• 98% of the 53 studies to date report positive effects (25 statistically significant in isolation).• Random effects meta-analysis with pooled effects shows 81% and 85% improvement for early treatment and prophylaxis (RR 0.19 [0.10-0.39] and 0.15 [0.09-0.25]). Results are similar after exclusion based sensitivity analysis, with 83% and 87% improvement (RR 0.17 [0.11-0.25] and 0.13 [0.07-0.25]).• 76% and 80% lower mortality is observed for all treatment delays and for early treatment (RR 0.24 [0.14-0.42] and 0.20 [0.06-0.60]).• 96% of the 27 Randomized Controlled Trials (RCTs) report positive effects, with an estimated improvement of 74% and 83% for early treatment and prophylaxis (RR 0.26 [0.16-0.42] and 0.17 [0.05-0.61]).• The probability that an ineffective treatment generated results as positive as the 53 studies to date is estimated to be 1 in 167 trillion (p = 0.000000000000006). It is extremely unlikely that the observed results could have occurred by chance.• Heterogeneity arises from many factors including treatment delay, patient population, the effect measured, variants, and treatment regimens. The consistency of positive results across a wide variety of cases is remarkable. Heterogeneity is low when looking at specific cases, for example early treatment mortality.• All data to reproduce this paper and the sources are in the appendix. See [Bryant, Hill, Kory, Lawrie, Nardelli] for other meta analyses, all with similar results confirming effectiveness.

	Improvement	Studies	Authors	Patients
Early treatment	81% [61‑90%]	19	177	1,962
Late treatment	43% [27‑56%]	20	143	6,831
Prophylaxis	85% [75‑91%]	14	108	8,789
Mortality	76% [58‑86%]	19	157	7,292
RCTs only	64% [48‑76%]	27	246	4,849
All studies	72% [64‑78%]	53	428	17,582

Evidence base used for other COVID-19 approvals
Medication	Studies	Patients	Improvement
Remdesivir (USA)	1	1,063	31%
Casiri/imdevimab (USA)	1	799	66%
Budesonide (UK)	1	1,779	17%
Ivermectin evidence	53	17,582	72% [64‑78%]