Yeah, I was thinking more in terms of 'end user does something stupid, now somebody gets to collect the plutonium dust' type problems. I suppose that the major advantage is that people are somewhat less likely to do dumb things to electronics that they'd need to cut open their abdomens to get at.
It's really the end-user/disposal problem that makes me nervous about nuclear batteries, not the 'will the engineers screw it up?' aspect. 'Sealed sources', containing various isotopes neatly packaged as radiation sources, are even simpler to implement than nuclear batteries; and generally aren't an engineering problem; but the DoE has gone to a lot of trouble
hunting down 'orphan sources' that have left responsible supervision for one reason or another; and it's hardly unheard of for those to end up in some 3rd world junkyard being crowbared open by people who have no idea what a mistake they are making.
Pacemakers have the advantage of a more or less automatic paper trail(since the diagnosis of cardiac abnormality and implantation surgery tend not to be handled in cash and off the books) and people don't tend to cut through their own bodies in order to do stupid things to their gadgets; but I'd be rather pessimistic about the possibility of sound lifecycle management for nuclear batteries in broader application.
It's too bad; because they'd be extremely useful for a variety of low power off-grid stuff; but when people can't even be bothered to separate their trash from their recyclables; it's hard to be optimistic about their safe disposal of nuclear batteries.