Article Text

Original research
Time trends and patterns in opioid prescription use following orthopaedic surgery in Ontario, Canada, from 2004/2005 to 2017/2018: a population-based study
  1. Mayilee Canizares1,
  2. J Denise Power1,
  3. Anthony V Perruccio1,
  4. Christian Veillette2,
  5. Nizar Mahomed2,
  6. Y Raja Rampersaud2
  1. 1Schroeder Arthritis Institute, University Health Network, Toronto, Ontario, Canada
  2. 2Department of Orthopaedic Surgery, University Health Network, Toronto, Ontario, Canada
  1. Correspondence to Dr Mayilee Canizares; mayilee.canizares{at}uhnresearch.ca

Abstract

Objectives Increased use of opioids and their associated harms have raised concerns around prescription opioid use for pain management following surgery. We examined trends and patterns of opioid prescribing following elective orthopaedic surgery.

Design Population-based study.

Setting Ontario, Canada.

Participants Ontario residents aged 66+ years who had elective orthopaedic surgery from April 2004 to March 2018.

Primary and secondary outcome measures Postoperative opioid use (short term: within 90 days of surgery, prolonged: within 180 days and chronic: within 1 year), specific opioids prescribed, average duration (days) and amount (morphine milligram equivalents) of the initial prescription by year of surgery.

Results We included 464 460 elective orthopaedic surgeries in 2004/2005–2017/2018: 80% of patients used opioids within 1 year of surgery—25.1% were chronic users. There was an 8% increase in opioid use within 1 year of surgery, from 75.1% in 2004/2005 to 80.9% in 2017/2018: a 29% increase in short-term use and a decline in prolonged (9%) and chronic (22%) use. After 2014/2015, prescribed opioid amounts initially declined sharply, while the duration of the initial prescription increased substantially. Across categories of use, there was a steady decline in coprescription of benzodiazepines and opioids.

Conclusions Most patients filled opioid prescriptions after surgery, and many continued filling prescriptions after 3 months. During a period of general increase in awareness of opioid harms and dissemination of guidelines/policies aimed at opioid prescribing for chronic pain, we found changes in prescribing practices following elective orthopaedic surgery. Findings illustrate the potential impact of guidelines/policies on shaping prescription patterns in the surgical population, even in the absence of specific guidelines for surgical prescribing.

  • surgery
  • health policy
  • health services for the aged
  • adult orthopaedics

Data availability statement

No data are available. The databases from this study are held securely in coded form at IC/ES. Although the data are not publicly available, access may be granted to those who meet prespecified criteria for confidential access, available at https://www.ices.on.ca/DAS.

http://creativecommons.org/licenses/by-nc/4.0/

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

STRENGTHS AND LIMITATIONS OF THIS STUDY

  • Our findings used data on dispensed drugs, but we are unable to determine whether patients used the medication. It is possible that some of the prescriptions captured may have been unused.

  • The dataset used lacks important details such as in-hospital drug use, reasons for opioid prescriptions and prescriptions that were indicated but never dispensed.

  • Our study focused on the population aged 66+ years; thus, our findings may not be generalisable to younger populations.

  • Given the number of initiatives and the timing of their implementation, it is difficult to attribute a particular initiative to the trends seen in our study.

  • The study reports an in-depth assessment of trends in postoperative opioid use, amount and type of opioids prescribed following orthopaedic surgery using comprehensive data spanning 14 years in a large population.

Introduction

Opioid use in North America and its associated harms remain a major public health concern among policymakers, clinicians and the public.1 In Canada, a series of initiatives aimed at restricting opioid availability and reducing harm began implementation in 2010, with a focus on opioid prescriptions for the treatment of chronic pain.2 Among them, new guidelines for the safe and effective use of prescription opioids were released in 2010.3 In 2012, most provinces delisted slow-release oxycodone formulations from their formularies,4 and the Narcotics Monitoring System—capturing prescriber, pharmacist and patient information for all narcotics and other controlled drugs dispensed—was implemented in Ontario.5 Between 2013 and 2015, measures to reduce opioid-related harms, efforts aimed at increasing the use of safer opioid formulations and limiting high-dose prescribing and numerous reports offering guidelines and recommendations for prescription opioids, all aimed at reducing harms, were released across Canada.2 In 2017, Canadian opioid prescribing guidelines were further updated, similar to the 2016 US Centres for Disease Control guidelines.6 Largely, these initiatives suggested the prescription of opioids in low dosage and for the shortest possible duration, and to avoid, whenever possible, the concurrent use of opioids and benzodiazepines.5 7–14 Mass media emphasis on the emerging opioid crisis during this period15 16 may have also contributed to broader public and professional practice awareness.

Opioid therapy is commonly prescribed to treat moderate to severe pain after major elective surgery. However, opioids are associated with potential harms, including risks of long-term use, negative surgical outcomes and potential diversion.17–21 US and Canadian studies have reported that about 10% of patients undergoing surgery are persistent opioid users within 1 year of surgery18 22 and that surgical patients generally use less opioids than prescribed, with the potential for diversion for non-medical use.23

Recognising these potential harms, initiatives and interventions to address the opioid crisis have mainly focused on regulatory changes for prescription opioids for chronic pain, with less clear guidelines/recommendations for postoperative opioid therapy.2 5 6 10 24 The use of opioids in surgical patients is challenging as physicians need to balance managing acute pain in the immediate postoperative period while curtailing the risks of persistent opioid use following the surgery. In Canada, there is a lack of surgery-specific guidelines and recommendations for opioid prescribing, and it is not clear if broad initiatives focused on prescribing opioids for chronic pain may influence orthopaedic surgeon’s prescribing practices. Therefore, in the absence of specific guidelines/recommendations for postoperative opioid therapy, examining whether broad initiatives focused on opioid prescribing for chronic pain nevertheless influence opioid prescribing following major elective surgery can inform alternative approaches to treating postsurgical pain.

Orthopaedic surgery is one of the most common elective procedures performed worldwide, particularly among the older population,25 and orthopaedic surgeons are among the top prescribers of opioid prescriptions among physicians.26 The examination of prescription opioid use following orthopaedic surgery over time may help to elucidate if evidence-based recommendations and public attention to the opioid crisis have influenced orthopaedic surgeons’ opioid prescription practices. Therefore, the goal of this study was to characterise time trends and patterns of postoperative prescription opioid use in a large cohort of patients undergoing elective orthopaedic surgery in Ontario (Canada’s most populous province with a population of 14.5 million people) from 2004/2005 to 2017/2018. We also examined trends over time in the amount prescribed, type of opioid prescribed and coprescription of opioids and benzodiazepines.

Methods

Study design, data sources and cohort assembly

This is a descriptive population-based cohort study of all Ontario residents who were aged 66+ years and had elective orthopaedic surgery between 1 April 2004 and 31 March 2018, with follow-up to 31 March 2019. Ontario has a publicly funded healthcare system that covers all medically necessary hospital and doctors’ services without out-of-pocket payments at the point of service.

We identified patients with inpatient and outpatient surgeries using the Canadian Institute for Health Information Discharge Abstract Database and the National Ambulatory Care Reporting System, respectively. We defined orthopaedic surgeries using the Canadian Classification of Interventions codes (1SA-1WZ). These databases included information on patient demographics and clinical characteristics.27 We included patients with one eligible surgery within each year.

We used the Ontario Drug Benefit database to extract information on prescriptions filled, including the date, number of days supplied and drug identification number.28 In Ontario, data on prescription drugs are only available for those aged 65 years or older and we used a preoperative time window of 1 year for ascertaining previous medication use, hence the lower age limit of 66 years.

These databases were linked using encrypted unique health card numbers in a protected environment by ICES analysts. ICES is a not-for-profit research institute encompassing a secure and accessible array of Ontario’s health-related data. We had access to de-identified patient-level data through the ICES Data and Analytic Services division (https://www.ices.on.ca/DAS).

Measures

The opioids included were codeine, morphine, oxycodone, hydromorphone, meperidine, oxymorphone, methadone and transdermal fentanyl.22 We examined the following outcomes:

  • Preoperative opioid use. Defined as ≥1 opioid prescription filled within 1–90 days before the index surgery.

  • Postoperative opioid use. Defined as short-term (≥1 opioid prescription filled within 1–90 days after the index surgery), prolonged (≥1 opioid prescription filled within 1–90 days after surgery along with ≥1 opioid prescription filled within 91–180 days after surgery) and chronic (≥1 opioid prescription filled within 1–90 days after surgery along with ≥1 opioid prescription filled within 180–365 days after surgery).18 22

  • Duration and amount of the initial prescription. Using the information on the initial prescription filled, we calculated (1) prescription duration (days) and identified short-duration prescriptions (≤7 days)6 14 and (2) the amount of opioid prescribed measured in morphine milligram equivalents (MMEs) using well-established methods and identified low dosage (1–50).29

  • Coprescription of opioids and benzodiazepines. We identified patients with at least one concurrent opioid and benzodiazepine prescription filled within 1 year following surgical discharge.

We described cohort participants by the following covariates: age, sex, income, rurality, surgical joint/site, comorbidities, preoperative medication use, previous surgery, surgery type and surgical setting. We present a detailed definition of these variables in online supplemental material.

Statistical analyses

We describe the study population and outcomes by year of surgery using descriptive statistics—percentage for categorical variables and mean and SD for continuous variables. We did not adjust for patient characteristics, as the purpose of the study was not to explain how patient factors influence opioid prescribing in orthopaedic surgery but rather to document trends and how opioid prescribing patterns may correspond to broad public health initiatives within a very large clinical population in which opioids are regularly prescribed. Furthermore, public health initiatives targeted the broad clinical population and healthcare providers.

We present results for the overall cohort and preoperative opioid use within 90 days before the index surgery.14 In supplementary analyses, we examined if changes over time in postoperative opioid use, duration and amount of initial prescription filled differed by surgical joint/site. All analyses were performed using SAS/STAT software V.9.4 (Copyright 2020 SAS Institute, Cary, NC, USA).

Patient and public involvement

This study used data routinely collected by the healthcare system; hence, no patients were engaged in the study design or implementation.

Results

Characteristics of the study population

Considering one elective orthopaedic surgery per patient per year, there were 464 460 surgeries performed on patients aged 66+ years between 2004/2005 and 2017/2018. Overall, 58.6% of patients were female and 16.9% were aged 80+ years. The majority of patients were from urban areas (83.8%) and higher-income neighbourhoods (22.7% in the top quintile vs 16.6% in the bottom quintile). By surgical joint/site, 48.5% had knee surgery, 20% hip, 12.7% hand/wrist, 7% shoulder/elbow, 5.9% spine and 5.9% foot/ankle. About one-third were outpatient surgeries, 1.1% were revision surgeries, 28.3% of patients had at least one comorbidity, 23.2% filled opioid prescriptions within 90 days before surgery, 10.7% filled prescriptions for benzodiazepines within 1 year before surgery and 72% had high polypharmacy (table 1). The distribution of patient characteristics did not vary substantially over time (online supplemental table S1).

Table 1

Characteristics of the study population and prevalence of opioid use by patient characteristics (Ontario, Canada, 2004/2005–2017/2018)

Postoperative opioid use

Within 1 year of surgery, 80% of patients used opioids: 48.9% were short term, 6% were prolonged and 25.1% were chronic users. By preoperative opioid use, 77.3% of non-users were postoperative users (55.8% short term, 5.6% prolonged and 15.9% chronic users) versus 88.8% of preoperative users (26% short term, 7.2% prolonged and 55.6% chronic users) (table 1). There was an overall 8% increase in opioid use within 1 year of surgery (75.1% in 2004/2005 to 80.9% in 2017/2018)—a 29% increase in short-term use and a decline in prolonged (9%) and chronic (22%) use (figure 1A). We observed similar trends among preoperative users and non-users (figure 1B,C, respectively).

Figure 1

Prevalence of postoperative prescription opioid use by year of index surgery: (A) overall, (B) preoperative non-users within 90 days and (C) preoperative users within 90 days (Ontario, Canada, 2004/2005–2017/2018).

Duration and amount of initial prescription

Across opioid users, the average duration of the initial prescription was relatively stable between 2004/2005 and 2014/2015, then increased in 2015/2016–2017/18 (figure 2A). Additionally, between 2004/2005 and 2014/2015, the proportion of patients with short duration (≤7 days) increased for all categories of opioid use, with an approximately 80% decline in the proportion in 2015/2016–2017/2018 (figure 2B). The duration of the initial prescription was generally longer among preoperative users, particularly for postoperative chronic users.

Figure 2

Duration and amount of opioids in the initial prescription by opioid user type and year of surgery (Ontario, Canada, 2004/2005–2017/2018).

The average amount of opioids (MME/day) initially prescribed did not differ across categories of opioid use. The average amount increased between 2004/2005 and 2014/2015, declining in 2015/2016–2017/2018 (figure 2C). Furthermore, there was a twofold increase in the proportion of patients with low-dose (<50 MMEs/day) prescriptions between 2004/2005–2014/2015 and 2015/2016–2017/2018 (figure 2D). This was mirrored by a decline in the proportion of high-dose (≥90 MMEs/day) opioids from 20.1% in 2004/2005–2014/2015 to 0.3% in 2015/2016–2017/2018 (data not shown).

Type of opioids prescribed

The distribution of the most common opioids prescribed by year of surgery is shown in figure 3. Among all patients receiving prescriptions for opioids following surgery, the most common opioids prescribed in 2004/2005 were codeine (77.2%) and oxycodone (39.9%), while in 2017/2018, hydromorphone was the most common opioid prescribed (49.9%), followed by oxycodone (38.6%). These changes represented a 61.1% reduction in codeine prescribing with a concomitant 15-fold increase (from 3.2% in 2004/2005 to 49.9% in 2017/2018) in hydromorphone prescribing over the study period.

Figure 3

Most commonly prescribed opioids by year of surgery (Ontario, Canada, 2004/2005–2017/2018).

Coprescription of opioids and benzodiazepines

Overall, 12.4% of the study cohort had coprescription of opioids and benzodiazepines within 1 year of surgery, with a 47% decline in the proportion from 2004/2005 to 2017/2018. This decline was seen across categories of opioid use. Of note, the prevalence of coprescription was lowest among preoperative non-users with short-term postoperative use and highest among preoperative users with chronic postoperative use (figure 4).

Figure 4

Coprescription of opioids and benzodiazepines following surgery by opioid user type and year of surgery (Ontario, Canada, 2004/2005–2017/2018).

Supplementary analyses

The prevalence of postoperative opioid use varied by surgical joints/sites: 84.1% of knee, 78.9% of hip, 78% of spine, 86.6% of shoulder/elbow, 64.1% of hand/wrist and 79.2% of foot/ankle. Short-term postoperative use was highest for knee, hip, shoulder/elbow and foot/ankle (50.7%, 51.4%, 51.9% and 51.5%, respectively), while chronic use was highest for spine, shoulder/elbow and knee (35%, 28.8% and 26.3%, respectively). Generally, across joints/sites, a small increase in the overall prevalence of postoperative opioid use, with an increase in short term and a decline in chronic use over time, was seen. However, there were few exceptions; for example, among non-preoperative users, postoperative opioid use did not vary substantially for hand/wrist, while chronic use did not decline for foot/ankle (online supplemental figures S1 and S2).

For all categories of opioid use, prescriptions for hip, knee and spine had a longer duration, while those for hand/wrist had the shortest duration, with the exception of chronic users among preoperative users (the duration was longer for the spine and hand/wrist compared with other joints/sites). Generally, across surgical joints/sites, the duration of the initial prescription increased substantially after 2014/2015 (online supplemental figure S3).

The amount of the initial prescription was lowest for hand/wrist across categories of opioid use. The decline in the average amount of the initial prescription after 2014/2015 was seen across joints/sites (online supplemental figure S4). The prevalence of benzodiazepine and opioid coprescription was lowest for hand/wrist surgeries and highest for spine surgeries. Coprescriptions declined over time across all surgical joints/sites (online supplemental figure S5).

Discussion

In this large population-based study, we found that 80% of older patients filled opioid prescriptions within 1 year following elective orthopaedic surgery in Ontario in the period 2004/2005–2017/2018, and 25.1% had chronic opioid use. This was during a period of increased awareness of the harms of opioids15 16 and the implementation of several interventions, guidelines and policies aimed at reducing opioid-related harms.5 8–10 24 30–33 While these initiatives focused primarily on opioid prescribing for chronic pain, we nevertheless found evidence that they likely influenced changes in patterns of opioid prescribing following orthopaedic surgery. We report a general decline in postoperative chronic use that was offset by an increase in short-term use, resulting in a slight increase in overall use over time. Furthermore, the amount prescribed (MMEs/day) in the initial prescription decreased sharply after 2014/2015. This was driven by an increase in low-dose (<50 MMEs/day) prescriptions and a decline of more than 80% in high-dose prescribing (>90 MMEs/day). However, the average duration of the initial prescription was substantially longer after 2014/2015 than in the previous period. We also found a steady decline in coprescription of opioids and benzodiazepines since 2004/2005 across categories of opioid use. Of note, we found that time trends and patterns of opioids dispensed were somewhat comparable across surgical joints/sites.

The slight increase in overall postoperative opioid use, particularly short-term use, rates over the study period is in contrast to the general decline in opioid prescription use in the general population in Canada.29 34 However, the decline in postoperative chronic opioid use over time is in line with recent studies of patients undergoing surgery in Ontario.35 36

The sharp decline in high-dose opioids observed in 2015/2016 occurred following 2 years of full implementation of the Narcotic Monitoring System in Ontario,5 10 the publication of national guidelines discouraging the use of high-dose opioids37 38 and the publication of studies indicating progressively increasing overdose risk with increasing MMEs/day.39 Likewise, the decline in concurrent opioid and benzodiazepine use is in line with Canadian guidelines for older adults24 and has been reported by others.34 40 41 At the same time, mass media attention to the emerging opioid crisis in Canada during this period15 16 may have also contributed to broader public awareness and changes in professional practices in regard to opioid use. The increasing use of hydromorphone and the decrease in the use of codeine and oxycodone following surgery also mirror general trends for the use of these drugs in Ontario. This shift reflects reimbursement changes addressing oxycodone use4 and guidelines cautioning against the use of codeine.42

In contrast to recommendations, we found an increase in the number of days of supply of the initial prescription for all categories of opioid use, particularly for persistent chronic users. This is concerning, as studies have reported that long prescription duration in the early postoperative period following surgery is associated with an increased risk of prolonged opioid use and worse surgical outcomes.43 44 We cannot ascertain the appropriateness of the prescribed duration, but a recent US study reported that approximately half of the patients undergoing major orthopaedic surgery did not use the total amount of their initial prescription.45 Therefore, future studies to determine the appropriate amount and duration of opioid therapy following elective orthopaedic procedures are warranted.

Limitations

Our study has limitations due to the nature of routinely collected administrative data, which limited the years we could study, as we had to rely on data availability. Nevertheless, since our goal was to explore whether changes in general prescribing guidelines might correspond with changes in practice observed at a system (or population) level using administrative data, including the timing between the two, it might point towards a direct influence and the potential benefits of crafting and implementing specific postoperative guidelines. Furthermore, we used data on dispensed drugs, but we were unable to determine whether patients used the medication. It is possible that some of the prescriptions captured may have been unused. Likewise, the datasets used lack important details such as in-hospital drug use, reasons for opioid prescriptions and prescriptions that were indicated but never dispensed. We are unable to determine the reasons (eg, persistent postsurgical pain) for persistent and/or chronic postoperative opioid use. There is no standardised definition of opioid use following surgery, and it remains unclear if prolonged/chronic opioid use represents ongoing surgical pain or the management of other symptoms. We used a broad definition of opioid use, which may obscure distinct patterns of opioid use in a surgical population. Our study focused on the population aged 66+ years; thus, our findings may not be generalisable to younger populations. Given the number of initiatives and the timing of their implementation, it is difficult to attribute a particular initiative to the trends seen in our study. Lastly, our study is based in Ontario, Canada, and it is unknown to what extent the findings can be generalised to other jurisdictions. Nevertheless, our study reports an in-depth assessment of trends in postoperative opioid use, amount and type of opioids prescribed following orthopaedic surgery using comprehensive data spanning 14 years from Ontario. A particular strength of our study is that we examined elective orthopaedic surgeries in all joints and in a population with universal access to care. Our results add to a growing body of literature on the impact of public initiatives and guidelines on reducing opioid prescriptions. Our findings highlight the importance of further research to understand opportunities to develop and implement surgery-specific prescribing guidelines and additional research into the provider-related factors that may influence prescribing patterns. Another area that warrants further attention is investigating alternative pain management strategies while recognising that opioids remain an important therapy for managing pain after surgery.

Conclusion

Our findings demonstrate changes in opioid prescribing practices following elective orthopaedic surgery in Ontario, particularly since 2015/2016—a period in which numerous reports offering guidelines and recommendations for prescription opioids aimed at reducing opioid-related harms were released and a general increase in awareness of the harms of opioids. In line with these recommendations, we report a reduction in postoperative chronic use, the amount of opioids initially prescribed and the concurrent use of opioids and benzodiazepines, but an increase in the duration of the initial prescription. Therefore, our study suggests that changes in both policies and guidelines may have played a role in shaping overall opioid prescription patterns in the orthopaedic surgical population, even in the absence of specific guidelines for surgical prescribing. However, as our findings are based on administrative data from which a direct correlation cannot be made, we cannot make a determination in this respect. Nevertheless, our findings highlight the need for high-quality evidence and initiatives to optimise opioid use in the surgical population, as well as additional research into the provider-related factors that may influence prescribing patterns. Another area that warrants further attention is investigating alternative pain management strategies while recognising that opioids remain an important therapy for managing pain after surgery.

Data availability statement

No data are available. The databases from this study are held securely in coded form at IC/ES. Although the data are not publicly available, access may be granted to those who meet prespecified criteria for confidential access, available at https://www.ices.on.ca/DAS.

Ethics statements

Patient consent for publication

Ethics approval

The research protocol was approved by the research ethics board of University Health Network (17-6124.4).

Acknowledgments

This study contracted IC/ES Data & Analytic Services (DAS) and used de-identified data from the IC/ES Data Repository, which is managed by IC/ES with support from its funders and partners: Canada’s Strategy for Patient-Oriented Research (SPOR), the Ontario SPOR Support Unit, the Canadian Institutes of Health Research and the Government of Ontario. The opinions, results and conclusions reported are those of the authors.

References

Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

Footnotes

  • Contributors MC, JDP, AVP, CV, NM and YRR were involved in the conception and design of the study. MC conducted the analysis and drafted the manuscript. All authors contributed to the interpretation of analysis results, critically revised the article, and approved the final version for submission. MC takes responsibility for the integrity of work as a whole. MC is the guarantor.

  • Funding This study was financially supported by the Toronto General & Western Hospital Foundation through the University Health Network-Schroeder Arthritis Institute. The funding source had no involvement in study design or manuscript preparation. AVP is supported by an award from Arthritis Society Canada (STAR-20-0000000012).

  • Competing interests None declared.

  • Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.