Mini Oral Australian Epidemiology Association ASM 2018

Using PHREDSS to monitor codeine-related emergency department presentations (#61)

Branislav Igic 1 , Claire Harper 2 , Therese Ryan 3 , Anna Do 3 , Melissa Irwin 3
  1. Biostatistics Trainee, Centre for Epidemiology and Evidence, NSW Ministry of Health, Sydney, NSW
  2. Emergency Response Coordination/Clinical Ethics and Policy, Office of the Chief Health Officer, NSW Ministry of Health, Sydney, NSW, Australia
  3. Rapid Surveillance, Population Health Intelligence Systems, Centre for Epidemiology and Evidence, NSW Ministry of Health, Sydney, NSW, Australia

Background

In February 2018, over-the-counter codeine medicines were up-scheduled across Australia to prescription only due to health risks of long-term use. How this policy has affected emergency departments (EDs) is unknown; however, it was suggested to have potentially increased ED presentations for pain and codeine withdrawal.

The NSW Public Health Rapid, Emergency, Disease and Syndromic Surveillance (PHREDSS) system monitors ED presentations for headaches and some illicit drugs in near-real time. However, monitoring the use and misuse of licit drugs, including codeine, is challenging.

Aim

Develop and test strategies for monitoring codeine-related ED presentations using PHREDSS.

Methods

Keyword searches on triage text and diagnosis description fields were used to identify ED presentations (a) mentioning codeine, including appropriate use, and (b) related to codeine addiction, withdrawal, and overdose. Codeine allergies were excluded. Using random sampling of presentations from 2017-2018, results of the keyword searches were reviewed and false positive rates calculated. Temporal trends for headaches (proxy for pain) and codeine-related ED presentations were examined for 60 EDs using Rapid Emergency Department Data.

Results

A minor initial increase in headache and codeine-related ED presentations following the up-scheduling was not sustained. The all codeine search (a) mostly identified the use (75%) and administration (14%) of codeine, with a false positive rate of 4%. The specific search (b) identified overdoses (74%), codeine withdrawal (7%) and addiction (5%), with a false positive rate of 13%.

Conclusion

PHREDSS in combination with text analytics has limitations but is timely and can be used to monitor codeine-related ED presentations.