Mathematically it depends on the usual S of attacks coming at them; an extra 1/6 chance of not being wounded combined with an extra 1/6 chance of failing the save is pretty much a wash. Looking at different S attacks, though...
S3 attacks (old T & shield) - wound 3/6, die 1/6 : 3/36 fatality
S3 attacks (old T & halb) - wound 3/6, die 1/6 : 3/36 fatality
S3 attacks (new T & shield) - wound 2/6, die 1/6 : 2/36 fatality
S3 attacks (new T & halb) - wound 2/6, die 1/6 : 2/36 fatality
S4 attacks (old T & shield) - wound 4/6, die 1/6 : 4/36 fatality
S4 attacks (old T & halb) - wound 4/6, die 2/6 : 8/36 fatality
S4 attacks (new T & shield) - wound 3/6, die 1/6 : 3/36 fatality
S4 attacks (new T & halb) - wound 3/6, die 2/6 : 6/36 fatality
S5 attacks (old T & shield) - wound 5/6, die 2/6 : 10/36 fatality
S5 attacks (old T & halb) - wound 5/6, die 3/6 : 15/36 fatality
S5 attacks (new T & shield) - wound 4/6, die 2/6 : 8/36 fatality
S5 attacks (new T & halb) - wound 4/6, die 3/6 : 12/36 fatality
So, the new T does not fully justify dropping the shield. It does, however, make it less of a bad choice in terms of survivrability - the difference between fatality levels is less pronounced.
Of course, if the majority of attacks coming at your DGKs are S3 or less, the halberd is an automatic choice because of the way a 1+ and 2+ AS are identical vs S3 attacks.