TLDR: There are issues with the way/how quickly we diagnose shaken baby syndrome without exploring other potential diagnoses, and with how severely/quickly we take kids away from their parents once we "think" we have a shaken baby
It's much worse. There's no evidence of any baby shaking. At all.
I think it's reasonable to say that billions of babies may have been shaken in the past, yet the vast majority (of the order of 99.9% or even much more? [2]) are not diagnosed with shaken baby syndrome, since this diagnosis has an incidence of ~1/3,000 among children 50%, it is striking to note that the SBS diagnosis actually does not exist, i.e. doctors are not trained to diagnose it.
On the other hand, actual cases of severe inflicted head injuries probably affect In other words, I've come to the likely conclusion that the intersection of babies who are effectively shaken, and babies who are labelled with "shaken baby syndrome", is an abysmal proportion of the first set, and a small proportion of the second.
[1] https://www.sciencedirect.com/science/article/pii/S074937970...
[2] Excerpt from the paper above: "the ratio of children hospitalized or dying from inflicted neurotrauma compared to the numbers of reported shaken children may be estimated at 1 to 152."