SAT0569 CHANGING TRENDS AND PRESCRIBING PATTERNS IN OPIOID-TREATED PRIMARY CARE PATIENTS WITH NON-CANCER PAIN

Background: The opioid epidemic in the U.S. has led to similar concerns about prescribed opioids in countries within Europe. In new users, the rate of escalation to more potent opioids is likely to contribute to long-term prescriptions, which in turn may be associated with opioid dependency, addiction and overdose. The scale of such escalation however is unclear in the U.K for non-cancer pain. Objectives: We sought to: (i) describe trends of prescribed opioids for non-cancer pain in the UK primary care setting over a 10-year period (ii) assess the sequential transition of opioid strength from index date over a 2-year period. Methods: We conducted a retrospective observational study over a 10-year period from 1/1/2006 to 31/12/2015 using the Clinical Practice Research Datalink (CPRD). CPRD collects de-identified patient data from a network of GP practices across the UK. New users of opioids, 18 years or over, without cancer in the 2 years prior to index date were included. The number of prescriptions for each drug were calculated by each calendar year accounting for the number of eligible patients registered in CPRD for that year. Sunburst plots were created to evaluate the sequential transition of opioids over time. A 4-state hidden Markov model was used to estimate the transition probability for individuals escalating to more potent opioids over a 2 year period. States were defined as (i) no drug (ii) weak opioid (codeine, dihydrocodeine) (iii) moderate opioid (tramadol) (iii) strong opioid (all others in CPRD). Methadone prescriptions were excluded for the purposes of this analysis. Results: 1,026,955 opioid users were included: mean age (SD) was 55(18) years; 58% being female. New users of opioids were most commonly prescribed codeine (n=723,102; 70.8%), followed by dihydrocodeine (n=179,831; 17.6%), tramadol (n=93,338, 9.1%) with n=24,808 (2.4%) strong opioid prescriptions. The rate of prescribing strong opioids/10,000 population increased 12 fold from 2006-2013, followed by a gradual decline till 2015 (Figure 1). This trend was most marked with certain opioids: morphine, oxycodone, buprenorphine and fentanyl (28.5 prescriptions per 10,000 population to a peak of 353.0 prescriptions per 10,000 population in 2013 and 303.1 prescriptions per 10,000 population in 2015). Using sunburst plots, of the new users prescribed weak opioids as their first prescription at index date, 5.5% transitioned to moderate opioids, 4.1% to strong opioids over 2 years (Figure 2). Transition probability of moving from weak to strong opioid at a given time point over 2 years was 0.001, whilst staying on a strong opioid (if first prescription at index date) was 0.97. Conclusion: Strong opioid prescribing increased till 2013-14 gradually decreasing following UK initiatives to improve monitoring and use of controlled drugs. Although less potent codeine prescriptions made up the majority of first prescriptions, the transition probability of staying on a strong opioid at 2 years remained high if prescribed first as a new user. Disclosure of Interests: None declared