I think that is a reasonable gloss on the following:
> And yet, although subdural and retinal hemorrhage may be caused by non-accidental trauma, especially when impact is involved, they simply are not specific for it: indeed, it has been demonstrated that a wide range of accidental events and medical conditions are plausible alternative causes. Particularly fragile infants may sustain severe head injuries following minor household falls. Others may suffer from genetic conditions, metabolic disorders, blood clotting abnormalities, or infections.
> But in practice, extremely few medical conditions are checked for and “excluded” before concluding a diagnosis of abuse – the great majority are not checked for at all. Very often, abuse is diagnosed “by default”, because no known alternative explanation was found (or even actively sought). This is extremely dangerous, as it seems to indicate that no further medical discovery need ever be made in the future.
> Overall, the clinical literature supporting the shaking hypothesis suffers from a number of severe methodological shortcomings. The main issue is circular reasoning. It is only in a small minority of “shaken baby” cases that actual shaking has been observed by independent witnesses, videotaped, or spontaneously confessed before police interrogation. Far more often, shaking is “inferred” after the observation of subdural and retinal hemorrhage in infants who are brought to the hospital by parents or caregivers. Physicians interpret these types of bleeding as markers of violent trauma. When asked about these findings, parents and caregivers generally do not provide “acceptable explanations” – but the only “acceptable explanations” today apart from shaking are multistory falls and high speed motor vehicle accidents. This being so, it is considered that parents and caregivers must be lying when they report non-traumatic events such as a sudden collapse, an unexplained respiratory arrest, or a minor fall – even though this happens in case after case