Patient safety monitoring initiative
Why we are implementing the Patient Safety Monitoring Initiative (PSMI)
There is a clear need to measure safety and performance in Victoria’s hospitals. Tracking improvement and monitoring patient safety event rates at hospital level is enabled by routinely collected administrative data in conjunction with a set of validated indicators.
A number of tools exist having potential to address this need. These are the Australian Patient Safety Indicators (AusPSI), Queensland Health’s clinical indicators (QCI) and the Variable Life Adjusted Display (VLAD) reporting methodology. These are discussed below.
Administrative data is relatively inexpensive, readily available, universally applied in Victoria and available in a single database, the Victorian Admitted Episodes Dataset (VAED).
What the indicators are and how are they being developed
The AusPSIs are a set of patient safety indicators providing information on potential in-hospital complications and adverse events following surgery, procedures, and childbirth.
The AusPSIs have been produced, through refinement and modification of a set of indicators (AHRQPSIs) developed in the United States of America (US) by the Agency for Healthcare Research and Quality. The VAED offers benefits not available in the equivalent US dataset.
In order to use these indicators in Victoria translation was undertaken, producing a usable set of indicators for application to the VAED. The AusPSI list is available.
A reference group of sector experts guided the translation of the indicators. Consultation has also occurred with other jurisdictions including the Qld team who manage the QCIs.
How the reports to health services will be presented
Queensland Health monitors clinical outcomes using a number of indicators (QCIs) covering mortality, complications, length of stay and readmission rates. Reports are presented to hospitals using VLAD.
VLAD is a reporting methodology providing a graphical overview of clinical outcomes over the course of a selected period using 'Expected minus Observed'.
When used with mortality as the outcome indicator, it displays 'estimated statistical lives gained' by plotting the cumulative difference between expected and actual outcomes. The expected number is a statistically generated figure based on the risk profile of the patient and organisation.
This technique enables rapid identification of trends at health service level not captured when reporting periodic averages. That is, every separation influences the data output.
A flagging mechanism can then be applied which indicates areas of significant change as an 'apparent outlier'. These parameters are calculated to detect if a particular hospital’s outcomes are higher or lower than expected. An apparent outlier may be either an area of concern or of exceptional practice.
- The proposed approach uses a set of 11 indicators chosen to address large portions of the sector, in areas of known concern, in sufficient numbers and that are amenable to change.
How the reports will be presented to health services
The selected indicators are provided to health services as VLAD reports with flags.
Health services would consider flags on a case-by-case basis depending on the level of flagging. Investigation of identified trends would expose potential areas for quality and safety improvement. To improve the usability and applicability of the PSI reports, the initiative actively seeks and welcomes feedback from health services.
The 11 indicators chosen to address large portions of the sector are:
- Death in low mortality DRGs (AusPSI)
- Complications of anaesthesia (AusPSI)
- In-hospital fracture (AusPSI)
- Postoperative haemorrhage or haematoma (AusPSI)
- Postoperative deep vein thrombosis/pulmonary embolus (AusPSI)
- Obstetric trauma – vaginal or caesarean delivery (AusPSI). Combined measure.
- Stroke in-hospital mortality (QCI)
- Heart failure in-hospital mortality (QCI)
- AMI (heart attack) in-hospital mortality (QCI)
- Pneumonia in-hospital mortality (QCI)
- Fractured neck of femur in-hospital mortality (QCI)
Sample report: Expected minus Observed
