Abstract
Aims
The aims of the current study were to determine the prevalence and incidence of prescription opioid analgesic use in Australia and compare the characteristics of people with and without cancer initiating prescription opioid analgesics.
Methods
A retrospective population‐based study was conducted using the random 10% sample of adults who were dispensed prescription opioid analgesics in Australia between July 2013 and June 2017 through the Pharmaceutical Benefits Scheme. Poisson regression was used to calculate rate ratios (RR) for opioid prevalence and incidence. The characteristics of people initiating opioids, including type of opioid initiated, total oral morphine equivalents dispensed, prescriber speciality, medical comorbidities, and past analgesic and benzodiazepine use, were compared for people with and without cancer.
Results
Opioid prevalence increased {RR = 1.006 [95% confidence interval (CI) 1.004, 1.008]}, while incidence decreased [RR = 0.977 (95% CI 0.975,0.979)] from 2013/2014 to 2016/2017. There were between 287 677 and 307 772 prevalent users each year. In total, 769 334 adults initiated opioids between 2013/2014 and 2016/2017, and half of these initiations were by general practitioners. Initiation with a strong opioid occurred in 55.8% of those with cancer and 28.2% of those without cancer.
Conclusion
Rates of opioid use have remained high since 2013, with approximately 3 million adults using opioids and over 1.9 million adults initiating opioids each year. Between 2013 and 2017, opioid prevalence has slightly increased but incidence has decreased. People without cancer account for the majority of opioid use and are more likely to be initiated on short‐acting and weak opioids. Initiation of strong opioids has increased over time, reinforcing concerns about increased use and the harms associated with strong opioids in the community.
Keywords: Australia, drug utilization, incidence, pain, prevalence, opioid analgesics
What is Already Known about this Subject
The steep increase in opioid use has mainly been attributed to use for chronic noncancer pain.
The increase in opioid use is associated with parallel increases in opioid‐related morbidity and mortality, including dependence and overdose.
It is unclear whether the prevalence and incidence of opioid use has changed in Australia in recent years.
What this Study Adds
The prevalence of opioid use increased 0.6% between 2013 and 2017, with approximately 3 million adults using opioids each year.
The incidence of opioid use decreased by 2.3% between 2013 and 2017, with approximately 1.9 million adults initiating opioids each year.
The initiation of strong opioids has increased, reinforcing concerns about increased use and the harms associated with strong opioids in the community.
Introduction
Opioid use has increased rapidly in Australia and internationally over the past two decades 1, 2. This increase is mainly attributable to greater opioid prescribing for chronic noncancer pain (CNCP) 3, 4. Although the effectiveness of opioids has been established for acute 5 and cancer 6 pain, the long‐term effectiveness in CNCP is uncertain 7. Increasing opioid use has been associated with parallel increases in opioid‐related morbidity and mortality, including dependence, hospitalizations and overdose 1, 8.
Opioid‐related deaths appear most common in people aged <50 years 8, 9, 10. Previous work examining strong opioid initiation among concessional beneficiaries in Australia identified that oxycodone was the most commonly initiated strong opioid, and that those initiating oxycodone were younger and less likely to have cancer 11. The Australian Pain and Opioids In Treatment (POINT) study has found that people with CNCP taking opioids have complex demographic and clinical profiles, and a long history of pain 12. Similar complex profiles have been described internationally among people with CNCP 13, 14, 15, 16. To date, no population‐based studies have examined the initiation of all opioid analgesics in the Australian general adult population.
Following recent declines in annual prescribing rates in the US 17, it is unclear whether the prevalence and incidence of opioid use has changed in Australia in recent years. Additionally, comparisons of Australians who initiate opioids with and without a history of cancer have not been extensively described. Therefore, the objectives of the present study were to: (i) determine the prevalence and incidence of prescription opioid analgesic use in Australia between July 2013 and June 2017; and (ii) compare the characteristics of people with and without cancer initiating prescription opioid analgesics.
Methods
Study design and setting
We identified a nationally representative cohort of prevalent and incident opioid users between July 2013 and June 2017. Data were sourced from a random 10% sample of people who accessed subsidized medicines through Australia's Pharmaceutical Benefits Scheme (PBS). The Department of Human Services (DHS) subsidizes PBS prescriptions for Australian citizens, permanent residents and foreign visitors from countries with reciprocal healthcare agreements 18. Two co‐payment thresholds are set each year: a general co‐payment (A$39.50 in 2018) and a concessional co‐payment amount (A$6.40 in 2018). General beneficiaries are entitled to subsidized rates on higher‐cost medicines priced above a set co‐payment, whereas concessional beneficiaries (e.g. pensioners and those earning a low income) pay a reduced co‐payment. Since July 2012, data have been available for all under co‐payment dispensings from approved community pharmacies and private hospitals, and for outpatient and discharge dispensings from public hospitals in most states and territories (except NSW and the ACT). As most opioids are under co‐payment for general beneficiaries, our study period commenced in July 2013, to obtain complete capture of opioid dispensings.
Opioid prevalence and incidence
All PBS‐listed opioids available during the study period were included and defined according to the World Health Organization's Anatomical Therapeutic Chemical (ATC) classification system (Appendix Table A1). To obtain a 10% population sample, population data for adults aged 18–99 years from the Australian Bureau of Statistics were divided by 10 for each financial year (01 July–30 June). This was used as the denominator to determine the prevalence and incidence of opioid use 19. To maintain confidentiality, we did not analyse data for people aged ≥100 years due to small numbers of people in this category. People with ≥1 opioid dispensing in each financial year were defined as prevalent opioid users. Therefore, prevalent users included both new (incident) users and continuous users from the previous financial years. To determine the prevalence of opioid use, the total number of opioid users was divided by the population‐based denominator for each financial year. People initiating opioids were defined as those with no opioid dispensings in the 12 months prior to the opioid index date of supply. Therefore, the number of people aged 18–99 years initiating opioids was divided by the population‐based denominator for each financial year to determine the incidence of use. As we used a 12‐month look‐back period to define initiation, it is possible that a person initiated opioids more than once throughout the 4‐year study period.
Characteristics of people initiating opioids
To explore further the characteristics of people initiating opioids, and to compare these among people with and without cancer, we also examined a range of variables, including opioid characteristics, prescriber specialty, medical comorbidities and other medicine use.
Opioid characteristics
Each person's initial opioid/s were further categorized as short acting, long acting or both; and weak opioid, strong opioid or both (Appendix Table A1). Individuals dispensed both a weak and strong opioid on the index date were categorized as having both strong and weak opioids. Similarly, individuals dispensed both a short‐acting and long‐acting opioid on the index date were categorized as having both short‐acting and long‐acting opioids. The administration route was categorized as oral, transdermal, other (intravenous, subcutaneous or intramuscular injections, buccal, rectal) and ≥2 routes (≥2 opioids dispensed on the index date with different routes). The initial opioid dispensings were categorized as concessional or general. The total oral morphine equivalents (OMEs), in mg, dispensed on the index date was calculated using the following formula: Total OME mg dispensed = (pack strength*OME conversion factor of opioid dispensed*quantity dispensed). The OME conversion factors were adapted from published values by Nielsen et al. 20. The total OME mg dispensed on the index date was calculated and categorized into five groups: total OME mg <100 [e.g. 10 tablets of oral oxycodone 5 mg (75 mg OME)]; 100–249 [e.g. 16 tablets of oral oxycodone 10 mg (240 mg OME)]; 250–499 [e.g. 30 tablets of oral oxycodone 10 mg (450 mg OME)]; 500–749 [e.g. 16 tablets of oral oxycodone 30 mg (720 mg OME)]; and ≥750 [e.g. 20 tablets of oxycodone 30 mg (900 mg OME)].
Prescriber specialty
The speciality of the prescriber for each opioid initiation was categorized consistently with the terminology used by the DHS 21. For each prescription, the DHS provides a prescriber speciality code from a list of 196 codes which are included in the 10% random PBS sample. We categorized these codes into prescriber speciality of interest: general practitioner (GP) (codes 130, 132–134, 178, 184, 201, 202, 450); anaesthetist (specialist and nonspecialist) (codes 51, 60, 216); surgical specialist (codes 32–37, 39, 62, 73, 231, 232, 411); oncology specialist (codes 17, 49, 97, 804); intern (codes 160); dental practitioner (codes 102, 112, 115, 139, 143, 155, 156, 159); nurse practitioner (codes 651); other or missing.
Medical comorbidities
Medicines dispensed 12 months prior to and on the date of initial opioid dispensing were used to identify medical conditions using the validated RxRisk‐V tool 22, 23. We mapped the RxRisk‐V tool to the International Classification of Diseases, Tenth Revision (ICD‐10). To group these disease categories, we matched them to ICD‐10 chapter groupings, as has been done in a previous study utilizing PBS data 11. The RxRisk‐V tool identifies fewer people with cancer than the Charlson's Comorbidity Index (43.2% vs. 67.2%) 22. Therefore, a more comprehensive indicator for cancer was used that included other antineoplastic therapies, such as hormonal cancer therapies, which were not included in the original and updated RxRisk‐V tool (Appendix Table A2).
Previous medicine use
Use of paracetamol, pregabalin and the PBS‐subsidized nonsteroidal anti‐inflammatory drugs (NSAIDs) were examined in the 3 months prior to and including the day of opioid initiation, in order to determine the recent non‐opioid analgesic treatment of pain. This definition is consistent with previous published studies 11. NSAIDs were further categorized into nonselective and cyclooxygenase 2 selective (Appendix Table A1). Recent use of PBS‐subsidized benzodiazepines (Appendix Table A1) was also investigated as benzodiazepines are commonly implicated in opioid overdose deaths 24.
Statistical analyses
Characteristics are summarized as frequencies and percentages, means and standard deviations (SDs), or medians and interquartile ranges (IQRs). Pearson's chi‐square test was used to compare proportions for categorical variables, and Student's t‐test was used for continuous variables. The percentage distribution of types of opioids prescribed at initiation by prescriber speciality and according to age groups were presented graphically. Poisson regression was used to calculate rate ratios (RRs) for the prevalence and incidence of opioid use, comparing the period from 2013/2014 to 2016/2017. All analyses were conducted using SAS Version 9.4 (SAS Institute Inc., Cary, NC, USA).
Ethical review
The Monash University Human Research Ethics Committee approved the study. The analysis plan and manuscript were approved by the Australian Government Department of Human Services.
Results
Opioid prevalence and incidence
There were 287 677 prevalent opioid users between July 2013 and June 2014, and 307 772 between July 2016 and June 2017 (Table 1). From 2013/2014 to 2016/2017, there was an overall 0.6% increase {RR = 1.006 [95% confidence interval (CI) 1.004, 1.008]} in the prevalence of opioid use. Over the study period, the prevalence of opioid use decreased among people with cancer [RR = 0.974 (95% CI 0.967, 0.981)] and increased among people without cancer [RR = 1.008 (95% CI 1.006, 1.009)]. A total of 194 478 people initiated opioids between July 2013 and June 2014, and 190 638 people between July 2016 and June 2017, equating to 10.7% and 10.0% of the total adult Australian population, respectively (Table 1). There was an overall 2.3% decrease [RR = 0.977 (95% CI 0.975, 0.979)] in opioid initiation from 2013/2014 to 2016/2017. Among people initiating opioids who did not have a history of cancer, there was a 2.4% decrease [RR = 0.976 (95% CI 0.974, 0.978)] in opioid initiation and a 2.2% [RR = 1.022 (95% CI 1.006, 1.038)] increase in opioid initiation among people with cancer over the same period.
Table 1.
Incidence and prevalence of opioid use in each financial year for people with and without cancer
| 2013/2014 | 2014/2015 | 2015/2016 | 2016/2017 | |
|---|---|---|---|---|
| N (%) a | ||||
| Incidence (people with cancer) | 3077 (0.17) | 3047 (0.16) | 3133 (0.17) | 3447 (0.18) |
| Incidence (people without cancer) | 191 401 (10.52) | 190 477 (10.31) | 187 561 (9.99) | 187 191 (9.80) |
| Incidence (all) | 194 478 (10.67) | 193 524 (10.47) | 190 694 (10.16) | 190 638 (9.98) |
| Prevalence (people with cancer) | 14 822 (0.81) | 15 026 (0.81) | 14 943 (0.80) | 14 320 (0.75) |
| Prevalence (people without cancer) | 272 855 (14.99) | 286 674 (15.51) | 292 509 (15.59) | 293 452 (15.37) |
| Prevalence (all) | 287 677 (15.81) | 301 700 (16.34) | 307 452 (16.38) | 307 772 (16.12) |
Percentage was determined using 10% of the Australian Bureau of Statistics population figures for people aged 18–99 years for the denominator for each financial year
Characteristics of people initiating opioids
Demographics
In total, 769 334 people initiated opioids between July 2013 and June 2017. The mean (SD) age of the cohort was 49.5 (18.9) years and 53.1% were female (Table 2). There were 756 630 people initiating opioids without cancer, with a mean (SD) age of 49.2 (18.8) years. There were 12 704 people with cancer who initiated opioids, with a mean (SD) age of 66.6 (14.8) years.
Table 2.
Baseline characteristics in the 365 days prior to, and on the day of, opioid initiation
| All (n = 769 334) | People without cancer (n = 756 630) | People with cancer (n = 12 704) | P‐valuea , b | |
|---|---|---|---|---|
| N (%) or mean ± standard deviation | ||||
| Demographic characteristics | ||||
| Age, years | 49.5 ± 18.9 | 49.2 ± 18.8 | 66.6 ± 14.8 | <0.001 |
| 18–29 | 135 850 (17.7) | 135 580 (17.9) | 270 (2.1) | <0.001c |
| 30–44 | 198 473 (25.8) | 197 623 (26.1) | 850 (6.7) | |
| 45–54 | 128 692 (16.7) | 127 424 (16.8) | 1268 (10.0) | |
| 55–64 | 119 340 (15.5) | 116 823 (15.4) | 2517 (19.8) | |
| 65–74 | 99 260 (12.9) | 95 586 (12.6) | 3674 (28.9) | |
| 75+ | 87 719 (11.4) | 83 594 (11.1) | 4125 (32.5) | |
| Sex, female | 408 281 (53.1) | 401 597 (53.1) | 6684 (52.6) | 0.299 |
| Concessional beneficiary | 300 249 (39.0) | 292 000 (38.6) | 8249 (64.9) | <0.001 |
| Comorbidities d | ||||
| Total number e | 2.7 ± 2.3 | 2.7 ± 2.3 | 4.7 ± 2.5 | <0.001 |
| Certain infectious and parasitic diseases | 1987 (0.3) | 1893 (0.3) | 94 (0.7) | <0.001 |
| Endocrine, nutritional and metabolic disorders | 97 005 (12.6) | 94 284 (12.5) | 2721 (21.4) | <0.001 |
| Mental and behavioural disorders | 203 610 (26.5) | 199 553 (26.4) | 4057 (31.9) | <0.001 |
| Alcohol dependence | 1495 (0.2) | 1481 (0.2) | 14 (0.1) | 0.030 |
| Anxiety | 62 570 (8.1) | 61 366 (8.1) | 1204 (9.5) | <0.001 |
| Bipolar disorder | 2297 (0.3) | 2266 (0.3) | 31 (0.2) | 0.256 |
| Dementia | 3168 (0.4) | 3079 (0.4) | 89 (0.7) | <0.001 |
| Depression | 150 706 (19.6) | 147 685 (19.5) | 3021 (23.8) | <0.001 |
| Nicotine dependence | 17 346 (2.3) | 17 135 (2.3) | 211 (1.7) | <0.001 |
| Psychotic illness | 21 875 (2.8) | 21 356 (2.8) | 519 (4.1) | <0.001 |
| Diseases of the nervous system | 36 102 (4.7) | 35 267 (4.7) | 835 (6.6) | <0.001 |
| Diseases of the circulatory system | 280 017 (36.4) | 271 370 (35.9) | 8647 (68.1) | <0.001 |
| Diseases of the respiratory system | 109 686 (14.3) | 107 234 (14.2) | 2452 (19.3) | <0.001 |
| Diseases of the digestive system | 187 767 (24.4) | 181 120 (23.9) | 6647 (52.3) | <0.001 |
| Diseases of the musculoskeletal system and connective tissue | 49 738 (6.5) | 46 926 (6.2) | 2812 (22.1) | <0.001 |
| Diseases of the genitourinary system | 8481 (1.1) | 8052 (1.1) | 429 (3.4) | <0.001 |
| Other | 282 405 (36.7) | 275 023 (36.4) | 7382 (58.1) | <0.001 |
Chi‐square test
Student's t‐test
A chi‐square comparison between all the age groups
Determined by the RxRisk‐V tool
Total number excluding pain and cancer
Comorbidities
The mean (SD) number of comorbidities for all people initiating opioids was 2.7 (2.3) (Table 2). Among people without cancer, the mean (SD) number of comorbidities was 2.7 (2.3), compared with 4.7 (2.5) among those with cancer (P < 0001). Among all people initiating opioids, diseases of the circulatory system (36.4%), mental and behavioural disorders (26.5%) and diseases of the digestive system (24.4%) were common. Mental and behavioural disorders were more common among people with vs. without cancer (31.9% vs. 26.4%; P < 0.001), as were disorders of the circulatory system (68.1% vs. 35.9%; P < 0001) and digestive system (52.3% vs. 23.9%; P < 0.001).
Characteristics of initial opioid(s)
Paracetamol/codeine was the most commonly dispensed opioid on initiation (54.7%), followed by single‐ingredient oxycodone (23.7%) and tramadol (10.8%) (Table 3). The distribution of prescribed opioids by opioid type in each financial year and by age group are depicted in Figures 1 and 2. A strong opioid was initiated in 28.7% of the cohort and 8.6% were initiated on a long‐acting opioid. The proportion of people initiating strong opioids increased from 26.0% to 31.6%, and the proportion of people initiating long‐acting opioids increased from 8.0% to 9.5% between 2013/2014 and 2016/2017. The majority of people (97.6%) were initiated on oral opioid formulations. Most people (60.1%) initiating opioids had less than 100 mg in total OMEs dispensed [e.g. 10 tablets of oxycodone 5 mg (75 OME mg) or 20 tablets of paracetamol/codeine 500/30 mg (60 OME mg)]. Over half of the people with cancer (55.8%) initiated a strong opioid compared with 28.2% of people without cancer (P < 0.001). People with cancer had a median (IQR) of 150 (60–210) mg OMEs dispensed on initiation, compared with a median (IQR) of 60 (60–150) mg OMEs among people without cancer. The proportion of people dispensed total OME <250 mg did not change over the 4‐year study period (Appendix Table A3). People with cancer were more commonly dispensed a long‐acting opioid (16.5% vs. 8.4%), oxycodone (37.2% vs. 23.4%) and oxycodone/naloxone (7.8% vs. 3.1%) than those without cancer (P < 0.001).
Table 3.
Characteristics of opioids dispensed on the date of initiation
| All (n = 769 334) | People without cancer (n = 756 630) | People with cancer (n = 12 704) | P‐valuea , b | |
|---|---|---|---|---|
| N (%) or median (interquartile range) | ||||
| Type of opioid | ||||
| Weak opioid | 530 660 (69.0) | 525 281 (69.4) | 5379 (42.3) | <0.001 |
| Strong opioid | 220 445 (28.7) | 213 363 (28.2) | 7082 (55.8) | |
| Both weak and strong opioid | 18 229 (2.4) | 17 986 (2.4) | 243 (1.9) | |
| Short acting | 676 790 (88.0) | 667 102 (88.2) | 9688 (76.3) | <0.001 |
| Long acting | 65 883 (8.6) | 63 784 (8.4) | 2099 (16.5) | |
| Both short and long acting | 26 661 (3.5) | 25 744 (3.4) | 917 (7.2) | |
| Route of administration | ||||
| Oral | 750 478 (97.6) | 738 751 (97.6) | 11 727 (92.3) | <0.001 |
| Transdermal | 13 035 (1.7) | 12 473 (1.7) | 562 (4.4) | |
| Other | 3652 (0.5) | 3422 (0.5) | 230 (1.8) | |
| ≥2 routes | 2169 (0.3) | 1984 (0.3) | 185 (1.5) | |
| Opioid dispensed | ||||
| Buprenorphine | 10 990 (1.4) | 10 622 (1.4) | 368 (2.9) | <0.001 |
| Codeine | 33 852 (4.4) | 33 207 (4.4) | 645 (5.1) | <0.001 |
| Fentanyl | 2178 (0.3) | 1972 (0.3) | 206 (1.6) | <0.001 |
| Hydromorphone | 954 (0.1) | 834 (0.1) | 120 (0.9) | <0.001 |
| Methadone | 143 (0.0) | 128 (0.0) | 15 (0.1) | <0.001 |
| Morphine | 6671 (0.9) | 5930 (0.8) | 741 (5.8) | <0.001 |
| Paracetamol/codeine | 420 766 (54.7) | 417 108 (55.1) | 3658 (28.8) | <0.001 |
| Oxycodone | 182 021 (23.7) | 177 293 (23.4) | 4728 (37.2) | <0.001 |
| Oxycodone/naloxone | 24 270 (3.2) | 23 282 (3.1) | 988 (7.8) | <0.001 |
| Tapentadol c | 4209 (0.6) | 4122 (0.5) | 87 (0.7) | 0.034 |
| Tramadol | 83 280 (10.8) | 82 132 (10.9) | 1148 (9.0) | <0.001 |
| Oral morphine equivalents (OME), in mg dispensed | ||||
| Total OME d | 60 (60–150) | 60 (60–150) | 150 (60–210) | <0.001 |
| Total OME <100 | 462 692 (60.1) | 458 275 (60.6) | 4417 (34.8) | <0.001 |
| Total OME 100–249 | 233 612 (30.4) | 228 097 (30.2) | 5515 (43.4) | |
| Total OME 250–499 | 49 005 (6.4) | 47 358 (6.3) | 1647 (13.0) | |
| Total OME 500–749 | 13 304 (1.7) | 12 881 (1.7) | 423 (3.3) | |
| Total OME ≥750 | 10 721 (1.4) | 10 019 (1.3) | 702 (5.5) | |
Chi‐square test
Student's t‐test
Tapentadol was Pharmaceutical Benefits Scheme listed in November 2013
Example of initial OME of ~60 mg: eight tablets of oxycodone 5 mg (60 OME mg) or 20 tablets of paracetamol/codeine 500/30 mg (60 OME mg); initial OME of ~240 mg: 16 tablets of oxycodone 10 mg (240 OME mg) or 80 tablets of paracetamol/codeine 500/30 mg (240 OME mg)
Figure 1.
Proportion of opioids dispensed on initiation for each financial year
Figure 2.
Percentage of each opioid type dispensed on initiation by age group between July 2013 and June 2017
Prior medicine use
Non‐opioid analgesics were used by 36.6% of people prior to opioid initiation, with NSAIDs being the most commonly (27.5%) used non‐opioid analgesic (Table 4). Prior paracetamol (23.1% vs. 12.2%; P < 0.001) and pregabalin (6.8% vs. 3.5%; P < 0.001) use was more common among people with cancer compared with those without cancer, whereas NSAID use (27.6% vs. 21.5%; P < 0.001) was more common among people without cancer. Benzodiazepines were used by 13.2% of the cohort. Prior benzodiazepine use was more common among people with cancer than without cancer (22.6% vs. 13.0%; P < 0.001).
Table 4.
Selected medicine use in the 3 months prior to, and on the day of, opioid initiation
| All (n = 769 334) | People without cancer (n = 756 630) | People with cancer (n = 12 704) | P‐valuea | |
|---|---|---|---|---|
| Benzodiazepines | 101 338 (13.2) | 98 473 (13.0) | 2865 (22.6) | <0.001 |
| Alprazolam | 3165 (0.4) | 3098 (0.4) | 67 (0.5) | 0.039 |
| Diazepam | 49 569 (6.4) | 48 765 (6.5) | 804 (6.3) | 0.596 |
| Nitrazepam | 3267 (0.4) | 3143 (0.4) | 124 (1.0) | <0.001 |
| Oxazepam | 12 891 (1.7) | 12 503 (1.7) | 388 (3.1) | <0.001 |
| Temazepam | 45 557 (5.9) | 43 687 (5.8) | 1870 (14.7) | <0.001 |
| Non‐opioid analgesics | 281 365 (36.6) | 276 069 (36.5) | 5296 (41.7) | <0.001 |
| Paracetamol b | 95 315 (12.4) | 92 382 (12.2) | 2933 (23.1) | <0.001 |
| Pregabalin | 27 401 (3.6) | 26 539 (3.5) | 862 (6.8) | <0.001 |
| NSAIDs | 211 428 (27.5) | 208 691 (27.6) | 2737 (21.5) | <0.001 |
| Selective COX‐2 inhibitors | 106 821 (13.9) | 105 184 (13.9) | 1637 (12.9) | 0.001 |
| Meloxicam | 62 434 (8.1) | 61 529 (8.1) | 905 (7.1) | <0.001 |
| Celecoxib | 48 563 (6.3) | 47 770 (6.3) | 793 (6.2) | 0.743 |
| Nonselective NSAIDs | 118 971 (15.5) | 117 704 (15.6) | 1267 (10.0) | <0.001 |
| Diclofenac | 42 001 (5.5) | 41 579 (5.5) | 422 (3.3) | <0.001 |
| Ibuprofen | 35 306 (4.6) | 35 002 (4.6) | 304 (2.4) | <0.001 |
| Naproxen | 26 518 (3.5) | 26 183 (3.5) | 335 (2.6) | <0.001 |
| Indomethacin | 13 901 (1.8) | 13 730 (1.8) | 171 (1.4) | <0.001 |
| Piroxicam | 3466 (0.5) | 3423 (0.5) | 43 (0.3) | 0.057 |
| Mefenamic acid | 2682 (0.4) | 2675 (0.4) | 7 (0.1) | <0.001 |
| Ketoprofen | 1627 (0.2) | 1596 (0.2) | 31 (0.2) | 0.421 |
COX, cyclooxygenase; NSAID, non‐steroidal anti‐inflammatory drug
Chi‐square test
Since January 2016, paracetamol has no longer been subsidized on the Pharmaceutical Benefits Scheme
Prescriber specialty
Approximately half of all opioid initiations were by GPs (51%) (Appendix Figure A1). Codeine was the most commonly prescribed opioid by dental practitioners (97.7%), GPs (67.9%) and nurse practitioners (61.7%). Oxycodone was the most commonly prescribed opioid by interns (65.6%), oncology specialists (57.2%), surgical specialists (47.0%) and anaesthetic specialists and nonspecialists (43.8%) (Figure 3).
Figure 3.
Percentage of opioids initiated by prescriber specialities of interest between July 2013 and June 2017. *Other category for oncology specialists:10.7% morphine, 4.8% fentanyl and 4.5% buprenorphine, tapentadol and methadone
Discussion
This is the first study to report the prevalence and incidence of opioid use in Australian adults. Approximately 16% of the adult population use opioids each financial year, and the prevalence of opioid use increased by 0.6% over the study period. The rate of opioid initiation decreased by 2.3% over the study period, with approximately 10% of adults initiating opioids each financial year. Paracetamol/codeine, followed by single‐ingredient oxycodone and tramadol, were the most commonly prescribed opioids on initiation, with codeine use declining in the previous 2 years and oxycodone use increasing. Approximately half of all opioid initiations were by GPs.
Extrapolating to the total population, each year over 1.9 million adults (10%) are estimated to initiate opioids and over 3 million adults (16%) are using opioids in Australia. The prevalence of opioid use in Australia is similar to the 14% prevalence in Ontario, Canada 25. Despite increasing prescribing rates between 2006 and 2010 in the US, annual prescribing rates declined from 81.2 to 70.6 per 100 persons between 2010 and 2015 17. Our results also demonstrate a decrease (2.3%) in overall opioid initiation in Australian adults since 2013/2014. In light of increasing concerns about opioid‐related harms, it is encouraging that the rate of opioid initiation appears to be decreasing. However, it is not known whether the decrease in opioid initiation has translated to a decrease in harm. The total OME amount dispensed on initiation has remained relatively stable over the 4‐year study period. However, the proportion of people initiating strong or long‐acting opioids has increased over the same period. This is concerning because strong and long‐acting opioids have been linked to harms in both the US 26 and Australia 9. Therefore, future studies should investigate the impact of recent changes in opioid use on opioid‐related harm in Australia. Our study supports recent calls from the Therapeutic Goods Administration for a review of current practice and policy related to opioid prescribing 27.
Previous research investigating opioid use in Australia and the US attributed the overall increase in opioid use to their increase in use for CNCP 3, 4, 28. In our study, previous non‐opioid analgesic use was higher among people with cancer compared with those without cancer. CNCP management is complex and requires individualization 29. Although multimodal strategies for CNCP are recommended 29, in our study only one‐third of people without cancer had used non‐opioid analgesics in the previous 3 months. Therefore, our results may suggest suboptimal treatment with non‐opioid analgesics, particularly in those with CNCP, where opioid use should be reserved for those who are unresponsive to optimized non‐opioid therapies 29, 30. However, data were not available on non‐opioid analgesics purchased over the counter and therefore may underestimate non‐opioid analgesic use. Additionally, no data were available on prescriptions that were issued by prescribers but never dispensed in pharmacies (e.g. people who successfully trialled non‐opioids first and decided not to fill their opioid prescription). Nevertheless, non‐opioid treatment modalities should be trialled first, to reduce inappropriate long‐term opioid use and minimize harm. A US study found that 5% of opioid initiators became long‐term users, defined as ≥6 opioid prescriptions in the 12 months following the initiation month 16. They also found that long‐term use was, in turn, associated with a higher risk of opioid‐related hospitalization 16. A recent study in Australia, using group‐based trajectory modelling to define persistent use, found that 2.6% of adults initiating opioids became persistent users over a 12‐month period 31.
In order to identify risk mitigation strategies for opioid‐related harm, we examined prescriber specialty and the characteristics of the initial opioid. We found that approximately half of all opioid initiations were by GPs. Similarly, US studies have reported that primary care specialities prescribed 44.5% of dispensed opioid prescriptions in 2012 32. As pain conditions account for at least 11% of all conditions managed by GPs in Australia 33, it is not surprising that half of all opioid initiations were by GPs. Yet, previous research in the US reported that of all specialities prescribing opioids, primary care specialities were most often associated with opioid fatalities 34. This is concerning as, although Australia has followed a similar trend to the US in opioid‐related harm, to date there have been minimal regulatory, policy or programmatic shifts in response 35. Although interns are responsible for 8.3% of all opioids initiated, we found that oxycodone was the most commonly prescribed opioid by this group (65.6%). Interns usually prescribe medicines on discharge from hospitals in Australia under the guidance of the specialty team. The proportion of interns prescribing opioids provides important information regarding opioid initiation from hospitals. However, this may be underestimated as prescriptions dispensed at hospitals in NSW and the ACT are not captured in this dataset.
Strengths and limitations
The key strength of our study was the use of data from a 10% random sample of people accessing medicines through the PBS over a period when under co‐payment data were captured. Although we captured more than 80% of prescription opioid use in Australia, data on private non‐PBS prescriptions and over‐the‐counter analgesics were not captured 27. The subsidy of paracetamol and tapentadol has changed throughout the study period. Since January 2016, paracetamol has no longer been subsidized on the PBS. This means that the estimate of prior paracetamol use since 2016 may be an underestimate of actual use. Tapentadol was approved for listing on the PBS in November 2013. Therefore, its lower utilization compared with other opioids may reflect a slower uptake by prescribers. Additionally, opioid initiation for hospital inpatients in all States and on discharge from hospitals in NSW and the ACT were not captured in this dataset. Therefore, our study does not necessarily reflect the characteristics of all adults initiating opioids in Australia. As PBS data are based on dispensing records, we do not know whether and how the medicine was taken. In the absence of comorbidity data, we used dispensing data to identify treatment for specific medical conditions. Although we used a comprehensive list of cancer medicines, some people with cancer may not have had cancer medicines dispensed and therefore may have been misclassified. Additionally, people may have been categorized into the noncancer group if they had a first dispensing of a cancer medicine after opioid initiation. Clinical information, including the indication for opioid initiation, dose, duration and severity of pain, was not available in our dataset.
Conclusion
Rates of opioid use in Australia have remained high since 2013, with approximately 3 million adults using opioids and over 1.9 million adults initiating opioids each year. There has been a slight increase in the prevalence of opioid use but a decrease in incidence in the 4‐year period between 2013 and 2017. People without cancer account for the majority of opioid use and are more likely to be initiated on short‐acting and weak opioids. The proportion of people initiating on strong opioids has increased over time, reinforcing concerns about the increase in the use and harms associated with strong opioids in the community.
Competing Interests
The authors have no known or potential conflicts of interest to declare. S.L. is supported through an Australian Government Research Training Program Scholarship. J.I. and N.G. are supported by the NHMRC Early Career Fellowship Scheme (grant numbers #1072137 and #1091878). J.S.B. is supported by a NHMRC Dementia Leadership Fellowship (grant number #1140298).
The authors would like to acknowledge the Australian Government, Department of Human Services for the provision of the data. The authors had full access to all of the data (including statistical reports and tables) in the study.
Appendix 1.
Table A1.
Name of medicines used in analyses and their corresponding Anatomical Therapeutic Classification codesa
| Medicine name | Anatomical Therapeutic Classification code |
|---|---|
| Paracetamol | N02BE01 |
| Pregabalin | N03AX16 |
|
NSAIDs
Nonselective • Diclofenac • Ibuprofen • Indomethacin • Ketoprofen • Mefenamic acid • Naproxen • Piroxicam COX‐2 selective • Celecoxib • Meloxicam |
M01AB01‐M01AH06 M01AB05 M01AE01 M01AB01 M01AE03 M01AG01 M01AE02 M01AC01 M01AH01 M01AC06 |
|
Benzodiazepines
• Diazepam • Oxazepam • Bromazepam • Alprazolam • Nitrazepam • Flunitrazepam • Temazepam • Midazolam |
N05BA, N05CD N05BA01 N05BA04 N05BA08 N05BA12 N05CD02 N05CD03 N05CD07 N05CD08 |
|
Opioids
Weaker opioids • Codeine • Combination codeine preparations • Tramadol • Tapentadol Stronger opioids • Morphine • Hydromorphone • Oxycodone • Oxycodone/naloxone • Fentanyl • Methadone • Buprenorphine |
NO2A, R05DA04 R05DA04 N02AA59, N02AJ06 N02AX02 N02AX06 N02AA01 N02AA03 N02AA05 N02AA55 N02AB03 N02AC52 N02AE01 |
COX, cyclooxygenase; NSAID, non‐steroidal anti‐inflammatory drug
ATC/DDD Index 2018: World Health Organization; 2017 [19/05/18]. Available from: https://www.whocc.no/atc_ddd_index/
Table A2.
RxRisk‐V categories and corresponding Anatomical Therapeutic Classification codes
| Disease category | Anatomical Therapeutic Classification codes |
|---|---|
| Alcohol dependence | N07BB01‐N07BB99 |
| Allergies | R01AC01‐R01AD60, R06AD02‐R06AX27, R06AB04 |
| Anticoagulation | B01AA03‐B01AB06, B01AE07, B01AF01, B01AF02, B01AX05 |
| Antiplatelet | B01AC04‐B01AC30 |
| Anxiety and tension | N05BA01‐N05BA56, N05BE01 |
| Arrhythmias | C01AA05, C01BA01‐C01BD01, C07AA07 |
| BPH | G04CA02‐G04CA03, G04CB01, G04CB02, C02CA01 |
| Bipolar disorder | N05AN01 |
| CHF/hypertension | [C03CA01‐C03CC01 AND (C09AA01‐C09AA16 OR, C09CA01 ‐ C09CX99)], C03DA04, C07AB07, C07AG02, C07AB12, C09DX04, C07AB02_2 |
| Dementia | N06DA02‐N06DA04, N06DX01 |
| Depression | N06AA01‐N06AG02, N06AX03‐N06AX11, N06AX13‐N06AX26 |
| Diabetes | A10AA01‐A10BX08 |
| End‐stage renal disease | B03XA01‐B03XA03, V03AE02, V03AE03, V03AE05 |
| Epilepsy | N03AA01‐N03AX30 |
| Gastric acid disorder | A02BA01‐A02BX77 |
| Glaucoma | S01EA01‐S01EB03, S01EC03‐S01EX02 |
| Gout | M04AA01‐M04AC01 |
| Hepatitis B | J05AF08, J05AF10, J05AF11 |
| Hepatitis C | J05AE11‐J05AE15, J05AX14‐J05AX68, J05AB04, L03AB11, L03AB60, L03AB61 |
| HIV | J05AE01‐J05AE10, J05AF01‐J05AF07, J05AF09, J05AF12‐J05AG05, J05AR01‐J05AR19, J05AX07‐J05AX09, J05AX12 |
| Hyperkalaemia | V03AE01 |
| Hyperlipidaemia | C10AA01‐C10BX12 |
| HTN | C03AA01‐C03BA11, C03BB04, C03DA01‐C03DA03, C03EA01‐C03EA14, C09BA02‐C09BA15, C09DA01‐C09DA09, C02AB01‐C02AC05, C02DB01‐C02DB04, C03DB01‐C03DB02 |
| Hypothyroidism | H03AA01‐H03AA05 |
| IHD/angina | C01DA02‐C01DA70, C01DX16, C08EX02 |
| IHD/HTN | C07AA01‐C07AA06,C07AG01, C08CA01‐C08DB01, C09DB01‐C09DB08, C09DX01‐C09DX03, C09BB02‐C09BB12, C07AB03, C07AB02_1 |
| IBD | A07EC01‐A07EC04,A07EA01‐A07EA02, A07EA06, L04AA33 |
| Liver failure | A07AA11 |
| Malignancy | L01AA01‐L01AX04, L01BA01_2, L01BA03‐ L01XX53, L02BG03, L02BG04, L02BG06, L02BB01‐L02BB04, L02BX01‐L02BX03, L04AX02, L04AX04, L04AX06, L02BA01_01, L02AE03_1, L02AE02_1 |
| Malnutrition | B05BA01‐B05BA10 |
| Migraine | N02CA01‐N02CX01 |
| Osteoporosis/Paget's disease | M05BA01‐M05BB08, M05BX03, M05BX04, H05AA02 |
| Pain a | N02AA01‐N02AX99, R05DA04 |
| Pain/inflammation | M01AB01‐M01AH06 |
| Pancreatic insufficiency | A09AA02 |
| Parkinson's disease | N04AA01‐N04BX03 |
| Psoriasis | D05AA, D05BB01‐D05BB02, D05AX02, D05AC01‐D05AC51, D05AX52 |
| Psychotic illness | N05AA01‐N05AB02, N05AB06‐N05AL07, N05AX01‐N05AX17 |
| Reactive airway disease | R03AC02‐R03DC03, R03DX05 |
| Nicotine dependence | N07BA01‐N07BA03, N06AX12 |
| Steroid‐responsive conditions | H02AB01‐H02AB17 |
| Transplant | L04AA06, L04AA10, L04AA18, L04AD01, L04AD02, L04AC02 |
| Tuberculosis | J04AB02, J04AC01‐J04AM06 |
BPH, benign prostatic hypertrophy; CHF, congestive heart failure; HTN, hypertension; IBD, inflammatory bowel disease; IHD, ischaemic heart disease; HIV, human immunodeficiency virus
Where the ATC code has _01 or_02 at the end, this can be used for multiple indications; therefore, to separate indications based on Pharmaceutical Benefits Scheme item codes, these additional digits were used.
Table A3.
Total oral morphine equivalents (OMEs), in mg, dispensed on initiation in each financial year
| 2013/2014 | 2014/2015 | 2015/2016 | 2016/2017 | |
|---|---|---|---|---|
| N (%) | ||||
| Total OME mg dispensed | ||||
| <100 | 120 304 (61.9) | 117 918 (60.9) | 113 531 (59.5) | 110 939 (58.2) |
| 100–249 | 56 489 (29.1) | 57 846 (29.9) | 58 917 (30.9) | 60 360 (31.7) |
| 250–499 | 11 704 (6.0) | 11 884 (6.1) | 12 364 (6.5) | 13 053 (6.9) |
| 500–749 | 3236 (1.7) | 3263 (1.7) | 3323 (1.7) | 3482 (1.8) |
| ≥750 | 2745 (1.4) | 2613 (1.4) | 2559 (1.3) | 2804 (1.5) |
Figure A1.
Proportion of opioids initiated during the study period by each prescriber speciality
Lalic, S. , Ilomäki, J. , Bell, J. S. , Korhonen, M. J. , and Gisev, N. (2019) Prevalence and incidence of prescription opioid analgesic use in Australia. Br J Clin Pharmacol, 85: 202–215. 10.1111/bcp.13792.
References
- 1. Blanch B, Pearson SA, Haber PS. An overview of the patterns of prescription opioid use, costs and related harms in Australia. Br J Clin Pharmacol 2014; 78: 1159–1166. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Berterame S, Erthal J, Thomas J, Fellner S, Vosse B, Clare P, et al Use of and barriers to access to opioid analgesics: a worldwide, regional, and national study. Lancet 2016; 387: 1644–1656. [DOI] [PubMed] [Google Scholar]
- 3. Karanges EA, Blanch B, Buckley NA, Pearson SA. Twenty‐five years of prescription opioid use in Australia: a whole‐of‐population analysis using pharmaceutical claims. Br J Clin Pharmacol 2016; 82: 255–267. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Boudreau D, Von Korff M, Rutter CM, Saunders K, Ray GT, Sullivan MD, et al Trends in long‐term opioid therapy for chronic non‐cancer pain. Pharmacoepidemiol Drug Saf 2009; 18: 1166–1175. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Schug, Palmer GM, Scott DA, Halliwell R, Trinca J. Acute Pain Management: Scientific Evidence, 4th edn. Melbourne: APM:SE Working Group of the Australian and New Zealand College of Anaesthetists and Faculty of Pain Medicine, 2015. [Google Scholar]
- 6. Caraceni A, Hanks G, Kaasa S, Bennett MI, Brunelli C, Cherny N, et al Use of opioid analgesics in the treatment of cancer pain: evidence‐based recommendations from the EAPC. Lancet Oncol 2012; 13: e58–e68. [DOI] [PubMed] [Google Scholar]
- 7. Chou R, Turner JA, Devine EB, Hansen RN, Sullivan SD, Blazina I, et al The effectiveness and risks of long‐term opioid therapy for chronic pain: a systematic review for a National Institutes of Health pathways to prevention workshop. Ann Intern Med 2015; 162: 276–286. [DOI] [PubMed] [Google Scholar]
- 8. Roxburgh A, Bruno R, Larance B, Burns L. Prescription of opioid analgesics and related harms in Australia. Med J Aust 2011; 195: 280–284. [DOI] [PubMed] [Google Scholar]
- 9. Pilgrim JL, Yafistham SP, Gaya S, Saar E, Drummer OH. An update on oxycodone: lessons for death investigators in Australia. Forensic Sci Med Pathol 2015; 11: 3–12. [DOI] [PubMed] [Google Scholar]
- 10. Gomes T, Tadrous M, Mamdani MM, Paterson J, Juurlink DN. The burden of opioid‐related mortality in the United States. JAMA Netw Open 2018; 1: e180217. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Gisev N, Pearson SA, Blanch B, Larance B, Dobbins T, Larney S, et al Initiation of strong prescription opioids in Australia: cohort characteristics and factors associated with the type of opioid initiated. Br J Clin Pharmacol 2016; 82: 1123–1133. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Campbell G, Nielsen S, Bruno R, Lintzeris N, Cohen M, Hall W, et al The pain and opioids IN treatment study: characteristics of a cohort using opioids to manage chronic non‐cancer pain. Pain 2015; 156: 231–242. [DOI] [PubMed] [Google Scholar]
- 13. Gureje O, Von Korff M, Kola L, Demyttenaere K, He Y, Posada‐Villa J, et al The relation between multiple pains and mental disorders: results from the world mental health surveys. Pain 2008; 135: 82–91. [DOI] [PubMed] [Google Scholar]
- 14. Mailis‐Gagnon A, Lakha SF, Ou T, Louffat A, Yegneswaran B, Umana M, et al Chronic noncancer pain: characteristics of patients prescribed opioids by community physicians and referred to a tertiary pain clinic. Can Fam Physician 2011; 57: e97–e105. [PMC free article] [PubMed] [Google Scholar]
- 15. Lakha SF, Louffat AF, Nicholson K, Deshpande A, Mailis‐Gagnon A. Characteristics of chronic noncancer pain patients assessed with the opioid risk tool in a Canadian tertiary care pain clinic. Pain Med 2014; 15: 1743–1749. [DOI] [PubMed] [Google Scholar]
- 16. Deyo RA, Hallvik SE, Hildebran C, Marino M, Dexter E, Irvine JM, et al Association between initial opioid prescribing patterns and subsequent long‐term use among opioid‐naive patients: a statewide retrospective cohort study. J Gen Intern Med 2017; 32: 21–27. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Guy GP Jr, Zhang K, Bohm MK, Losby J, Lewis B, Young R, et al Vital signs: changes in opioid prescribing in the United States, 2006–2015. MMWR Morb Mortal Wkly Rep 2017. [online]. Available at https://www.cdc.gov/mmwr/volumes/66/wr/mm6626a4.htm (last accessed 8 October 2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Mellish L, Karanges EA, Litchfield MJ, Schaffer AL, Blanch B, Daniels BJ, et al The Australian Pharmaceutical Benefits scheme data collection: a practical guide for researchers. BMC Res Notes 2015; 8: 634. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Australian Bureau of Statistics. Australian demographic statistics; 2017. [online]. Available at http://www.abs.gov.au/AUSSTATS/abs@.nsf/DetailsPage/3101.0Jun%202017?OpenDocument (last accessed 7 October 2018).
- 20. Nielsen S, Degenhardt L, Hoban B, Gisev N. A synthesis of oral morphine equivalents (OME) for opioid utilisation studies. Pharmacoepidemiol Drug Saf 2016; 25: 733–737. [DOI] [PubMed] [Google Scholar]
- 21. Australian Institute of Health and Welfare, Australian Government . Pharmaceutical Benefits Scheme prescription – specialty of prescriber [online]. Available at http://meteor.aihw.gov.au/content/index.phtml/itemId/600679 (last accessed 7 October 2018).
- 22. Lu CY, Barratt J, Vitry A, Roughead E. Charlson and Rx‐risk comorbidity indices were predictive of mortality in the Australian health care setting. J Clin Epidemiol 2011; 64: 223–228. [DOI] [PubMed] [Google Scholar]
- 23. Pratt NL, Kerr M, Barratt JD, Kemp‐Casey A, Kalisch Ellett LM, Ramsay E, et al The validity of the Rx‐risk comorbidity index using medicines mapped to the anatomical therapeutic chemical (ATC) classification system. BMJ Open 2018; 8: e021122. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Sun EC, Dixit A, Humphreys K, Darnall BD, Baker LC, Mackey S. Association between concurrent use of prescription opioids and benzodiazepines and overdose: retrospective analysis. BMJ 2017; 356: j760. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Health Quality Ontario . 9 million prescriptions: what we know about the growing use of prescription opioids in Ontario [online]. Available at http://www.hqontario.ca/portals/0/Documents/system‐performance/9‐million‐prescriptions‐en.pdf (last accessed 2 October 2018).
- 26. Hirsch A, Proescholdbell SK, Bronson W, Dasgupta N. Prescription histories and dose strengths associated with overdose deaths. Pain Med 2014; 15: 1187–1195. [DOI] [PubMed] [Google Scholar]
- 27. Gisev N, Pearson SA, Karanges EA, Larance B, Buckley NA, Larney S, et al To what extent do data from pharmaceutical claims under‐estimate opioid analgesic utilisation in Australia? Pharmacoepidemiol Drug Saf 2018; 27: 550–555. [DOI] [PubMed] [Google Scholar]
- 28. Sullivan MD, Edlund MJ, Fan MY, Devries A, Brennan Braden J, Martin BC. Trends in use of opioids for non‐cancer pain conditions 2000‐2005 in commercial and Medicaid insurance plans: the TROUP study. Pain 2008; 138: 440–449. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Dowell D, Haegerich TM, Chou R. CDC guideline for prescribing opioids for chronic pain – United States, 2016. JAMA 2016; 315: 1624–1645. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Busse JW, Craigie S, Juurlink DN, Buckley DN, Wang L, Couban RJ, et al Guideline for opioid therapy and chronic noncancer pain. CMAJ 2017; 189: E659–E666. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Lalic S, Gisev N, Bell JS, Korhonen MJ, Ilomaki J. Predictors of persistent prescription opioid analgesic use among people without cancer in Australia. Br J Clin Pharmacol 2018; 84: 1267–1278. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Levy B, Paulozzi L, Mack KA, Jones CM. Trends in opioid analgesic‐prescribing rates by specialty, US, 2007–2012. Am J Prev Med 2015; 49: 409–413. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Britt H, Miller GC, Henderson J, Bayram C, Harrison C, Valenti L, et al General practice activity in Australia 2015–16, General practice series no. 40 Sydney: Sydney University Press, 2016. [Google Scholar]
- 34. Porucznik CA, Johnson EM, Rolfs RT, Sauer BC. Specialty of prescribers associated with prescription opioid fatalities in Utah, 2002–2010. Pain Med 2014; 15: 73–78. [DOI] [PubMed] [Google Scholar]
- 35. Larance B, Degenhardt L, Peacock A, Gisev N, Mattick R, Colledge S, et al Pharmaceutical opioid use and harm in Australia: the need for proactive and preventative responses. Drug Alcohol Rev 2018; 37 (Suppl. 1): S203–S205. [DOI] [PubMed] [Google Scholar]