Ten years ago, Queen Mary had a study programme in Discrete Mathematics. We were one of the first universities in Britain to do this, and I believe they invented a special UCAS code for the course. The study programme was designed by Rosemary Bailey, Leonard Soicher and me.
The study programme had three strands, with emphasis on pure mathematics, statistics, and computer science respectively, and students could mix and match to some extent. It cost nothing, since we used existing modules in Mathematics and Computer Science, but it gave us a sense of being ahead of the game.
It didn’t last long. Although it was a good recruitment tool for undergraduates, not many of them took this particular study programme, and the administrators decided it was not worth keeping.
Recently I stumbled on the slides I used in publicity for the study programme. For historical interest, I have posted them here.
Peter Cameron's Blog 
I suspect the problem with the “Discrete Mathematics: programme (G111, if I remember correctly) was that many students who find that they like discrete mathematics don’t really realise this until they’re actually studying mathematics at university: unless they do one of the “Decision Maths” modules at A-level, any exposure they have to discrete mathematics is likely to be very limited. (In my own experience–which is, however, now 15 years out of date–all we did was permutations and combinations, and we only did that in order to do basic probability and the binomial theorem, so we all regarded it as part of “stats”.) Then, as all the G111 options were available to BSc Mathematics (G100) students, there was no point going through the layer of bureaucracy needed to switch from G100 to G111.