In the wake of our questions, the Auckland District Health Board has changed the way violence is reported so it can find how badly its staff are being treated.
Evidence has also emerged of health managers discussing ways to avoid providing unfavourable information.
During our investigation, health minister Tony Ryall and associate health minister Jonathan Coleman - responsible for mental health - have begun to ask questions.
The Te Whetu Tawera unit, based at Auckland City Hospital, has suffered through the most unfortunate of existences.
It has room at most for 58 of the most acutely unwell of Auckland's 3000 mental health patients.
In its troubled past, bad practice at the unit has been blamed for patients killing themselves and others. It has gone through extensive review.
The health managers who oversee it are among the most senior staff at Auckland health board. They include one of the most senior health bureaucrats, Fionnagh Dougan, general manager of clinical services.
Most directly, oversight also involves mental health director Dr Clive Bensemann and the nursing director for mental health services Anna Schofield.
The Herald on Sunday uncovered the assaults on staff after receiving a letter from Lauren Meraw, a mental health nurse from Canada who had quit the unit saying it was too dangerous to work there.
Meraw joined Te Whetu Tawera on a working holiday to New Zealand and lasted eight months before becoming a whistleblower. She said she was "appalled at the conditions that nurses are expected to work in".
Meraw said there needed to be a review of staff safety and "the lack of management support" for staff. She wrote how patients in the unit had no fear of being "secluded" in a secure room, in physical restraints or with medication because management were focused on "keeping seclusion rates down".
Staff and patients were at a "serious risk of harm", she wrote.
In May, Bensemann dismissed her complaint, describing Meraw as a "relatively inexperienced staff member". Asked if assaults could happen, he said "it can happen on occasion".
Now, documents show Meraw's description was more accurate than Bensemann's.
The day after the Herald on Sunday published Bensemann's dismissal of claims by Meraw, he tried to find out what the exact level of assaults were.
An email trail shows Bensemann and Schofield were ignorant of how many staff were impacted - and discussed whether they should find out.
Schofield asked whether it would be better to simply release a broad figure saying how many assaults there had been.
Bensemann replied: "We must be factually accurate but if possible avoid enlarging the story."
"I therefore do think we should consider doing the manual exercise separating out patients vs staff [assaults] at least. It might be quite informative too."
In fact, it became clear the exercise in finding out how many staff were being assaulted exposed the flaws in the data held.
An analyst told Schofield and Bensemann the number of assaults appeared to have tripled between 2009 and 2010 but that no one knew why.
He also explained there was a surge of assaults in March this year and suggested altering the data by removing one patient who had created a series of "very low-level incidents and seeing what impact removing them has".
Data extracted by the Herald on Sunday has found the "low-level incidents" actually referred to the patient repeatedly punching staff.
The documents show the incidents were not all reported and at least 50 assaults were filed as only seven incidents.
The board's data mining exercise was overseen by another health bureaucrat, clinical effectiveness adviser Annette Shea. She wrote on June 2 to Bensemann, Schofield and Dougan to say there were "limitations" on the data.
"Not all incidents are reported," she wrote.
She also said the "influence of senior staff" could affect whether assaults were reported.
And sometimes, she wrote, staff filled out the official forms recording an assault had taken place "to make a point".
"Reporting by staff is often issue-driven," she said.
Senior management agonised over the findings.
Bensemann, Schofield and "communications specialist" Matt Rogers debated by email whether they could release just the staff assault figures of 131, rather than the total figure.
Rogers wrote, "I think full disclosure is the way to go here", and prepared a press release which he revised 15 times.
He currently refuses to make available any drafts, saying they are not public documents. He also told Dougan to warn the health minister the figures were to be made public.
Schofield wrote that she wanted to get the release out to staff first to "manage the nursing staff unrest".
Roger's final press release revealed 213 assaults had happened at the unit. No one heeded Shea's warning about the unreliable data by warning that the figures were likely to be higher.
While Bensemann and Schofield were ignorant of the numbers of figures, they were meant to be notified of serious assaults. They were assured by senior staff at Te Whetu Tawera that 95 per cent of assaults were "minor" - and didn't realise the system was flawed, which the health board now concedes.
The Herald on Sunday studied 12 months of incident reports and found the "minor" classification was used to describe assaults featuring biting, kicking, punching, scratching and strangulation.
Examples of "minor" attacks include one patient attacking an elderly patient by "kicking him forcefully in the head then proceeding to strangle him in an unprovoked attack".
An incident described as being of "minimal" severity saw a staff member knocked to the ground by the force of a kick to the stomach.
The incident report forms also show some "serious" assaults were downgraded to "minor" assaults by managers.
The papers also raise safety issues which have an impact on staff across the health board.
One failure exposed in the documents affects a "safety net" meant to be in place for all health board staff - the Occupational Health and Safety Committee. The group, chaired by Dougan, is meant to allow staff the opportunity to voice concerns over issues.
But, around the time Meraw was finding the violence intolerable, the committee failed to meet because not enough managers turned up.
"No meeting held, quorum not achieved," a document on February 14 stated. It scheduled the next meeting for more than two months later, just before Meraw quit.
The documents show those present at the meeting didn't even know whether the staff representatives still worked for the health board. A note reads: "Occupation Health and Safety to check and see if reps current on payroll systems before sending out invites."
Meraw is now back in Canada. She said she was pleased the issues she raised were being investigated but was saddened that managers appeared slow to react to the safety and wellbeing of staff.
She said any concerns about the release of information should be a "red flag". "I hope investigations continue to get to the root of the dangerous work environment of TWT and that positive changes are made."
'GENUINE CONCERN' OVER STAFF ASSAULTS
Fionnagh Dougan said there was genuine concern among management about every assault on staff.
"We are endeavouring to create as safe an environment as we can for them to work in."
She said senior managers outside the unit did not need to know how many assaults took place and only needed to know about the most serious assaults.
But she also conceded the system for notifying senior managers of serious assaults was flawed. It had failed to correctly categorise the seriousness of assaults.
Dr Clive Bensemann said staff welfare was important to him but the inquiry into staff assaults had revealed flaws in the system.
"As part of this process over the last couple of months we have looked at our data collection system and we have improved it. This whole process for us has highlighted the limitations of the data."
