So far I’ve found 6 vaccine candidates that are reporting positive results from clinical trials.
Moderna  has been testing an mRNA vaccine called mRNA-1273, consisting of the RNA encoding a portion of the COVID-19 virus’s spike protein, wrapped in a lipid nanoparticle. Their Phase 1 study on 45 healthy adults aged 18-55 found that there was a dose-dependent antibody response (relative to baseline) lasting until the end of the 60-day study, and that all participants’ antibodies were able to neutralize at least 80% of the COVID-19 virus in vitro. The antibodies were as effective at neutralizing virus as those in convalescent plasma from patients who had recovered.
AstraZeneca  in collaboration with researchers from Oxford University, has been testing a vaccine called ChAdOx1, and has released interim results from a Phase 1 / 2 study on 1077 healthy adults aged 18-55. This vaccine is an adenovirus vector expressing the spike protein of the COVID-19 virus. The trial found that the antibody levels were higher in vaccinated subjects than controls, with a comparable magnitude of antibody response to convalescent plasma, and with the elevated response lasting throughout the 60-day trial. Of the 35 vaccinated patients tested for a neutralizing response, all were able to neutralized COVID-19, while none of the control patients were.
A Chinese group  also tested a vaccine consisting of an adenovirus vector expressing a part of the COVID-19 virus’ spike protein, on 508 adults aged 18-60. The vaccinated subjects had significantly more antibody response and neutralizing ability than the placebo group. 96-97% of vaccinated subjects had antibodies against spike protein; 59% had neutralizing antibodies. The vaccine also produced significantly more T-cell response than placebo.
A German company, BioNTek, produced another vaccine candidate based on mRNA of a portion of the spike protein, this one attached to a “foldon” peptide to improve immunogenicity. In an un-peer-reviewed preprint  they report an uncontrolled trial on 60 healthy adults age 18-55. Their average antibody response to the virus was significantly higher and more effective at neutralizing COVID-19 than the antibodies in convalescent plasma.
Two other biotech companies, Sinovac  and Inovio , have announced in press releases that their initial human vaccine trials had positive results; Sinovac claims 90% of its vaccinated subjects had neutralizing antibodies and Inovio claims 94% of its vaccinated subjects had an antibody response to the vaccine. Neither company has released a paper or additional data.
Ok, so what does this mean?
Looking for antibody response is a pretty standard way to test vaccines. The purpose of a vaccine is to stimulate the immune system to produce antibodies (and potentially other responses) to the virus, so that when you get the actual virus you’ll be able to produce the right antibodies faster, in greater quantities, and for a greater length of time. So, given that experimental ethics rules make it hard to actually infect people with the virus and seeing if the vaccine protects them, researchers usually use antibody response as a proxy.
Do the vaccinated subjects produce antibodies that actually bind to the viral antigen? How much do they produce? And do those antibodies neutralize the virus, i.e. prevent it from spreading in a petri dish full of cells?
Like all proxy metrics, antibody response is not a perfect substitute for finding out what happens in a real world scenario – in this case, how well the vaccine will protect against infection. But I think it’s a reasonable proxy to use.
The base rates of vaccines that enter clinical trials ultimately getting approved by the FDA are actually not bad – quite a bit higher than the clinical success rates of other drugs.
Based on a dataset of clinical trials between 2000 and 2015,  it turns out that vaccines for infectious diseases have an overall success rate of 33.7% from Phase I trials to approval; that is, once a vaccine _enters _ clinical trials, it has about a ⅓ chance of being found safe and effective enough to satisfy the FDA. For all drugs and vaccines taken as a whole, that number is more like 1/20.
Another study on a clinical trial database ranging from 1995 to 2017  reached a similar conclusion: a vaccine for infectious disease that enters clinical trials has a 31% chance of ultimately reaching approval.
The riskiest stage in this process is going from phase 2 to phase 3, which is a 61% probability; Moderna has already passed that test, having begun its phase 3 trial, and now, on priors alone, it has an 80% chance of approval.
There’s a more pessimistic number floating around, that says vaccine candidates have only a 6% chance of success, but I don’t think it’s applicable to the current COVID-19 situation. The number comes from a study  that looks at the probability that a _preclinical _vaccine candidate will ultimately be approved – that means, candidates that have only been tested on animals so far, which are obviously riskier than candidates that showed positive enough results to start testing in humans. Moreover, that study included vaccines for _noninfectious _diseases in that group, including cancer vaccines, which have a dismal track record.
If we’re thinking about how hopeful to be about COVID-19 vaccine candidates, we shouldn’t be lumping them in with chronic disease “vaccines”, which don’t work the same way and aren’t nearly as effective.
With as many vaccine candidates as are currently in the pipeline, I think chances are quite good that we’re going to see an approval somewhere.
The other big question, though, which _isn’t _going to show up in the clinical success rates, is how well this vaccine is going to work on the elderly.
COVID-19 is deadlier the older you get. Also, because immune system function declines with age, vaccines are less effective the older you get. Flu vaccines are basically ineffective on the old. And you don’t need to prove that a vaccine is effective on the elderly to get it approved.
Both the Moderna and Oxford vaccine teams are enrolling older adults in trials, so we’ll actually find out how well they work soon enough.
 Jackson, Lisa A., et al. “An mRNA Vaccine against SARS-CoV-2—Preliminary Report.” New England Journal of Medicine (202)
Folegatti, Pedro M. et al. “Safety and immunogenicity of the ChAdOx1 nCoV-19 vaccine against SARS-CoV-2: a preliminary report of a phase 1/2, single-blind, randomised controlled trial” The Lancet (2020)
 Zhu, Feng-Cai, et al. “Safety, tolerability, and immunogenicity of a recombinant adenovirus type-5 vectored COVID-19 vaccine: a dose-escalation, open-label, non-randomised, first-in-human trial.” The Lancet (2020).
Sahin, Ugur, et al. “Concurrent human antibody and TH1 type T-cell responses elicited by a COVID-19 RNA vaccine.” medRxiv, 2020
Wong, Chi Heem, Kien Wei Siah, and Andrew W. Lo. “Estimation of clinical trial success rates and related parameters.” Biostatistics 20.2 (2019): 273-286.
DiMasi, Joseph A., et al. “Development times and approval success rates for drugs to treat infectious diseases.” Clinical Pharmacology & Therapeutics 107.2 (2020): 324-332
Pronker, Esther S., et al. “Risk in vaccine research and development quantified.” PloS one 8.3 (2013): e57755.