The robustness of Starship statistics and data in general

If the seemingly unimpeachable Starship hospital is going to use its admission statistics to lobby for legislation (which they are increasingly doing) they should ensure its consistency or provide additional descriptive information. Take these two examples for instance:

A leading authority on shaken baby syndrome says the level of child abuse in New Zealand is staggering, with nearly one child a week admitted to Auckland's Starship Hospital with serious physical injuries. Hospital data, released to the Herald on Sunday under the Official Information Act, shows that since 2001, 181 children have been hospitalised as a result of "suspected or definite" child abuse. July 2007.

Then:

"...the largest group of injuries treated at the Starship's intensive care unit were suffered by juvenile car passengers.

They accounted for 150 admissions in 11 years since 2000, ahead of 105 child pedestrians knocked down in streets and driveways, and 95 victims of domestic violence. May 2011."

Reconciling the two statements poses difficulty. 95 victims of domestic violence over 11 years is considerably less than 181 over 6 years. 8.6 per year versus 30 per year. The qualifying definitions may be different but that should be noted. One may relate to any admission while the other relates to admission to intensive care. Though I would have thought as Starship tends to the most serious injuries the difference wouldn't be that great.

Hospital statistics are vital because they represent actual incidents as opposed to reported or suspected events. However as the first statistic does contain the qualifier for the cause as, " suspected or definite", the second leads me to believe many of the susepected causes turn out to be unprovable or wrong.

The increasing tendency for Starship practitioners to lobby government however also casts doubt on the data they provide. It is natural for lobbyists to build their case with the most compelling picture. This may even result in conflict amongst the practitioners themselves. Especially if they were lobbying for extra resources. Not very helpful really. (Notwithstanding, this anomaly may just be a failure or omission by the journalist.) Additionally, if a problem being over-stated there is an accompanying tendency towards over-suspicion.

I can't remember where but I read about the suggestion or proposal that a neutral group should be formed to scrutinise statistical information. They exist in other jurisdictions? Perhaps it would need to be formed from unpaid, possibly retired folk with the necessary credentials.