An international network study with Danish participation has examined the demographics, health conditions and medication use of 34,128 adults hospitalized with COVID-19.
New comprehensive research among 34,128 adults hospitalized with COVID-19 in Spain, the United States and South Korea, shows common comorbidities such as lung cancer, chronic obstructive pulmonary disease, cardiovascular diseases and diabetes. Many are taking medication for these diseases and have other infections when admitted.
The results are not surprising, but the tool researchers used to derive the conclusions opens up unique and interesting possibilities for monitoring the development of the pandemic.
“Many data are available on COVID-19 but also on many other diseases. However, they are rarely assessed because they are not in the same data format and therefore difficult to compare and analyse. Using a Common Data Model enables us to much more easily analyse the results and study the characteristics of people with COVID-19,” explains the Danish contributor to the major international study, PhD student Benjamin Skov Kaas-Hansen, Clinical Pharmacology Unit, Zealand University Hospital, Roskilde, Denmark and the Novo Nordisk Foundation Center for Protein Research, University of Copenhagen.
The research results have been published in Nature Communications.
Comparing people with COVID-19 internationally
The research shows that more men than women have been hospitalized with COVID-19 in Spain and the United States, whereas women have been hardest hit in South Korea.
About 40% had cardiovascular disease, a similar proportion had a metabolic disorder and about 75% had a lung disease.
The researchers also compared the characteristics of people hospitalized with COVID-19 in 2020 with those of people hospitalized with influenza from 2014 to 2019.
Those hospitalized with COVID-19 were more often men, generally younger, had fewer comorbidities and used less medication than those hospitalized with influenza.
The results indicate that a treatment protocol for people with COVID-19 that is similar to that for people with influenza may generally be beneficial, but the groups still differ and the protocol should reflect the specific characteristics of people with COVID-19.
“These data can also be explored in greater depth to determine whether some patients have a higher risk of dying or being hospitalized for a long time – even though this was not the purpose of the current study. Adapting a treatment protocol to account for specific characteristics may make sense,” says Benjamin Skov Kaas-Hansen.
Unique approach to processing data
The results are based on data obtained from March and April 2020, and doctors are already aware of the basic conclusions today.
But according to Benjamin Skov Kaas-Hansen, the method the researchers used to process the data is at least as interesting.
When researchers collect data, such as health data, the data format often varies, and the data are therefore difficult to compare with other data sets.
The Observational Health Data Sciences and Informatics (OHDSI) consortium has therefore created a constantly evolving open-source ecosystem to process and analyse health data: the Common Data Model of the Observational Medical Outcomes Partnership (OMOP).
Researchers who collect or harmonize their health data through the Common Data Model format have access to many existing tools for analysing their data to confirm or refute their results based on data from other researchers and institutions.
Benjamin Skov Kaas-Hansen explains that the system is based on a privacy-by-design approach, in which data analysis protects sensitive personal information, because the researchers only exchange their analyses and aggregated results.
“Researchers who find interesting links or signals in their data almost always benefit from testing the results in another population, and OHDSI and the OMOP Common Data Model make this easy. In this case, we could quickly and easily compare people with COVID-19 versus influenza across countries and regions, because the data were already in the Common Data Model format and we only had to design and exchange the analysis,” adds Benjamin Skov Kaas-Hansen.
Monitoring characteristics over time
Benjamin Skov Kaas-Hansen says that the Common Data Model enables specific data to be extracted easily, and this provides interesting perspectives.
For example, most of the work to produce the results in the new study happened during a 3-day virtual symposium in April.
This easy access to compare data across countries or over time can make the OHDSI ecosystem a valuable partner in the future to monitor both the COVID-19 pandemic and future outbreaks of other diseases.
“We can monitor the characteristics of people with a specific disease over time. For example, people with COVID-19 had certain comorbidities in March and April, such as diabetes and lung diseases, but future COVID-19 patients might have other characteristics and the treatment protocol may therefore need to be adjusted. These tools enable us to regularly monitor whether and how these characteristics change and, if necessary, change the general approach to treatment,” concludes Benjamin Skov Kaas-Hansen.
“Deep phenotyping of 34,128 adult patients hospitalised with COVID-19 in an international network study” has been published in Nature Communications. Benjamin Skov Kaas-Hansen is affiliated with the Novo Nordisk Foundation Center for Protein Research, University of Copenhagen.