Individual-level predictive models for management of postoperative pain

Opioids are a first-line treatment of postoperative pain, and as a result, this perioperative opioid exposure 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 US.
Patterns of opioid use are likely to be associated with patient characteristics and patient history, and patterns of opioid prescription are also likely
to vary across practices.
HMS is interested in using 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. It has requested DHCRC to engage a PhD Candidate through Stanford University to conduct this
project. DHCRC will also contract with University of Canberra to co-supervise the selected PhD candidate.
The PhD candidates (1.5 PhDs) working on this project will need to have a solid machine learning or advanced statistics background. Traditional
off-the-shelf methods of biostatistics such as logits and survival models are unlikely to be adequate for this project. The candidates will also need to
be comfortable with analyzing and processing very large data sets, with millions of records and hundreds of variables. The candidates will need to
be knowledgeable of epidemiology and health care services and will need to be able to read current research in opioid misuse and utilize that
knowledge to build state-of-the-art predictive models.