(87) Understanding Opioid Use Prior to and Immediately Following Cancer Diagnosis Using Group-Based Trajectory Models (GBTM)

Sunday, August 28, 2022

8:00 AM – 1:45 PM CEST

Location: Hall C

Publication Number: 1132

Presenting Author(s)

Benjamin Daniels
– Centre for Big Data Research in Health, Kensington, New South Wales, Australia

Co-Author(s)

Tim Luckett
– Faculty of Health, University of Technology, Sydney, Australia, Australia
Michael O. Falster
– Dr, University of New South Wales, UNSW Sydney, Sydney, New South Wales, Australia
Winston Liauw
– UNSW, South East Sydney Local Health District, Australia
Natasa Gisev, PhD
– University of New South Wales, Randwick, New South Wales, Australia
Sallie-Anne Pearson, PhD
– Director, Centre of Research Excellence in Medicines Intelligence, Centre for Big Data Research in Health, Faculty of Medicine and Health, University of New South Wales, Sydney, Australia, New South Wales, Australia

Background: Opioid use is common among people with cancer. Although opioid use prior to cancer diagnosis is a risk factor for ongoing use, little is known about how pre-diagnosis utilization patterns influence longer-term patterns of use after diagnosis.

Objectives: To describe trajectories of opioid use in the 2 years prior to and following cancer diagnosis as a first step to understanding the characteristics of the people, their cancers and their treatment within different opioid treatment regimes.

Methods: Data: dispensing records linked to cancer registry and hospitalization data for all New South Wales residents dispensed at least one opioid medicine between 2003–2018. Cohort: 266,782 whose first cancer was diagnosed between 2005–2016. Analyses: GBTM using binary indicators of monthly opioid dispensing two years pre-/post-diagnosis. Within each pre-diagnosis trajectory group we used GBTM, accounting for non-random dropout due to death, to further cluster patients based on post-diagnosis opioid use. We describe use in each trajectory in terms of oral morphine equivalents (OMEs).

Results: Among the 41,976 people dispensed opioids prior to diagnosis, 6 distinct opioid use trajectories were identified: 1) high OMEs (7%); 2) mid OMEs (8%); 3) steadily increasing OMEs (4%); 4) late OME increase (20%); 5) low OMEs, increasing (36%); and 6) low OMEs, decreasing (25%). Pre-diagnosis opioid treatment in all groups was primarily comprised of weak opioids. All groups experienced an increase in dispensed OMEs during the month of cancer diagnosis, with the largest increases occuringin groups 3 and 4. GBTM identified 4–5 post-diagnosis trajectory groups within each pre-diagnosis trajectory group. Post-diagnosis trajectories varied considerably by pre-diagnosis trajectory. Broadly, people in the ‘high’ and ‘mid’ groups retained similar levels of pre-diagnosis use; those in all other groups had increases during the month of diagnosis, but heterogeneous trajectories thereafter. Post-diagnosis opioid treatment was primarily comprised of strong opioids in all groups.

Conclusions: There is substantial heterogeneity in opioid use prior to cancer diagnosis, which may influence long-term patterns of post-cancer opioid treatment. In general, a cancer diagnosis is accompanied by an immediate increase in opioid use and a change from weak to strong opioids. Further research will explore the characteristics of people, their cancers, and their treatments in different opioid regime trajectories.