Opioid medicines (like morphine, oxycodone, codeine, tramadol) are important for treating acute/short-term illnesses that are not related to cancer.

Several chronic non-cancer pain conditions may be treated with opioids. These include, but are not limited to, osteoarthritis, low back pain, musculoskeletal pain, fibromyalgia, and neuropathic pain. In addition, certain acute pain conditions such as post-operative pain, traumatic injuries, and severe musculoskeletal pain, can lead to long-term or high-dose opioid prescribing, particularly if pain becomes persistent or poorly managed.

However, there is little evidence that they help with long-term pain and there are concerns about side effects including opioid use disorder(OUD), which means becoming dependent on these drugs or illicit substances like heroin.

Higher and longer doses of opioids, having multiple medical conditions, taking multiple medications, lower socioeconomic status, and a history of substance use problems are all known risk factors for OUD in people with chronic non-cancer pain. However, it is unclear how these factors interact and how OUD develops from the initial prescription.

A major challenge for doctors and pharmacists is gathering and understanding information from patient records stored in different places to identify those at risk of OUD, especially during time-constrained consultations and with patients who have complex health needs.

This research project aims to identify risk factors for OUD in patients’ electronic health records who were initially prescribed opioids for pain relief in Scotland. This information, along with input from patients, clinicians, and data scientists, will be used to develop a new risk prediction tool for OUD that could be used in digital systems across Scotland.

The project also involves people with lived experiences, senior academics and clinicians with national and international expertise in addiction, pain management and data sciences for the Advisory Research Group. This group will not have access to PHS data.