How well do diagnosis-related groups perform in the case of extremely low birthweight neonates?

J Paediatr Child Health. 2003 Nov;39(8):602-5. doi: 10.1046/j.1440-1754.2003.00214.x.


Objective: To determine how well diagnosis related groups (DRG) perform for extremely low birthweight neonates, by examining length of stay (LOS) data for babies with birthweight of 750-999 g and managed in the seven New South Wales maternity hospital neonatal intensive care units (NICU).

Methodology: Neonates with an admission weight of 750-999 g and admitted to one of the seven NICU were studied. Those who died or were transferred in the first 5 days were excluded, leaving the group of neonates defined by DRG P62Z. In order to allow comparisons between the different NICU, neonates whose care was split between two or more NICU were also excluded, leaving those who either died, were transferred to a nursery closer to their home or were discharged directly to their home. The relationship between LOS and a number of perinatal factors was studied for the total group, and LOS data were compared for individual hospitals.

Results: Length of stay was most strongly associated with mode of separation. It was also significantly associated with gestational age, time ventilated, oxygen need at 36 weeks postconceptual age, and retinopathy of prematurity. Neonatal intensive care units with the highest proportion of neonates being transferred had the lowest average LOS, and vice versa.

Conclusions: The LOS of neonates in DRG P62Z is more strongly related to mode of separation (died, transferred or discharged home) than to factors that reflect best practice. A funding model that rewards NICU for a short LOS may unfairly disadvantage some units while favouring others.

Publication types

  • Multicenter Study

MeSH terms

  • Benchmarking*
  • Diagnosis-Related Groups / statistics & numerical data*
  • Hospitals, Maternity / statistics & numerical data
  • Humans
  • Infant, Newborn
  • Infant, Very Low Birth Weight*
  • Intensive Care Units, Neonatal / statistics & numerical data*
  • Length of Stay / statistics & numerical data*
  • Linear Models
  • Multivariate Analysis
  • New South Wales
  • Utilization Review