Patterns of opioid dispensing and associated wage replacement duration in workers with accepted claims for low back pain: a retrospective cohort study

Trends in opioid dispensing to injured workers following codeine scheduling changes in Australia: a retrospective cohort study

OBJECTIVES

To describe the prevalence and patterns of opioid analgesic and pain medicine dispenses, and the impact of up-scheduling of low-dose (≤15 mg) codeine-containing products to Australians with accepted workers' compensation time loss claims for musculoskeletal conditions between 2010 and 2019.

DESIGN

Interrupted time series.

SETTING

Workers' compensation scheme in Victoria, Australia.

POPULATION

Australians with accepted workers' compensation time loss claims for musculoskeletal conditions between 2010 and 2019.

MAIN OUTCOME MEASURES

Number and proportion of workers dispensed pain medicines in the first year of claim and the monthly number, percentage of pain medicine dispenses and mean morphine equivalent dispense dose.

RESULTS

Nearly one-third (28.4%, n=22 807) of our sample of 80 324 workers were dispensed any opioid in the first year since the workers' compensation insurer received their claim. There were no significant step or trend changes in the number or percentage of pain medicines dispensed of up-scheduled low-dose codeine. Only 2.9% of workers were ever dispensed up-scheduled low-dose codeine, specifically 2.5% after up-scheduling (1 February 2018). After up-scheduling of low-dose codeine, workers were more likely to be dispensed opioids (excluding codeine) (prevalence ratio (PR) 1.21, 99% CI 1.13, 1.31) or other pain medicines (eg, pregabalin, paracetamol) (PR 1.11, 99% CI 1.03, 1.19) compared with the year prior. There was a significant 28.5% (99% CI 16.3, 41.9) step increase (ie, increase immediately after up-scheduling) in high-dose (>15 mg) codeine with a significant trend decrease (-1.3%, 99% CI -2.5, -0.2).

CONCLUSION

Up-scheduling low-dose codeine to prescription-only medicines did not significantly change the dispensing of low-dose codeine-containing products to workers with accepted workers' compensation time loss claims for musculoskeletal conditions.

Can General Practitioner Opioid Prescribing to Compensated Workers with Low Back Pain Be Detected Using Administrative Payments Data? An Exploratory Study

BACKGROUND

Approximately one third of Australians with accepted time loss workers' compensation claims for low back pain (LBP) are dispensed opioid analgesics. Structured administrative payments data is scalable but does not directly link opioids to prescribers. We sought to determine whether opioid prescribing by general practitioners (GPs) to workers with workers' compensation claims for LBP can be detected in structured administrative payments data.

METHODS

We used a sample of workers with accepted time loss workers' compensation claims for low back pain from 2011 to 2015 from the Australian states of Victoria and South Australia. We structured administrative data to test the assumption that opioid dispenses that occurred immediately after a GP encounter in sequence and occurred on the same date as the GP encounter are likely to be related. We measured the number and proportion of opioid dispenses with a GP encounter prior and the days between a GP encounter and opioid dispense.

RESULTS

Nearly one third of workers (32.2%, N = 4,128) in our sample (n = 12,816) were dispensed opioids a median of five times (interquartile range 2, 17). There were 43,324 opioid dispenses to included workers. 30,263 (69.9%) of opioid dispenses were immediately preceded by a GP encounter. Of those dispenses, 51.0% (n = 15,443) occurred on the same day as the GP encounter.

CONCLUSION

At least one third of opioids dispensed to workers with claims for LBP can be potentially linked to GP prescribing using workers' compensation structured administrative payments data. This approach could have potential applications in supporting monitoring and audit and feedback systems. Future research should test this approach with a more diverse array of pain medicines and medical practitioners.

Patterns of antidepressant use in people with low back pain: A retrospective study using workers' compensation data

BACKGROUND

Antidepressants are commonly used to treat low back pain (LBP), but little is known about patterns of antidepressant use in this population. This study aimed to identify patterns of antidepressant dispensing and switching in Australian workers with an accepted workers' compensation claim for LBP, and to investigate factors associated with dispensing and switching.

METHODS

This retrospective study included data from accepted workers' compensation time loss claims for LBP between 2010 and 2018 with a 2-year follow-up. We described the type of antidepressant dispensed, the time for the index antidepressant to be dispensed, and whether people in the study switched to a different antidepressant. Logistic regression models investigated factors associated with antidepressants being dispensed for the first time and for switching to a different antidepressant.

RESULTS

Antidepressants were dispensed to 2476 people with LBP (14%) at least once after a median (IQR) of 28 (10.9-54.7) weeks. Amitriptyline was the most dispensed antidepressant at any one point (47.8%), and the most common index antidepressant (42.9%). Also, 32.7% of people switched to a different antidepressant at least once. Sex, age, having been dispensed opioids, gabapentinoids or diazepam prior to antidepressants being dispensed, having used psychological services, and socioeconomic disadvantage were associated with antidepressant dispensing.

CONCLUSION

One in seven people with LBP were dispensed an antidepressant, most commonly amitriptyline. Antidepressants were commonly used in combination with other pain medicines such as opioids, gabapentinoids and diazepam.

SIGNIFICANCE

Antidepressants were dispensed to one in seven people with low back pain, most commonly amitriptyline followed by duloxetine. Antidepressants were typically dispensed after 6 months and after other medicines such as opioids, gabapentinoids and diazepam had been dispensed. Due to the administrative nature of the data, the study cannot infer whether antidepressants were dispensed to treat pain or other health problems, such as a mental health condition.

Patterns of physiotherapy attendance in compensated Australian workers with low back pain: a retrospective cohort study

ABSTRACT

Workers with low back pain (LBP) frequently seek care from physiotherapists. We sought to identify patterns of physiotherapy attendance and factors associated with these patterns in Australian workers with accepted compensation claims for LBP. We included workers with accepted workers' compensation claims for LBP from 4 Australian states between 2011 and 2015. We used trajectory modelling to identify distinct groups of workers based on the number (ie, volume) of monthly physiotherapy attendances over a 2-year period from claim acceptance. Descriptive statistics and logistic regression models were used to compare the characteristics of the groups. A small but significant proportion attend numerous times over a long period. 79.0% of the sample (N = 22,767) attended physiotherapy at least once in the 2 years after claim acceptance. Among these, trajectory modelling identified 4 distinct patterns of attendance. Most (N = 11,808, 51.9%) recorded a short-term low-volume pattern, 26.8% (n = 6089) recorded a short-term high-volume pattern, 14.3% (n = 3255) recorded a long-term low-volume pattern, and 7.1% (n = 1615) recorded a long-term high-volume pattern. Workers from Victoria (OR 0.34, 99% CI 0.31, 0.37), South Australia (OR 0.69, 99% CI 0.60, 0.80), and Western Australia (OR 0.79, 99% CI 0.69, 0.88) were significantly less likely to attend physiotherapy than workers from Queensland. Victorian workers were significantly more likely to be in one of the 2 long-term trajectory groups (OR 8.17, 99% CI 6.86, 9.73; OR 18.68, 99% CI 13.57, 25.70). In conclusion, most compensated Australian workers with LBP attend physiotherapy. Significant interjurisdictional differences between attendance patterns suggests that policy may play an important role in healthcare delivery.

Trajectories of opioid use among patients with low back pain: Association to work absence

BACKGROUND

Low back pain (LBP) is a leading reason for opioid use and a closer examination of opioid use and productivity losses among these patients is needed. We identify opioid use trajectories using a group-based trajectory model (GBTM) and estimate productivity losses across the trajectories.

METHODS

Patients diagnosed with LBP in Swedish specialty care between 2011 and 2015, between the ages of 20 and 60, were included. Two GBTMs were estimated on monthly opioid use (converted to oral morphine equivalents) during the two 12-month periods preceding and following diagnosis. Productivity losses were estimated using the human-capital approach.

RESULTS

In total, 147,035 patients were included. The mean age at diagnosis was 43 years of age and 49% of the patients were male. A qualitative assessment of the identified groups in the GBTM models was made based on the patterns of opioid use. We chose three pre-diagnosis groups characterized as 'Pre-low' (N = 109,492), 'Pre-increase' (N = 27,336) and 'Pre-high' (N = 10,207). Similarly, four post-diagnosis groups were chosen and characterized as 'Post-low' (N = 73,287), 'Post-decrease' (N = 39,446), 'Post-moderate' (N = 20,001) and 'Post-high' (N = 13,595). Only 50% of the patients in the 'Pre-high' group were in the 'Post-high' group. The total productivity losses by the pre-diagnosis groups were more than 2.7 billion Euros over the total 6-year study period.

CONCLUSION

This study highlights how patients with LBP and high use of opioids are highly correlated before and after diagnosis. Patients with high use of opioids also exhibit high work absence and productivity losses.

SIGNIFICANCE STATEMENT

This was the first study to estimate trajectories of opioids in the two time periods before and after a diagnosis of low back pain. For the first time, productivity losses were also estimated across the identified opioid use trajectories.

The Association of Physiotherapy and Opioid Use With Duration of Compensated Time Loss for Workers With Low Back Pain

OBJECTIVE

To determine patterns of physiotherapy and opioid use among compensated workers with low back pain (LBP), factors associated with these, and their association with time loss.

METHODS

Accepted Victorian and South Australian workers' compensation claims, services, and medicines data for LBP claims lodged June 30, 2010-July 1, 2015. Descriptive statistics, multinomial logistic, and Cox regression were used to determine usage groups, their predictors, and effect of these on time loss.

RESULTS

Of 15,728 claims, 24.4% received no services, 3.6% received opioids only, 43.3% received physiotherapy only, and 28.8% received both opioids and physiotherapy. Sex, age, occupation, remoteness, jurisdiction, and socioeconomic status were significantly associated with usage groups. Using opioids and physiotherapy had the longest time loss.

CONCLUSIONS

Any services/medicine usage was associated with increased time loss and was longest for combined physiotherapy and opioids.

Impact of Prescription Medicines on Work-Related Outcomes in Workers with Musculoskeletal Disorders or Injuries: A Systematic Scoping Review

PURPOSE

Medicines are often prescribed to workers with musculoskeletal disorders (MSDs) and injuries to relieve pain and facilitate their recovery and return to work. However, there is a growing concern that prescription medicines may have adverse effects on work function. This scoping review aimed to summarize the existing empirical evidence on prescription medicine use by workers with MSD or injury and its relationship with work-related outcomes.

METHODS

We identified studies through structured searching of MEDLINE, EMBASE, PsycINFO, CINAHL Plus, Scopus, Web of Science and Cochrane library databases, and via searching of dissertations, theses, and grey literature databases. Studies that examined the association between prescription medicine and work-related outcomes in working age people with injury or MSDs, and were published in English after the year 2000 were eligible.

RESULTS

From the 4884 records identified, 65 studies were included for review. Back disorders and opioids were the most commonly studied musculoskeletal conditions and prescription medicines, respectively. Most studies showed a negative relationship between prescription medicines and work outcomes. Opioids, psychotropics and their combination were the most common medicines associated with adverse work outcomes. Opioid prescriptions with early initiation, long-term use, strong and/or high dose and extended pre- and post-operative use in workers' compensation setting were consistently associated with adverse work function. We found emerging but inconsistent evidence that skeletal muscle relaxants and non-steroidal anti-inflammatory drugs were associated with unfavorable work outcomes.

CONCLUSION

Opioids and other prescription medicines might be associated with adverse work outcomes. However, the evidence is conflicting and there were relatively fewer studies on non-opioid medicines. Further studies with more robust design are required to enable more definitive exploration of causal relationships and settle inconsistent evidence.

Dose Escalations Among Workers' Compensation Claimants Using Opioid Medications-An 8 Year Postinjury Follow up Study

OBJECTIVE

The aim of the study is to determine the morphine equivalent dose in milligrams (MED)/day escalation trend after initial utilization.

METHODS

A total of 25,108 lost time claims filed between 1998 and 2007 were followed for 8 years from injury date. Claims were stratified by initial MED/day at 3 months after injury into four groups (0, 1 to < 15, 15 to < 30, and ≥30 MED/day). The slopes in MED/year of opioid dose escalation were determined for each initial MED/day group.

RESULTS

The slopes of MED/day escalation by initial MED categories were similar ( P ≥ 0.05) ranging from 5.38 to 7.76 MED annually. On average, MED/day increased in a liner pattern with a slope at 6.28 MED/year ( P < 0.01).

CONCLUSIONS

Opioid MED/day increased in a linear pattern, regardless of initial MED/day dose.

Development and initial application of a harmonised multi-jurisdiction work injury compensation database

Objectives

Workers' compensation schemes provide funding for wage replacement and healthcare for injured and ill workers. In Australia, workers' compensation schemes operate independently in different jurisdictions, making comparisons of health service use challenging. We sought to develop and deploy a new database of health service and income support data, harmonising data from multiple Australian workers' compensation jurisdictions.

Methods

We worked with workers' compensation authorities from six Australian jurisdictions to combine claims, healthcare, medicines and wage replacement data for a sample of compensated workers with claims for musculoskeletal conditions. We designed a structured relational database and developed a bespoke health services coding scheme to harmonise data across jurisdictions.

Results

The Multi-Jurisdiction Workers' Compensation Database contains four data sets: claims, services, medicines and wage replacement. The claims data set contains 158,946 claims for low back pain (49.6%), limb fracture (23.8%) and non-specific limb conditions (26.7%). The services data set contains a total of 4.2 million cleaned and harmonised services including doctors (29.9%), physical therapists (56.3%), psychological therapists (2.8%), diagnostic procedures (5.5%) and examinations and assessments (5.6%). The medicines data set contains 524,380 medicine dispenses, with 208,504 (39.8%) dispenses for opioid analgesics.

Conclusions

The development of this database presents potential opportunities to gain a greater understanding of health service use in the Australian workers' compensation sector, to measure the impact of policy change on health services and to provide a method for further data harmonisation. Future efforts could seek to conduct linkage with other data sources.