Back to accessibility links

Maternity Services of NSW fact of birth


Variation in hospital maternal average length of stay (ALOS) has been raised as an issue on both a national and state level.

A preliminary population-based record linkage1 study of maternity services in NSW public hospitals has been undertaken.  The overarching aims were to:

  • quantify variation in maternal ALOS taking into account differences in patient and hospital characteristics.
  • determine how a hospital’s Average LOS compared to that predicted at the statewide level for the kinds of women seen at the individual hospital.


The study population included all separations for women admitted to a NSW public hospital for childbirth between 1 July 2009 and 30 June 2012 with a one-year look-back prior to birth.

Multi-level models were used to examine variation in maternal ALOS among NSW public hospitals and to adjust for differences in patient and hospital characteristics.Risk-standardised scores were derived and utilised to assign hospitals into three categories: higher than expected, no different from expected and lower than expected.


For the period 2009/10-2011/12 there were nearly 215,000 hospital separations for maternity services (72.6% vaginal births and 27.4% deliveries by caesarean section) utilising around 700,000 bed days with an average annual cost of over $400M.

After adjustment for patient and hospital characteristics, variation in hospital maternal ALOS was evident (2 selected Diagnosis Related Groups are shown below).

Next steps

Additional consultation with clinicians to add context to what factors may be influencing variation and to determine what is clinically significant.

The perinatal data was only available until 2012. Additional analysis using post 2012 data for all other linked data will be undertaken following clinical input.

Two graphs showing average length of hospital stay for Vaginal Delivery (28 higher than expected, 22 no different to expected and 17 lower than expected) and Caesarean delivery (20 higher than expected, 23 no different to expected and 22 lower than expected)

Hospital and birth records, from the NSW Admitted Patient Data Collection and the NSW Perinatal Data Collection respectively, were obtained from the Maternal and Child Health Public Health Register (Source: SAPHaRI, Centre for Epidemiology and Evidence, NSW Ministry of Health) and linked by the Centre for Health Record Linkage.