For question 2, I don' think anyone knows the admittance rate since you will need to know the number of people with either no symptom or very mild symptoms. Since neither of these people will go to the hospital, we will never find out.
Ref Point 2. Wouldn't you get the overall hospital admittance rate for serious cases with a large scale sampling of the Wuhan (or Guangdong) population with a blood antibody test? This will give you the overall number of infected. Presumably this is happening very soon, if not already.
For question 3, most people with pneumonia will develop calcified granulomas in the lungs. Younger people will eventually fully recover, while older people may never fully recover. The COVID-19 will be similar. As to people with mild symptoms, no one knows at this point. We will find out later.
3a. Given a very healthy cohort of 18-40 year old subjects with no pre-existing conditions, what sort of timescales are we typically looking at for a full recovery from pneumonia? And is pneumonia the only realistic complication?
3b. Are there any comparable coronaviruses which would give us an indication on whether that healthy 18-40 cohort with mild symptoms would suffer any long-term effects?
For question 4, that would be yes since the test is based on detecting antibodies in the blood. For Coronaviruses, the antibodies typically last about 6 months because the viruses keep mutating. Eventually, they will mutate too much and your antibodies won't recognize the virus.
Say that healthy 18-40 cohort has been infected by a particular strain of the Coronavirus, how long could we expect immunity to that same strain to last for:
4a. Assuming they weren't exposed again?
4b. Assuming they were to be continually exposed to low levels of the same coronavirus strain?