Healthy Communities: Potentially preventable hospitalisations in 2013–14 - Technical Supplement - Interpretation of data

Healthy Communities: Potentially preventable hospitalisations in 2013–14

Interpretation of data

Principal and additional diagnoses

The identification of most potentially preventable hospitalisations was based on principal diagnosis only. However, four conditions were identified using principal and additional diagnoses, these were:

  • Pneumonia and influenza (vaccine-preventable)
  • Pneumonia (not vaccine-preventable)
  • Other vaccine-preventable conditions
  • Gangrene.

This means, for example, that for the same hospitalisation, pneumonia and gangrene may be reported with pneumonia as principal and gangrene as an additional diagnosis. As a result, components may not add to totals.

Emergency department only patients

The Authority, through its Jurisdictional Advisory Committee, was advised that two states in Australia have a process of administratively admitting a portion of their emergency department (ED) patients (this includes patients who died). These patients are included in this report.

Comparison over time

The report does not compare 2012–13 and 2013–14 data with 2011–12 data published in November 2013, as the specification changed for some conditions. In addition, Activity Based Funding (ABF) was introduced Australia wide after the 2011–12 report. For this reason, the Authority recommends that comparisons should not be undertaken across years in this release of potentially preventable hospitalisations data.

Age-standardising the data

The report includes potentially preventable hospitalisations expressed as both crude and age-standardised rates per 100,000 population. Crude rates are the number of potentially preventable hospitalisations in an area divided by the total ABS estimated resident population (ERP) for that area multiplied by 100,000. The ERP at 30 June 2012 was used for 2012–13 data and 30 June 2013 ERP for 2013–14 data.

Age-standardised rates are hypothetical rates that would have been observed if the populations studied had the same age distribution as the standard population, while all other factors remained unchanged. Age-standardised rates were derived by calculating crude rates within an area for each five year age group (0–4, 5–9, 10–14, …, 80–84, 85+). These rates were then given a weight that reflected the age composition of the standard population. The current standard population is the ERP for Australia as at 30 June 2001.

When comparing rates adjusted for age, any remaining observed differences between the populations cannot be attributed to confounding by age.

The method for calculating age-standardised rates for both PHNs and SA3s was the same.

Where the age for an individual patient was clearly invalid, it was classified as a data error and excluded in performing the age-standardisation process.

Suppression of data

Suppression of data is used to protect individual confidentiality. It is also used to produce reliable results by not calculating statistics using insufficient numbers of observations. For this report, there were several levels of suppression. These rules have been applied to both PHN areas and SA3s if the:

  • Number of separations are <5 then all data for that geographic area are suppressed
  • Number of separations is <20, and/or the geography’s population is <2,500, and/or the population of an age group in a particular geography is <30, then all data are suppressed, except the number of hospitalisations. The <30 rule was relaxed for Barkley and East Arnhem SA3s (both in the Northern Territory), after consultation with the jurisdiction
  • Number of same-day hospitalisations is <5 then all data pertaining to length of stay are suppressed.

There are several reasons why national totals may not correspond to the sum of lower-level statistics:

  • Rounding
  • Omission of results where postcodes did not match to SA3s or PHN areas
  • Suppression of results for SA3s and/or PHN areas.

Potentially preventable hospitalisations for Aboriginal and Torres Strait Islander people

The Authority have undertaken exploratory analyses of admitted patient care data focussing on potentially preventable hospitalisations for Aboriginal and Torres Strait Islander people. Rates of potentially preventable hospitalisations for Aboriginal and Torres Strait Islander people were not able to be reported in ways that allowed fair and robust comparisons at the local level, due to the following:

  • Under-estimates due to incomplete recording of Indigenous status in the Admitted Patient Care Dataset
  • Scale-up factors being insufficiently robust to allow data to be reported and compared fairly, particularly at smaller levels of geography
  • Unavailability of Estimated Resident Population data for Aboriginal and Torres Strait Islander people at PHN and SA3 geographic areas.