We are all aware of the effect COVID-19 has had on our day-to-day lives, but how about COVID’s effects on academia? The COVID-19 pandemic caused an outbreak of scientific evidence. As of April 2021, the number of unique publications indexed by our living evidence database, created to retrieve COVID-19 publications, exceeded 160,000. In our research, we compared the patterns of scientific publications for the two infections, the Zika virus emergence in 2016 and SARS-CoV-2, to assess the evolution of the evidence.
The living evidence database was created initially for retrieving publications on Zika virus. In 2016, we included around 50-60 new Zika virus articles per week in our database. We thought it would be an easy transition to switch to indexing COVID-19 articles in late 2019. However, we were shocked by the sheer volume of published articles retrieved, which came to around 2,500 articles per week at the end of April 2021.
To assess the evidence evaluation for Zika virus, we identified and classified 2,286 publications in 2016. However, for SARS-CoV-2, although we recruited a group of international volunteer scientists with a background in health sciences to cope with the volume, we were only able to analyze a random sample of 5,294 (24%) out of 21,990 articles that were published before 24 May 2020.
A substantial proportion of articles indexed were non-original.
Proportions of epidemiological study designs
Our findings showed that a substantial proportion of articles indexed were non-original (commentaries, basic reviews, opinion papers, etc.) for both Zika virus (55%) and SARS-CoV-2 (34%, [95% Confidence interval (Cl): 33-35]). The role of preprints was more prominent at the start of the SARS-CoV-2 pandemic than the Zika virus epidemic.
Case reports and case series accounted for approximately 10% of the total body of evidence for both SARS-CoV-2 (10.7% [95% CI: 9.9-11.6]) and Zika virus (9.7%) research.
Case-control and cohort studies accounted for 4.0% [95% CI: 3.5-4.6] for SARS-CoV-2 and 0.8% for Zika virus.
Trials emerged in smaller numbers (27/5,294 for SARS -CoV-2, and 1/2,286 for Zika virus) and at the beginning of the outbreak, there were more mathematical modelling studies for SARS-CoV-2 (10.1%, [95% CI: 9.3-11.0]), when compared with Zika virus (3.2%).
Time-study type trends
Case reports, case series and cross-sectional studies were the first epidemiological study designs to be reported, together with non-original articles and reviews. Case-control and cohort studies followed later; this delay was more prominent in the Zika virus research.
In vivo and in vitro laboratory studies followed between case reports and controlled observational studies. Trials were the last type of study to be published.
What does it all mean?
Emerging infectious diseases are accompanied by a vast quantity of published evidence quickly, which is a significant challenge that clinicians, scientists, researchers, and even students have to face. Keeping up with the available evidence becomes a complicated task.
The speed of SARS-CoV-2 evidence accumulation was unprecedented. Although we recruited a large team of experienced scientists, we reached a point where we could not categorize all the evidence retrieved by our database within the first few months of the pandemic.
To tackle this issue, we believe the use of natural language processing methods to help classify and categorize evidence seems like a promising approach for the triage of publication types not only for SARS-CoV-2 but for future emerging diseases. Also, collaborative crowd-sourcing within scientists in the field could increase researchers’ efficiency and prevent research waste.
Assessing the evidence during emerging infections can help us identify which types of public health questions we can answer and when.
Assessing the evidence during emerging infections can help us identify which types of public health questions we can answer and when. Further research on evidence evaluation during emerging outbreaks could help improve the public health response. With the accumulation of evidence during particular situations such as the COVID-19 pandemic, the use of specific resources can save time.
Due to the amount of evidence published every day on this non-ending pandemic, students, clinicians, and stakeholders need to approach the published studies with caution. Some of the studies available might be providing wrong findings or conclusions, and people might ‘cherry pick’ what aligns with their beliefs. We need critical eyes to scrutinize all available evidence.