Technical notes

ICD-10-AM (8th edition) codes used

External cause of injury

  • V00–V99 Transport accidents
  • W00–W19 Falls
  • W20–W49 Contact with objects (Exposure to inanimate mechanical forces)
  • W50–W64 Contact with living things (Exposure to animate mechanical forces)
  • X60–X84 Intentional self-harm
  • X85–Y09 Assault
  • Other external causes – includes all external causes of injury not included in the above groups. Note that this group contains Complications of medical and surgical care, which may not have been associated with the TBI. This may particularly affect initial hospital admissions where the TBI was an additional diagnosis.

Place of occurrence

  • Y92.0 Home
  • Y92.1 Residential institution
  • Y92.21 Educational setting (School, college, university, day care, kindergarten)
  • Y92.22 Health service setting
  • Y92.3 Sports and athletics setting
  • Y92.4 Street and highway
  • Y92.5 Trade and service setting
  • Y92.6 Industrial and construction setting
  • Y92.7 Farm
  • Y92.8–Y92.9 Other specified and unspecified place

Activity when injured

  • U50.0 Football (all codes)
  • U50.1–U50.9, U51–U53 Team sports (excluding football)
  • U65 Motor sports
  • U63 Equestrian sports
  • U54–U62, U64, U66–U71 Non-team sports
  • U72 Leisure activity
  • U73.0 Paid work
  • U73.1 Unpaid work
  • U73.2 Essential activities (resting, sleeping, eating)
  • U73.8–U73.9 Other specified/unspecified activity.

Main limitations of NIHSI AA (v0.5)

Reference period

The NIHSI AA v0.5 is limited to data from 1 June 2010 to 30 July 2017.

Data gaps

There are known data gaps within some of the Australian health care data landscape that affect the NIHSI AA.

Emergency department presentation data

The initial proposal for this project included the identification of TBI cases presenting to emergency departments. However, this approach was abandoned due to difficulties in defining TBI cases in the ED data due to:

  • lack of specific diagnosis information in the ED – the underlying diagnosis may not be known at the end of the ED episode and the reported diagnosis may include general signs or symptoms or external causes. For example, a TBI case may be reported using a less specific code such as Headache, Multiple head injuries or Other and unspecified head injury
  • lack of uniformity in classification systems used to describe the principal diagnosis or presenting problem – for example, New South Wales used SNOMED-CT-AU to code ED principal diagnoses, rather than ICD-10-AM.

These issues may result in an under- or over-estimate of TBI cases presenting to emergency departments. Therefore, ED presentations were not used to define the cohort for this study.

Other data gaps

Other current data gaps in the Australian health care system include health care services provided in primary care settings, such as general practice; allied health care settings, such as private practice physiotherapy or occupational therapy; and outpatient specialist clinic settings. No information on services provided in these settings is currently included in the NIHSI AA.

Data quality issues

Not all records with an initial TBI admission in the NHMD between 1 July 2013 and 30 June 2015 were linked in the NIHSI AA. Some records were discarded from the analysis due to lack of data needed for data linkage.

In the NDI data, one person’s death was reported more than 2 years before the TBI event. Therefore, linkage of later hospitalisations, MBS and PBS data for this person was probably due to a false match during the data linkage process. The record was retained in the cohort, but the death was not counted as an outcome.

One person died on the day before the initial admission (a same-day hospitalisation “Died” as the separation mode). The disparity between the two dates was taken to be an artefact of the calculation of the admission date difference in the hospital content table in NIHSI AA. This record was retained in the cohort, and the death was counted as an outcome.