Individual-level predictive models for management of postoperative pain
Flagship Program: Intelligent Decision Support to Improve Value and Efficiency
Opioids are a first-line treatment of pain following surgery, but this may be a gateway to opioid misuse and addiction. Over the past decade, opioid misuse and abuse has become a major epidemic crisis in the USA.
Most surgical patients receive opioids regardless of co-morbidities, prior opioid-related problems, or possible drug-drug interactions.
In addition, the success of treatment using opioids is likely to be a complex function of a variety of individual and societal level factors, which are currently not well understood.
This project will use linked claims data – sourced from a de-identified Medicaid (USA) dataset made available by HMS in conjunction with the Digital Health CRC – to gain a deep understanding of patterns of opioid use and prescription, and develop new insights into what constitutes successful treatment.
It will accomplish this utilising novel machine learning methods to extract interpretable patterns and associations in opioid prescribing and use.
Through the project, a PhD student embedded at Stanford University (and co-supervised by the University of Canberra) will work on research addressing the following aims:
- Allow the early identification, education/outreach and monitoring of patients who may be at risk of developing opioid dependency.
- Allow the identification of prescribers who are not necessarily fraudulent (pill mills) but show patterns of careless or outdated overprescribing, and thus contribute to patients’ opioid misuse disorders. Identification of these prescribers will allow educational outreach and the monitoring of providers who create excess risk.
- Use analytic modeling to determine the characteristics of successful treatment that can inform the development of evidence-based treatment regimens.
- Allow the early identification of patients with a history of opioid misuse disorders who have a period of abstinence but then relapse into a pattern of addiction. This will enable more focused outreach, education and monitoring for patients with higher risk.
The research undertaken will incorporate the following tasks:
- Understand the nature and trends of pain management following major elective surgery (including dental) – including defining, understanding and measuring the extent of opioid misuse and abuse in the population and across different sub-populations (eg. racial/ethnic categories, gender, and other social determinants for health whenever available). This includes understanding the use of alternative pain management strategies, multimodal regimes, and opioid use (including pre-, peri-, and post-operative use).
- Develop individual-level models to predict the risk of developing an adverse pain outcome. Outcomes of interest will include the first episode of opioid misuse, prolonged opioid use, opioid misuse/abuse, and 30-day readmissions. The models will include opioid tolerance, mental health status and other known associated patient and clinical risk factors.
- Develop personalised treatment regimens for postoperative pain management that yield higher quality of life outcomes. Using artificial intelligence technologies, the project team will identify the best pain management strategies for diverse patient populations.
- Develop predictive models for the identification of providers (via reviews of opioid prescribing claims) who have at-risk opioid prescribing patterns and may be creating excess risk for their patients, using machine learning methods to identify opioid prescribing patterns associated with opioid misuse.
This project is highly relevant to a range of stakeholders, both in Australia and overseas.
While the findings and the delivered machine learning tools will be specific to the Medicaid (USA) population, the methodology will be easily portable to the Australian setting.
This project will use linked claims data to gain a deep understanding of patterns of opioid use and prescription as well as developing new insights in what constitutes successful treatment.