Concerns About Methodology on "Curbing COVID-19 progression and mortality with traditional Chinese medicine among hospitalized patients with COVID-19".

Dear Editor,

In their recent article published in Pharmacological Research, Tseng et al. investigating the effect of the Traditional Chinese Medication (TCM) formulae NRICM101 and NRICM102 on the clinical outcomes of unvaccinated and hospitalized patients with laboratory confirmed mild-to-severe COVID-19 [1]. Although Tseng et al. demonstrated that oral TCM NRICM101 and NRICM102 plus usual care provide additional benefits in reducing the risk of intubation, intensive care unit admission and death than the controls, [1] we have several concerns about some residual confounders.

First, despite the similar distribution of baseline characteristics (age, sex, BMI, smoking, alcohol consumption, and most comorbidities) between the TCM and control groups, the two groups did not have balanced proportions of chronic heart disease and chronic lung disease. The TCM group NRICM101 even had more participants receiving remdesivir and dexamethasone than the control group, which may have biased their reporting [2].

Second, for retrospective cohort study based on hierarchical data in survival analysis, it is crucial to consider the problem of correlation within the same stratum (eg. data from the same hospital will be more similar, while the data from hospital to hospital will be different). It is therefore recommended to consider a Cox model with frailty, taking into account the correlation (described by random effects) within clusters at each levels of hospital [3].

Third, as the authors mentioned in their study, they intentionally excluded patients who received the COVID-19 vaccine and those who were prescribed Paxlovid during their study. The readers may be very interested in further studying the interactions between these medications to more closely match the real-world situation.

Conflict of interest disclosures

None reported.