Check out the accepted paper here

This project was a real-world stress test for the OPALS peri-op dataset we’ve been building. When the MHRA withdrew the licence for modified-release morphine and oxycodone in post-surgical patients, we wanted to see how quickly we could generate the usage data. Moving from a policy alert to a full data-driven evaluation in under four weeks felt like a true realisation of the Learning Health System (LHS).

Aims and Background

In March 2025, the MHRA issued a public report regarding the risks of modified-release (MR) opioids in post-operative settings.

Our goal was to repurpose the data pipeline during the OPALS project to rapidly quantify the prescribing prevalence of these medications. We focused on identifying “new” prescriptions—those not present in the patient’s admission history—to specifically capture post-surgical initiation.

Analysis

We analysed all of the elective surgery encounters between 2019 and 2025. This efficiency is a primary benefit of maintaining a “research-ready” dataset rather than starting from scratch for every clinical query.

We were also able to extract prescriber-level data, identifying who and in which divisions the most prolific prescribers were.

Results

In meetings with the Pain and Opioid Stewardship teams, they said they had been looking for this data for over a year. We were able to generate the initial analysis within an afternoon, and within two weeks we were very close to our final analysis produced in the paper.

Stay tuned for the next steps to intervention!

Reflections

This started out a mini-project that was intented to a be quick and easy paper (the academics oxymoron). Whilst it took almost a year in editorial (for what was supposed to be a letter) the final paper allowed us to pen alots of the soft issues we encountered. Important issues like the silo’ed working between teams and differint data lineage. As a first paper of the PhD this will lay some nice ground for the work to come!