I wanted to comment on the numbers.
When scientists plan a study, if they are careful, they take into account factors like how many people they need to test to see a difference or a result, etc. These of course are often guesses but it helps them calculate the sample size needed. Depending on the factors and condition looked at, you can get great results with small numbers or big numbers of subjects. There is no arbitrary set number like 50, 500, 1000, etc. that is viewed as "large." For a common condition and a drug which shows dramatic effects for example, you might need say 100 subjects whereas for for a rare condition with a drug that doesn't have dramatic effects, 1000 might not be enough.
I think the WPI tried to take homebound folks into account by arranging for homedraws. If say 30% of people were homebound and 30 of 100 drawn were homebound, it would be pretty representative. If it turns out after all draws completed that only 10% were homebound, then it might not be representative.
All I'm saying is that we can't take a single number and say it's too small without taking into account multiple factors. The other point also is that the priority of initial studies is not always representativeness of the condition study; in this case, all WPI needs to do is prove XMRV positives in the UK first to encourage more studies.