A recent study by researchers in the Department of Preventive and Social Medicine at the University of Otago’s Dunedin School of Medicine has found that diagnoses of small-for-gestational-age (SGA) are substantially under-recorded in the National Minimum Dataset (hospitalisations) and Mortality Collection.
This validation study was undertaken in response to an earlier observation that there were fewer births with a recorded SGA diagnosis than expected. Co-author Dr Sarah Donald said, “Given that SGA refers to the smallest 10% of babies for their gestational age, we were surprised in one of our previous studies when just over 1 per cent of infants had this diagnosis coded in the National Minimum Dataset or Mortality Collection.”
The current study analysed routinely-collected hospital and birth records for nearly 890,000 infants. Using recorded birthweight, gestational age at birth, infant sex, and other variables, the researchers used two different calculators to estimate birthweight centiles to determine which infants were SGA. Coded diagnoses of SGA in the National Minimum Dataset and Mortality Collection were then compared against their calculated SGA status to see how well they aligned.
Only 1.8 per cent of the infants had a coded diagnosis, whereas 9.1 per cent and 15.6 per cent of infants were SGA according to the two calculators. While it was rare for an infant who was not SGA to have a diagnosis recorded, around 90 per cent of the infants who were SGA were missing a diagnosis. Certain groups of infants were more likely than others to have a diagnosis recorded, including those with severe SGA, those born prematurely, and those whose mother was aged 30 years or older.
Due to this substantial under-recording, the researchers advise against using SGA diagnoses in the NMDS and Mortality Collection for research purposes. Instead, they recommend calculating birthweight centiles from other routinely collected health data for more accurate identification of SGA babies.
Publication details
Blank ML, Donald S, Parkin, L. Small for gestational age coded diagnoses in Aotearoa New Zealand's administrative health datasets: a validation study. Health Science Reports. 2025;8: e70610. doi.org/10.1002/hsr2.70610.