Issues in COVID-19 research and statistical analyses (Part X)

As we have heard or know, COVID-19 research, especially in early 2020, was filled with observational studies, many of which were potentially underpowered, but it turns out, even randomized controlled trials. Sample size estimation went out the window. Even Lee (2020) published two articles about this situation. He reviewed articles published between January 1 to March 25, 2020. Of the 374 articles he reviewed, only 4 were clinical trials. Even amongst those trials, the sample size estimations were found to be unacceptable. His suggestion was for these studies to include a statistician. He stated sample size calculations were omitted from 40% of papers published on randomized controlled trials during the period of time when journals were expediting publication of research articles on COVID-19.

In another article, Dominic (2020) suggested that small N studies may not produce adequate results simply by not having enough power for their study. They suggest that perhaps instead of focusing on studies only within one’s own institution that it could be more beneficial to collaborate across academic medical centers or institutions. What is not discussed though is how one can ideally do that during a pandemic when one needs to get answers out quickly. Furthermore, If one is a statistician on these trials, how is one to handle being approached with data already collected, and no sample size planning done a priori. Should the statistician still analyze this data? These are questions to be still be answered and understood for pandemic planning purposes.

Keywords: COVID-19, sample size calculation. Observational studies, randomized controlled trials, Usha Govindarajulu

References:

Dominic P. Corsello, Deepa B. Gotur, Christopher L. Carroll, Faisal N. Masud, Steven Q. Simpson, Impact of Small-N Studies During a Pandemic, Chest, Volume 158, Issue 4, 2020, Pages 1338–1340, ISSN 0012–3692, https://doi.org/10.1016/j.chest.2020.05.581.

Lee PH. Sample sizes in COVID-19-related research. CMAJ. 2020;192(17):E461. doi:10.1503/cmaj.75308

Lee PH. The quality of the reported sample size calculation in clinical trials on COVID-19 patients indexed in PubMed. Eur J Intern Med. 2020;77:139–140. doi:10.1016/j.ejim.2020.04.057

https://www.statsols.com/hubfs/Resources_/How-to-calculate-sample-size-in-5-steps/step-1-plan-your-study-sample-size-calculator.png

Usha Govindarajulu is a writer and biostatistician . www.UshaGovindarajulu.com