If you look at the viral load across time where remdesivir bounces slightly higher or lower than placebo it seems remdesivir is no better than placebo at controlling viral loads. If true than the quicker time to recovery is either a statistical anomaly, a study design flaw or manipulated data.