COVID-19: Persons tested, confirmed cases and associated hospitalizations by education and income
Report
|Published
This report demonstrates that the COVID-19 pandemic has hit people with low education and household income disproportionately, although adjustment for age, sex, municipality of residence and country of birth decreases the differences.
Key message
Persons with low education and persons with low household income have been hit harder by the COVID-19 pandemic than the rest of the population. Groups with low education and income have in general been overrepresented in confirmed cases, hospitalizations, invasive ventilation, and deaths related to COVID-19, compared to groups with higher education and higher household income. Persons with higher education and persons with higher household income were overrepresented in confirmed cases and related hospitalizations in the beginning of the first wave of the pandemic, while persons with low education and low household income have had the highest relative numbers of confirmed cases and related hospitalizations during the rest of the study period (until May 2021). There was a positive correlation between increasing education from below upper secondary education to university/college education and test rate. For income, the relationship after 15th June 2020 was slightly S-shaped, with generally higher test rate in the higher income deciles compared to the lower income deciles.
When we adjust for age, sex, municipality of residence and country of birth, the differences in confirmed cases and related hospitalizations decrease, however are rarely eliminated. People with lower education remain slightly overrepresented in confirmed cases and related hospitalizations also after adjustment. However, the overrepresentation of confirmed cases among people with lower household income was fully attenuated after adjustments. The highest income groups still had a somewhat lower hospitalisation rate. The positive relationship between test rate and education (from below upper secondary education to university/college education), and between testing rate and household income remained after adjustments. Overall, adjustment for country of birth made the most difference to the results, especially for the category undisclosed/no education, which was predominantly made up of immigrants.
This report demonstrates that the COVID-19 pandemic has hit people with low education and household income disproportionately, although adjustment for age, sex, municipality of residence and country of birth decreases the differences. We have not fully examined possible reasons why some groups are overrepresented, but family and household structure, employment, health literacy, delayed access to health services or underlying diseases are discussed as underlying reasons. It will be important to obtain more knowledge about these or other mechanisms behind the overrepresentation in future studies.
Summary
Introduction
Reports from several countries indicate that people with low education and income are more likely to be affected by COVID-19. In this report we investigate how education level and household income are related to test activity, confirmed cases, hospitalizations, invasive ventilation, and death related to COVID-19 in Norway.
Methods
We have utilized the nation-wide emergency preparedness register, BeredtC19, which contains individual-level data covering the entire Norwegian population. The study population consists of individuals with a personal identification number, alive and registered as residents per 1st March 2020 and are aged 25 years or older. The study period was February 2020 to May 2021. The exposure variables, highest attained education and household income after tax (equivalent income), are operationalised, respectively, into five categories for education, and into deciles for household income. Outcomes studied were test activity (being tested and share of positive tests among those tested), laboratory confirmed cases, related hospitalizations, use of invasive ventilation and death. We have used both descriptive analyses and linear regressions. For the outcomes test activity, confirmed cases, and hospitalization, we adjust for age, sex, municipality of residence and country of birth.
Results
The sample comprised 3 882 249 persons, of which 1 864 860 were tested at least once, 75 698 were confirmed cases, 8 222 hospitalizations, 821 received invasive ventilation, and 710 died. In total, 20 % had attained below upper secondary education and 37 % university or college education. The median household income was 644 143 NOK. The lowest income decile and the undisclosed/no education group were on average younger compared to the other categories. Immigrants made up 93 % of those in the undisclosed/no education group and 40 % of the 1st (lowest) income decile.
There was a positive correlation between increasing education from below upper secondary education to university/college education and test rate, which remained after adjustments. For household income, the relationship after 15th June 2020 was slightly S-shaped, with generally higher testing rates in the higher income deciles compared to the lower income deciles. After adjustment there was a clear positive relationship between household income and testing rate.
Both lower education level and household income were associated with elevated risk of infection and with more severe disease (hospitalization, invasive ventilation, and death). Persons in the three lowest household income deciles and those with a below upper secondary education or undisclosed/no education have been hit hardest. The differences between the different education and income groups in confirmed cases and related hospitalizations decrease when we adjust for age, sex, municipality of residence and country of birth. For confirmed cases, the differences between household income deciles were fully attenuated. For other outcomes, differences were somewhat decreased, however there were still clear trends even after adjustment. Adjusting for immigrant status (country of birth) made the largest impact overall, especially for the category undisclosed/no education, which was predominantly immigrants.
Discussion
These findings correspond with other reports and studies on the relationship between COVID-19 (and health in general) and education and income. This report does not empirically examine why people with low education and low household income have been disproportionally affected, but factors such as working conditions, living conditions, health literacy, delayed access to health services or underlying diseases are suggested as possible mechanisms.
Conclusion
Individuals with lower education and household income have had higher rates of confirmed cases, hospitalization, invasive ventilation and deaths related to COVID-19. When we adjust for age, sex, municipality of residence and country of birth, differences between the groups are reduced, but rarely eliminated. Of the factors included, country of birth had the largest impact in accounting for the differences between the education and income groups. Going forward, it will be important to obtain more knowledge about the causes and mechanisms behind the observed overrepresentation.
