I think the problem is that the prosecution, in presenting their analysis, ignored the deaths that occurred when she wasn't on shift. No idea whether she did it or not but I think there's a need for analysis by an independent bayesian statistical expert.
It's starting to look like her defence were negligent. However, to provide one strand of mitigation, getting real experts to provide their expertise as a defence witness is, as I understand it, hugely difficult because it can and has been shown to ruin careers. Seems the medical profession don't like people talking. Once this is investigated, properly, I think that, rather than a retrial, it is just as likely that there won't be a trial at all because the prosecution will withdraw the charges or the judge will rule no case to answer.
I think the publicity following the guilty verdict has prompted further investigation by experts who might not have looked at the case before it went to trial. If you do eliminate some of the suspected cases it becomes less of a pattern and much more circumstantial.
