ABSTRACT
Interpretable risk assessment of SARS-CoV-2 positive patients can aid clinicians to implement precision medicine. Here we trained a machine learning model to predict mortality within 12 weeks of a first positive SARS-CoV-2 test. By leveraging data on 33,928 confirmed SARS-CoV-2 cases in eastern Denmark, we considered 2,723 variables extracted from electronic health records (EHR) including demographics, diagnoses, medications, laboratory test results and vital parameters. A discrete-time framework for survival modelling enabled us to predict personalized survival curves and explain individual risk factors. Performances of weighted concordance index 0.95 and precision-recall area under the curve 0.71 were measured on the test set. Age, sex, number of medications, previous hospitalizations and lymphocyte counts were identified as top mortality risk factors. Our explainable survival model developed on EHR data also revealed temporal dynamics of the 22 selected risk factors. Upon further validation, this model may allow direct reporting of personalized survival probabilities in routine care.
ABSTRACT
BackgroundNational data on the spread of SARS-CoV-2 infection and knowledge on associated risk factors are important for understanding the course of the pandemic. "Testing Denmark" is a national large-scale epidemiological surveillance study of SARS-CoV-2 in the Danish population. MethodsBetween September and October 2020, approximately 1.3 million of 5.8 million Danish citizens (age > 15 years) were randomly invited to fill in an electronic questionnaire covering COVID-19 exposures and symptoms. The prevalence of SARS-CoV-2 antibodies was determined by Point-of Care rapid Test (POCT) distributed to participants home addresses. FindingsIn total 318,552 participants (24.5% invitees) completed the questionnaire and provided the result of the POCT. Of these, 2,519 (0.79%) were seropositive (median age 55 years) and women were more often seropositive than men, interquartile range (IQR) 42-64, 40.2% males. Of participants with a prior positive Polymerase Chain Reaction (PCR) test (n=1,828), 29.1% were seropositive in the POCT. Seropositivity increased with age irrespective of sex. Elderly participants (>61 years) reported less symptoms and had less frequently been tested for SARS-CoV-2 compared to younger participants. Seropositivity was associated with physical contact with SARS-CoV-2 infected individuals (Risk ratio (RR) 7.43, 95% CI: 6.57-8.41) and in particular household members (RR 17.70, 95% CI: 15.60-20.10). Home care workers had a higher risk of seropositivity (RR 2.09 (95% CI: 1.58-2.78) as compared to office workers. Geographic population density was not associated to seropositivity. A high degree of compliance with national preventive recommendations was reported (e.g., > 80% use of face masks), but no difference was found between seropositive and seronegative participants. InterpretationThis study provides insight into the immunity of the Danish population seven to eight months after the first COVID-19 case in Denmark. The seroprevalence was lower than expected probably due to a low sensitivity of the POCT used or due to challenges relating to the reading of test results. Occupation or exposure in local communities were major routes of infection. As elderly participants were more often seropositive despite fewer symptoms and less PCR tests performed, more emphasis should be placed on testing this age group.
Subject(s)
COVID-19 , Severe Acute Respiratory SyndromeABSTRACT
Background COVID-19 is suggested to be more prevalent among ethnic minorities and individuals with low socioeconomic status. We aimed to investigate the prevalence of SARS-CoV-2 antibodies during the COVID-19 pandemic among citizens 15 years or older in Denmark living in social housing (SH) areas . Methods As part of Testing Denmark, a nationwide sero-epidemiological surveillance survey, we conducted a study between January 8th and January 31st, 2021 with recruitment in 13 selected SH areas in Denmark. Participants were offered a point-of-care rapid SARS-CoV-2 IgM and IgG antibody test and a questionnaire concerning previous testing (viral throat- and nasopharyngeal swab or antibody test), test results for COVID-19, demographics, household characteristics, employment, risk factors for SARS-CoV-2 infection and history of symptoms associated with COVID-19. Data on seroprevalence from Danish blood donors in same period using a total Ig ELISA assay were used as a proxy for the general Danish population. Findings Of the 13,279 included participants, 2,296 (17.3%) were seropositive (mean age 46.6 (SD 16.4) years, 54.2% female), which was 3 times higher than in the general Danish population (mean age 41.7 (SD 14.1) years, 48.5% female) in the same period (5.8%, risk ratios (RR) 2.96, 95% CI 2.78-3.16, p>0.001). Seropositivity was higher among males than females (RR 1.1, 95% CI 1.05-1.22%, p=0.001) and increased with age, with an OR seropositivity of 1.03 for each 10-year increase in age (95% CI 1.00-1.06, p=0.031). Close contact with COVID-19-infected individuals was associated with a higher risk of infection, especially among members of the same households (OR 5.0, 95% CI 4.1-6.2 p<0,001). Adjusted for age, gender and region living at least 4 people in a household significantly increased the OR of seropositivity (OR 1.3, 95% CI 1.1-1.6, p=0.02) as did living in a multi-generational household (OR 1.3 per generation, 95% CI 1.1-1.5, p=0.007). Only 1.6% of participants reported not following any of the national COVID-19 recommendations. Anosmia (RR 3.2 95% CI 2.8-3.7, p<0.001) and ageusia (RR 3.3, 95% CI 2.9-3.8, p<0.001) were strongest associated with seropositivity. Interpretation Danish citizens living in SH areas of low socioeconomic status had a three times higher SARS-CoV-2 seroprevalence compared to the general Danish population. The seroprevalence was significantly higher in males and increased with age. Living in multiple generations or more than four persons in a household was an independent risk factor for being seropositive. Results of this study can be used for future consideration of the need for preventive measures in the populations living in SH areas.
Subject(s)
COVID-19 , AgeusiaABSTRACT
Background People experiencing homelessness (PEH) and associated shelter workers may be at higher risk of infection with 'Severe acute respiratory syndrome coronavirus 2' (SARS-CoV-2). The aim of this study was to determine the prevalence of SARS-CoV-2 among PEH and shelter workers in Denmark. Design and methods In November 2020, we conducted a nationwide cross-sectional seroprevalence study among PEH and shelter workers at 21 recruitment sites in Denmark. The assessment included a point-of-care test for antibodies against SARS-CoV-2, followed by a questionnaire. The seroprevalence was compared to that of geographically matched blood donors considered as a proxy for the background population, tested using a total Ig ELISA assay. Results We included 827 participants in the study, of whom 819 provided their SARS-CoV-2 antibody results. Of those, 628 were PEH (median age 50.8 (IQR 40.9-59.1) years, 35.5% female) and 191 were shelter workers (median age 46.6 (IQR 36.1-55.0) years and 74.5% female). The overall seroprevalence was 6.7% and was similar among PEH and shelter workers (6.8% vs 6.3%, p=0.87); and 12% among all participants who engaged in sex work. The overall participant seroprevalence was significantly higher than that of the background population (2,9%, p <0.001). Participants engaging in sex work had a significantly increased risk of seropositivity compared to other participants (95% CI: 0.86-2.36, p=0.02). Seropositive and seronegative participants reported a similar presence of at least one SARS-CoV-2 associated symptom (49% and 54%, respectively). Interpretations The prevalence of SARS-CoV-2 antibodies was more than twice as high among PEH and associated shelter workers, compared to the background population. The subset of the study participants who were also sex workers were at particularly high risk of COVID-19 infection. Funding Tryg and HelseFonden.