Health Survey error led to overestimate of obesity numbers

New Zealand has 7,000 fewer obese children than previously thought following identification of an error which overstated obesity estimates.

The Ministry found that a measurement error by one interviewer in the New Zealand Health Survey led to obesity prevalence for children aged 2–14 years being overstated by around one percentage point for the last 2-3 years.

Although the interviewer concerned completed a small proportion of all surveys, once the results are scaled up to the population level it equates to about 7,000 children being classified as obese when they weren’t.

After correcting the data we now estimate there are 94,000 obese children, down from the 101,000 we estimated previously.

The error was due to mismeasurement of height and weight in some children living in the Taranaki and Whanganui regions.

The error affects three years of data from 2015/16 to 2017/18, although the 2015/16 year was less affected.

The national data for the three affected years is being revised, with the revised national results expected to be published in November.

Regional data will be more difficult to correct given the relatively smaller sample sizes. The Ministry will be working with the DHBs on how best to manage the results affecting them.

From a statistical perspective, the error is relatively small, as the changes will be within the margin of error already published for the data.

However, as this data is relied on by researchers, policy analysts and planners, it’s important it’s corrected. Having our data as correct as possible helps ensure that any subsequent use of the data can rely on a solid base.

The Ministry and survey provider (CBG Health Research) have taken steps to ensure this error cannot reoccur, including retraining of survey staff and additional data monitoring.

/Public Release. View in full here.