Prevalence of gold stages are wrong

[aminadibi]

report_copd_by_ctime() returns this:

Which is almost the opposite of what it should be. Gold IV should have the lowest prevalence of all, and Gold I the highest.

[aminadibi]

This is how it looks in epicR v0.3.1, which shows slightly higher prevalence of GOLD II compared to Gold I.

Here are calculated GOLD stages for incidence cases:
> run(1000000)
[1] 0
> all_events_raw <- Cget_all_events_matrix()
> all_events <- as.data.frame(all_events_raw)
> COPD_events <- subset(all_events, event==4)
> 
> 
> table(COPD_events[,"gold"])/sum(table(COPD_events[,"gold"]))

         1          2          3          4 
0.39599318 0.50639386 0.07757886 0.02003410 

And same for prevalent cases:

> table(start_events[,"gold"])/sum(table(start_events[,"gold"]))

         1          2          3          4 
0.39910979 0.47873393 0.09000989 0.03214639 

To put it in perspective, these are CanCOLD GOLD stages for prevalent COPD cases:

and CanCOLD GOLD stages for incidence cases:

Mohsen believes that epicR numbers for GOLD stages are coherent, and the difference between CanCOLD and epicR are not significant, most likely due to a difference in population.

So, we're good!

[aminadibi]

``