NIH COVID-19 treatment analysis

Authority review was extremely slow and overlooked most research. Official treatment recommendations were often made based on a small fraction of the evidence, with substantial bias in the selection of studies that were reviewed. Just as 78% of experts polled regarding the origin of COVID-19 were not familiar with the DEFUSE protocol1, experts on COVID-19 treatments were not familiar with much of the key evidence. In some cases, recommendations were made without reviewing any clinical evidence, for example the US FDA recommended against ivermectin in an article that stated "The FDA has not reviewed data to support use of ivermectin in COVID-19 patients to treat or to prevent COVID-19"2. In cases where the people involved in recommendations are known, they were often very busy professionals that could not realistically have time to carefully review all of the evidence3.

NIH reviews. NIH treatment reviews were done by a panel of 40+ external experts with extensive conflicts of interest towards high-profit treatments3,4. Out of 8,965 proposed treatments5, they found only three high-profit drugs from top lobbying companies to be beneficial for early treatment(a). Most others received either "insufficient evidence" or no review at all, despite extensive unreviewed evidence.

For vitamin D, the NIH claims "There is insufficient evidence for the Panel to recommend either for or against the use of vitamin D for the prevention or treatment of COVID-19"6. No authors are listed, they provide only a brief narrative review with no quantitative analysis, and they reference only 5 of the 125 controlled studies (30 RCTs). The state is similar or worse for other treatments as shown in Table 1. The NIH provides no quantitative analysis for any of the treatments listed.

Considering RCTs for vitamin D, the NIH references only 4 trials7-10, missing 26 other RCTs11-36 as shown in Figure 137. The NIH selection does not correspond to the most relevant and highest quality studies, for example including Murai et al.8 and Mariani et al.9, which study very late treatment using bolus doses of cholecalciferol—a trifecta of poor design and irrelevance compared to recommended use: early treatment is better, continuous dosing is more effective than bolus doses, and calcifediol, calcitriol, or analogs avoid very long delays in conversion. They include none of the early treatment RCTs. We provide similar analysis for vitamin C38 and zinc39.

Across all low-cost treatments that reduce risk (within the 121 we cover), there are 1,817 studies including 511 RCTs, of which the NIH only reviewed 2% of the studies (6% of RCTs).

Waiting for RCTs costs lives. While the NIH examined few of the RCTs, they dismissed almost all observational studies. RCTs can provide significant advantages, however they show no benefit for COVID-19 and they introduce significantly delay. For the 121 treatments analyzed, there is no difference in results between RCTs and observational studies, RR 1.00 [0.93‑1.08] as shown in Figure 4. Observational studies do not systematically overestimate or underestimate efficacy, as found in previous research40,41, and they provided confirmation of efficacy 7+ months faster. RCTs aim to equalize study groups, but add their own biases. For acute diseases with strong benefits for earlier treatment, the typical increased treatment delay adds a major confounding factor. Both RCTs and observational studies span the bias spectrum, from minimal to extreme. Studies must be evaluated individually.

Waiting for specific outcomes costs lives. While patented treatments may be approved with a single non-clinical result42, authorities often dismiss results for low-cost treatments where studies report different outcomes. To avoid delay and unnecessary mortality, we must use all available information. Logically, minimizing viral replication will minimize serious outcomes. Singh et al.43 confirm, showing that higher viral clearance was significantly associated with lower hospitalization/death. We confirm this across all 121 treatments covered, finding higher viral clearance strongly associated with lower serious outcomes, p = 0.0000000094. For more detailed discussion see44.

Novel treatments are high risk, existing treatments have a critical safety advantage. Existing treatments have a strong advantage with known pharmacokinetics and safety profiles. New agents are more risky. By definition, long-term risks cannot be known, and known risks may not be acknowledged for some time. For example, molnupiravir's potential risks include the creation of dangerous variants45-49, mutagenicity, carcinogenicity, teratogenicity, and embryotoxicity50-63. While the risk in this case always exceeded the benefits within the context of all treatments, and confirmation of harm continues to accumulate, the treatment is still used in 2025 in some locations. Novel treatments saw limited use during the pandemic64,65, partly due to prescription requirements, limited availability, high cost, and, for certain treatments, administration requirements and the need to evaluate drug interactions.