linkage of primary care, laboratory, hospital and 7 death data enabled us to examine a range of COVID-19 outcomes and study a range of 8 exposure variables including the overall risk categories and their individual criteria. The 9 datasets were linked using exact, rather than probabilistic, matching. We were able to adjust 10 for potential sociodemographic confounders. The exposure data were collected prior to the 11 outcomes occurring avoiding potential reverse causation and recall or recording bias. We did 12 not have data on potential lifestyle confounding factors such as smoking and obesity, or 13 ethnicity. We did not analyse multimorbidity or other potential risk factors not currently 14 included in the high or moderate risk categories as the aim was to evaluate the current 15 strategy. The shielding and moderate-risk criteria were correct at the time of extracting data 16 but may be revised over time. 17 18 Implications of findings 19 20 Our findings suggest that our attempts to shield those at highest risk have not been as 21 successful as hoped, with those advised to shield experiencing higher rates of infection and 22

death. ICU provision has been successfully protected but via systematic exclusion of those 23 with worse prognosis, rather than prevention of infection in those highest at risk. For 24 shielding to be effective as a population level strategy, the current criteria would need to be 25 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this this version posted September 21, 2020. . expanded since three-quarters of deaths were associated with moderate risk criteria for which 1 shielding has not hitherto been recommended. In our study, more than one-quarter of the 2 general population would have needed to be effectively shielded to prevent over 80% of 3 deaths. Since this is unlikely to be acceptable at a time when governments are under pressure 4 to ease lock-down restrictions, shielding is probably best viewed as an individual-level 5 intervention to be used alongside other population-wide interventions such as physical 6 distancing, face coverings and hand hygiene. 7 8 9 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this this version posted September 21, 2020. .

The dataset supporting the conclusions of this article is available in the Glasgow Safe Haven 2 (https://www.nhsggc.org.uk/about-us/professional-support-sites/safe-haven/services/). . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this this version posted September 21, 2020. . BJ, FH, and JP serves as the guarantor of the manuscript and accepts full responsibility for 1 the work and/or the conduct of the study, had access to the data, and controlled the decision 2 to publish. The corresponding author attests that all listed authors meet authorship criteria 3 and that no others meeting the criteria have been omitted. 4 CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint

The copyright holder for this this version posted September 21, 2020. . 34. Minotti C, Tirelli F, Barbieri E, Giaquinto C, D D. How Is Immunosuppressive Status 8

Affecting Children and Adults in SARS-CoV-2 Infection? A Systematic Review. J 9

Infect. 2020;S0163-4453. doi:10.1016/J.JINF.2020.04.026 10 11 12 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint

The copyright holder for this this version posted September 21, 2020. . . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint

The copyright holder for this this version posted September 21, 2020. . 

28) 6,767 (0.72) Moderate

Moderate risk criteria Chronic respiratory disease 149,325 (98.33) 2,540 (1.67) <0.0001 Heart disease

) 1,174 (55.83) 929 (44.17) Sex