Vaccines, Drugs, and Myths About the Novel Coronavirus: Understanding the Boundaries of Evidence

The conclusion in one sentence

A sound assessment of evidence on the novel coronavirus must evaluate the prevention of infection, prevention of severe disease, shortened recovery, reduced mortality, and individual safety signals as separate outcomes. A single case, a mechanistic hypothesis, or a social-media slogan cannot substitute for clinical evidence. PMID: 38850519

Vaccines are not treatments, but they are far from “useless”

The primary role of COVID-19 vaccines is to prevent deterioration and hospitalization after infection, not to treat someone who has already been diagnosed. Calling a preventive measure a treatment and then rejecting it because it does not treat established disease confuses two different questions. PMID: 42224418

Randomized trials and systematic reviews support the efficacy of mRNA vaccines in preventing COVID-19 and show no overall difference in serious adverse events. Claims that vaccines offer no protection, or that their harms must exceed those of the virus, are therefore not supported by the evidence. PMID: 35061702

During the early Omicron period, an mRNA booster remained associated with a lower risk of symptomatic infection and also reduced related emergency-department visits, urgent care, and hospitalization. Protection may change as the virus evolves, but that does not mean it has disappeared altogether. PMID: 35060999

Mixed schedules and special populations: one rule does not fit everyone

Heterologous vaccination can elicit strong neutralizing-antibody and T-cell responses, with immunogenicity often no lower than that of homologous schedules. Local or systemic reactions may be more frequent, but overall reactogenicity remains tolerable. PMID: 35562753

Immunogenicity, however, is a surrogate endpoint; it does not automatically amount to direct proof that infection and hospitalization are prevented. Later studies of clinical effectiveness found heterologous schedules to be no less effective than homologous ones, providing a more complete boundary as the science evolved. PMID: 39346662

Patients receiving dialysis should not be considered ineligible for vaccination solely because they undergo dialysis. Studies show that they can mount an immune response without a distinctive safety signal, although the actual plan must still be tailored to their clinical condition. PMID: 34753894

Most people with autoimmune rheumatic diseases can still be vaccinated, but the certainty of the evidence behind early recommendations was low. Disease activity and immunomodulatory treatment should be assessed individually by the clinical team. PMID: 33993119

Small prospective studies in pregnant and lactating people found mRNA vaccines to be immunogenic and detected antibodies in cord blood and breast milk. This finding alone, however, cannot quantify the infant’s actual clinical protection. PMID: 33983379

Vaccine adverse events: a report is not proof of causation

The medical literature does contain reports of systemic lupus erythematosus flares or new diagnoses after vaccination, so it would be inaccurate to say that no such reports exist. Case reports primarily establish a temporal association; they cannot directly prove that the vaccine was the sole cause. PMID: 35936557

For people with systemic lupus erythematosus, the current literature continues to support an overall benefit of vaccination that exceeds its risk. An informed assessment should accommodate both rare safety signals and the harms of infection itself, rather than selecting only one side of the evidence. PMID: 35936557

On claims that mRNA vaccines have been confirmed to cause Alzheimer’s disease or other neurodegenerative disorders, the literature has advanced: subsequent observational research has proposed a possible association, but an observational association is still not causal confirmation. PMID: 38806183

Remdesivir: recovery time and mortality are different questions

Remdesivir is a nucleotide-analogue prodrug that inhibits viral RNA-dependent RNA polymerase and interferes with RNA replication. Its pharmacologic mechanism explains why the drug merits study; mechanism alone cannot establish clinical efficacy. PMID: 32943188

Reviews of controlled trials support the conclusion that remdesivir can shorten recovery time and increase the chance of clinical improvement, but overall mortality findings have been inconsistent. “May shorten recovery” is therefore not equivalent to “has reliably reduced mortality in every hospitalized patient.” PMID: 41766239

If observational data show longer hospital stays among treated patients, one also cannot immediately conclude that the drug prolonged hospitalization. Patients with more severe illness are more likely to receive treatment, and the treatment course itself may affect discharge planning. Controlled comparisons should remain the priority for causal inference. PMID: 41766239

Ivermectin and other “miracle cures”: large trials are decisive

Reliable outpatient research has not shown that ivermectin shortens recovery, and there is insufficient evidence that it prevents infection or reduces hospitalization or death. It should therefore not be marketed as a proven cure for the novel coronavirus. PMID: 40124821

Small studies, multidrug regimens, and reports of questionable quality can magnify an apparent effect. A study bearing the same name was later the subject of a retraction notice, underscoring why a retracted paper cannot continue to serve as evidence of efficacy. PMID: 39781906

An outpatient trial of high-dose zinc and vitamin C did not support symptom improvement, while an inpatient trial of high-dose vitamin D did not improve the primary clinical outcome. These nutrients therefore cannot be presented as effective substitutes for vaccination and standard treatment. PMID: 33576820

A coordinated randomized trial in hospitalized patients with COVID-19 likewise found a low probability that intravenous vitamin C improved the composite outcome of organ support and survival, offering no support for treating it as an established therapy. PMID: 37877585

Origins and transmission: evidence favoring a direction is not final proof

Bats are an important source of the original-host hypothesis, and pangolins have been proposed as a possible host. Yet “there may have been an intermediate host” cannot be rewritten as “all evidence proves that a particular intermediate host must have been involved.” PMID: 32218527

Subsequent research provided further support for the Huanan market in Wuhan as an early epicenter, making the claim that the outbreak originated in the United States even less scientifically grounded. The updated literature does not, however, mean that upstream events, the exact host, and the complete chain of origin have received final proof. PMID: 35881010

A method readers can take away

When one patient improves after receiving a drug, the change may reflect the natural course of illness, the immune response, or another concurrent treatment. Without a control group, sequence in time cannot be treated as causation. PMID: 38850519

A rise in antibodies can support the presence of an immune response, but it does not automatically tell us whether hospitalization or death is reduced. A case report can flag a safety signal, but it cannot by itself establish population-level causation. An observational association can generate a hypothesis, but controlled studies are still needed to address confounding. These boundaries both preserve the historical place of Professor Lin’s original article within the evidence available at the time and allow conclusions affected by updated literature to be redrawn through scientific progress.