To the Editor In a recent cohort study, Aldred et al1 found that early tecovirimat treatment within 7 days of symptoms was associated with reduced progression of mpox disease in people with HIV. The authors recommend initiating early tecovirimat treatment for all patients with HIV and immunocompromised status or mucosal site involvement.
The propensity score matching procedure was not clearly justified in the Methods.1 Although the greedy nearest-neighbor 1:1 propensity score matching ensured minimal exclusion of treatment participants, this may produce poor-quality matches. Setting a caliper width may improve matching quality by reducing poor matches despite risking an increase in unmatched individuals. Therefore, as an alternative to matching, we suggest implementing weights, such as standardized mortality ratio or inverse-probability-of-treatment weights to include all available data. For the purpose of this study,1 standardized mortality ratio may be a more suitable weighting scheme, comparing the effect of early tecovirimat to late or no tecovirimat administration (average treatment effect of the treated). We also suggest the authors consider stratification to explore heterogeneity of the effect estimate across different propensity score quantiles.2