Our best chance to fight CV is to create a vaccine as soon as possible. The irony is that we probably already have a vaccine but we can’t prove that it works, until animal safety test, 1 and 2 stage of the clinical trials, which will take 12-18 months at least. How we could accelerate the creation of the vaccine?

There are several ideas, which are mostly inspired by the research of anti-aging drugs, which suffer from the same problem: the need for very long clinical trials. See More Dakka post by Sara Constantin. 

1. Immediate human trials. If we have the vaccine in 3 months, not 18 months, we will save millions of lives. So, based on the trolley problem logic, we may risk the health of a few thousand people to achieve these goals, especially if they were volunteers. Thus we need to start a human test of the vaccine candidates immediately, even before the end of animal trials, which are still needed. here we get acceleration by performing in parallel the actions which are typically done sequentially.
2. Test safety and efficiency simultaneously. We also should combine 1 stage and 2 stages oа tests, that is safety and efficiency, by giving the vaccine to people who are already under potential exposure to CV, like nurses, or old people in nurseries. 
3. Test on large groups. We should test a vaccine on a large group of people, like 10 000, so any finding will quickly get statistical significance. If the number of infection will decline relative to the control group, we could see it in one month.
4. Test all vaccine candidates. All said above in 1-3 should be done with each of a dozen vaccine-candidates which are currently under developing. As a result, in one month we could know which vaccine candidate is the strongest and safest. 
5. Manufacturing in advance. Simultaneously, we should start large scale production of all vaccine-candidates, if it is possible. After the best (maybe 3 or 5) vaccine-candidates are validated, the stockpile of failed vaccine-candidates is destroyed, and the best vaccines are delivered to the population. This will help to fight the production delay.
6. Give people different best vaccines. There is still could be long-term detrimental effects of some of the vaccines, so it may be better not to give everybody just one best vaccine - what if it makes everybody will become sterile in 1 year?
7. Combine best vaccines. To increase protection, we could give each person a combination of several (but not all to ensure point 6) of the best vaccines. Here I assume that detrimental interaction between vaccines is typically unlikely, but more technical analysis is needed. 
8. Establish biomarkers of a good vaccine (e.g. antibodies). Biomarkers are important to check efficiency of a clinical trial before the final outcome is known.
9. Try other approaches, like DRACO, distancing, coconut oil, a large dose of vitamin C etc and test them in the same accelerated way.

Several human trials already started: In China military volunteers are testing an experimental vaccine, and there are clinical trials of Moderna vaccine in the US.