Research spotlight: Using RWE to evaluate complications of COVID-19
William Murk, Ph.D., M.P.H.
Jacobs School of Medicine & Biomedical Sciences
While COVID-19 is primarily a respiratory disease, its impact has also been associated with a range of other, non-respiratory complications. The risks of developing such complications have, in many cases, been difficult to assess and measure. A new study published in the Canadian Medical Association Journal uses real-world data (RWD) to better understand the risk of COVID-19 complications.
Many studies reporting on complications of COVID-19 have relied on small cohort or case series studies. As a result, researchers have not been able to establish a causal association between many of these complications and COVID-19, or provide risk estimates across different care settings.
Therefore, a team of researchers, including myself and co-authors from Aetion, HealthVerity, and the University of Toronto, sought to use RWD to study all possible complications of COVID-19 in a large group of COVID-positive patients.
Read on to learn more about the study, and what real-world evidence reveals about complications of COVID-19.
Goal of the study
Our research team aimed to evaluate every available International Classification of Diseases, 10th Revision, Clinical Modification (ICD-10-CM) diagnosis code to identify as many short-term COVID-19 complications as possible.
Process for the analysis
We used health insurance data from HealthVerity’s database of de-identified medical claims to compare the frequency of all available diagnosis codes. This came out to about 65,000 diagnosis codes, which we examined in an exposure-crossover study design. Specifically, we compared the frequency of codes before and after COVID-19 diagnosis. Codes that significantly increased in frequency within 30 days after COVID-19 diagnosis were defined as COVID-19 complications. We also estimated the risk of complications in different populations, grouped by age, sex, and whether or not they were hospitalized.
Our study included patients who received a COVID-19 diagnosis in the months of March and April 2020.
Key findings from the research
After identifying approximately 70,000 patients with COVID-19 in the data, we analyzed 1,724 ICD-10-CM codes. Of these, 69 codes were significantly associated with COVID-19.
Not surprisingly, we found that the most common complications were respiratory in nature, including pneumonia (with an overall risk in the studied population of 27.6 percent) and respiratory failure (with a 22.6 percent risk). Acute kidney failure and systemic inflammation were also common, with an 11.8 percent and 10.4 percent risk respectively.
We also confirmed that COVID-19 was associated with a variety of non-respiratory complications, including heart inflammation, heart attack, arrhythmia, blood clotting disorders, collapsed lung, and brain dysfunction. However, the overall risks of these complications were low—generally less than one or two percent.
It’s important to note that the risk estimates reported in the study only pertain to patients who sought medical care related to COVID-19. The research does not include people whose disease was mild enough that they chose to not seek care. Therefore, it’s likely that the risk estimates we provided are higher than the true risks in the full population of those with COVID-19.
Implications on COVID-19 understanding and research
Although the spectrum of disease for COVID-19 is wide, this study confirms that it is primarily a respiratory disorder for most patients. When we hear about unusual manifestations of the disease, it is important to understand the risks to appreciate how relevant those manifestations actually are. According to our study, the risks for most non-respiratory complications are generally low in the general population.
Future RWD analyses on long-term complications of COVID-19 can help us understand the lingering effects of the disease.
The Evidence Digest
Subscribe to our newsletter for insights and updates on research and applications for real-world evidence in life sciences and value-based health care.