KM figures cut off at ends with visr(x_ticks=)

[ddsjoberg]

When using visr.survfit(x_ticks=) to reduce the range of values shown, the curves a cut off early, rather than being zoomed in on. In the example below, the KM curves are cut off, but you can see from the risk table that them are still patients at risk at 10 years and the lines do not appear.

Perhaps we should use coord zoom rather than xlim?

[timtreis]

Played around with it hoping for a quick fix and it looks a bit more complex than a 30 min fix. It seems the line until the end is primarily motivated by the next drop. Would it even be numerically correct to extend the line since at that point we have no data basis for it

[ddsjoberg]

I think if we switch from xlim to coord_cartesian(), the figure will zoom-in to the specified limits after the graph is constructed, rather than delete data outside the limits then build graph. @timtreis and I agree, not a simple fix!