Associate Professor Lucy Burns, Professor Louisa Degenhardt, Doctor Timothy Dobbins, Doctor Natasa Gisev, Doctor Sarah Larney, Doctor Mark Montebello, Professor Alison Ritter, Doctor Marian Shanahan


Record linkage studies are increasingly being used to examine harms associated with opioid use and injection. Cohorts examined in record linkage studies are often defined using opioid treatment databases, such as the PHDAS. A criticism of this approach is that it does not include people who are experiencing opioid- and injecting-related harms, but have yet to seek treatment. This project will use emergency department and inpatient hospital records to identify this latter group, allowing for an examination of the extent to which an opioid treatment database captures the population experiencing opioid-related harm.

Recent studies have demonstrated that the opioid use disorders population is ageing. Opioid-related hospital separations and deaths among older Australians have consistently increased over the past decade, even as they have fluctuated among younger age groups, and the proportion of opioid treatment clients aged over 50 years more than doubled between 2006 and 2012, from 8% to 18%. Opioid dependent people in their forties and fifties have health profiles typically only seen in much older people, and as the ageing trend continues, increasing numbers of this group will develop serious acute and chronic diseases. Management of health conditions in this group will be complicated by ongoing injecting drug use and high levels of physical and psychiatric comorbidity. There is a pressing need for current information on patterns of emergency department use and hospitalisation of people with opioid use disorders, particularly in relation to ageing. 


Aim 1: Assess the benefits of using multiple health databases (as opposed to a single opioid treatment database) to identify individuals experiencing opioid- and injecting-related harm.

Aim 2: Compare health service use of people with and without opioid use disorders.

Who Will Benefit?

This project will address an important methodological issue in data linkage research on opioid- and injecting-related harm. It will also provide evidence to inform the care of people with opioid use disorders by identifying target conditions for early intervention.


This study will link three routinely collected administrative datasets to assemble a retrospective observational cohort of people with experiencing opioid-related health harms (the base cohort). For specific analyses, we will identify matched controls with no evidence of opioid-related harm for comparison to the base cohort (control cohorts).

The databases to be linked are the Pharmaceutical Drugs of Addiction System (PHDAS), which contains fully-identified records of all persons prescribed methadone or buprenorphine for the treatment of opioid dependence in NSW since 1985; the Emergency Department Data Collection (EDDC), which includes presentations to all major metropolitan and non-metropolitan emergency departments in NSW; and the Admitted Patient Data Collection (APDC), which contains records of all inpatient hospital separations (discharges) in NSW. Linkage will be undertaken by the Centre for Health Record Linkage.

The base cohort will be defined as those individuals who appear in the PHDAS, or have an emergency department presentation or hospital separation with a diagnostic code indicating opioid-related harm. Two control cohorts with no evidence of opioid-related harm will be identified: a control cohort taken from the EDDC, matched to the members of the base cohort with EDDC records, and a control cohort taken from the APDC, matched to the members of the base cohort with APDC records. Two controls will be selected per base cohort case. Matching will be based on case sex, the age of the case at the index opioid event, and the year of the index opioid event.

Using the base cohort dataset, we will examine the value of using multiple datasets to identify people experiencing opioid-related harm. Health service use of the base cohort will be described. We will compare the subset of the base cohort with EDDC records to the EDDC control cohort, and the subset of the base cohort with APDC records to the APDC control cohort. 


2014 – 2015


For more information relating to this project, please contact Sarah Larney:


  • Thumbnail_img_logo-ndarc_1