Transforming Opioid Poisoning Surveillance Through Novel Technologies

Investigators: Dr Ting Xia, Dr Joanna Dipnall, Dr Jane Hayman, Dr Tina Lam, Prof Suzanne Nielsen, Assoc Prof Richard Beare, Assoc Prof Nadine Andrew, Prof Paul Dietze, Assoc Prof Nick Scott, Prof Mark Stoove, Dr Amanda Roxburgh

Funder: NHMRC

Opioid-related mortality in Australia has doubled in the past decade. In response, a range of policychanges were made to urgently address growing opioid-related harm. Emergency Departments (EDs) have enormouspotential for drug harm and injury surveillance; however, inadequate coding of ED data prevents its use to understand policyoutcomes.

In the absence of good quality, routinely collected structured ED data, opioid harm cannot be accurately monitored and it cannot be determined whether policy interventions that have been deployed are having their intended, or unintended, effects on the harms in our communities.

In this project, we will transform the coding of ED data using innovative natural language processing approaches. This will enable automated, reliable surveillance of opioid-related harms. We will quickly demonstrate impact through planned policy outcome analyses.

Specifically, this project will: (i) transform opioid coding in ED data using novel natural language processing applications for text extraction using existing government ED data; (ii) determine if ED data can be further optimised using additional data fields not currently standard to ED data sets; and, (iii) use the improved data to conduct urgently needed analyses on the outcomes of recent opioid policies.

To achieve these aims, we will first develop and deploy a natural language processing model on 15 years of ED data. Next, we will extend language processing tools within Peninsula Health, through the National Centre for Healthy Aging data platform, and Alfred Health, using more advanced methods and additional data. Finally, detailed policy analysis using causal inference methods will be used to determine policy outcomes.

The outcomes of this study will (i) transform the way opioid harm is coded and monitored in emergency departments, and (ii) provide policy makers and clinicians with timely evidence on the impact of recent opioid policies

This project is a collaboration between Monash University and Burnet Institute.

Publications

Xia, T., Lam, T., Dipnall, J., et al. Transforming Opioid Poisoning Surveillance Through Novel Technologies: Rationale and Methodological Protocol for Applying Natural Language Processing to Emergency Department Data. Drug and Alcohol Review. 2026; e70117. https://doi.org/10.1111/dar.70117