The short answer is "we don't know." In other words, at this moment, there is no scientific evidence that allows us to assert that any treatment has more benefits than risks.

Under normal conditions, it would be advisable to patiently wait for information to reach us through our doctor or the relevant authority. However, considering that we are not in normal conditions, many people want to understand a bit more about the claims that are circulating or the differences of opinion among many authorities or experts.

To understand the arguments behind each of the suggestions that have been circulating and will continue to do so in the coming days, it is necessary to understand some concepts:

Direct evidence: Scientific information of sufficient quality to be incorporated into decision-making in humans, derived from studies evaluating the intervention (treatment or action) of interest directly in the disease in question, that is, "Coronavirus Disease 2019 (COVID-19)". So far, there are very few completed studies, and the ones available constitute evidence of very low certainty. For example, a systematic review evaluating the role of masks in preventing the transmission of COVID-19 concluded that there is no direct evidence in this population, so we can only extrapolate information from reasonably similar diseases, which is indirect evidence.

Indirect evidence: Scientific information from studies that do not directly assess the intervention of interest in COVID-19 but were conducted in similar diseases. Usually, diseases are categorized as more direct or less direct. For example, evidence from studies on SARS-CoV or MERS-CoV (epidemic diseases caused by coronaviruses) is considered more direct than evidence from influenza, which, in turn, is less indirect than evidence from any respiratory disease.

Preliminary evidence: Studies that are not yet relevant for decision-making, such as studies on animals or in cell cultures. Initial studies in humans that do not meet certain criteria are also considered preliminary evidence. The aim of this type of evidence is to generate hypotheses so that researchers can continue to advance towards human use. While this information is not relevant for decision-making, it is common for the media to report on it, not always taking the necessary precautions to make it clear that it is preliminary evidence. For example, the Twitter account @justsaysinmice focuses on pointing out news based on evidence from studies in mice.

Systematic reviews: This type of evidence is considered the basis for decision-making today. It is a type of scientific article in which an exhaustive and transparent effort is made to collect all the scientific information on a specific question (Does intervention X work for disease Y?). Thousands of these reviews are produced every week. The Epistemonikos database is the world's largest repository of systematic reviews in health. The L.OVE platform takes this information and organizes it by topic and question, using both artificial intelligence and a network of expert collaborators who validate the information. Any systematic review assessing the evidence directly for a treatment in COVID-19 will be immediately identified in the Coronavirus infection section of the L.OVE platform.

Certainty of the evidence: The set of evidence (ideally obtained from a systematic review) can be more or less reliable. Research provides very low certainty about what will happen in reality if the studies conducted have few patients or design flaws (i.e., risk of bias). As a general rule, low or very low certainty evidence does not allow for strong recommendations, so decisions will often be more influenced by other factors. For example, the evidence of the benefits of gargling with saltwater to prevent COVID-19 transmission is of very low certainty, so it is expected that different experts may have conflicting opinions on this measure, especially considering that the risks and costs of this measure are minimal.

Publication bias: Publication bias corresponds to the publication or non-publication of research results based on the nature and/or direction of their findings. Typically, studies that show more spectacular results are disclosed first. Systematic reviews counteract this effect by transparently analyzing the existing evidence.