A report looking at information from 97 serious case reviews that Cafcass contributed to between 2009 and 2016 found that domestic abuse was the most common risk factor, occurring in 71 per cent of the cases examined.
But in 33 of the 69 cases (48 per cent) that involved domestic abuse, the person suspected of killing or seriously harming the child was not the alleged domestic abuser.
Richard Green, national child care policy manager at Cafcass, said these cases suggested that domestic violence can sometimes overshadow other less-evident threats to children.
"We found many different contexts in which the incident took place, presenting significant challenges for those charged with safeguarding children," said Green.
"The suspected perpetrator of the serious incident was not necessarily the person that was thought to pose the greatest risk during proceedings.
"Some children were harmed or died in an alternative placement, or at the hands of the parent considered to be the ‘safe' one."
In 14 of the 33 cases where the domestic abuser was not responsible, the source of harm was the young person themselves.
In its analysis report Cafcass said: "In one example the risk posed by the very violent male was thought to have masked other less evident risk factors connected to the mother (self-harm, her volatility and outbursts of anger, possible drug misuse) and the child (the impact of witnessing domestic abuse and her anxiety in the company of men)."
The report, however, noted that the dangers obscured by the risk from the violent adult only became clear with hindsight. This, it said, is why few SCRs conclude that a child's death or serious harm could have been predicted or prevented.
In the cases analysed by Cafcass, 28 per cent involved domestic abuse, substance misuse, and parental mental illness - the so-called "toxic trio". At least one toxic trio factor was present in 89 per cent of the cases.
Cafcass's report said that the vast majority of SCRs do not offer new lessons and instead reiterate what is already known about the difference between strong and weak safeguarding practice.