Abstract
Background
Evidence on the relationship between socioeconomic position (SEP) and infections with SARS-CoV-2 is still limited as most of the available studies are ecological in nature. This is the first German nationwide study to examine differences in the risk of SARS-CoV-2 infections according to SEP at the individual level.
Methods
The ‘CORONA-MONITORING bundesweit’ (RKI-SOEP) study is a seroepidemiological survey among a dynamic cohort of the German adult population (n=15 122; October 2020–February 2021). Dried blood samples were tested for SARS-CoV-2 antibodies and oral-nasal swabs for viral RNA. SEP was measured by education and income. Robust logistic regression was used to examine adjusted associations of SARS-CoV-2 infections with SEP.
Results
288 participants were seropositive, PCR positive or self-reported a previous laboratory-confirmed SARS-CoV-2 infection. The adjusted odds of SARS-CoV-2 infection were 1.87-fold (95% CI 1.06 to 3.29) higher among low-educated than highly educated adults. Evidence was weaker for income differences in infections (OR=1.65; 95% CI 0.89 to 3.05). Highly educated adults had lower odds of undetected infection.
Conclusion
The results indicate an increased risk of SARS-CoV-2 infection in low-educated groups. To promote health equity in the pandemic and beyond, social determinants should be addressed more in infection protection and pandemic planning.
Keywords: infections, health inequalities, social class, COVID-19, communicable diseases
Introduction
Initial research on the pandemic indicates that SARS-CoV-2 infections occur more frequently in socioeconomically deprived areas.1 Infection risks may accordingly be higher for people in disadvantaged living and certain working conditions, especially during more advanced stages of the pandemic.2 3 For example, essential workers employed in the logistic, healthcare, retailing or public transport sector, who tend to have lower incomes than non-essential workers,4 often work in conditions involving closer physical proximity to others.5 The possibility to work remotely is, by contrast, more available to people with higher incomes and qualifications.6 Furthermore, crowded living conditions and limited access to effective personal protective equipment may increase the risk of viral transmissions and thereby produce inequalities in infections.7
Research on socioeconomic differences in SARS-CoV-2 infections has been based mostly on ecological studies correlating area-level infection rates with area-level socioeconomic indicators.1 2 8 9 The advantage of ecological studies is that area-level data are available relatively quickly, for example, from existing surveillance systems. Ecological studies can thus be an expedient starting point for exploring the phenomenon. However, their findings are prone to ecological fallacy, areas can be very heterogeneous in terms of their residents’ socioeconomic position (SEP) and inequalities are probably underestimated. Seroepidemiological individual-level studies are therefore needed to examine the relationship between SEP and SARS-CoV-2. In this study, we introduced serological and PCR testing along with a self-report questionnaire on previously conducted PCR tests into an existing German cohort and investigated whether SARS-CoV-2 infections were associated with SEP at the individual level.
Methods
Study design
The ‘CORONA-MONITORING bundesweit’ study (RKI-SOEP study) is based on the nationwide population-based random samples administered by the German Socio-Economic Panel (SOEP), a dynamic cohort from Germany’s resident population in private households.10 11 A gross sample of 31 675 adult cohort members was invited to participate, of which 15 122 individuals (response: 48%) participated in the study. The net sample comprised participants in 400 of the 401 German districts (Nomenclature of Territorial Units for Statistics [NUTS]-3 level), with an average of 38 participants per district.
Between October 2020 and February 2021, biospecimens and interview data were collected once from each participant. Participants provided a dried capillary blood sample obtained by finger prick to serologically detect antibodies from a previous SARS-CoV-2 infection by ELISA. To test for a current infection, participants provided an oral-nasal swab sample for PCR testing. Both specimens were collected by the participants themselves using CE-certified sample collection and submission kits sent by post along with written, pictorial and video instructions. A one-page paper questionnaire included questions on previous throat swab laboratory tests for SARS-CoV-2. Further details can be found in the study protocol.11
Measures and definitions
We used three infection metrics based on combinations of the following: seropositivity for SARS-CoV-2 IgG antibodies from dried blood samples (Euroimmun SARS-CoV-2-S1 IgG-ELISA; cut-off adapted for dried blood spot testing: ratio ≥ 0.94),12 a positive SARS-CoV-2-PCR test result in the study and any self-reported SARS-CoV-2-positive throat swab laboratory test conducted prior to study participation. Cases were defined as those meeting at least one of these criteria (ie, if participants had missing data or were not positive on one criterion but were positive on another, they were considered as cases). Undetected infection with SARS-CoV-2 was defined as having tested seropositive or PCR-positive during the study but self-reported never having had a SARS-CoV-2-positive swab test before.
SEP was measured using participants’ self-reported prepandemic (and hence exogenous) education and income from SOEP wave 2019 or the latest available from earlier waves. Using the CASMIN (Comparative Analysis of Social Mobility in Industrial Nations) educational classification,13 participants’ highest school and vocational qualifications were classified into low (no, primary or low secondary education), middle (intermediate/high secondary education) and high (tertiary education). Equivalised current disposable household income was calculated using the OECD (Organisation for Economic Co-operation and Development)-modified equivalence scale and categorised into low (<60% of median), middle (60%−<150% of median) and high (≥150% of median).14
Statistical analysis
The prevalence of SARS-CoV-2 infections was estimated using weighting factors to compensate for sampling design and non-random non-response. The weighting factors result from complex non-response modelling at the person and household level and calibration of the sample to match the official German population statistics by age, sex, federal state, municipality size, household size and owner-occupied housing. Standard error (SE) estimation was performed using Stata’s survey data commands (V.17.0, StataCorp LLC, College Station, Texas, USA) accounting for weighting and household clustering. Logistic regression models with household-clustered SEs were used to estimate odds ratios (ORs) for SARS-CoV-2 infections by SEP, adjusted for a set of covariates: age, sex, household size, migrant background, urban–rural residence, region (east/west), date of participation and dummy variables for missing values.
Results
Among the 15 122 participants aged 18–99 years, more than 80% participated in October and November 2020 (median participation date: 11 November 2020). Overall, 192 participants were seropositive, 51 were PCR positive and 146 self-reported having had a laboratory-confirmed SARS-CoV-2 infection before study participation (table 1). At least one of these criteria was met by 288 participants.
Table 1.
Characteristics of the study population (n=15 122)
| n | %* | |
| Sex | ||
| Women | 8099 | 50.9 |
| Men | 7023 | 49.1 |
| Age group (years) | ||
| 18–34 | 2805 | 25.0 |
| 35–49 | 3553 | 22.7 |
| 50–64 | 4945 | 27.6 |
| 65–79 | 3126 | 18.4 |
| 80+ | 693 | 6.4 |
| Education | ||
| Low | 3267 | 31.0 |
| Middle | 6326 | 41.9 |
| High | 4916 | 27.1 |
| Missing | 613 | – |
| Income | ||
| <60% of median | 1524 | 15.2 |
| 60%−<150% of median | 8406 | 57.6 |
| ≥150% of median | 5065 | 27.2 |
| Missing | 127 | – |
| Date of participation | ||
| October 2020 | 5532 | (36.6) |
| November 2020 | 6748 | (44.6) |
| December 2020 | 2090 | (13.8) |
| January 2021 | 595 | (3.9) |
| February 2021 | 157 | (1.0) |
| SARS-CoV-2 IgG antibodies | ||
| Seropositive | 192 | 1.3 |
| Seronegative | 14 589 | 98.7 |
| Missing | 341 | – |
| SARS-CoV-2 RNA | ||
| PCR positive | 51 | 0.4 |
| PCR negative | 14 638 | 99.6 |
| Missing | 433 | – |
| Previously tested positive for SARS-CoV-2† | ||
| Yes | 146 | 1.1 |
| No | 14 771 | 98.9 |
| Missing | 205 | – |
*Weighted percentage (unweighted percentage in brackets).
†Self-reported positive throat swab test before study participation.
Table 2 shows the prevalence and adjusted ORs for SARS-CoV-2 infection by SEP. Seropositivity, seropositivity or PCR positivity and a self-reported positive swab test were each most prevalent in the lowest education and income groups. After adjusting for covariates, low education remained associated with seropositivity and the combined infection indicator of measured seropositivity or PCR positivity. A previously detected infection with SARS-CoV-2 as indicated by a self-reported positive swab test prior to the study showed no consistent association with either education or income. When all three infection parameters were combined (seropositive/PCR positive or previously tested positive), the OR for SARS-CoV-2 infection was 1.87-fold higher in the low than high education group. Evidence was weaker for income differences in infections. With regard to undetected infections, highly educated adults had lower odds of being seropositive or PCR positive without previously having received a positive swab test result compared with adults with low as well as those with medium education, net of all covariates. The adjusted OR comparing high versus low/middle education was 0.45 (95% CI 0.22 to 0.93, p=0.031).
Table 2.
SARS-CoV-2 infection by socioeconomic position among adults in Germany, October 2020–February 2021
| % (95% CI)* | OR (95% CI)† | p value | |
| Seropositive | |||
| Education | |||
| Low | 1.80 (1.19 to 2.71) | 2.32 (1.18 to 4.53) | 0.014 |
| Middle | 0.98 (0.67 to 1.43) | 1.20 (0.69 to 2.08) | 0.527 |
| High | 1.01 (0.69 to 1.50) | Ref. | |
| Income | |||
| <60% of median | 2.04 (1.12 to 3.70) | 1.56 (0.77 to 3.17) | 0.220 |
| 60%−<150% of median | 1.20 (0.87 to 1.66) | 1.02 (0.58 to 1.79) | 0.945 |
| ≥150% of median | 1.02 (0.69 to 1.52) | Ref. | |
| Seropositive/PCR positive | |||
| Education | |||
| Low | 2.05 (1.37 to 3.07) | 2.03 (1.10 to 3.75) | 0.024 |
| Middle | 1.46 (1.04 to 2.05) | 1.45 (0.89 to 2.37) | 0.140 |
| High | 1.21 (0.85 to 1.72) | Ref. | |
| Income | |||
| <60% of median | 2.69 (1.29 to 5.51) | 1.58 (0.77 to 3.26) | 0.213 |
| 60%−<150% of median | 1.53 (1.15 to 2.04) | 1.09 (0.65 to 1.82) | 0.746 |
| ≥150% of median | 1.18 (0.82 to 1.70) | Ref. | |
| Previously tested positive ‡ | |||
| Education | |||
| Low | 1.39 (0.82 to 2.35) | 1.68 (0.76 to 3.69) | 0.197 |
| Middle | 0.84 (0.54 to 1.29) | 0.87 (0.43 to 1.76) | 0.694 |
| High | 1.17 (0.74 to 1.86) | Ref. | |
| Income | |||
| <60% of median | 1.99 (0.94 to 4.17) | 1.42 (0.61 to 3.30) | 0.418 |
| 60%−<150% of median | 0.98 (0.68 to 1.42) | 0.82 (0.40 to 1.69) | 0.596 |
| ≥150% of median | 1.06 (0.59 to 1.88) | Ref. | |
| Seropositive/PCR positive or previously tested positive | |||
| Education | |||
| Low | 2.44 (1.71 to 3.48) | 1.87 (1.06 to 3.29) | 0.029 |
| Middle | 1.78 (1.31 to 2.42) | 1.29 (0.78 to 2.14) | 0.315 |
| High | 1.68 (1.18 to 2.39) | Ref. | |
| Income | |||
| <60% of median | 3.64 (2.10 to 6.24) | 1.65 (0.89 to 3.05) | 0.112 |
| 60%−<150% of median | 1.81 (1.39 to 2.33) | 0.95 (0.57 to 1.56) | 0.827 |
| ≥150% of median | 1.63 (1.09 to 2.43) | Ref. | |
| Undetected infection | |||
| Education | |||
| Low | 1.00 (0.60 to 1.66) | 2.17 (0.90 to 5.25) | 0.085 |
| Middle | 0.91 (0.58 to 1.43) | 2.27 (1.03 to 5.00) | 0.042 |
| High | 0.51 (0.29 to 0.89) | Ref. | |
| Income | |||
| <60% of median | 1.56 (0.75 to 3.24) | 1.77 (0.81 to 3.83) | 0.150 |
| 60%−<150% of median | 0.77 (0.52 to 1.13) | 1.03 (0.54 to 1.96) | 0.855 |
| ≥150% of median | 0.64 (0.41 to 1.00) | Ref. | |
*Prevalence (weighted).
†Adjusted for age, sex, household size, migrant background, urban–rural residence, region (east/west), date of participation and dummies for missing values (separate models for education and income).
‡Self-reported positive throat swab test before study participation.
Ref, reference group.
Discussion
This is the first German nationwide study of SEP differences in SARS-CoV-2 infections based on individual-level data from the pandemic. The results indicate an increased risk of infection among low-qualified adults and that undetected infections were least common among the highly educated.
The seroepidemiological design enabled the detection of known infections and infections that had previously gone undetected, for example, in asymptomatic cases. Moreover, adding PCR to serological testing was especially relevant because the specimens were collected during the ongoing second wave of SARS-CoV-2 infections in Germany, in which seroconversion from infections in this wave was still in progress. Completely non-random non-response could be counteracted by complex weighting, which enabled extrapolation to Germany’s adult population in private households. As both SEP and infection variables were measured at the individual level, this study can contribute to overcoming the limitations of previous ecological studies, such as the possibility of ecological fallacy.
An important limitation is that the sample was restricted to residents in private households. Institutionalised people, such as nursing home residents or people living in shared accommodations for homeless people, migrant workers or asylum seekers, are not represented. Assuming increased infection risks in these groups, this study may have underestimated socioeconomic differences in infections. Selection bias may have occurred if individuals with SARS-CoV-2 infection were more or less willing to participate in the study and if this varied by SEP. Depending on the direction of this potential effect, it may have led to overestimation or underestimation of SEP differences in infections. Conditions that may alter immune response, for example, immunosuppressive therapy or obesity, could not be controlled for in the serological testing, which may have been another source of bias.
Our findings support evidence from earlier pandemics of viral respiratory pathogens, such as influenza, suggesting higher levels of viral exposure in low-SEP settings.7 Consistent with our findings, higher levels of SARS-CoV-2 infections in low-educated individuals have been found in the UK Biobank and a seroprevalence survey in five German regions.15 16 However, individual-level nationwide findings on SEP differences in SARS-CoV-2 infections during the pandemic are still scarce and sometimes contrary.17 18 Important mediators in the relationship between education and infections may be occupational working conditions.19 People in medium-skilled to low-skilled occupations such as nursing, retailing, production or logistics, had few opportunities to reduce occupational contact and mobility by working remotely during the pandemic and are associated with more contact and proximity with others. Moreover, campaigns to educate people about infection protection may have been less effective in reaching low-literate groups, and highly educated people may have had better opportunities in their everyday circumstances to implement infection control measures, such as physical distancing or regular testing.
Our findings indicate an increased risk of SARS-CoV-2 infection in educationally disadvantaged groups and suggest a higher detection of infections among highly educated adults. Infection control strategies should provide universal access to testing that is independent of socioeconomic background from early on. To promote health equity in the pandemic and beyond, social determinants should be given more recognition in infection protection and pandemic planning.
What is already known on this subject.
Initial research on the pandemic indicates that SARS-CoV-2 infections are more common in socioeconomically deprived areas.
Individual-level studies on socioeconomic differences in SARS-CoV-2 infections during the pandemic are still scarce, and their findings are sometimes inconsistent or contradictory.
What this study adds.
This is the first German nationwide study of socioeconomic differences in SARS-CoV-2 infections based on individual-level data from the pandemic. We found a higher risk of infection among adults with low education. Infections were least likely to go undetected among the highly educated. Our analysis identified low-educated adults as an important target group for infection protection, testing and control strategies during pandemics involving novel viral respiratory pathogens.
Acknowledgments
We would like to thank all our colleagues at the Robert Koch Institute and the Socio-Economic Panel at the German Institute for Economic Research for their support and cooperation. Special thanks are due to Angelika Schaffrath Rosario for statistical advice and Hans Walter Steinhauer for his contribution to weighting the sample. Our sincere thanks for the laboratory analyses are due in particular to Janine Michel, Martin Schlaud, Silke Stahlberg, Antje Kneuer and Andreas Nitsche. We would also like to thank the employees of Kantar GmbH who contributed to the field work and data collection. We sincerely thank all study participants for their willingness to participate.
Footnotes
Contributors: All authors of this work meet the ICMJE Authorship Criteria. MMG, SZ, SL and JH made substantial contributions to the design of the study. JH conducted the statistical analysis and drafted the manuscript with substantial contributions from SZ, MMG, BW, CS, SH and HN. All authors substantially contributed to the interpretation of data, edited the manuscript and revised it critically for important intellectual content. All authors approved the final manuscript.
Funding: The RKI-SOEP study was funded by the German Federal Ministry of Health (reference number: ZMVI1-2520COR402). JH has received funding from the German Research Foundation for the project ‘Socioeconomic inequalities in health during the COVID-19 pandemic (INHECOV): empirical analyses and implications for pandemic preparedness’ (reference number: HO 6565/2–1). SZ, CS and CH acknowledge funding by the German Research Foundation for the project ‘The consequences of SARS-CoV-2 for societal inequalities’ (reference numbers: ZI 1535/1–1, SCHR 1498/7–1, HO 7021/1–1).
Competing interests: None declared.
Provenance and peer review: Not commissioned; externally peer reviewed.
Ethics statements
Patient consent for publication
Not applicable.
Ethics approval
The study was approved by the Ethics Committee of the Berlin Chamber of Physicians (Eth-33/20) in compliance with the Declaration of Helsinki.
References
- 1. Wachtler B, Michalski N, Nowossadeck E, et al. Socioeconomic inequalities and COVID-19: a review of the current international literature. J Health Monit 2020;5:3–17 10.25646/7059 10.25646/7059 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Hoebel J, Michalski N, Wachtler B, et al. Socioeconomic differences in the risk of infection during the second SARS-CoV-2 wave in Germany. Dtsch Arztebl Int 2021;118:269–70. 10.3238/arztebl.m2021.0188 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Yang B, Wu P, Lau EHY, et al. Changing disparities in COVID-19 burden in the ethnically homogeneous population of Hong Kong through pandemic waves: an observational study. Clin Infect Dis 2021. 10.1093/cid/ciab002. [Epub ahead of print: 06 Jan 2021]. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Blundell R, Costa Dias M, Joyce R, et al. COVID‐19 and Inequalities. Fisc Stud 2020;41:291–319. 10.1111/1475-5890.12232 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. United Nations Department of Ecomic and Social Affairs . World economic situation and prospects: April 2020 Briefing, no. 136, 2020. Available: https://www.un.org/development/desa/dpad/publication/world-economic-situation-and-prospects-april-2020-briefing-no-136/ [Accessed 30 Apr 2020].
- 6. Schröder C, Entringer T, Goebel J. COVID-19 is not affecting all working people equally. SOEPpapers on multidisciplinary panel data research 1083. Berlin: DIW, 2020. [Google Scholar]
- 7. Quinn SC, Kumar S. Health inequalities and infectious disease epidemics: a challenge for global health security. Biosecur Bioterror 2014;12:263–73. 10.1089/bsp.2014.0032 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Clouston SAP, Natale G, Link BG. Socioeconomic inequalities in the spread of coronavirus-19 in the United States: a examination of the emergence of social inequalities. Soc Sci Med 2021;268:113554. 10.1016/j.socscimed.2020.113554 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. De Ridder D, Sandoval J, Vuilleumier N, et al. Socioeconomically disadvantaged neighborhoods face increased persistence of SARS-CoV-2 clusters. Front Public Health 2021;8:626090. 10.3389/fpubh.2020.626090 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Goebel J, Grabka MM, Liebig S, et al. The German socio-economic panel (SOEP). J Econ Stat 2019;239:345–60. 10.1515/jbnst-2018-0022 [DOI] [Google Scholar]
- 11. Hoebel J, Busch MA, Grabka MM, et al. Seroepidemiological study on the spread of SARS-CoV-2 in Germany: Study protocol of the ‘CORONA-MONITORING bundesweit’ study (RKI-SOEP study). J Health Monit 2021;6:2–16. 10.25646/7853 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Robert Koch Institute . Application of the Euroimmun anti-SARS-CoV-2-S1-IgG ELISA antibody test to dried blood spots, 2021. 10.25646/8671 10.25646/8671 [DOI] [Google Scholar]
- 13. Brauns H, Scherer S, Steinmann S. The CASMIN educational classification in international comparative research. In: Hoffmeyer-Zlotnik JHP, Wolf C, eds. Advances in cross-national comparison: an European working book for demographic and socio-economic variables. New York: Kluwer, 2003: 196–221. [Google Scholar]
- 14. Eurostat . Statictics explained – glossary: At-risk-of-poverty rate, 2021. Available: https://ec.europa.eu/eurostat/statistics-explained/index.php?title=Glossary:At-risk-of-poverty_rate [Accessed Aug 2021].
- 15. Niedzwiedz CL, O'Donnell CA, Jani BD, et al. Ethnic and socioeconomic differences in SARS-CoV-2 infection: prospective cohort study using UK Biobank. BMC Med 2020;18:160. 10.1186/s12916-020-01640-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Gornyk D, Harries M, Glöckner S, et al. SARS-CoV-2 seroprevalence in Germany: a population based sequential study in five regions. medRxiv 2021. 10.1101/2021.05.04.21256597 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Šmigelskas K, Petrikonis K, Kasiulevičius V, Kęstutis P, et al. SARS-CoV-2 seroprevalence in Lithuania: results of national population survey. Acta Med Litu 2021;28:48–58. 10.15388/Amed.2020.28.1.2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Alsuwaidi AR, Al Hosani FI, Al Memari S, et al. Seroprevalence of COVID-19 infection in the Emirate of abu dhabi, United Arab Emirates: a population-based cross-sectional study. Int J Epidemiol 2021;50:1077–90. 10.1093/ije/dyab077 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Khalatbari-Soltani S, Cumming RC, Delpierre C, et al. Importance of collecting data on socioeconomic determinants from the early stage of the COVID-19 outbreak onwards. J Epidemiol Community Health 2020;74:620–3. 10.1136/jech-2020-214297 [DOI] [PMC free article] [PubMed] [Google Scholar]