The Ministry of Health has admitted to undercounting the number of people hospitalised for Covid-19 by more than 30 per cent.
In a tweet, the ministry said this has seen the number hospitalised since the start of the pandemic increase by over 5000 people, from 14,043 to 19,476.
The ministry said it was due to a "coding issue".
The ministry said the miscounting had "no impact" on the advice it provided about Covid-19 settings. This was because the daily hospitalisation data, captured by a different system, was used to inform the ministry's advice.
"The issue largely affected data about people hospitalised for Covid-19 who had relatively short stays of between one and three days, and who recovered after they were discharged. It had no impact on the care received by patients in our hospitals. "
The ministry also shared that it will continue to review and improve Covid-19 data collection in order to "gain greater insights from complex data sources."
Epidemioligist Michael Baker told Newstalk ZB the correction was "complicated" and said this was more likely an interpretation issue rather than a computer error.
"When they quoted hospitalisations, it was actually hospital occupation," Baker explained.
Baker went on to say that patients often spent less than seven days in the hospital - sometimes less than three - so when the weekly hospitalisations numbers were calculated, these patients may not have been counted.
He referred to the ministry's method of only releasing the number of bed occupations rather than actual hospitalisations as "strange", but that doesn't mean the old number was necessarily wrong.
"It's more classification - it's actually quite useful, obviously, managing a hospital, because you only have a certain number of beds in ICU positions," Baker said.
"So, it's not like they've lost those cases, it's just that they've been a bit lost in translation."
- More to come