She had no way of knowing when she was asked that question that the test used was not proof that the babies had synthetic insulin in their blood streams. That's come out since. All she was doing was agreeing that if the babies had artificial insulin in their blood streams, somebody put it there.
Her defence team had access to the evidence being presented and had the opportunity to look into it and query it beforehand. Which surely would have been the natural course of action if you know you're innocent but they're saying this is a deliberate poisoning.
I can't recall where I saw this (perhaps the C5 doc or one of the articles), but someone said recently that they googled the insulin test and the test company states on their website that the test shouldn't be relied on for forensic purposes. Are we really to believe that the defence team for a huge high profile murder trial did less research on the test than random bloggers?
Even if it had been there, that answer wouldn't make her guilty.
I get that answer in itself doesn't make her guilty, but since she agreed that evidence as an actual fact of the case, the jury has to accept that her version of events is that she's simply unlucky and that also there is someone else going round poisoning babies, but it wasn't her. That position is a lot harder to accept as a plausible explanation.